National tuberculosis prevalence surveys 2007-2016

NATIONAL TUBERCULOSIS PREVALENCE SURVEYS 2007-2016 National tuberculosis prevalence surveys 2007-2016 ISBN 978-92-4-002243-0 (electronic version) ISBN 978-92-4-002244-7 (print version) © World Health Organization 2021 Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo). Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work, you should add the following disclaimer along with the suggested citation: “This translation was not created by the World Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English edition shall be the binding and authentic edition”. Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the World Intellectual Property Organization (http://www.wipo.int/amc/en/mediation/rules/). Suggested citation. National tuberculosis prevalence surveys 2007-2016. Geneva: World Health Organization; 2021. Licence: CC BY-NC-SA 3.0 IGO. Cataloguing-in-Publication (CIP) data. CIP data are available at http://apps.who.int/iris. Sales, rights and licensing. To purchase WHO publications, see http://apps.who.int/bookorders. To submit requests for commercial use and queries on rights and licensing, see http://www.who.int/about/licensing. Third-party materials. If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user. General disclaimers. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. iii Preface v Acknowledgements vii Abbreviations xv Part I. An overview of the 25 surveys implemented 2007–2016 1 Introduction 3 Methods 15 Results and their implications 33 Success, challenges and lessons learned 53 Future direction 61 Part II. Country-by-country survey profiles 71 Bangladesh 73 Cambodia 81 China 89 Democratic People’s Republic of Korea 97 Ethiopia 105 Gambia 113 Ghana 121 Indonesia 129 Kenya 137 Lao People’s Democratic Republic 145 Malawi 153 Mongolia 161 Myanmar 169 Nigeria 177 Pakistan 185 Philippines 2007 193 Philippines 2016 201 Rwanda 209 Contents iv Contents Sudan 217 Thailand 225 Uganda 233 United Republic of Tanzania 241 Viet Nam 249 Zambia 257 Zimbabwe 265 vPreface At the time of publication of this book in early 2021, tuberculosis (TB) remains a major cause of ill health and one of the top causes of death worldwide. During the period 2000–2015, global and national efforts to reduce the burden of TB disease had the aim of achieving global TB targets that were set as part of the United Nations (UN) Millennium Development Goals (MDGs), the World Health Organization’s (WHO) Stop TB Strategy (2006–2015) and the Stop TB Partnership’s Global Plan to Stop TB (2006–2015). Three targets were set: to halt and reverse TB incidence by 2015; to halve the TB mortality rate by 2015 compared with 1990; and to halve the prevalence of TB disease by 2015 compared with 1990. In 2006, WHO established a Global Task Force on TB Impact Measurement, convened by the TB monitoring, evaluation and strategic information (TME) unit of WHO’s Global Tuberculosis Programme. The Task Force’s aim was to ensure a robust, rigorous and consensus-based assessment of whether the 2015 TB targets were achieved at global, regional and national levels. At its second meeting, held in 2007, the Task Force agreed on three strategic areas of work for the period 2007– 2015: strengthening of routine national surveillance systems (notification and vital registration) in all countries; national TB prevalence surveys in 22 global focus countries; and periodic review of the methods used by WHO to translate surveillance and survey data into estimates of TB disease burden. The 22 global focus countries were a prioritised subset of 53 countries considered eligible to implement a national TB prevalence survey: 13 in Africa and 9 in Asia. Global recognition of the importance of national TB prevalence surveys was reinforced and supported by considerable national interest in and commitment to implementing such surveys, which had started to grow and intensify in many countries during the early-mid 2000s. In 2007, however, the goal of completing a large number of national TB prevalence surveys in a relatively short period of time was a daunting task. The number of recent national surveys was small, and global and national experience and expertise in their design, implementation and analysis was scarce. Between 1990 and 2006, only a handful of countries in Asia successfully completed a national TB prevalence survey. No national survey had been attempted in the WHO African Region since the 1950s, with the sole exception of a survey in Eritrea in 2005 that was limited by the diagnostic methods used to detect people with TB. What followed was an unprecedented national, regional and global effort to implement national TB prevalence surveys. Between 2007 and the end of 2016, 24 countries implemented a total of 25 national surveys using methods recommended by WHO. The 24 countries comprised 18 of the 22 global focus countries and six other countries. The 25 surveys consisted of 13 in Asia and 12 in Africa. The outcome is a wealth of new data. These data were crucial to WHO’s assessment of whether the 2015 TB targets were met at global, regional and country levels, by providing a much better understanding of the burden of TB disease, including its distribution by age and sex, and reliable evidence about trends in countries where a repeat survey was done. The data have also provided new evidence about the symptoms experienced by people with undiagnosed TB in the community, the extent of gaps between the number of people with TB in the community and the number of people officially detected with TB, and health care-seeking behaviour in the public and private sectors, in turn shining new light on reasons for delays in diagnosing people with TB and for the underreporting of people diagnosed with TB to national authorities. Collectively, survey findings have informed the policies, plans and programmatic actions needed to address gaps in TB diagnosis and treatment and to reduce the burden of TB disease. Finally, the 24 countries have a robust baseline for assessing progress towards new global targets set in the UN Sustainable Development Goals (2016–2030) and WHO’s End TB Strategy (2016–2035). vi Preface At the global level, efforts to support the design, implementation, analysis and reporting of national TB prevalence surveys between 2007 and 2016 were led and coordinated by a subgroup of the WHO Global Task Force on TB Impact Measurement. This subgroup was led by staff in WHO’s TME unit. In 2016, it was our collective view that the methods, results, successes achieved, challenges faced and lessons learned from the 25 national surveys implemented 2007–2016 should be comprehensively documented in a book. We viewed such a product as a global public good, that should be available to all those with an interest in and commitment to using survey findings, now and in the future. As with implementation of the 25 surveys themselves, the book is the result of a major global, regional and national collaborative and collective effort, with more than 450 contributors from all around the world. We are proud of the final product, wholeheartedly thank all those who made it possible, and hope that it will be a valuable resource for many people for many years to come. March 2021 Jaap Broekmans Chair WHO Global Task Force on TB Impact Measurement Katherine Floyd Unit Head TB monitoring, evaluation and strategic information unit, WHO Global Tuberculosis Programme Philippe Glaziou Senior epidemiologist TB monitoring, evaluation and strategic information unit, WHO Global Tuberculosis Programme Ikushi Onozaki Medical officer TB monitoring, evaluation and strategic information unit, WHO Global Tuberculosis Programme Leader WHO Global Task Force on TB Impact Measurement subgroup on national TB prevalence surveys, 2007–2016 Irwin Law Technical officer TB monitoring, evaluation and strategic information unit, WHO Global Tuberculosis Programme Charalambos Sismanidis Team Lead for TB surveillance and surveys TB monitoring, evaluation and strategic information unit, WHO Global Tuberculosis Programme vii Acknowledgements This book represents a global collaborative effort of more than 450 people. A core team of six people at WHO headquarters conceptualized the book and led and coordinated its production: Katherine Floyd, Philippe Glaziou, Sayori Kobayashi, Irwin Law, Ikushi Onozaki and Charalambos Sismanidis. This core team was led by Katherine Floyd. The core team prepared the five cross-cutting chapters that form Part I of the book. Chapter 1 was prepared by Katherine Floyd, Philippe Glaziou, Irwin Law and Ikushi Onozaki. Chapter 2 and Chapter 3 were prepared by Irwin Law, with contributions from Katherine Floyd, Philippe Glaziou, Sayori Kobayashi, Ikushi Onozaki and Marina Tadolini (WHO consultant). Chapter 4 was prepared by Katherine Floyd. Chapter 5 was prepared by Katherine Floyd, Irwin Law and Charalambos Sismanidis, with contributions from Philippe Glaziou. The country-specific chapters for the 25 national TB prevalence surveys that were completed 2007–2016, which form Part II of the book, were prepared by the WHO core team together with key members of the national survey teams and people who provided technical assistance to these teams. Sayori Kobayashi and Irwin Law produced and checked the final datasets and reports that were used for each chapter in close collaboration with national survey teams, analysed the data and produced the standard sets of figures and tables that are featured in each chapter. The text of each chapter was drafted by national survey teams in collaboration with the WHO core team; all chapters were reviewed and finalized by Katherine Floyd and Irwin Law. Key contributors from the national survey teams to the preparation of the country-specific chapters were as follows: • Bangladesh: Vikarunnessa Begum, Mourad Gumusboga, Mohammad Mushtuq Husain, Mahmudur Rahman, Mohammed Sayeedur Rahman, Ahmad Raihan Sharif. • Cambodia: Mao Tan Eang. • China: Zhang Hui, Yanni Sun, Lixia Wang, Xia Yin Yin. • Democratic People’s Republic of Korea: Nam Ju O, Kamar Rezwan, O Hyang Son. • Ethiopia: Zeleke Alebachew. • Gambia: Ifedayo Adetifa. • Ghana: Zeleke Alebachew, Frank Bonsu. • Indonesia: Maria Christian, Bintari Dwihardiani, Dina Lolong. • Kenya: Hillary Kipruto, Enos Okumu Masini, Nkirote Mwirigi, Jane Nabongo, Geoffrey Okallo. • Lao People’s Democratic Republic: Phonenaly Chittamany, Jacque Sebert. • Malawi: Rhoda Banda, James Mpungu. • Mongolia: Tsolmon Boldoo, Yasunori Ichimura, Naranzul Dambaa, Narantuya Jadambaa. • Myanmar: Thandar Lwin, Norio Yamada, Kiyohiko Izumi. • Nigeria: Philip Patrobas Dashi. • Pakistan: Razia Fatima, Mahboob ul haq, Nasir Mahmood, Ejaz Qadeer, Sabira Tahseen, Aashifa Yaqoob. • Philippines: Bicbic Amarillo, Celina Garfin, Mary Ann Lansang, Leilani Naval, Olivia Sison, Rajendra Yadav. • Rwanda: Patrick Migambi. • Sudan: Asrar Mohammed Abdelsalam, Igbal Ahmed Elbasheer, Heba Kamal Hamed Elneel, Mai Mohammed Eltigany. viii Acknowledgements • Thailand: Sirinapa Jittamanee, Sriprapa Nateniyom. • Uganda: Elizeus Rutebemberwa, Frank Mugabe Rwabinumi, Samuel Kasozi. • United Republic of Tanzania: Armand Van Deun, Basra Doulla, Deusdedit V. Kamara, Sayoki Mfinanga, Beatrice Mutayoba, Mbazi Senkoro. • Viet Nam: Nguyen Binh Hoa. • Zambia: Nathan Kapata, Pascalina Kapata. • Zimbabwe: Joconiah Chirenda, Patrick Hazangwe, Ronnie Matambo, Junior Mutsvangwa, Charles Sandy. The WHO core team is grateful to the government of Japan and the United States Agency for International Development (USAID) for providing the funding that was necessary to produce the book. A full list of all contributors organized by country, followed by a list of contributors who provided international technical assistance organized according to their institutional affiliation, is provided below. * Principal investigator Cambodia Democratic People’s Republic of Korea China National survey team: Vikarunnessa Begum, Md Ehteshamul Huq Choudhury, Meerjady Sabrina Flora, Rouseli Haq, Md Mozammel Haque, Akter Hossain, Ashaque Husain, Mohammad Mushtuq Husain, Md Quamrul Islam, Shahid Md Sadiqul Islam, S.M. Mostofa Kamal, Iqbal Ansary Khan, Ahmed Hussain Khan, Asif Mujtoba Mahmud, Mahbubur Rahman, Mahmudur Rahman*, Mohammed Sayeedur Rahman, Md Jahangir Alam Sarker, Ahmad Raihan Sharif, Md Ashraf Uddin. Technical assistance: J. Sean Cavanaugh, Shua Chai, Mourad Gumusboga, Susumu Hirao, Sayori Kobayashi, Irwin Law, Ikushi Onozaki. National survey team: Koy Bonamy, Mao Tan Eang*, Chea Manith, Saint Saly, Peou Satha, Tieng Sivanna, Pheng Sok, Keo Sokonth, Kouet Pichenda. Technical assistance: Emily Bloss, Sian Floyd, Philippe Glaziou, Yutaka Hoshino, Kunihiko Ito, Hiroko Matsumoto, Hiroyuki Nishiyama, Kosuke Okada, Ikushi Onozaki, Masaki Ota, Charalampos Sismanidis, Tetsuhito Sugamoto, Sara Whitehead, Rajendra Yadav, Norio Yamada, Kiyomi Yamamoto, Takashi Yoshiyama. National survey team: Choe Tal Bom, Choe Tong Chol, Jo Won Ryong, Kim Hyong Hun, Ko Jin Hyok, Ri Chan Hyok*, Ri Jong Chan, Rim Gye Tong, Yun Jong Chol. Technical assistance: Mubeen Aslam, M. Bintari Dwihardiani, Philippe Glaziou, Partha Pratim Mandal, Charalampos Sismanidis. National survey team: Cao Jiping, Chen Mingting, Chen Wei, Chen Yude, Cheng Shiming, Du Xin, Duanmu Hongjin, He Guangxue, Jiang Shiwen, Jin Shuigao, Li Jun, Li Renzhong, Pan Yuxuan, Qian Yuanfu, Ruan Yunzhou, Shi Hongsheng, Tang Danlin, Tu Dehua, Wang Lixia*, Wang Shengfen, Wang Xiexiu, Wang Zhongren, Wu Zhenglai, Xia Yinyin, Xu Caihong, Xu Weiguo, Zhang Hui, Zhang Zongde, Zhao Fengzeng, Zhao Yanlin, Zheng Suhua, Zhou Lin, Zhou Xinhua, Zhu Guilin, Zhu Lizhen, Zou Jiqian. Technical assistance: Philippe Glaziou, Ikushi Onozaki, Charalampos Sismanidis. List of contributors by country Bangladesh ixAcknowledgements * Principal investigator Ethiopia Gambia Ghana Indonesia Kenya Lao People’s Democratic Republic National survey team: Almaz Abebe, Mulualem Agonafer, Zeleke Alebachew, Menelik Balcha, Tibebu Biniam, Feleke Dana, Molla Endale, Gashawtena Fantu, Tedla Fiseha, Amha Kebede*, Eshetu Lema, Shewalem Negash, Fasil Tsegaye, Sale Workneh. Technical assistance: Wilfred Nkhoma, Ikushi Onozaki, Peou Satha, Charalampos Sismanidis, Marina Tadolini, Hazim Timimi. National survey team: Ifedayo Adetifa*, Beatrice dei Alorse, William dei Alorse, Martin Antonio, Adedapo Bashorun, Elina Cole, Edward Demba, Simon Donkor, Umberto D’Alessandro, David Jeffries, Lindsay Kendall, Ma Ansu Kinteh, Christopher Linda, Ramatoulie Manne, Kodjovi Mlaga, Catherine Bi Okoi, Semeeh Omoleke. Technical assistance: Bimbo Fasan, Sian Floyd, Jan van den Hombergh, John Mayanda, Ikushi Onozaki, Charalampos Sismanidis, Marina Tadolini, Etienne Leroy Terquiem. National survey team: Kwasi Addo*, Robertson Adiei, Sauda Ahmed, Jane Amponsah, Herve Awako, Prince Boni, Frank Bonsu*, Ellis Owusu Dabo, Francisca Dzata, Raymond Yaw Gockah, John Gyapong, Kwadwo Koram, Nii Nortey Hanson Nortey, Michael Omari, Augustina Badu Peprah. Technical assistance: Zeleke Alebachew, Irwin Law, Wilfred Nkhoma, Ikushi Onozaki, Charalampos Sismanidis, Marina Tadolini. National survey team: Narendro Arifia, Elisabeth Bernadeth, Retno Kusuma Dewi, M. Bintari Dwihardiani, Darmawati, Irfan Ediyanto, M.N. Farid, Aziza G. Icksan, Jubaedi, Dina Bisara Lolong*, Lamria Pangaribuan, Pandu Riono, Risnawati Ainur Rofiq, Safrizal, Ade Yoska Tilla Serihati, Francisca Srioetami, Laura Valeria. Technical assistance: Philippe Glaziou, Irwin Law, Ikushi Onozaki, Charalampos Sismanidis, Marina Tadolini. National survey team: Janet Agaya, Jeremiah Chakaya, Martin Githiomi, Joel Kangangi, Maureen Kamene Kimenye, Richard Kiplimo, Hillary Kipruto*, Dickson Kirathe, Bernard Langat, Maurice Maina, Veronica Manduku, Enos Masini, Brenda Mungai, Rose Mwirigi, Margaret Ndisha, Amos Ndombi, Faith Ngari, James Ng’ang’a, Janice Njoroge, Obadiah Njuguna, Drusilla Nyaboke, Geoffrey Okallo, Jane Ong’ang’o*, Joseph Sitienei*, Anja Vant’Hoog, Josephine Wahogo. Technical assistance: Emily Bloss, Martien W. Borgdorff, Kevin Cain, Julia Ershova, Sayori Kobayashi, Irwin Law, Wilfred Nkhoma, Peou Satha, Marina Tadolini, Hazim Timimi. National survey team: Soth Bounmala, Phonenaly Chittamany, Bounkong Fongosa, Donekham Inthavong, Irwin Law, Vatthana Nanthana, Fulgence Nzabintwali, Manikhone Ouanephongchaleune, Phimpha Paboriboune, Vongvilay Paphatsalang, Thavone Phengsavatdy, Liene Phonekeo, Oroth Rajphol, Saveang Saisongkham, Jacques Sebert, Phasouk Senephansiri, Phannasinh Sylavanh*, Phouvang Vangvichit. Technical assistance: Etienne Leroy-Terquem, Pierre L’Her, Ikushi Onozaki, Sang Jae Kim, Peou Satha, Charalampos Sismanidis. x Acknowledgements Mongolia Myanmar Nigeria Pakistan Philippines 2016 Philippines 2007 National survey team: Puntsag Banzragch, Tsolmon Boldoo, Buyankhishig Burneebaatar, Naranzul Dambaa, Narantuya Jadambaa, Naranbat Nyamadawa, Soe Nyunt-U, Bayasgalan Purev, Tugsdelger Sovd*, Oyuntuya Tumenbayar. Technical assistance: M. Bintari Dwihardiani, M.N. Farid, Yasunori Ichimura, Satoshi Mitarai, Ikushi Onozaki, Norio Yamada. National survey team: Si Thu Aung, Tin Mi Mi Khaing, Thandar Lwin, Tin Tin Mar, Win Maung*, Bo Myint, Hnin Wai, Lwin Myo, Wint Wint Nyunt, Htar Htar Oo, San San Shein, Htay Lwin, Thandar Thwin, Ti Ti, Moe Zaw, Myo Zaw. Technical assistance: Eva Nathanson, Kosuke Okada, Ikushi Onozaki, Norio Yamada. National survey team: Haruna Adamu, Babalola Akin, Awe Ayodele, Osakwe Puis Chijioke, Emmanuel Idigbe, Daniel Olusoji James, Jose Michael Madu, Chukwueme Nkemdilim, Joshua Obasanya*, Osahon Ogbweiwe, Samuel Ogiri, Moses Onoh, Philip Patrobas, Abiola Tubi, Gideon Zaphania. Technical assistance: Julia Ershova, Daniella Cirillo, Eugene McCray, Wilfred Nkhoma, Ikushi Onozaki, Narayan Pendse, Charalampos Sismanidis. National survey team: Riaz Ahmed, Mohammad Asif, Razia Fatima*, Zulfiqar Ul Hassan, Abdul Mannan, Aisha Mariam, Ejaz Qadeer*, Sabir Rehman, Alamdar Hussain Rizvi, Zia Samad, Ghulam Nabi Shaikh, Arshad Shamsi, Sabira Tahseen*. Technical assistance: Amal Bassili, Nico Kalisvaart, Masja Straetemans, Edine Tiemersma. National survey team: Marissa Alejandria*, Maria Lourdes Amarillo, Concepcion Ang, Luis Anos, Joseph Adrian Buensalido, Johanna Patricia Cañal, Jose Rene Cruz, Anjo Benedict Fabellon, Nori Jane Galagar, Anna Marie Celina Garfin*, Noel Juban, Mary Ann Lansang*, Jacinto Blas Mantaring III, Myrna Mendoza*, Allison Noel, Rodelia Pascua, Sonia Salamat, Olivia Sison, Aser Sisona. Technical assistance: M. Bintari Dwihardiani, Julia Ershova, Yasunori Ichimura, Irwin Law, Hiroko Matsumoto, Ikushi Onozaki, Tetsuhiro Sugamoto, Marina Tadolini. National survey team: Lena Ablis, Virgil Belen, Gerardo Beltran, Jennifer Chua, Albert Angelo Concepcion, Grace Egos, Maricar Galipot, Ruffy Guilatco, Joselito Legaspi, Vivian Lofranco, Nellie Mangubat, Onofre Edwin Merilles, Leilani Naval, Ruth Orillaza-Chi, Ma. Imelda Quelapio*, Nona Rachel Mira, Sistla Radhakrishna, Genesis Ramos, Jesus Sarol, Thelma E. Tupasi*. Malawi National survey team: Rhoda Banda, Masy Chiocha, Isaiah Dambe, Andrew Dimba, Henry Kanyerere, Damson Kathyola, Sidon Konyani, Charles Mandambwe, Lameck Mlauzi, James Mpunga*, Alister Munthali*, Suzgo Mzumara, Daniel Nyangulu, Ishmael Nyasulu, George B. Samuti. Technical assistance: Julia Ershova, Sian Floyd, Irwin Law, Patrick Moonan, Wilfred Nkhoma, Ikushi Onozaki, Peou Satha. * Principal investigator xiAcknowledgements Sudan Thailand Uganda United Republic of Tanzania Viet Nam Zambia National survey team: Abdelaeem, Sawsan Mustafa Abdalla, Nahid Abdelgader, Ahmed Elhaj Ali, Fatih Alrahaman Ali, Abdel-rahaman, Asrar Mohammed Abdelsalam, Alfakie, Ayyed Muneam El-Dulaimi, Igbal Ahmed Elbasheer*, Hozifa Omer Eljak, Heba Kamal Hamed Elneel, Majda Elsayed, Mai Mohammed Eltigany, Sami Abdel Hameed, Hashim Salah Hamza, Nazar Alnoor Ibrahim, Sumia Yousif Mohammed, Mona Hassen Mustafa, Mustafa, Mohammed Osman, Hasham Alamin Salem. Technical assistance: Amal Bassili, Kiyohiko Izumi, Ikushi Onozaki, Sabira Tahseen, Fasil Tsegaye, Norio Yamada. National survey team: Sakchai Chaiamahapurk, Autagorn Chunmathong, Pavasuth Chutjuntaravong, Wilawan Dangsaart, Ratree Dokkabowt, Ornnipa Iamsamang, Kamonwan Imduang, Sirinapha Jittimanee, Phalin Kamolwat, Wiriya Madasin, Nuntaporn Meksawasdichai, Sriprapa Nateniyom*, Auyporn Petborisuit, Pattana Pokaew, Hataichanok Pukcharern, Walaya Sitti, Saijai Smithtikarn, Runjuan Sukkavee, Supaporn Wattanatoan, Narong Wongba. Technical assistance: Ikushi Onozaki, Norio Yamada. National survey team: Ronald Anguzu, Moses Joloba, Samuel Kasozi, Bruce Kirenga, Harriet Kisembo, Frank Mugabe*, Kenneth Musisi, Annet Nagudi, Abel Nkolo, Okot Martin Nwang, Elizeus Rutebemberwa*, Rogers Sekibira, Racheal Tumwebaze, William Worodria. Technical assistance: Emily Bloss, Julia Ershova, Wilfred Nkhoma, Ikushi Onozaki, Peou Satha, Charalampos Sismanidis, Marina Tadolini. National survey team: Basra Doulla, Said M. Egwaga*, Lulu Fundikira, Rahim Ishumi, Msaki John, Deusdedit V. Kamara, Ahmed Khatib, Senkoro Mbazi, Godfrey S. Mfinanga*, Lugano Mtafya, Blasdus F. Njako, Moses Ringo, Raymond P. Shirima. Technical assistance: Frank van Leth, Wilfred Nkhoma, Ikushi Onozaki, Charalampos Sismanidis. National survey team: Bao Thuyet, Chu Manh Dung, Dinh Ngoc Sy*, Do Trong Nghia, Ha Thuc Van, Nguyen Binh Hoa, Nguyen Cong Chi, Nguyen Van Cu, Nguyen Van Hung, Nguyen Viet Nhung*, Pham Vuong Khac, Thai Anh Sam, Tran Ngoc Thach, Vu Ngoc Tuan. Technical assistance: Martien W. Borgdorff, Frank G.J. Cobelens, Edine Tiemersma, Nico Kalisvaart, Agnes Gebhard, Marleen Vree. National survey team: Mashina Chomba, Pascalina Chanda-Kapata*, Nathan Kapata*, Patrick Katemangwe, Mazyanga Mazuba Liwewe, Chitani Mbewe, Mine Metitiri, Sam Msariri, Lutinala Nalomba, William Ngosa, Jane Shawa, Chris Silavwe, Veronica Sunkuntu, Mathias Tembo. Technical assistance: Julia Ershova, Nico Kalisvaart, Eveline Klinkenberg, Wilfred Nkhoma, Ikushi Onozaki, Charalampos Sismanidis. Rwanda National survey team: Pauline Basinga*, Michel Gasana*, Louise Kalisa, Elaine Kamanzi, Ndeziki Mashengesho, Patrick Migambi, Julie Mugabekazi, Calvin Mugabo, Jules Kamugunga Mulinzi, Védaste Ndahindwa, Alaine Umubyeyi Nyaruhirira, Liliane Umutesi, Claude Bernard Uwizeye*. Technical assistance: Nico Kalisvaart, Eveline Klinkenberg*, Ikushi Onozaki, Peou Satha. * Principal investigator xii Acknowledgements List of contributors who provided international technical assistance, by institutional affiliation Academic Medical Center, University of Amsterdam Frank Cobelens, Martien Borgdorff. Korean Institute of Tuberculosis, Republic of Korea Sang Jae Kim. KNCV Tuberculosis Foundation Agnes Genhard, Nico Kalisvaart, Eveline Klinkenberg, Frank van Leth, Masja Straetemans, Edine Tiemersma. Landsteiner Institute, Medical Center Haaglanden, The Hague, The Netherlands Marleen Vree. London School of Hygiene and Tropical Medicine, UK Sian Floyd. The National Centre for TB and Leprosy Control, Cambodia Peou Satha. PharmAccess, United Republic of Tanzania Jan van den Hombergh, John Mayanda. Research Institute of Tuberculosis, Japan Anti-Tuberculosis Association (RIT/JATA), Japan Susumu Hirao, Kunihiko Ito, Kiyohiko Izumi, Satoshi Mitarai, Hiroko Matsumoto, Kosuke Okada, Tetsuhiro Sugamoto, Norio Yamada. Soutien Pneumologique International, France Etienne Leroy-Terquem, Pierre L’Her. Supranational reference laboratory for TB, San Raffaele Scientific Institute, Milan, Italy Daniella Cirillo. Zimbabwe National survey team: Joconiah Chirenda, Patrick Hazangwe, Martin Mapfurira, Eve Marima, Ronnie Matambo, Ellen Munemo, Junior Mutsvangwa*, Hebert Mutunzi, Dumisani Ndlovu, Mkhokeli Ngwenya, Charles Sandy*, Peter Shiri, Nicholas Siziba. Technical assistance: Mourad Gumusboga, Kunihiko Ito, Fasil Tsegaye Kassa, Wilfred Nkhoma, Kosuke Okada, Ikushi Onozaki, Marina Tadolini, Hazim Timimi, Norio Yamada. * Principal investigator xiiiAcknowledgements Supranational reference laboratory for TB, Institute of Tropical Medicine, Antwerp, Belgium Mourad Gumusboga, Armand Van Deun. UNICEF, Democratic People’s Republic of Korea Mubeen Aslam. US Centers for Disease Control and Prevention Emily Bloss, Martien Borgdorff, Kevin Cain, J. Sean Cavanaugh, Shua Chai, Julia Ershova, Eugene McCray, Patrick Moonan, Sara Whitehead. World Health Organization, headquarters Katherine Floyd, Philippe Glaziou, Sayori Kobayashi, Irwin Law, Ikushi Onozaki, Narayan Pendse (STOP-TB), Charalambos Sismanidis, Hazim Timimi. World Health Organization, Regional Office for Africa (AFRO) Wilfred Nkhoma. World Health Organization, Eastern Mediterranean Regional Office (EMRO) Amal Basilli. World Health Organization, South-East Asia Regional Office (SEARO) Partha Pratim Mandal, Mukta Sharma. World Health Organization, Western Pacific Regional Office (WPRO) Philippe Glaziou. World Health Organization, Country Office for Indonesia M. Bintari Dwihardiani. World Health Organization, Country Office for Myanmar Eva Nathanson. Independent consultants Zeleke Alebachew, Ethiopia; M.N. Farid, Indonesia; Bimbo Fasan, Nigeria; Yasunori Ichimura, Japan; Marina Tadolini, Italy; Sabira Tahseen, Pakistan; Etienne Leroy Terquiem, France; Fasil Tsegaye, Ethiopia. xiv xv Abbreviations AFB acid-fast bacilli BCG Bacille Calmette-Guérin C&NCD Communicable and Noncommunicable Disease Administration CAD computer aided detection CDC Centers for Disease Control and Prevention CI confidence interval CXR chest X-ray DDR direct digital radiography DOTS directly observed treatment, short course FATA Federally Administered Tribal Areas FIND Foundation for Innovative New Diagnostics FM fluorescence microscopy GCP good clinical practice GDMP good data management practices GFC global focus countries Global Fund Global Fund to Fight AIDS, Tuberculosis and Malaria HBC high TB burden country HEPA high efficiency particulate air HIV human immunodeficiency virus IGRA interferon gamma release assay JATA Japan Anti-Tuberculosis Association JICA Japanese International Cooperation Agency k coefficient of between cluster variation LED light-emitting diode LJ Löwenstein–Jensen media LPA line-probe assay MDG Millennium Development Goal MDR-TB multidrug-resistant tuberculosis MGIT mycobacteria growth indicator tube MMR mass miniature radiography MoH ministry of health MRCG Medical Research Council Unit, The Gambia MTB Mycobacterium tuberculosis NTM nontuberculous mycobacteria NTLP National Tuberculosis and Leprosy Program NTP national tuberculosis programme NTRL national TB reference laboratory ODPC Office of Disease Prevention and Control xvi Abbreviations P:N prevalence:notification PIN personal identification number PNB para-nitrobenzoic acid PPS probability proportional to size RIF rifampin RIT Research Institute of Tuberculosis SDG Sustainable Development Goal SOP standard operating procedure SRL supranational reference laboratory TB tuberculosis TBMU TB management unit UI uncertainty interval UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS US United States WHO World Health Organization ZIMSTAT Zimbabwe National Statistics Agency ZN Ziehl-Neelsen 1PART I An overview of the 25 surveys implemented 2007–2016 2Photo credit: Yasunori Ichimura Examining chest X-rays in the field during the 2010–2011 national TB prevalance survey of Ethiopia. 3This book is about national surveys of the prevalence of tuberculosis (TB) disease that were completed between 2007 and 2016. During this 10-year period there was an unprecedented national, regional and global effort to implement such surveys. Particular attention was given to a group of 22 global focus countries (GFCs) in Asia and Africa that, in 2007, were selected by the World Health Organization (WHO) Global Task Force on TB Impact Measurement (1). Between 2007 and the end of 2016, 24 countries implemented a total of 25 national TB prevalence surveys1 using methods recommended by WHO (2); the 24 countries comprised 18 of the 22 GFCs, and six other countries. The book documents the survey methods used, results and their implications, successes achieved, challenges faced, and lessons learned for future surveys. It ends with a discussion of prevalence surveys post-2016. This opening chapter explains the rationale for conducting national TB prevalence surveys; provides a historical overview of where, when and how they were implemented in the years up to 2007; and describes why prevalence surveys in 22 GFCs became one of three strategic areas of work pursued by the WHO Global Task Force on TB Impact Measurement between 2007 and 2015. It ends with a summary of the 25 national TB prevalence surveys that were completed between 2007 and 2016, which are the subject of the rest of the book. 1.1 Why are national TB prevalence surveys needed? Dr Robert Koch announced that he had discovered Mycobacterium tuberculosis (M. tuberculosis) as the cause of TB on 24 March 1882, an event now marked annually as World TB Day (3). At that time, TB was one of the leading causes of death in European countries, with cause-of-death data from national vital registration (VR) systems showing mortality rates of over 100 per 100 000 population per year. National laws that made reporting Chapter 1 Introduction of TB cases compulsory were introduced in these and various other industrializing (and now high-income) countries. In combination, these national notification and VR systems have allowed the burden of TB disease (in terms of numbers of cases and deaths each year) to be reliably monitored using routine health information systems for several decades, with a few time series covering a span of more than 100 years (Fig. 1.1). The ultimate goal is that all countries can reliably track their TB epidemics using national notification and VR systems. However, although all countries have national notification systems for TB, and report notification data to WHO on an annual basis (4), the number of notified cases each year is generally not a good proxy for the actual number of new cases. This is due to (1) underreporting, especially in countries with large private sectors or in which people with TB seek care in public facilities that are not linked to the national TB programme (NTP) and associated reporting systems; and (2) underdiagnosis, especially in countries where there are geographic or financial barriers to seeking health care. In the early 2000s, national VR systems of high quality and coverage had not been established in many parts of the world (including most countries with a high burden of TB), and there was limited progress between 2000 and 2012 (5). In the absence of national notification and VR systems of high quality and coverage, national population-based surveys of the prevalence of TB disease provide an alternative way of measuring the burden of TB disease. Such surveys allow direct measurement of the number of TB cases in the population at a given point in time, and the distribution of cases by age and sex. Repeat surveys allow assessment of trends, and of the impact of interventions to reduce the burden of disease in the period since the last survey. Other benefits of surveys include documentation of health care seeking behaviour in the public and private sectors; identification of reasons why people with TB were not diagnosed before the survey or officially reported 1 The country that implemented two surveys was the Philippines. National TB prevalence surveys 2007–20164 to national authorities (or both); and development or improvement of strategies and interventions for TB case finding, diagnosis and treatment. 1.2 The first wave of national TB prevalence surveys: 1953–1960 The first national TB prevalence surveys were implemented in the 1950s, in East Asia and Africa (Fig. 1.2).1 The first survey was implemented in Japan in 1953, followed by a repeat survey in 1958 (12-14). These surveys used mass miniature radiography (MMR) as an initial screening tool for pulmonary TB; diagnosis was based on smear microscopy and culture. MMR was developed in 1936, and free-of-charge MMR was one of 12 interventions for TB control recommended by the first World Health Assembly in July 1948 (15). Results from the first survey in Japan alarmed national authorities by revealing a high prevalence of radiologically active and bacteriologically confirmed cases (3.4% and 0.75% of the VR: vital registration. Sources: Public Health England (2017) (6), The Research Institute of Tuberculosis/JATA (2018) (7), National Institute for Public Health and the Environment, Ministry of Health, Welfare and Sport (2016) (8) and Centers for Disease Control and Prevention (9). Fig. 1.1 Trends in TB incidence (solid line) and TB mortality (dashed line) based on data from national notification and national VR systems, selected countries population, respectively). They also showed that most cases (79%) had not been diagnosed before the survey and were in those aged 30 years or over. As a result, a systematic programme of MMR screening, previously restricted to those under the age of 30 years, was expanded to cover the whole population. Registration and case- holding systems were also introduced. WHO implemented a series of surveys in 12 African countries between 1955 and 1960 (16). Of these, 11 were national surveys: Basutoland (Lesotho), Bechuanaland (Botswana), Gambia, Ghana, Liberia, Nigeria, Sierra Leone, Swaziland (Eswatini), Tanganyika (United Republic of Tanzania), Uganda and Zanzibar. A survey was also implemented in Kenya, excluding Nairobi and the country’s northern province. Survey investigations were based on a technical guide published by WHO (17), and included tuberculin skin tests, chest X-rays and sputum examination by direct microscopy. In five countries, X-ray examination was not possible because the necessary equipment was not available. Culture examinations were generally done for all smear-positive specimens and a random sample of smear-negative specimens. The estimated prevalence of bacteriologically 1 A full historical listing of all surveys implemented up to 2012 in Asia is provided in Onozaki et al. (2015) (10). For surveys in Africa, a listing is provided in WHO (2007) (11). PART I: An overview of the 25 surveys implemented 2007–2016 5 confirmed TB was 1.5%, with men (2%) having a higher prevalence than women (0.7%); however, there were proportionally more smear-positive cases in women (0.4%) than in men (0.5%). Pakistan conducted a survey in 1959, but from the available information it is not clear whether this was a national or subnational survey (18). In 1955–1959, India carried out subnational TB prevalence surveys (19). 1.3 National TB prevalence surveys in Asia in the 1960s and 1970s In the 1960s and 1970s, 14 countries implemented a total of 19 national TB prevalence surveys (Fig. 1.3). Multiple surveys were implemented in Japan (three), the Republic of Korea (three) and Thailand (two).1 After surveys in 1953 and 1958, repeat surveys were implemented in Japan in 1963, 1968 and 1973. These showed a rapid reduction in the number of cases, with an annual decline estimated at about 10% per year up to the late 1970s – one of the fastest national declines in TB disease burden ever recorded (Fig. 1.1). 1 A full historical listing of all surveys implemented up to 2012 in Asia is provided in Onozaki et al. (2015) (10). For surveys in Africa, a listing is provided in WHO (2007) (11). Grey, not applicable. Fig. 1.2 Countries that implemented a national TB prevalence survey, 1953–1960 Elsewhere in the WHO Western Pacific Region, national surveys were implemented in China (1979), Malaysia (1970), the Republic of Korea (1965, 1970 and 1975), Samoa (1975) and Singapore (1975). Outside the WHO Western Pacific Region, surveys were conducted in Bangladesh (1964), Indonesia (1979‒1982), Iraq (1970), Libya (1976), Myanmar (1972), Netherlands (1970), Sri Lanka (1970) and Thailand (1960‒1964 and 1977). Most of these surveys used MMR for screening, and sputum and culture examination for diagnosis. Several subnational surveys in south India were also implemented under the leadership of the National Tuberculosis Institute in Bangalore from 1961 to 1968 (20). 1.4 The 1980s and 1990s: A period of few national TB prevalence surveys Few national TB prevalence surveys were implemented in the 1980s and 1990s; six countries implemented a total of 11 surveys (Fig. 1.4).1 In Asia, the series of surveys at 5-year intervals that started in the Republic of Korea in 1965 was continued, National TB prevalence surveys 2007–20166 Fig. 1.3 Countries that implemented a national TB prevalence survey in the 1960s and 1970s with a further four surveys completed in 1980, 1985, 1990 and 1995. The survey in 1995 was the last to be carried out in the country. The other five countries in Asia that implemented surveys were Bangladesh (1987), China (1984 and 1990), Pakistan (1987), Philippines (1981 and 1997) and Thailand (1991–1992). Reasons for the relatively small number of surveys included: • In 1974, WHO recommended that MMR should not be used for TB case finding (21). This recommendation affected investments in mobile radiographic equipment, and NTPs prioritized bacille Calmette-Guérin (BCG) vaccination and the diagnosis and treatment of people seeking medical care over systematic screening programmes in the general population. • There were declines in disease burden in countries that had previously implemented surveys. In countries such as Japan, the Republic of Korea, Singapore and Sri Lanka, these declines meant that increasingly large sample sizes would be required, which were prohibitive for logistical and cost reasons. • In some countries, increased urbanization and improved living conditions led to reduced willingness among the general population to participate in surveys. For example, in the Republic of Korea, where seven national surveys were conducted every 5 years from 1965 to 1995, urban participants progressively increased from 34% (1965) to 74% (1995) of the total, and survey participation rates declined from 96% (1965) to 87% (1995) (22, 23). Similarly, in the five national surveys in Japan from 1953 to 1973, participation declined from 99% to 89% (24). • In a growing number of countries, the quality and coverage of routine TB surveillance data improved. These data provided most if not all of the information needed to monitor the TB epidemic and inform TB policy, strategy and planning. 1.5 The early to mid 2000s: Few national TB prevalence surveys, but growing interest in them From 2000 to 2006, six surveys that used a variety of methods were implemented (Table 1.1). Grey, not applicable. PART I: An overview of the 25 surveys implemented 2007–2016 7 Only two of these surveys (China, 2000 and Cambodia, 2002) used both smear microscopy and culture for diagnostic testing, and achieved a sufficiently high participation rate for results to be nationally representative. The surveys in Eritrea (2004) and Indonesia (2004) used smear microscopy only (25); the survey in Malaysia (2003) had a very low participation rate in urban areas; and the survey in Thailand collected sputum samples only from those who reported symptoms, owing to delays in reading MMRs. Recognizing the value of updated guidance and greater standardization in survey methods, the WHO Regional Office for the Western Pacific took the initiative to develop a handbook on national TB prevalence surveys (11); this later became known as the red book. Recommendations drew heavily on the 2002 survey in Cambodia. Although the number of national TB prevalence surveys implemented during this period was small, there was growing interest in them. This occurred in the context of a series of developments that started in the early to mid 1990s: • The World Health Assembly agreed the first-ever targets for global TB control in 1991 (26). The targets were to detect 70% of new smear-positive Fig. 1.4 Countries that implemented a national TB prevalence survey in the 1980s and 1990s pulmonary TB cases and to successfully treat 85% of these cases by 2000, a target date that was later reset to 2005 (27). The first indicator became commonly known as the case detection rate. The numerator was the annual number of new cases of smear-positive TB notified to national authorities in a year, and the denominator was an estimate of the incidence of smear-positive TB (i.e. the number of new cases of smear-positive TB) in the same year. • WHO declared TB a global health emergency in 1993 (28). • In 1994, WHO published a framework for effective TB control (29). This was subsequently branded as the DOTS strategy, which was WHO’s recommended approach to TB control until the end of 2005.1 The DOTS strategy had five components,2 1 The DOTS strategy was succeeded by the Stop TB Strategy (30) in 2006 (which included an updated version of DOTS as its first component) and by the End TB Strategy in 2016. 2 The five components of the DOTS strategy were; political commitment; diagnosis by quality-assured sputum smear microscopy; standardized short-course chemotherapy with direct observation of treatment (DOT); a regular and uninterrupted supply of high-quality anti-TB drugs; and a standardized system for recording and reporting of cases and their treatment outcomes. Grey, not applicable. National TB prevalence surveys 2007–20168 and its main aim was to achieve the global targets of a 70% case detection rate and an 85% treatment success rate for smear-positive pulmonary TB cases. By 2000, almost all WHO Member States had adopted the DOTS strategy (31). • In 1999, TB was declared a crisis in the WHO Western Pacific Region. In response, WHO established the Stop TB Special Project, which aimed to halve 2000 levels of TB prevalence and mortality by 2010. This was the first time that a regional target for TB prevalence had been set (32). • The United Nations (UN) Millennium Development Goals (MDGs) were adopted by all UN Member States in 2000 (33). Targets were defined for each of the eight MDGs. One of the targets under MDG 6 was that the TB incidence rate (new cases per 100  000 population per year) should be declining by 2015. The MDG framework also included four other TB indicators: prevalence per 100 000 population, the mortality rate, the case detection rate and the treatment success rate. • In the context of the MDGs, regional targets set in the WHO Western Pacific Region and a resolution passed at a summit of the Group of Eight (G8) countries in Okinawa, Japan, the Stop TB Partnership set global targets to halve TB prevalence and mortality (per 100 000 population) by 2015 compared with levels in 1990 (34). This was the first time that a global target for TB prevalence was set. Table 1.1 National TB prevalence surveys implemented in 2000–2006 CXR: chest X-ray; MMR: mass miniature radiography. Country Year Specific survey characteristics China 2000 The last national survey to include all age groups (>3 months of age); fluoroscopy used for screening; smear microscopy and culture used for diagnostic testing. Cambodia 2002 Screening done using an interview about symptoms and direct CXR with onsite full-size film development using portable equipment; smear microscopy and culture used for diagnostic testing. Malaysia 2003 Symptom screening done at home and CXR screening at a health facility; smear microscopy and culture used for diagnostic testing; survey results were not usable owing to a low participation rate in urban areas. Eritrea 2004 Sputum smear specimens taken from all participants but no CXR screening; smear microscopy used for diagnostic testing (no culture). Indonesia 2004 Implemented as part of a national health demographic survey. Sputum specimens taken from any participant reporting a productive cough of any duration; no CXR screening; smear microscopy used for diagnostic testing (no culture). Thailand 2006 An interview about symptoms and MMR were used for screening. Survey results were not usable because sputum samples were not collected from participants who had an abnormal MMR but did not report symptoms, due to delays in reading MMRs and providing feedback about results. From 2000 until 2015, national, regional and global efforts in TB control focused on achievement of the targets set by the World Health Assembly, the UN and the Stop TB Partnership. Up to 2005, the greatest attention was given to the World Health Assembly targets of a 70% case detection and 85% treatment success rate. There was considerable interest in estimates of TB incidence, because it was the denominator of the first target. After a series of consultations, the first estimates of TB incidence produced by WHO for the national, regional and global levels were for 1997 (35). Subsequently, WHO published updated estimates annually in its global TB report. Given that notification data in many countries were not a good proxy for TB incidence (owing to underreporting and underdiagnosis), and in the frequent absence of other direct measurements of TB disease (e.g. prevalence surveys or cause-of-death data from national VR systems), these estimates relied heavily on two things: expert opinion about the gap between notifications and the true level of TB incidence, and tuberculin survey data. National authorities, including ministries of health and their NTPs whose performance in making progress towards the World Health Assembly targets was being regularly assessed and reported, became increasingly interested in improving the evidence available to inform estimates of TB incidence, in particular through the implementation of a national TB prevalence survey. This interest was reinforced by growing evidence and PART I: An overview of the 25 surveys implemented 2007–2016 9 consensus that methods used to estimate TB incidence from tuberculin survey data were problematic (36, 37); the inclusion of TB prevalence as an MDG indicator; the setting of regional and global targets for reductions in TB prevalence; and the launch of the Stop TB Strategy. 1.6 The decade 2007–2016: a period of unprecedented national, regional and global efforts to implement national TB prevalence surveys In 2006, the DOTS strategy was succeeded by the Stop TB Strategy (30) which, in line with the MDGs, had an end date of 2015. The new strategy had three targets for 2015, all of which were related to reductions in TB disease burden. The targets were that TB incidence should be falling (in line with the TB target under MDG 6), and that 1990 levels of TB prevalence and mortality (per 100 000 population) should be halved by 2015 (thus incorporating the targets that had been set by the Stop TB Partnership for the MDG indicators of prevalence and mortality). CDR: case detection rate; HIV: human immunodeficiency virus; NTP: national TB programme. When the criteria were applied in December 2007, the sources of data used were WHO (2007) (41), the WHO global TB database and UNAIDS/WHO (2006) (42). Table 1.2 The four groups of criteria used to identify countries in which national surveys of the prevalence of TB disease could be justified in the period up to 2015 From 2006 to 2015, efforts in TB control at national, regional and global levels were focused on achieving the three “impact” targets of the Stop TB Strategy. In June 2006, WHO established a Global Task Force on TB Impact Measurement (1). The main aim of the Task Force was to ensure that assessment of whether the 2015 targets were achieved at global, regional and national levels was robust, rigorous and consensus-based (38). After its first meeting in 2006, which focused on a review of available methods to estimate TB disease burden (39), the second meeting in December 2007 was used to discuss and reach agreement on strategic areas of work to be pursued by the Task Force between 2008 and 2015. Three strategic areas of work were defined: strengthening of routine surveillance systems (notification and VR) in all countries; implementing national TB prevalence surveys in 22 GFCs; and periodic review of the methods used to translate surveillance and survey data into estimates of TB disease burden (40). The 22 GFCs for national TB prevalence surveys (13 in Africa and 9 in Asia) were selected based on four major Group of criteria Explanation Group 1 → 1. Estimated prevalence of smear-positive TB ≥100 per 100 000 population and 2. Accounts for ≥1% of the estimated total number of smear-positive TB cases globally and 3. CDR for smear-positive TB ≤50% or >100% • Major contribution to the global burden of TB disease. Sample size small enough to make surveys feasible in terms of cost and logistics. • Excludes countries whose contribution to the global burden of TB disease is insignificant for the purposes of global and regional assessments of burden and impact. • CDR ≤50% or >100% indicates weak reporting systems and problematic TB estimates, respectively. Group 2 → 1. Estimated prevalence of smear-positive TB ≥70 per 100 000 population and 2. Accounts for ≥1% of the estimated total number of smear-positive TB cases globally and 3. Estimated HIV prevalence rate in the adult population (15–49 years) ≥1% • Less stringent criteria for the TB prevalence rate, but incorporates countries with high HIV prevalence and therefore where there is potential for a rapid increase in TB incidence and prevalence rates. Group 3 → 1. Estimated prevalence of smear-positive TB ≥200 per 100 000 population and 2. Accounts for ≥0.5% of the estimated total number of smear-positive TB cases globally • Less stringent criteria for a country’s contribution to the global burden of TB disease, but incorporates countries with particularly high TB prevalence per 100 000 population. Group 4 → 1. Nationwide survey implemented between 2000 and 2007 or 2. Nationwide survey planned before 2010 • Prior survey data allow monitoring of trends. • High motivation of NTP to conduct a survey. National TB prevalence surveys 2007–201610 groups of criteria (Table 1.2) and are shown in Table 1.3 (2). A total of 53 countries met one of the four groups of criteria shown in Table 1.2. There were two major reasons for selecting a subset of 22 GFCs. The first was that providing the necessary technical support to all of the 53 countries would be challenging if not impossible, given the relatively limited expertise at that time in the design and implementation of prevalence surveys at both global and country levels. The second was that, in combination, the GFCs accounted for a substantial share 1 The other two WHO regions – the European Region and the Region of the Americas – already had relatively strong notification and VR systems. GFC: global focus country; WHO: World Health Organization. a “High burden” refers to the 22 high TB burden countries (HBCs) that were defined by WHO for the period 1998–2015. The 22 HBCs were the countries that ranked first to 22nd in terms of their estimated number of incident cases of TB per year. In 2015, WHO reviewed and updated the definition and a list of 30 HBCs was defined for the period 2016–2020. Region and country Criteria met (group number as defined in Table 1.2) High-burden country?a Data from baseline survey conducted between around 1990 and 2008? WHO African Region Ethiopia 1,3 Yes No Ghana 1,2 No No Kenya 2,4 Yes No Malawi 1,2,3,4 No No Mali 1,2,3,4 No No Mozambique 1,2,3 Yes No Nigeria 1,2,3,4 Yes No Rwanda 1,2,3 No No Sierra Leone 1,2,3 No No South Africa 2,3 Yes No Uganda 1,2,3,4 Yes No United Republic of Tanzania 1,2,3,4 Yes No Zambia 2,3 No No WHO Eastern Mediterranean Region Pakistan 1,4 Yes Yes (1987) WHO South-East Asia Region Bangladesh 4 Yes Yes (2008–2009) Indonesia 4 Yes Yes (2004) Myanmar 4 Yes Yes (1994) Thailand 2,4 Yes Yes (1991, 2006) WHO Western Pacific Region Cambodia 2,3 Yes Yes (2002) China 4 Yes Yes (1990, 2000) Philippines 4 Yes Yes (1981, 1997, 2007) Viet Nam 4 Yes Yes (2007) Table 1.3 The 22 GFCs for TB prevalence surveys selected by the WHO Global Task Force on TB Impact Measurement of the estimated number of TB cases in each of the four WHO regions where routine surveillance systems were weakest (i.e. the African, Eastern Mediterranean, South- East Asia and Western Pacific regions).1 From the beginning of 2008, substantial efforts were made to design and implement national TB prevalence surveys, and to analyse and report results. At the global level, these efforts were led and coordinated under the PART I: An overview of the 25 surveys implemented 2007–2016 11 umbrella of the WHO Global Task Force on TB Impact Measurement, and more specifically by a Task Force subgroup on national TB prevalence surveys that was led by WHO staff in the Global Tuberculosis Programme’s TB monitoring and evaluation unit. Examples of key actions, activities and products of the Task Force subgroup on national TB prevalence surveys included: • sending high-level letters from WHO to the ministers of health of each GFC, to explain why the country had been selected as a priority for a national TB prevalence survey and to offer guidance and support from the Task Force; • organization of multicountry workshops for protocol development; • development of updated guidance on standardized methods for undertaking national TB prevalence surveys in the form of a handbook, which became known as the lime book (2). This was used as the foundation of all national TB prevalence surveys implemented from 2010 to 2016 and was produced as a major collaborative effort among technical Grey, not applicable. GFC: global focus country; WHO: World Health Organization. Fig. 1.5 The 24 countries that implemented a national TB prevalence survey in 2007–2016 and that are the subject of this book Of the 13 GFCs in Africa, nine completed a survey (blue); all nine GFCs in Asia completed at least one survey (blue). Six other countries completed a survey but were not GFCs (red). agencies, financial partners and lead investigators involved in surveys implemented in the 1990s and 2000s, with a total of 50 co-authors; • expert reviews of protocols, using a checklist based on guidance provided in the lime book (2); • organization of multicountry workshops hosted by countries that had recently launched survey field operations (Ethiopia in October 2010, Cambodia in July 2011 and Ghana in May 2013), to enable a mixture of support for all countries combined with an opportunity to observe a survey at first hand; • organization of study tours to countries where surveys were being implemented for countries that were in the preparation phase; and • coordination and provision of technical assistance to all countries, with an emphasis on Asia–Asia, Asia–Africa and Africa–Africa (AA) collaboration. This global effort reinforced and supported the considerable interest in and commitment to implementing a national TB prevalence survey that had been growing and intensifying in many countries during the early 2000s (including in those outside the list of 22 GFCs). National TB prevalence surveys 2007–201612 GFC: global focus country. Country Year of survey GFC? Bangladesh 2015–2016 Yes Cambodia 2010–2011 Yes China 2010 Yes Democratic People’s Republic of Korea 2016 No Ethiopia 2010–2011 Yes Gambia 2012 No Ghana 2013 Yes Indonesia 2013–2014 Yes Kenya 2015–2016 Yes Lao People’s Democratic Republic 2010–2011 No Malawi 2013–2014 Yes Mongolia 2014–2015 No Myanmar 2009–2010 Yes Nigeria 2012 Yes Pakistan 2011 Yes Philippines 2007 Yes Philippines 2016 Yes Rwanda 2012 Yes Sudan 2013–2014 No Thailand 2012–2013 Yes Uganda 2014–2015 Yes United Republic of Tanzania 2012 Yes Viet Nam 2007 Yes Zambia 2013–2014 Yes Zimbabwe 2014 No Table 1.4 The 25 national TB prevalence surveys implemented in 2007– 2016, which are the subject of this book 1 Further details about survey budgets and sources of funding are provided in Chapter 2 (see in particular Table 2.5). 2 The only survey that relied on domestic funding alone was the one in China. 3 There was one survey implemented during this period that did not use the screening algorithm recommended in the lime book. This was a 2008 survey in Bangladesh, which took sputum samples from the entire eligible population (without screening based on symptoms and chest X-ray). Results appeared to considerably understate the true burden of TB disease, likely probably due to challenges in processing large numbers of samples; this was confirmed by results from the 2015 survey. 4 The four GFCs that had not implemented a survey by the end of 2016 were Mali, Mozambique, Sierra Leone and South Africa. Mozambique implemented a survey in 2018–2019 and South Africa implemented a survey in 2017–2019. implemented in 24 countries in 2007–2016 according to methods set out in the lime book are shown in Table 1.4 and Fig. 1.5;3 they included 18 of the 22 GFCs.4 For context, Fig. 1.6 shows a timeline of all surveys conducted between the 1950s and 2016. This book provides comprehensive documentation of the 25 national TB prevalence surveys implemented in 2007–2016 that are listed in Table 1.4. Part I includes cross-cutting chapters: methods (Chapter 2); results and their implications (Chapter 3); successes, challenges and lessons learned (Chapter 4); and a discussion of prevalence surveys post-2016 (Chapter 5). Part II contains 25 country-specific profiles, which provide details about each survey in a standardized format. The book represents the collective effort and contribution of more than 450 people, with leadership and coordination provided by WHO. References 1. WHO Global Task Force on TB Impact Measurement [website]. Geneva: World Health Organization; 2019 (https://www.who.int/ tb/areas-of-work/monitoring-evaluation/impact_measurement_ taskforce/en/, accessed 12 June 2019). 2. Tuberculosis prevalence surveys: a handbook. Geneva: World Health Organization; 2011 (https://apps.who.int/iris/bitstream/ handle/10665/44481/9789241548168_eng.pdf, accessed 22 November 2019). 3. Sakula A. Robert Koch: centenary of the discovery of the tubercle bacillus, 1882. Thorax. 1982;37(4):246–51 (https://www.ncbi.nlm. nih.gov/pubmed/6180494, accessed 22 November 2019). 4. Global tuberculosis report. Geneva: World Health Organization; 2018 (https://apps.who.int/iris/bitstream/hand le/10665/274453/9789241565646-eng.pdf?ua=1, accessed 22 November 2019). 5. Mikkelsen L, Phillips DE, AbouZahr C, Setel PW, de Savigny D, Lozano R et al. A global assessment of civil registration and vital statistics systems: monitoring data quality and progress. Lancet. 2015;386(10001):1395–406 (https://www.ncbi.nlm.nih.gov/ pubmed/25971218, accessed 22 November 2019). The creation of a new source of financing in 2002, in the form of the Global Fund to Fight AIDS, Tuberculosis and Malaria (the Global Fund), helped to turn this interest and commitment into reality. The Global Fund helped to finance 22 surveys implemented between 2007 and 2016.1 Domestic resources contributed to funding for 12 surveys,2 the United States (US) government contributed to funding for nine surveys, and other donors contributed to funding for 13 surveys. All of this interest and commitment culminated in a decade of unprecedented global, regional and national efforts to implement national TB prevalence surveys, with particular attention to 22 GFCs. The 25 surveys that were PART I: An overview of the 25 surveys implemented 2007–2016 13 Fig. 1.6 National surveys of the prevalence of TB disease, 1950–2016 DPR Korea: Democratic People’s Republic of Korea; Lao PDR: Lao People’s Democratic Republic; UR Tanzania: United Republic of Tanzania. a The survey listed is the one in Bangladesh in 2015, which is featured in this book. There was also a survey in 2008, but this did not use the screening and diagnostic algorithm recommended in the lime book and for this reason is not counted in the total of 25 surveys. 6. Public Health England. Historical TB notification data tables to end December 2017 (Table no. 1). 2018 (https://www.gov.uk/ government/publications/tuberculosis-tb-annual-notifications- 1913-onwards, accessed 19 June 2019). 7. Statistics of TB [website]. The Tuberculosis Surveillance Center; 2018 (https://www.jata.or.jp/rit/ekigaku/en/statistics-of-tb/, accessed 19 July 2019). 8. National Institute for Public Health and the Environment; Ministry of Health WaS. Curvekaart tuberculose in Nederland 1901– 2015. Netherlands: 2016 (https://www.rivm.nl/documenten/ curvekaart-tuberculose-in-nederland1901-2015, accessed 24 July 2019). 9. Centers for Disease Control and Prevention (CDC). Reported tuberculosis in the United States, 2017. Atlanta, GA: US Department of Health and Human Services, CDC; 2018 (https:// www.cdc.gov/nchs/nvss/mortality_historical_data.htm, accessed 19 June 2019). 10. Onozaki I, Law I, Sismanidis C, Zignol M, Glaziou P, Floyd K. National tuberculosis prevalence surveys in Asia, 1990–2012: an overview of results and lessons learned. Trop Med Int Health. 2015;20(9):1128–45 (https://www.ncbi.nlm.nih.gov/ pubmed/25943163, accessed 22 November 2019). 11. World Health Organization. Assessing tuberculosis prevalence through population-based surveys. Geneva, Switzerland 2007 (https://apps.who.int/iris/handle/10665/206962, accessed 20 November 2019). 12. Yamaguchi M. Survey of tuberculosis prevalence in Japan, 1953. Bull World Health Organ. 1955;13(6):1041–73 (https://www.ncbi. nlm.nih.gov/pubmed/13284568, accessed 22 November 2019). National TB prevalence surveys 2007–201614 13. Yamaguchi M, Oka H, Kumabe H, Yosano H. Survey of tuberculosis prevalence in Japan, 1954: trends in tuberculosis from 1953 to 1954. Bull World Health Organ. 1959;21:145–59 (https://www.ncbi.nlm.nih.gov/pubmed/13846395, accessed 22 November 2019). 14. Shimao T. [Surveys of tuberculosis prevalence]. Bull Int Union Tuberc. 1982;57(2):127–33 (https://www.ncbi.nlm.nih.gov/ pubmed/6983377, accessed 22 November 2019). 15. First world health assembly. Draft resolution on tuberculosis proposed by the delegation of Czechoslovakia. Geneva: World Health Organization; 1948 (https://apps.who.int/iris/ bitstream/10665/98189/1/WHA1_Prog-29_eng.pdf, accessed 6 May 2019). 16. Roelsgaard E, Iversen E, Blocher C. Tuberculosis in tropical Africa. An epidemiological study. Bull World Health Organ. 1964;30:459– 518 (https://www.ncbi.nlm.nih.gov/pubmed/14178027, accessed 22 November 2019). 17. Technical guide for tuberculosis survey team. WHO/TUB/ Technical Guide/1, Geneva: World Health Organization; 1958 (https://apps.who.int/iris/bitstream/10665/75224/1/WHO_TUB_ Techn.Guide_1_1958_eng.pdf, accessed 22 November 2019). 18. Dolin P. WHO report of a visit to Pakistan. December 5–19. Geneva: World Health Organization; 1997. 19. Indian Council of Medical Research. Tuberculosis Sub- Committee. Tuberculosis in India: a sample survey, 1955–58. New Delhi:1959. 20. National Tuberculosis Institute Bangalore. Tuberculosis in a rural population of South India: a five-year epidemiological study. Bull World Health Organ. 1974;51(5):473–88 (https://www.ncbi.nlm. nih.gov/pubmed/4549498, accessed 22 November 2019). 21. WHO expert committee on tuberculosis: ninth report. WHO Technical Report Series, No. 552, Geneva: World Health Organization; 1974 (https://apps.who.int/iris/ bitstream/10665/41095/1/WHO_TRS_552_eng.pdf, accessed 19 May 2019). 22. Hong YP, Kim SJ, Lew WJ, Lee EK, Han YC. The seventh nationwide tuberculosis prevalence survey in Korea, 1995. Int J Tuberc Lung Dis. 1998;2(1):27–36 (https://www.ncbi.nlm.nih. gov/pubmed/9562108, accessed 22 November 2019). 23. Ministry of Health and Social Affairs, Korean National Tuberculosis Association. Report on the tuberculosis prevalence survey in Korea. 1965. 24. Shimao T. [Tuberculosis prevalence survey in Japan]. Kekkaku. 2009;84(11):713–20 (https://www.ncbi.nlm.nih.gov/ pubmed/19999593, accessed 22 November 2019). 25. Sebhatu M, Kiflom B, Seyoum M, Kassim N, Negash T, Tesfazion A et al. Determining the burden of tuberculosis in Eritrea: a new approach. Bull World Health Organ. 2007;85(8):593–9 (https://www.ncbi.nlm.nih.gov/pubmed/17768517, accessed 22 November 2019). 26. Resolution WHA44.8: tuberculosis control programme, 3rd ed, Handbook of resolutions and decisions of the World Health Assembly and the Executive Board, Geneva, World Health Organization. 1993 (https://www.who.int/tb/publications/ tbresolution_wha44_8_1991.pdf?ua=1, accessed 22 November 2019). 27. Stop tuberculosis initiative. Report by the director-general. Fifty- third world health assembly, Geneva: World Health Organization; 2000 (https://apps.who.int/iris/handle/10665/260193, accessed 22 November 2019). 28. TB: a global emergency, WHO report on the TB epidemic (WHO/ TB/94.177) [website]. Geneva 1994 (https://apps.who.int/iris/ handle/10665/58749, accessed 21 June 2019). 29. Framework for effective TB control. WHO/TB/1994.179, Geneva: World Health Organization; 1994 (https://apps.who.int/iris/ handle/10665/58717, accessed 22 November 2019). 30. Raviglione MC, Uplekar MW. WHO’s new Stop TB Strategy. Lancet. 2006;367(9514):952–5 (https://www.ncbi.nlm.nih.gov/ pubmed/16546550, accessed 22 November 2019). 31. Global tuberculosis report. Geneva: World Health Organization; 2014 (https://apps.who.int/iris/ bitstream/10665/137094/1/9789241564809_eng.pdf ?ua=1, accessed 22 November 2019). 32. van Maaren P, Tomas B, Glaziou P, Kasai T, Ahn D. Reaching the global tuberculosis control targets in the Western Pacific Region. Bull World Health Organ. 2007;85(5):360–3 (https://www.ncbi. nlm.nih.gov/pubmed/17639220, accessed 22 November 2019). 33. Millennium development goals [website]. (https://www.un.org/ millenniumgoals/, accessed 3 May 2019). 34. Global tuberculosis report. Geneva: World Health Organization; 2006 (https://apps.who.int/iris/bitstream/ handle/10665/144567/9241563141_eng.pdf, accessed 12 June 2019). 35. Dye C, Scheele S, Dolin P, Pathania V, Raviglione MC. Consensus statement. Global burden of tuberculosis: estimated incidence, prevalence, and mortality by country. WHO Global Surveillance and Monitoring Project. JAMA. 1999;282(7):677–86 (https:// www.ncbi.nlm.nih.gov/pubmed/10517722, accessed 22 November 2019). 36. van Leth F, van der Werf MJ, Borgdorff MW. Prevalence of tuberculous infection and incidence of tuberculosis: a re-assessment of the Styblo rule. Bull World Health Organ. 2008;86(1):20–6 (https://www.ncbi.nlm.nih.gov/pubmed/18235886, accessed 22 November 2019). 37. Styblo K. The relationship between the risk of tuberculous infection and the risk of developing infectious tuberculosis. Bull Int Union Tuberc Lung Dis. 1985;60:117–9. 38. Global tuberculosis report. Geneva: World Health Organization; 2015 (https://apps.who.int/iris/bitstream/handle/10665/191102/ 9789241565059_eng.pdf, accessed 12 June 2019). 39. Dye C, Bassili A, Bierrenbach AL, Broekmans JF, Chadha VK, Glaziou P et al. Measuring tuberculosis burden, trends, and the impact of control programmes. Lancet Infect Dis. 2008;8(4):233– 43 (https://www.ncbi.nlm.nih.gov/pubmed/18201929, accessed 22 November 2019). 40. TB impact measurement policy and recommendations for how to assess the epidemiological burden of TB and the impact of TB control. Stop TB policy paper no. 2, Geneva: World Health Organization; 2009 (https://apps.who.int/iris/bitstream/ handle/10665/44231/9789241598828_eng.pdf, accessed 22 November 2019). 41. Global tuberculosis control: surveillance, planning, financing  : WHO report 2007. Geneva: World Health Organization; 2007 (https://apps.who.int/iris/handle/10665/43629, accessed 22 November 2019). 42. UNAIDS, World Health Organization. 2006 report on the global AIDS epidemic. Geneva: UNAIDS and World Health Organization; 2006 (https://data.unaids.org/pub/report/2006/2006_gr_en.pdf, accessed 22 November 2019). 15 Chapter 2 Methods Background Guidance on national surveys of the prevalence of TB disease was published by WHO in 2007 (1) and 2011 (2), with the 2011 edition becoming known as the lime book. The 2011 handbook was a major collaborative effort of technical agencies, financial partners and lead investigators involved in surveys implemented in the 1990s and 2000s, and had a total of 50 co-authors. Examples of important changes in the 2011 edition were a definitive recommendation on which screening strategy to use (as opposed to the first edition, which provided four options); improved guidance on sampling design, data management and analysis, and reporting of survey results; a new chapter on repeat surveys; and more country case studies from recent surveys to illustrate what the guidance meant in practice. The national surveys featured in this book are the 25 surveys implemented between 2007 and 2016 that followed the methods set out in the lime book.1 This chapter provides an overview of the key methods used, structured according to 13 major topics: survey objectives, eligibility criteria, definition of a prevalent survey TB case, screening and diagnostic testing strategies, sampling design, field operations, additional testing for HIV infection and drug susceptibility, central- level activities, data management and analysis, additional studies, reporting and dissemination of results, ethics approval, and budgeting and financing. 1 Only one national survey implemented over the period 2007–2016 was not included. This was the 2008 survey in Bangladesh, which did not use the screening strategy recommended in the lime book. Instead, sputum samples were taken from all individuals considered eligible based on age and residency. 2.1 Survey objectives The primary objective of all national TB prevalence surveys implemented in the period 2007‒2016 was to estimate the burden of disease caused by TB; specifically, the national prevalence of smear-positive and bacteriologically positive pulmonary TB among the general population aged 15 years and above.2 Over time, increasing emphasis was given to the prevalence of bacteriologically confirmed TB, especially following a 2013 update to WHO-recommended case definitions, and associated recording and reporting of cases (3). Other objectives included measuring trends in the burden of disease caused by TB (e.g. the surveys in Cambodia, China, Myanmar and the Philippines, which were repeat surveys); and to use survey results, alongside an in-depth analysis of surveillance and programmatic data, as the basis for a comprehensive update of estimates of disease burden (incidence and mortality as well as prevalence). Most surveys implemented in the period 2007–2016 also collected data on the health care seeking behaviour of symptomatic participants and TB cases, to assess whether care had been sought and, if so, where (e.g. in the public or private health care sectors, and in which types of facilities or services). Some surveys further investigated reasons why cases were not diagnosed before the survey, and the extent to which people with TB were being treated by health care providers that were not linked to the NTP. All surveys collected data on those who were on (or had a past history of) anti-TB treatment at the time of the survey, and on the type of health facility in which treatment was provided. These data were used to evaluate case finding and care policies, as well as the performance of routine TB surveillance. 2 The Philippines used a 10-year-old eligibility threshold for its 2007 survey. National TB prevalence surveys 2007–201616 2.2 Eligibility criteria Eligibility to participate in a national TB prevalence survey was based on two criteria: age and residency (Table 2.1). Of the 25 surveys implemented in the period 2007–2016, 24 used the age criterion recommended in the lime book (i.e. individuals aged ≥15 years).1 The exception was the 2007 survey in the Philippines (implemented before the publication of WHO guidance on national TB prevalence 1 Diagnosis of TB among children is difficult with the diagnostic tools used in prevalence surveys. For example, it is hard for children to produce sputum samples (especially given the paucibacillary nature of TB in children) and chest X-rays are not suitable for use in healthy children with a low risk of TB disease. A further problem is the larger sample size needed to estimate the number of cases among children. Country Year Age of eligibility (years) Residency criteria Bangladesh 2015‒2016 ≥15 Lived in the cluster for ≥2 weeks before the census Cambodia 2010‒2011 ≥15 Lived in the household for ≥2 weeks before the census China 2010 ≥15 Lived in the household for ≥6 months before the census DPR Korea 2015‒2016 ≥15 Registered in the living administrative unit for ≥2 weeks before the census Ethiopia 2010‒2011 ≥15 Permanent residents who stayed in the household for ≥1 night in the 14 days before the census, and temporary visitors who stayed in the household for ≥14 days before the census Gambia 2011‒2013 ≥15 Residents who spent ≥1 night in the household in the 4 weeks before the census day; visitors who arrived in the household ≥4 weeks before the census Ghana 2013 ≥15 Permanent residents who lived in the household for ≥1 day in the past 14 days, or visitors who lived in the household for ≥7 days in the past 14 days Indonesia 2013‒2014 ≥15 Lived in the household for ≥1 month before the census Kenya 2015‒2016 ≥15 Lived in the selected cluster for ≥30 days before the census Lao PDR 2010‒2012 ≥15 Slept in the household for 14 days before the census Malawi 2013‒2014 ≥15 Slept in the household for ≥14 days before the census Mongolia 2014‒2015 ≥15 Slept in the household for 14 days before the census Myanmar 2009‒2010 ≥15 Lived in the household for ≥2 weeks before the census Nigeria 2012 ≥15 Slept in the household for ≥14 days before the census Pakistan 2010‒2011 ≥15 Slept in the household the night before the census Philippines 2007 ≥10 No residency criteria Philippines 2016 ≥15 Lived in the household for ≥2 weeks before the census Rwanda 2012 ≥15 Lived in the household for ≥1 month before the interview Sudan 2013‒2014 ≥15 Household members resident in the selected household for the past 6 months, and visitors who spent ≥3 weeks in the household before the census Thailand 2012‒2013 ≥15 Permanent residents based on household registration, or temporary residents or nonresidents who had slept in the household for ≥2 weeks before the census Uganda 2014‒2015 ≥15 Permanent residents who stayed ≥1 night in the household in the past 2 weeks; temporary visitors who arrived ≥2 weeks before census UR Tanzania 2011‒2012 ≥15 Slept in the household for 2 weeks before the census Viet Nam 2006‒2007 ≥15 Lived in the household for ≥3 months before the census Zambia 2013‒2014 ≥15 Slept in the household 24 hours before the census Zimbabwe 2014 ≥15 Permanent residents who had slept ≥1 night in the 14 days before the census; non- residents who had slept in the household for ≥14 days before the census DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. surveys), in which all those aged 10 years and above were considered eligible. In most surveys, a resident was defined as someone who had lived in the household for Table 2.1 Eligibility criteria to participate in a national TB prevalence survey, 2007‒2016 PART I: An overview of the 25 surveys implemented 2007–2016 17 at least 2‒4 weeks at the time of the survey census.1 The exceptions were the surveys in Pakistan and Zambia (which classified a resident as someone who had slept in the household the night before the census), and those in China (6 months residency), Sudan (6 months residency) and Viet Nam (3 months residency). 2.3 Definition of a prevalent survey TB case In 2007, a prevalent case of TB was defined as follows: • a definite case of smear-positive pulmonary TB: at least one specimen acid-fast bacilli (AFB) positive by smear microscopy and culture-positive for M. tuberculosis; • a probable case of smear-positive pulmonary TB: at least one specimen AFB positive by smear microscopy and chest X-ray consistent with TB disease according to the reading by the central radiology team, and culture-negative or not available; and • a case of smear-negative culture-positive pulmo- nary TB: two smear-negative slides and culture- positive for M. tuberculosis. Prevalent survey cases of smear-positive TB and smear-negative culture-positive TB were both classified as bacteriologically confirmed TB (1). In December 2010, WHO endorsed the rapid molecular test Xpert®MTB/RIF for the simultaneous diagnosis of TB and rifampicin-resistant TB (4), and in 2013, WHO reviewed its recommended routine case definitions for TB and issued an update (3). In the context of prevalence surveys documented in this book, bacteriologically confirmed TB was defined as a positive culture and/or positive Xpert MTB/RIF result for M. tuberculosis. Smear was not used to define a definite case of TB; rather, it was used to disaggregate cases according to their smear status. Smear-positive TB was defined as a bacteriologically confirmed case (by culture and/or Xpert MTB/RIF) with at least one AFB-positive smear result. Careful review of laboratory and chest X-ray results by a diagnostic panel, before finalizing the list of survey TB cases, was a standard part of national TB prevalence surveys. This was done not only to ensure the quality of survey results, but also for clinical management of those with positive laboratory results (not all of whom were eventually considered to have TB disease or were included 1 The aim of residency criteria is to exclude individuals who intentionally move into the household in anticipation of receiving health care from the survey team, thus potentially biasing results. in the list of prevalent TB cases). Participants with missing culture or Xpert results (or both) would require use of other evidence (e.g. smear and chest X-ray results) for them to be defined as a TB case. Instances of misdiagnosis or overdiagnosis could arise through data management errors, cross-contamination in the laboratory or false- positive laboratory results. In particular, participants with a single positive bacteriological result but no other supportive evidence of TB disease required special attention. For example, survey participants with a single scanty culture-positive result (i.e. <5 or 10 M. tuberculosis colonies on solid media) or a positive Xpert MTB/RIF result from a centrifuged sediment were not categorized as prevalent TB cases unless there was chest X-ray evidence of TB disease. Once the final list of survey cases was available, two categories were defined for the purposes of analysis and presentation of results: smear-positive pulmonary TB and bacteriologically confirmed pulmonary TB. Given the diagnostic technologies currently available and the logistics of population-based surveys, prevalence surveys focus on the measurement of active pulmonary TB disease in adults. Surveys cannot be used to directly measure the prevalence of extrapulmonary disease in adults or the prevalence of TB disease in children. 2.4 Screening and diagnostic testing strategies The screening and diagnostic testing strategies used in surveys implemented in the period 2007‒2016 are summarized in Table 2.2. 2.4.1 Screening Most surveys used two screening tools: an interview about TB symptoms and chest X-ray. Generally, individuals with symptoms that met screening criteria and/or a chest X-ray showing any lung shadow or findings suggestive of TB were considered eligible for sputum examination. Participants that screened negative on both interview and chest X-ray were categorized as not eligible for sputum examination, and were therefore assumed not to have TB. The main symptom screening criterion was a chronic cough (i.e. ≥2 weeks in most surveys), since this has been the primary screening criterion for TB in routine health services. Nine countries used cough ≥2 weeks as the only symptom screening criterion. The symptom screening criteria in seven other countries was cough ≥2 weeks or haemoptysis, or both. A few surveys considered individuals to be screen positive if they reported a history National TB prevalence surveys 2007–201618 Table 2.2 Screening methods used in national TB prevalence surveys, 2007‒2016 Country Symptom screening Radiography screening Other screening criteria Bangladesh Scoring system: eligible if the total score was ≥3 a Direct digital Chest X-ray exempted a Cambodia Cough for ≥2 weeks or haemoptysis (or both) Conventional Chest X-ray exempted China Cough for ≥2 weeks or haemoptysis of any duration (or both) Conventional Participants with known active pulmonary TB with normal chest X-ray, and those who were chest X-ray exempted DPR Korea Cough for ≥15 days or haemoptysis (or both) Conventional None Ethiopia Cough for ≥2 weeks Conventional Participants who were exempt from or declined chest X-ray but met one of the following criteria: weight loss ≥3 kg in the past month, night sweats ≥2 weeks, fever ≥2 weeks or contact with a TB patient in the past year Gambia Cough for ≥2 weeks, or cough <2 weeks with ≥2 other symptoms, or no cough with ≥3 other symptoms: chest pain, night sweats, shortness of breath, loss of appetite, weight loss, fever or haemoptysis Direct digital Chest X-ray exempted Ghana Cough for ≥2 weeks Direct digital Chest X-ray exempted Indonesia Cough for ≥2 weeks or haemoptysis (or both) Direct digital Chest X-ray exempted but had at least one of the following symptoms: cough, haemoptysis, fever, chest pain, night sweats, loss of appetite or shortness of breath Kenya Cough for ≥2 weeks Direct digital Chest X-ray exempted Lao PDR Cough for ≥2 weeks within the past month or haemoptysis within the past month (or both) Conventional None Malawi Any of the following symptoms for at least 1 week: cough, sputum production, haemoptysis, chest pain, weight loss, night sweats, fatigue, fever or shortness of breath Conventional None Mongolia Cough for ≥2 weeks Direct digital Chest X-ray exempted Myanmar Cough for ≥3 weeks or haemoptysis (or both) Conventional Chest X-ray exempted Nigeria Cough for ≥2 weeks Computed radiography None Pakistan Cough for ≥2 weeks, or cough of any duration if there was no available chest X-ray result Direct digital Participants on TB treatment at the time of the survey Philippines (2007) N/A b Conventional None Philippines (2016) Cough for ≥2 weeks or haemoptysis (or both) Direct digital Chest X-ray exempted Rwanda Cough of any duration Direct digital Chest X-ray exempted Sudan Cough for ≥2 weeks Direct digital Chest X-ray exempted or a participant was currently on TB treatment Thailand Scoring system: eligible if the total score was ≥3 c Direct digital Chest X-ray exempted c Uganda Cough for ≥2 weeks Conventional Chest X-ray exempted UR Tanzania Any of the following symptoms: cough for ≥2 weeks, haemoptysis, fever for ≥2 weeks, weight loss or excessive night sweats Computed radiography None PART I: An overview of the 25 surveys implemented 2007–2016 19 of TB, even in the absence of symptoms and chest X-ray abnormalities. In nine surveys, the sensitivity of symptom screening was increased by broadening criteria, including combinations of cough of any duration, loss of body weight, chest pain, night sweats and fever. In the 2007 survey in the Philippines, eligibility for sputum submission was based only on chest X-ray screening. In Bangladesh and Thailand, a points-scoring system based on reported symptoms was used.1 Since professional reading by a radiologist was not possible in most field sites, reporting of any chest X-ray abnormalities (especially lung abnormalities that were consistent with TB) was encouraged in all surveys, to increase the sensitivity of screening. In 17 surveys, participants who declined chest X-ray investigation or were exempt from having a chest X-ray were automatically eligible for sputum submission; in three countries (Ethiopia, Indonesia and Pakistan) this only applied if the participant reported symptoms. 1 In Bangladesh and Thailand, a participant with a symptom score of 3 points or more was eligible for sputum submission: a cough of ≥2 weeks or more (3 points), a cough of less than 2 weeks (1 point), haemoptysis in the past month (3 points), weight loss in the past month (1 point) fever of ≥1 week in the past month (1 point) and night sweats in the past month (1 point). In Thailand, if participants did not have a chest X-ray, then a score of 1 or more made them eligible for sputum submission. 2 Given the required dose of radiation and the lower quality of images that are produced, WHO does not recommend either MMR or fluoroscopy. Table 2.2 Continued Country Symptom screening Radiography screening Other screening criteria Viet Nam Productive cough for ≥2 weeks Digital scan onsite and mass miniature radiography (70x70 mm) Chest X-ray exempted or currently on anti-TB treatment or history of TB in preceding 2 years Zambia Any one of the following symptoms for ≥2 weeks: cough, fever or chest pain Direct digital None Zimbabwe Any one of the following symptoms: cough of any duration, haemoptysis in the past 12 months or drenching night sweats Direct digital Chest X-ray exempted DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a In Bangladesh, a participant was eligible if their total score was 3 points or more: cough ≥2 weeks (3 points), cough <2 weeks (1 point), haemoptysis in the past month (3 points), weight loss in the past month (1 point), fever ≥1 week in the past month (1 point) and night sweats in the past month (1 point). If the chest X-ray was exempted, then a clinical score of 1 or 2 classified a participant as symptom-screen positive. b In the Philippines, symptom screening was not used as a selection criterion for sputum submission, but an interview about TB symptoms and TB history was done for participants aged 20 years or more. c In Thailand, a participant was eligible if their total score was 3 or more (or ≥1 with chest X-ray exempted): cough for ≥2 weeks (3 points), haemoptysis over the past month (3 points), cough <2 weeks (2 points), weight loss in the past month (1 point), fever ≥1 week within the past 2 weeks (1 point) and night sweats in the past month (1 point). A single posterior–anterior (PA) image by radiography was used for chest X-ray screening in all surveys.2 From 2007, there was a transition from conventional to digital chest X-ray imaging systems. In nine surveys, film images were developed using an automatic film processor (a standard practice in surveys before 2007), but other surveys deployed digital X-ray systems. Advantages of digital systems included no requirement for removal of chemicals; immediate availability of the images for chest X-ray reading in the field; more efficient transmission of images to a central unit; and simpler image archiving and retrieval. Computer radiography with an imaging plate and image reader was used in the surveys in Nigeria and the United Republic of Tanzania, whereas direct digital radiography (DDR) with a flat panel detector subsequently became the standard technology for other countries. Depending on the accessibility of cluster sites and available funding, countries selected a variety of chest X-ray delivery options, including X-ray vans, X-ray containers loaded on a truck or portable X-ray units, or combinations of these options. Computer-aided detection for reading chest X-ray images was tested in the context of national TB prevalence surveys. However, as of the end of 2016, their performance was not considered satisfactory, especially for diagnosis (5). National TB prevalence surveys 2007–201620 1 Ghana used both LJ and MGIT, but only MGIT was used to define a survey TB case. 2 Owing to laboratory capacity constraints, two sputum samples were obtained from participants in 52 clusters (33%), and one sample from the remaining 104 clusters. 2.4.2 Diagnostic testing Sputum specimens were collected from all participants who screened positive according to the screening strategy described in Section 2.4.1. In general, two smear examinations and two culture examinations (or at least one culture examination when laboratory capacity was limited) were undertaken for each participant. From 2013, in all but one country (Sudan), Xpert MTB/RIF was systematically used in addition to culture, rather than as a substitute for culture (Table 2.3). Direct Ziehl-Neelsen (ZN) light microscopy was the standard technology used for smear examinations in most surveys, consistent with its use in routine clinical practice. However, following WHO’s 2011 recommendation to use light-emitting diode (LED) microscopy, this method was used in some of the later surveys (6). In both cases, the direct smear method was used in most surveys (20/25), in preference to smear from concentrated sediment (interpretation of results from smears using the concentrated method was a challenge in Ghana and Malawi owing to possible cross-contamination while making the smears and inoculating culture). In most surveys, at least two sputum specimens were examined (the exception was the Philippines in 2016), and this was usually done in one or more designated laboratories. In Pakistan, Rwanda and the United Republic of Tanzania, testing was carried out onsite or at the nearest hospital laboratory. In two repeat surveys in Asia (those in Cambodia and the Philippines), a simple primary culture method (i.e. Ogawa-Kudoh method), without centrifugation of specimens, was used to enable direct comparisons with previous survey results. Most other countries used the more sensitive concentration method with Löwenstein- Jensen (LJ media), in line with the latest WHO recommendations. Only Gambia, Ghana,1 Zambia and Zimbabwe had sufficient resources to use liquid culture – that is, mycobacteria growth indicator tube (MGIT) – for primary culture. The use of a second culture increased the number of positive results by almost 20%, suggesting that the testing of only one culture was a limitation of the surveys in Ethiopia, Kenya, Indonesia,2 Pakistan, the United Republic of Tanzania and Viet Nam. For identification of M.  tuberculosis complex, rapid immunochromatographic assays (strip tests for speciation) to identify cultured isolates are recommended. These assays provide a definitive identification of all types of M. tuberculosis. Capilia or SD MPT64TB Ag kits were used in several surveys. However, biochemical testing such as niacin production, nitrate reduction and growth on p-nitrobenzoic acid  were used in settings where national TB reference laboratories had not yet introduced rapid identification tests. Where Xpert MTB/RIF (or line probe assays [LPA]) were used to systematically confirm cases – that is, in Bangladesh, Ghana, Indonesia, Kenya, Malawi, Mongolia, Pakistan, the Philippines (2016), Sudan (LPA), Uganda, Zambia and Zimbabwe3 – a large proportion of Xpert-negative or LPA-negative individuals was observed among those with positive AFB microscopy results (and negative culture results or no culture results) (Table 3.7). Hence, confirmatory testing of smear-positive specimens using Xpert  MTB/RIF (or LPA) was encouraged. Xpert MTB/RIF was used in all screen-positive individuals (in addition to culture) in the surveys in Bangladesh, Kenya and the Philippines (2016). Given the risk of DNA cross-contamination, Xpert MTB/RIF testing of a direct sputum specimen was recommended (testing of a concentrated sputum sediment was only recommended as a confirmatory test for smear positivity, in place of culture). 2.4.3 Screening methods in repeat prevalence surveys Repeat surveys are typically undertaken every 7–10 years. During that time interval, screening and diagnostic practices can change with the adoption of improved techniques and technologies. Therefore, differences in screening and diagnostic methods between consecutive surveys can potentially generate biases that need to be accounted for when interpreting results. In Thailand, the 2012 survey used digital chest X-ray as opposed to the less sensitive MMR used in the 1991 and 2006 surveys. In China, the 1990 survey used chest fluoroscopy for symptomatic individuals, and sputum samples were only taken if this test was abnormal. In contrast, in the 2000 and 2010 surveys in China, participants with symptoms but normal fluoroscopic examination (2000) or normal chest radiography (2010) were also asked to submit sputum samples. The 1994 survey in Myanmar and the 2004 survey in Indonesia did not systematically perform 3 Pakistan was the first country to use Xpert MTB/RIF in a national TB prevalence survey, but it was only used for specimens that were smear positive with undetermined culture results. PART I: An overview of the 25 surveys implemented 2007–2016 21 Ta bl e 2. 3 Di ag no st ic m et ho ds u se d in n at io na l T B pr ev al en ce s ur ve ys , 2 00 7‒ 20 16 Co un tr y Sm ea r Pr im ar y cu ltu re Xp er t M TB /R IF M TB id en tifi ca tio n te st fo r po si tiv e cu ltu re s H IV te st in g Dr ug s us ce pt ib ili ty te st in g N um be r of sa m pl es Ty pe N um be r of sa m pl es Ty pe B an gl ad es h 2 D ir ec t FM 2 C on ce nt ra te d LJ Ye s, f or a ll pa rt ic ip an ts w ho sc re en ed p os it iv e C ap ili a N o Ye s C am bo di a 2 D ir ec t FM 2 D ir ec t O ga w a N o C ap ili a N o Ye s C hi na 3 D ir ec t ZN 2 D ir ec t LJ N o P N B N o Ye s D P R K or ea 2 C on ce nt ra te d FM 2 C on ce nt ra te d LJ N o M P T6 4 N o N o E th io pi a 2 D ir ec t FM 1 C on ce nt ra te d LJ N o C ap ili a N o O nl y as a p os t- su rv ey a ct iv it y G am bi a 2 D ir ec t FM 2 C on ce nt ra te d M G IT N o a M G IT ™ T B c Id en ti fic at io n Te st N o O nl y as a p os t- su rv ey a ct iv it y a G ha na 2 C on ce nt ra te d ZN 2 C on ce nt ra te d M G IT b Ye s, f or sm ea r- po si ti ve sp ec im en s, a nd if c ul tu re s w er e co nt am in at ed P N B , ca pi lia N o Ye s c In do ne si a 2 D ir ec t ZN 2 s am pl es fo r 5 2 cl us te rs , 1 sa m pl e fo r 1 0 4 c lu st er s C on ce nt ra te d LJ Ye s, f or sm ea r- po si ti ve sp ec im en s or no n- co nc lu si ve cu lt ur es M P T6 4 , ni ac in N o N o K en ya 2 D ir ec t FM 2 C on ce nt ra te d LJ Ye s, f or a ll pa rt ic ip an ts w ho sc re en ed p os it iv e M P T6 4 N o d Ye s La o P D R 2 D ir ec t ZN 2 D ir ec t O ga w a N o P N B , LP A N o Ye s M al aw i 2 C on ce nt ra te d FM 2 C on ce nt ra te d LJ Ye s, f or sm ea r- po si ti ve sp ec im en s, a nd if c ul tu re s w er e co nt am in at ed C ap ili a N o e Ye s c M on go lia 2 D ir ec t FM 2 D ir ec t O ga w a Ye s, f or sm ea r- po si ti ve sp ec im en s P N B , ni ac in N o Ye s M ya nm ar 2 D ir ec t FM 2 D ir ec t O ga w a N o N ia ci n, P N B , ca pi lia N o N o N ig er ia 2 D ir ec t ZN 2 C on ce nt ra te d LJ N o M P T6 4 N o N o National TB prevalence surveys 2007–201622 Ta bl e 2. 3 Co nt in ue d Co un tr y Sm ea r Pr im ar y cu ltu re Xp er t M TB /R IF M TB id en tifi ca tio n te st fo r po si tiv e cu ltu re s H IV te st in g Dr ug s us ce pt ib ili ty te st in g N um be r of sa m pl es Ty pe N um be r of sa m pl es Ty pe P ak is ta n 2 ( 1 o ns it e, 1 a t ce nt ra l) D ir ec t ZN 1 D ir ec t O ga w a Ye s, f or sm ea r- po si ti ve sp ec im en s w it ho ut c ul tu re co nfi rm at io n P N B , M P B 6 4 : al l c ul tu re - po si ti ve s pe ci m en s, a nd LP A o r Xp er t M TB /R IF us ed w it h sm ea r- po si ti ve sp ec im en s w it ho ut c ul tu re co nfi rm at io n. N o Ye s P hi lip pi ne s (2 0 0 7 ) 3 D ir ec t FM 3 C on ce nt ra te d LJ an d di re ct O ga w a N o N ia ci n, c at al as e, n it ra te te st in g N o Ye s P hi lip pi ne s (2 0 1 6 ) 1 f D ir ec t FM 1 f D ir ec t O ga w a Ye s, f or a ll pa rt ic ip an ts w ho sc re en ed p os it iv e M P T6 4 N o Ye s R w an da 2 D ir ec t FM 2 C on ce nt ra te d LJ N o M P T6 4 Ye s Ye s S ud an 2 D ir ec t FM 2 D ir ec t O ga w a N o C ap ili a, L PA g N o N o Th ai la nd 2 D ir ec t ZN 2 D ir ec t O ga w a N o h S im pl e im m un oc hr om at o- gr ap hi c as sa y N o N o U ga nd a 2 D ir ec t ZN 2 C on ce nt ra te d LJ Ye s, f or sm ea r- po si ti ve sp ec im en s, a nd if c ul tu re s w er e co nt am in at ed M P T6 4 Ye s O nl y as a p os t- su rv ey a ct iv it y U R T an za ni a 3 ( 2 o n si te , 1 a t ce nt ra l) D ir ec t FM 1 C on ce nt ra te d LJ Ye s i P N B Ye s O nl y as a p os t- su rv ey a ct iv it y Vi et N am 3 D ir ec t ZN 1 C on ce nt ra te d LJ N o N ia ci n N o A ll po si ti ve is ol at es w er e te st ed b ut re su lt s w er e no t of fic ia lly r ep or te d Za m bi a 2 C on ce nt ra te d FM 2 C on ce nt ra te d M G IT Ye s, f or a ll pa rt ic ip an ts w ho sc re en ed p os it iv e j C ap ili a Ye s N o Zi m ba bw e 2 C on ce nt ra te d FM 2 C on ce nt ra te d LJ an d M G IT Ye s, f or a ll sm ea r- po si ti ve sp ec im en s k M P T6 4 N o d Ye s c D P R K or ea , D em oc ra ti c P eo pl e’ s R ep ub lic o f K or ea ; FM , flu or es ce nc e m ic ro sc op y; H IV , hu m an i m m un od efi ci en cy v ir us ; La o P D R , La o P eo pl e’ s D em oc ra ti c R ep ub lic ; LJ , Lö w en st ei n- Je ns en m ed ia ; LP A , lin e pr ob e as sa ys ; M G IT , m yc ob ac te ri al g ro w th in di ca to r tu be ; M TB , M yc ob ac te riu m t ub er cu lo si s; P N B , pa ra -n it ro be nz oi c ac id ; TB , tu be rc ul os is ; U R T an za ni a, U ni te d R ep ub lic o f Ta nz an ia ; ZN , Zi eh l- N ee ls en s ta in . a In G am bi a, X pe rt M TB /R IF w as u se d to d et er m in e if s ur ve y ca se s w er e ri fa m pi ci n- re si st an t bu t no t as p ar t of t he s ur ve y. b In G ha na , co nc en tr at ed L J an d M G IT w er e bo th u se d in t he s ur ve y, b ut o nl y M G IT w as u se d to d efi ne a T B s ur ve y ca se . c In G ha na , M al aw i a nd Z im ba bw e, r if am pi ci n re si st an ce w as d et ec te d us in g Xp er t M TB /R IF o nl y. d In K en ya a nd Z im ba bw e, T B c as es d et ec te d by t he s ur ve y w er e of fe re d H IV c ou ns el lin g an d te st in g as p ar t of r ou ti ne t re at m en t m an ag em en t bu t w er e no t di re ct ly t es te d as p ar t of t he s ur ve y. e In M al aw i, in st ea d of H IV t es ti ng , al l p ar ti ci pa nt s w er e as ke d if t he y ha d ev er b ee n te st ed f or H IV a nd , if w ill in g, t o di sc lo se t he ir s ta tu s. f In t he P hi lip pi ne s (2 0 1 6 ), if t he fi rs t sa m pl e w as n ot o f su ffi ci en t vo lu m e, a s ec on d sa m pl e w as a ls o us ed . PART I: An overview of the 25 surveys implemented 2007–2016 23 culture examination, but culture was used in the repeat surveys in 2009 and in 2014, respectively. In Myanmar, the 1994 survey did not include chest X-ray screening but it was used in the 2009−2010 survey. In Cambodia, the 2002 and 2011 surveys used similar screening and diagnostic methods. In the Philippines, the 1997 and 2007 surveys used only chest X-ray for screening. Although the 2016 Philippines survey used Xpert MTB/RIF for diagnostic confirmation, comparisons with the 2007 survey results could still be made because the same culture method (Ogawa) was used in both surveys. 2.5 Sampling design A comprehensive description of the recommended sampling design is outlined in Chapter 5 of the lime book (2). 2.5.1 Sample size Until the advent of rapid molecular tests (in particular, Xpert MTB/RIF in 2010), smear examination was the main test used for TB diagnosis in most countries. From the mid-1990s until the mid-2000s, routine reporting of notified cases of smear-positive pulmonary TB and their treatment outcomes was a core component of WHO’s recommended global TB strategy, global TB monitoring undertaken by WHO and national TB surveillance systems. Hence, up to 2015, sample size calculations were based on the expected national prevalence of smear- positive pulmonary TB among adults. The expected prevalence was generally based on the assumption of a prevalence to notification ratio of 2:1. For repeat surveys, the sample size calculation was based on the expected decline in the prevalence of smear-positive pulmonary TB since the previous survey (7). After 2015, following WHO’s 2013 update to TB case definitions, sample size was calculated based on the expected prevalence of bacteriologically confirmed pulmonary TB in adults. Table 2.3 Continued g In Sudan, LPA was used to test all culture-positive and all smear-positive samples. h In Thailand, Xpert MTB/RIF was used after the study for quality assurance purposes for smear-positive, culture-negative samples. i In UR Tanzania, Xpert MTB/RIF was only used on smear-positive slides to confirm the presence of MTB at the Antwerp SRL, but was not part of the original protocol. j In Zambia, Xpert MTB/RIF was also performed on some smear-negative culture contaminated samples or smear-negative culture indeterminate samples if the chest X-ray was suggestive of TB. k In Zimbabwe, in addition to smear-positive samples all culture-positive samples were also tested for rifampicin resistance using Xpert MTB/RIF. The calculated samples sizes for surveys implemented in the period 2007‒2016 are shown in Table 2.4. They ranged from 30 000 in the Philippines (2007) to 264 000 in China (2010). Of note, Indonesia and Mongolia aimed to obtain subnational estimates, and Thailand’s survey was designed as two independent surveys: one for the Bangkok area and another for areas outside Bangkok. 2.5.2 Cluster number and size Both logistical and statistical issues are relevant when determining the number and size of clusters to be sampled. At least 50 clusters are strongly recommended, as a compromise between minimizing sampling design effects (which requires more and smaller clusters) and reducing logistical constraints (by having fewer clusters). All surveys implemented in the period 2007‒2016 had 50 or more clusters (Table 2.4). Cluster sizes of 400‒800 were generally recommended, because this size makes it possible to complete chest X-ray screening within 7‒10 days. Most surveys had a cluster size of 500‒900 people, apart from those in China (1500 people), Pakistan (1400) and Viet Nam (1500). The introduction of high-capacity direct digital chest X-ray units made it feasible to screen 250‒300 people per day, thus enabling completion of field operations in each cluster in fewer than 5 working days. 2.5.3 Stratification Most surveys used stratified designs to increase sampling efficiency, such as urban versus rural strata, or geographically defined strata (Table 2.4). Probability proportional to size (PPS) sampling was applied to the selection of primary sampling units (regions, states, zones or provinces), followed by smaller secondary sampling units (districts, townships, subdistricts and municipalities), and so on until reaching the level of geographical area that comprised the population size of a cluster. The last stage of cluster selection sometimes used simple random sampling. 2.5.4 Sampling frame The sampling frame defines the areas of the country from which clusters are selected. Ideally, all clusters should be included in the sampling frame to ensure optimal national representativeness. However, certain areas were excluded in several surveys because of security concerns or geographic inaccessibility. Excluded areas generally covered less than 5% of the total population (Table 2.4). In several surveys, some clusters that were initially National TB prevalence surveys 2007–201624 Table 2.4 Sampling and survey design, 2007‒2016 Country Planned sample size Planned number of clusters Cluster size Stratified sampling Geographical areas excluded initially from sampling frame Geographical areas excluded during field operations Bangladesh 100 000 125 800 Urban, rural None One cluster was replaced for security reasons Cambodia 39 680 62 640 Urban, rural, others None None China 264 000 176 1500 Urban, rural None None DPR Korea 69 442 100 700 Urban, rural None Five clusters in Anpyon, Kyongsong and Pukchang county were replaced by five others in the same counties due to inaccessibility Ethiopia 46 514 85 550 Urban, rural, pastoralist 37 out of 810 woredas (3% of the national population) were excluded from the sampling frame for security reasons and due to logistical challenges; two clusters (kebele) were replaced before field operations started due to logistical challenges None Gambia 55 281 80 700 Not stratified None Three clusters were replaced due to a large uninhabited area in the urban area around the capital (one cluster), military installations and areas around the president’s residence (two clusters) Ghana 63 905 98 650 Urban, rural None None Indonesia 78 000 156 500 Sumatra, Java-Bali and others, with each stratified into urban/rural None None Kenya 72 000 100 720 Urban, rural None One cluster in Mandera was excluded for security reasons Lao PDR 40 000 50 800 Not stratified None None Malawi 37 200 74 500 Urban, semi- urban, rural None None Mongolia 49 000 98 600 (city) / 500 (other) City, provincial centre, rural None None Myanmar 49 690 70 710 Region, state 32 townships were excluded for security reasons Five townships (Bokepyin, Kunlon, Kyarinnseikkyi, Mindat and Nattalin) were replaced by others within the same township during the pre-visit, owing to security and transportation problems, and an insufficient population aged 15 years and above PART I: An overview of the 25 surveys implemented 2007–2016 25 Table 2.4 Continued Country Planned sample size Planned number of clusters Cluster size Stratified sampling Geographical areas excluded initially from sampling frame Geographical areas excluded during field operations Nigeria 49 000 70 700 Six zones None Three clusters in the states of Borno and Yobe were excluded for security reasons; these were replaced in the states of Adamawa, Bauchi and Gombe Pakistan 133 000 95 1400 Not stratified The Federally Administered Tribal Areas, district Dera Bugti in Balochistan and 17 tehsils of Khyber Pakhtunkhwa were excluded for security reasons; this accounted for 6.4% of the national population Three clusters from Balochistan (Awaran, Lehri and Quetta) were replaced by other clusters (Hub in Balochistan, Khan Pur in Punjab, and Sharda in Azad- Jammu and Kashmir) for security reasons Philippines (2007) 30 000 50 600 Metro Manila, other urban, rural Four barangays in other urban strata and 14 barangays in rural strata were excluded for security reasons and due to inaccessibility None Philippines (2016) 54 000 108 500 National Capital Region, region 3 and 4A; Rest of Luzon; Visayas; Mindanao Before field operations started, one cluster in Basilan province was excluded for security reasons (this accounted for <1% of the national population) Three clusters (Maco, Madaya and Sipangkot barangays) were replaced for security reasons, and one cluster (Holy Spirit barangay) was dropped because the local authorities refused house-to-house mobilization and interviews Rwanda 44 500 73 610 Not stratified None None Sudan 91 131 114 800 Urban, rural None Four clusters (two in Darfur State, one in Gazira and one in South Kordofan) were cancelled for security reasons, and one was removed due to a protocol violation Thailand a 74 700 83 900 Urban, rural None None Uganda 40 180 70 580 Urban, rural None None UR Tanzania 46 792 62 750 Urban, semi- urban, rural, Zanzibar None None Viet Nam 105 000 70 1500 Urban, rural, remote None None Zambia 54 400 66 825 Urban, rural None None Zimbabwe 44 951 75 600 Urban, rural None Two clusters (Chiredzi and Macheke) were replaced due to logistical issues (e.g. weather, equipment failure) DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The Bangkok metropolitan area was excluded. National TB prevalence surveys 2007–201626 selected were excluded after the sampling stage had been completed, owing to security concerns or natural disasters. 2.6 Field operations The main activities conducted during field operations include a census in each cluster, screening of participants, and the collection and transportation of sputum specimens. The survey census and collection of sputum specimens are summarized below (screening methods are described in Section 2.4.1). 2.6.1 Survey census In each survey cluster, a population listing was typically obtained by local volunteers 1‒2 weeks in advance of the arrival of the survey investigators. In some surveys, the survey investigators (or staff from the bureau of statistics) undertook the census. At the beginning of field operations, the survey investigators would confirm the population listing, and assess each enumerated person’s eligibility to participate, based on their age and residential status (Table 2.1). During the census, data on household assets were collected in several surveys to measure socioeconomic status (in Kenya, Malawi, Mongolia, Myanmar, the Philippines, Rwanda, the United Republic of Tanzania, Viet  Nam and Zambia). In some of the surveys, it was possible to evaluate the relationship between household poverty and TB disease (8).1 2.6.2 Sputum collection and transportation Typically, two sputum samples (spot and the following morning) were collected. It was often a challenge to obtain quality sputum samples, compared with routine sputum collection for coughing patients who are seeking medical care. Despite a WHO recommendation to take two spot specimens 1  hour apart on the same day (i.e. front loading) (9), in the setting of prevalence surveys, a spot sample followed by a morning sample the next day was generally advised. An additional second spot sample (i.e. a third specimen) was collected in some surveys, especially when the quality of the first specimen was poor. 1 This study combined individual-level data from some of these countries, and found no relationship between household socioeconomic level and TB disease. However, because of the small numbers of TB cases usually detected, prevalence surveys are not an efficient study design for investigating TB risk factors. Sputum specimens were transported via cold chain to a designated laboratory, ideally within 3 days of specimen collection to allow for rapid culturing and to avoid contamination. A maximum processing time of 7  days after collection in the field was recommended, provided that the cold chain was maintained. 2.7 Additional testing for HIV infection and drug susceptibility 2.7.1 HIV testing Information about the HIV status of TB patients is essential both for individual patient care and for understanding the epidemiology of TB. However, HIV testing was not usually done as part of survey field operations owing to logistical constraints (Table 2.3). Only seven of the 25 surveys collected data about HIV status, and all seven of these were in Africa. In Zambia, HIV testing was offered in the field to every survey participant as part of the survey; in Rwanda, Uganda and the United Republic of Tanzania, HIV testing was offered as part of the survey to all participants that screened positive based on symptom screening or chest X-ray criteria (or both), with an opt-out modality. When incorporated in the survey, HIV testing was implemented according to national guidelines, and included pre- and post-test counselling. In Malawi, given the high population coverage of HIV testing, all survey participants were asked to report their HIV status to survey investigators. In Kenya and Zimbabwe, the HIV status of survey cases was obtained from linkage with available records in routine disease information systems. 2.7.2 Testing for drug susceptibility National TB prevalence surveys are not designed to precisely estimate the prevalence of drug-resistant TB, owing to the small number of survey cases. However, drug susceptibility testing was usually done for all survey cases to inform case management (Table 2.3). In some surveys that used Xpert MTB/RIF, rifampicin-susceptibility status was recorded. 2.8 Central-level activities Apart from the organisational and logistical aspects of surveys, the main activities conducted at the central level (as opposed to in the field) were the confirmatory reading of chest X-rays and the review of participants with positive laboratory results. PART I: An overview of the 25 surveys implemented 2007–2016 27 2.8.1 Central chest X-ray reading A second reading of chest X-rays taken in the field was done centrally by trained radiologists, to provide quality assurance of field chest X-ray readings, and a formal interpretation that could be used in determining the final list of survey cases. In surveys undertaken before the use of Xpert MTB/RIF, probable TB survey cases were defined using positive smear results and chest X-ray readings, especially when culture was negative or not available. In the later surveys done in Bangladesh, Kenya and the Philippines, central chest X-ray readings were also used to define a case when culture positivity was weak1 and there was no other positive evidence on Xpert MTB/RIF or smear. In most surveys, all chest X-rays were reread; however, in countries with limited capacity, all abnormal chest X-rays and 10‒20% of normal chest X-rays were reread. Some surveys attempted to have the central reading undertaken at the same time as field operations, but since this required major logistical organization and strong internet connectivity, it rarely happened. 2.8.2 Central review of participants with positive laboratory results Each survey conducted a review of all cases by a panel that typically comprised the survey coordinator, a radiologist, a medical officer, head of laboratory and the data manager. The panel was responsible for the final interpretation of radiographic and laboratory results for all participants with any positive laboratory results (e.g. smear, culture or Xpert MTB/RIF). The panel had two objectives: to define and confirm the status of TB survey cases; and to refer patients for further investigations and treatment, as needed. Typically, the panel reviewed only one to three cases each week. All panel decisions were documented. 2.9 Data management and analysis Given the sample size of a typical TB prevalence survey and the need to enter data from different sources (census, household surveys, symptom screening, field and central chest X-ray readings, and laboratory and final diagnostic panel decisions), data management is a crucial, and often underestimated component of a survey (as discussed in Chapter 4). In surveys implemented in the period 2007‒2014, data were mostly collected on paper and then entered into a database at the central level. Subsequently, the digitalization of survey data management increased with the use of computers, personal digital assistants, tablets, digital chest X-rays, barcoding and internet- connectivity in the field. The survey in Ghana (in 2013) was the first to rely predominantly on electronic data entry, and the survey in Zambia was the first to be virtually “paper-free”. The growing use of digital technologies increased the speed and efficiency with which data could be cleaned and analysed, and helped to improve data quality. It also required additional investment in equipment and, in particular, staff with specialist information technology skills. In areas with poor internet connectivity and unreliable power supply, complete reliance on digital systems was not possible. Furthermore, although such technologies have many advantages, overreliance on digital systems occasionally led to insufficient attention to data quality checks. Thus, systems using paper remain relevant, especially for data quality assurance and back- up purposes. Following data cleaning, analysis of survey results usually required specialist technical assistance to ensure the correct application of best-practice methods (10). Prevalence estimates were produced using three statistical approaches (cluster-based analysis, and two models based on individual-level analysis and multiple imputation for missing data). Multiple imputation of missing data and inverse probability weighting was the recommended method to report final results, unless there was a clear and documented justification to use one of the other two methods. With one exception, all national surveys implemented in the period 2007‒2016 were analysed using the recommended methods.2 2.10 Additional studies In surveys conducted before 2007 it was common to implement, in parallel, a tuberculin survey; the Viet Nam survey (2007) was the last survey to do this (11). The practice was discontinued following updated WHO policy guidance in 2009 about the limited usefulness of tuberculin surveys (12). Data about diseases or health conditions other than TB – for example, smoking, chronic obstructive pulmonary 1 Weak positive culture is defined as one to nine colonies of M. tuberculosis. 2 The 2007 survey in Viet Nam was analysed before the development and publication of these methods. The 2007 survey in the Philippines was initially not analysed using the recommended methods, but was reanalysed using these methods in 2009. National TB prevalence surveys 2007–201628 disease, obesity (body mass index was measured) and diabetes – were collected in a few surveys to assess TB risk factors. These data were not systematically collected or analysed across all surveys. However, data on the health care seeking behaviour of survey participants with symptoms suggestive of TB (e.g. cough ≥2 weeks) in the United Republic of Tanzania, Viet Nam and Zambia were published (13-15). These data highlighted the location where care was initially sought, and therefore the missed opportunities to diagnose TB, but they also highlighted that many symptomatic participants did not seek care. 1 Ghana and Rwanda submitted a paper at the time of writing. Thailand produced one report in Thai only. 2.11 Reporting and dissemination of results A report was produced for all 25 surveys, and results from 11 surveys were published in a peer-reviewed journal (7, 16-52).1 The process took about 1.3 years (and sometimes up to 3 years) from the time of completing field operations to official dissemination of results or publication of findings. Some survey investigators published results that extended beyond the primary objective of estimating national TB prevalence. Examples include the health care Table 2.5 Total budget and sources of funding for national TB prevalence surveys, 2007‒2016 Country Total budget (US$ millions) Global Fund US government Domestic funding Other Bangladesh 3.6 ● ● – – Cambodia 1.0 ● ● – ● China 5.6 – – ● – DPR Korea 1.4 ● – ● – Ethiopia 2.8 ● – ● ● Gambia 1.9 ● – – ● Ghana 2.2 ● – – ● Indonesia 4.6 ● ● – – Kenya 5.2 ● ● – ● Lao PDR 1.3 ● ● – – Malawi 2.2 ● – ● – Mongolia 1.1 ● – ● ● Myanmar 0.9 – ● – ● Nigeria 3.1 ● – ● ● Pakistan 4.4 – ● – ● Philippines (2007) Not known ● – – ● Philippines (2016) 2.4 ● – ● – Rwanda 2.4 ● – – ● Sudan 1.9 ● – ● – Thailand 1.9 ● – ● – Uganda 2.8 ● – – – UR Tanzania 3.4 ● ● ● ● Viet Nam 1.1 ● – ● ● Zambia 5.4 ● ● – ● Zimbabwe 3.5 ● – – – DPR Korea, Democratic People’s Republic of Korea; Global Fund, Global Fund to Fight AIDS, Tuberculosis and Malaria; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania; US, United States of America. ● yes; – no PART I: An overview of the 25 surveys implemented 2007–2016 29 seeking behaviour of survey participants in the United Republic of Tanzania, Viet Nam and Zambia (13-15); the characteristics of participants with non-tuberculosis mycobacteria and the use of computer-aided reading of chest X-rays in Zambia (53, 54); the diagnosis and treatment of TB in the private sector, and the association between TB and household expenditure in Viet Nam (55, 56). Pooled survey data have been used to help understand differences in TB burden by sex, and the effect of household poverty on TB (8, 57). 2.12 Ethics approval All surveys were approved by their respective national ethical review boards, and all protocols were reviewed and approved by partner agencies (e.g. those providing technical assistance) and the WHO Global Task Force on TB impact measurement. 2.13 Budgeting and financing As reported by the survey teams, the Global Fund was a crucial source of financing for all but three surveys (Table  2.5; further details in individual country profiles). Other international funders, especially the US government, also made major contributions to survey funding. Some countries were able to fully or partially fund their surveys from domestic resources. Most of the international technical assistance for the 25 surveys was funded by the US government and the Japan International Cooperation Agency. References 1. World Health Organization. Assessing tuberculosis prevalence through population-based surveys. Geneva, Switzerland 2007 (https://apps.who.int/iris/handle/10665/206962, accessed 20 November 2019). 2. World Health Organization. Tuberculosis prevalence surveys: a handbook. Geneva, Switzerland 2011 (https://apps.who.int/ iris/bitstream/handle/10665/44481/9789241548168_eng.pdf, accessed 27 November 2019). 3. World Health Organization. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF system. Policy statement. Geneva, Switzerland 2011 (https://apps.who. int/iris/bitstream/handle/10665/44586/9789241501545_eng.pdf, accessed 28 November 2019). 4. World Health Organization. Definitions and reporting framework for tuberculosis (updated 2014). Geneva, Switzerland 2013 (www. who.int/iris/bitstream/10665/79199/1/9789241505345_eng.pdf, accessed 28 November 2019). 5. World Health Organization. Chest radiography in tuberculosis detection: summary of current WHO recommendations and guidance on programmatic approaches. Geneva, Switzerland 2016 (https://apps.who.int/iris/bitstre am/10665/252424/1/9789241511506-eng.pdf, accessed 28 November 2019). 6. World Health Organization. Fluorescent light-emitting diode (LED) microscopy for diagnosis of tuberculosis policy. Policy statement. Geneva, Switzerland 2011 (https://apps.who.int/ iris/bitstream/handle/10665/44602/9789241501613_eng.pdf, accessed 28 November 2019). 7. Mao TE, Okada K, Yamada N, Peou S, Ota M, Saint S et al. Cross- sectional studies of tuberculosis prevalence in Cambodia between 2002 and 2011. Bull World Health Organ. 2014;92(8):573–81 (https://www.ncbi.nlm.nih.gov/pubmed/25177072, accessed 28 November 2019). 8. Siroka A, Law I, Macinko J, Floyd K, Banda RP, Hoa NB et al. The effect of household poverty on tuberculosis. Int J Tuberc Lung Dis. 2016;20(12):1603–8 (https://www.ncbi.nlm.nih.gov/ pubmed/27931334, accessed 28 November 2019). 9. TB diagnostics and laboratory strengthening - WHO policy [website]. 2007 (https://www.who.int/tb/areas-of-work/ laboratory/policy_diagnosis_pulmonary_tb/en/, accessed 26 April 2019). 10. Floyd S, Sismanidis C, Yamada N, Daniel R, Lagahid J, Mecatti F et al. Analysis of tuberculosis prevalence surveys: new guidance on best-practice methods. Emerg Themes Epidemiol. 2013;10(1):10 (https://www.ncbi.nlm.nih.gov/pubmed/24074436, accessed 28 November 2019). 11. Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, Tiemersma EW. First national tuberculin survey in Viet Nam: characteristics and association with tuberculosis prevalence. Int J Tuberc Lung Dis. 2013;17(6):738–44 (https://www.ncbi.nlm.nih. gov/pubmed/23676155, accessed 28 November 2019). 12. TB impact measurement policy and recommendations for how to assess the epidemiological burden of TB and the impact of TB control. Stop TB policy paper no. 2, Geneva: World Health Organization; 2009 (https://apps.who.int/iris/bitstream/ handle/10665/44231/9789241598828_eng.pdf, accessed 22 November 2019). 13. Chanda-Kapata P, Kapata N, Masiye F, Maboshe M, Klinkenberg E, Cobelens F et al. Health seeking behaviour among individuals with presumptive tuberculosis in Zambia. PLoS One. 2016;11(10):e0163975 (https://www.ncbi.nlm.nih.gov/ pubmed/27711170, accessed 28 November 2019). 14. Hoa NB, Tiemersma EW, Sy DN, Nhung NV, Vree M, Borgdorff MW et al. Health-seeking behaviour among adults with prolonged cough in Vietnam. Trop Med Int Health. 2011;16(10):1260–7 (https://www.ncbi.nlm.nih.gov/pubmed/21692960, accessed 28 November 2019). 15. Senkoro M, Hinderaker SG, Mfinanga SG, Range N, Kamara DV, Egwaga S et al. Health care-seeking behaviour among people with cough in Tanzania: findings from a tuberculosis prevalence survey. Int J Tuberc Lung Dis. 2015;19(6):640–6 (https://www.ncbi.nlm. nih.gov/pubmed/25946352, accessed 28 November 2019). 16. Adetifa IM, Kendall L, Bashorun A, Linda C, Omoleke S, Jeffries D et al. A tuberculosis nationwide prevalence survey in Gambia, 2012. Bull World Health Organ. 2016;94(6):433–41 (https://www. ncbi.nlm.nih.gov/pubmed/27274595, accessed 28 November 2019). 17. Bonsu F, Addo KK, Alebachew Z, Gyapong J, Badu-Peprah A, Gockah R et al. National population-based tuberculosis prevalence survey in Ghana, 2013. Int J Tuberc Lung Dis. 2020 Mar 1;24(3):321-328. (https://pubmed.ncbi.nlm.nih.gov/32228763/, accessed 20 March 2020). National TB prevalence surveys 2007–201630 18. Department of Public Health Federal Republic of Nigeria. Report of the first national TB prevalence survey 2012, Nigeria. 2014 (https://www.who.int/tb/publications/NigeriaReport_WEB_ NEW.pdf, accessed 9 October 2017). 19. Disease Control Bureau of the Ministry of Health – Chinese Center for Disease Control and Prevention. Report on the 5th national tuberculosis epidemiological survey in China – 2010. Beijing, China: 2011. 20. Enos M, Sitienei J, Ong’ang’o J, Mungai B, Kamene M, Wambugu J et al. Kenya tuberculosis prevalence survey 2016: challenges and opportunities of ending TB in Kenya. PLoS One. 2018;13(12):e0209098 (https://www.ncbi.nlm.nih.gov/ pubmed/30586448, accessed 28 November 2019). 21. Hoa NB, Sy DN, Nhung NV, Tiemersma EW, Borgdorff MW, Cobelens FG. National survey of tuberculosis prevalence in Viet Nam. Bull World Health Organ. 2010;88(4):273–80 (https://www. ncbi.nlm.nih.gov/pubmed/20431791, accessed 28 November 2019). 22. Institute of Epidemiology Disease Control & Research, Ministry of Health & Family Welfare Government of the People’s Republic of Bangladesh. National tuberculosis prevalence survey, Bangladesh 2015–2016. Dhaka, Bangladesh: 2017. 23. Kapata N, Chanda-Kapata P, Ngosa W, Metitiri M, Klinkenberg E, Kalisvaart N et al. The prevalence of tuberculosis in Zambia: results from the first national TB prevalence survey, 2013–2014. PLoS One. 2016;11(1):e0146392 (https://www.ncbi.nlm.nih.gov/ pubmed/26771588, accessed 28 November 2019). 24. Kebede AH, Alebachew Z, Tsegaye F, Lemma E, Abebe A, Agonafir M et al. The first population-based national tuberculosis prevalence survey in Ethiopia, 2010–2011. Int J Tuberc Lung Dis. 2014;18(6):635–9 (https://www.ncbi.nlm.nih.gov/ pubmed/24903931, accessed 28 November 2019). 25. Law I, Sylavanh P, Bounmala S, Nzabintwali F, Paboriboune P, Iem V et al. The first national tuberculosis prevalence survey of Lao PDR (2010–2011). Trop Med Int Health. 2015;20(9):1146–54 (https://www.ncbi.nlm.nih.gov/pubmed/25939366, accessed 28 November 2019). 26. Migambi P, Gasana M, Uwizeye CB, Kamanzi E, Ndahindwa V, Kalisvaart N et al. Prevalence of tuberculosis in Rwanda: Results of the first nationwide survey in 2012 yielded important lessons for TB control. PLoS One. 2020 Apr 23;15(4):e0231372. https:// pubmed.ncbi.nlm.nih.gov/32324750/, accessed 20 April 2020). 27. Ministry of Health. Technical report of the Zambia national tuberculosis prevalence survey (2013–2014). 2015 (https://www. moh.gov.zm/docs/reports/zntbs13-14_final.pdf, accessed 12 October 2017). 28. Ministry of Health – Cambodia. Report of the second national tuberculosis prevalence survey, 2011. 2012 (http://open_ jicareport.jica.go.jp/pdf/12120325.pdf, accessed 28 November 2019). 29. Ministry of Health – Federal Democratic Republic of Ethiopia. First Ethiopian national population based tuberculosis prevalence survey. Ethiopian Health and Nutrition Research Institute; 2011. 30. Ministry of Health – Mongolia. Report of the first national tuberculosis prevalence survey in Mongolia (2014–2015). Ulaanbaatar, Mongolia: 2016. 31. Ministry of Health – Myanmar. Report on national TB prevalence survey 2009–2010. Yangon, Myanmar: 2010. 32. Ministry of Health – Republic of Indonesia, National Institute of Health Research and Development. Indonesia tuberculosis prevalence survey 2013–2014. Jakarta, Indonesia: 2015. 33. Ministry of Health – Thailand. Fourth tuberculosis prevalence survey in Thailand, 2012. Bangkok, Thailand: 2015. 34. Ministry of Health – Viet Nam. National tuberculosis prevalence survey in Vietnam, 2006–2007. Hanoi, Viet Nam: 2008. 35. Ministry of Health (national tuberculosis leprosy and lung disease program). National tuberculosis prevalence survey of Kenya 2015–2016. 2016. 36. Ministry of Health (national tuberculosis program). National tuberculosis prevalence survey of Sudan 2013–2014 (report). 2018. 37. Ministry of Health and Child Care. The Zimbabwe national population based tuberculosis prevalence survey. 2014. 38. Ministry of Health and Social Welfare. The first national tuberculosis prevalence survey in the United Republic of Tanzania. 2013. 39. Ministry of Health and Social Welfare. The Gambian survey of tuberculosis prevalence report. Medical Research Council Unit, The Gambia; 2014. 40. Ministry of Health National TB Control Programme. Technical report: Malawi tuberculosis prevalence survey (2013–2014). 2016. 41. Ministry of Health Uganda and Makerere University School of Public Health. Population-based survey of prevalence of tuberculosis disease in Uganda 2014–15 (report). Makerere University School of Public Health; 2016. 42. Ministry of Public Health – DPRK. Report of DPRK national TB prevalence survey 2015–2016. Pyongyang, Democratic People’s Republic of Korea: 2017. 43. National TB control program – Pakistan. Prevalence of pulmonary tuberculosis among the adult population of Pakistan, 2010–2011. Islamabad, Pakistan: 2013. 44. National TB Control Programme (Ghana Health Service). Ghanaian national population based tuberculosis prevalence survey in 2013 (report). 2016. 45. National Tuberculosis Centre – Ministry of Health – Lao PDR. Report of the first report of the first national tuberculosis prevalence survey of Lao PDR (2010–2011). Vientiane, Lao PDR: 2013. 46. Philippines: Department of Health Republic of the Philippines, Foundation for the Advancement of Clinical Epidemiology Inc, Philippine Council for Health Research and Development. National tuberculosis prevalence survey 2016. Manila, Philippines: 2018. 47. Qadeer E, Fatima R, Yaqoob A, Tahseen S, Ul Haq M, Ghafoor A et al. Population based national tuberculosis prevalence survey among adults (>15 years) in Pakistan, 2010–2011. PLoS One. 2016;11(2):e0148293 (https://www.ncbi.nlm.nih.gov/ pubmed/26863617, accessed 28 November 2019). 48. Rwanda Ministry of Health. The first national tuberculosis prevalence survey in Rwanda (2012). 2015. 49. Senkoro M, Mfinanga S, Egwaga S, Mtandu R, Kamara DV, Basra D et al. Prevalence of pulmonary tuberculosis in adult population of Tanzania: a national survey, 2012. Int J Tuberc Lung Dis. 2016;20(8):1014–21 (https://www.ncbi.nlm.nih.gov/ pubmed/27393533, accessed 28 November 2019). 50. Tropical Disease Foundation Inc and Department of Health. Final report of the nationwide tuberculosis prevalence survey 2007. Navotas City, Philippines: 2008. 51. Tupasi TE, Radhakrishna S, Chua JA, Mangubat NV, Guilatco R, Galipot M et al. Significant decline in the tuberculosis burden in the Philippines ten years after initiating DOTS. Int J Tuberc Lung Dis. 2009;13(10):1224–30 (https://www.ncbi.nlm.nih.gov/ pubmed/19793426, accessed 28 November 2019). PART I: An overview of the 25 surveys implemented 2007–2016 31 52. Wang L, Zhang H, Ruan Y, Chin DP, Xia Y, Cheng S et al. Tuberculosis prevalence in China, 1990–2010; a longitudinal analysis of national survey data. Lancet. 2014;383(9934):2057–64 (https://www.ncbi.nlm.nih.gov/pubmed/24650955, accessed 28 November 2019). 53. Chanda-Kapata P, Kapata N, Klinkenberg E, Mulenga L, Tembo M, Katemangwe P et al. Non-tuberculous mycobacteria (NTM) in Zambia: prevalence, clinical, radiological and microbiological characteristics. BMC Infect Dis. 2015;15:500 (https://www.ncbi. nlm.nih.gov/pubmed/26545357, accessed 28 November 2019). 54. Melendez J, Philipsen R, Chanda-Kapata P, Sunkutu V, Kapata N, van Ginneken B. Automatic versus human reading of chest X-rays in the Zambia national tuberculosis prevalence survey. Int J Tuberc Lung Dis. 2017;21(8):880–6 (https://www.ncbi.nlm.nih. gov/pubmed/28786796, accessed 28 November 2019). 55. Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, EW T. Diagnosis and treatment of tuberculosis in the private sector, Vietnam. Emerg Infect Dis. 2011;17(3):562–4 (https://www.ncbi. nlm.nih.gov/pubmed/21392464, accessed 28 November 2019). 56. Hoa NB, Tiemersma EW, Sy DN, Nhung NV, Gebhard A, Borgdorff MW et al. Household expenditure and tuberculosis prevalence in VietNam: prediction by a set of household indicators. Int J Tuberc Lung Dis. 2011;15(1):32–7 (https://www.ncbi.nlm.nih.gov/ pubmed/21276293, accessed 28 November 2019). 57. Horton KC, MacPherson P, Houben RM, White RG, Corbett EL. Sex differences in tuberculosis burden and notifications in low- and middle-income countries: a systematic review and meta- analysis. PLoS Med. 2016;13(9):e1002119 (https://www.ncbi.nlm. nih.gov/pubmed/27598345, accessed 28 November 2019). 32 Photo credit: Jane Rahedi Ong’ang’o / KEMRI Use of a digital chest X-ray during the 2015-2016 national TB prevalence survey of Kenya. 33 Chapter 3 Results and their implications 3.1 Survey population, enrolment and participation Table 3.1 shows the size of the planned sample population in national tuberculosis (TB) prevalence surveys implemented in 2007–2016. The table also shows the actual size of the eligible population, the number of people who participated, the participation rate and the number of participants who screened positive for sputum examination. 3.1.1 Participation The participation rate was high in most surveys, at ≥80% of the eligible population in 19 of 25 surveys (Fig. 3.1 and Table 3.1). The six countries with lower participation rates were Gambia, Nigeria, the Philippines (in 2016), Thailand, the United Republic of Tanzania and Zimbabwe. In general, participation rates were higher among females, and middle and older age groups, compared with males and younger age groups (see Part II for details). Reasons for non-participation were not routinely documented, but included previous work-related health assessments or ease of access to health facilities (both of which reduced the incentive to participate for the purposes of having a chest X-ray examination), as well as lack of time. Achieving high levels of participation in highly urban settings, especially capital cities, was challenging in almost all countries. The most extreme example was the Bangkok metropolitan area of Thailand, in which only 26% of the eligible population participated. Results from Bangkok were subsequently excluded from the final analysis. In the Republic of Korea, the repetition of prevalence surveys every 5 years was discontinued after 1995 because of declining participation (in the context of an increasingly urbanized and modern environment) and a reduction in disease burden, which would have necessitated much larger sample sizes (1). In countries that were not able to achieve a high participation rate in surveys implemented in 2007–2016, careful consideration of whether a future survey should be attempted will be needed, especially if there are further increases in the proportion of the population living in urban or more economically developed areas. This is discussed further in Chapter 5. 3.1.2 Eligibility for sputum examination The proportion of participants who were eligible for sputum examination averaged 16%, ranging from a low of 4% of screened participants in China to a high of 40% in the 2016 survey in the Philippines (Table 3.1). The proportion was more than 20% in Bangladesh, Indonesia, Mongolia, Myanmar, the Philippines (2007 and 2016) and Sudan, due to high yields from chest X-ray screening. In 15 of 25 surveys, chest X-ray screening identified more participants eligible for sputum examination than symptom screening (Table 3.1). However, the opposite applied in Malawi, the United Republic of Tanzania and Zambia; these African countries used a broader range of symptoms with the aim of increasing the sensitivity of the screening algorithm in a high HIV prevalence setting.1 Of the other seven surveys, screening yields were similar, and one survey (Philippines 2007) did not systematically use symptoms for screening purposes. 3.2 TB prevalence and updated estimates of TB disease burden 3.2.1 Prevalence of pulmonary TB disease Surveys showed that the estimated prevalence of pulmonary TB per 100 000 population was high in many countries, but there was also considerable variation 1 The symptom screening criteria used in Malawi were any symptom for at least 1 week, including cough, sputum production, haemoptysis, chest pain, weight loss, night sweats, fatigue, fever or shortness of breath; in the United Republic of Tanzania, cough of ≥2 weeks, or haemoptysis or fever of ≥2 weeks, or weight loss or excessive night sweats; and in Zambia, cough of ≥2 weeks, or fever of ≥2 weeks, or chest pain of ≥2 weeks. National TB prevalence surveys 2007–201634 Ta bl e 3. 1 Su m m ar y of s am pl in g po pu la tio n, s ur ve y pa rt ic ip an ts a nd s cr ee ni ng o ut co m es C XR , ch es t X- ra y; D P R K or ea , D em oc ra ti c P eo pl e’ s R ep ub lic o f K or ea ; La o P D R , La o P eo pl e’ s D em oc ra ti c R ep ub lic ; N /A , no t ap pl ic ab le ; S ym , sy m pt om ; U R T an za ni a, U ni te d R ep ub lic o f Ta nz an ia ; + , po si ti ve ; -, n eg at iv e. Co un tr y Tim ef ra m e of fi el d op er at io ns Pl an ne d sa m pl e po pu la tio n Nu m be r of p eo pl e el ig ib le to pa rt ic ip at e Su rv ey pa rt ic ip an ts Nu m be r o f p ar tic ip an ts e lig ib le fo r s pu tu m e xa m in at io n Nu m be r R at e (% ) Sy m +, CX R+ % Sy m +, CX R- / N /A % Sy m -, CX R+ % Ot he rs % An y Sy m + % An y CX R+ % To ta l el ig ib le % Af ri ca E th io pi a 2 0 1 0 –2 0 1 1 4 6 5 1 4 5 1 6 6 7 4 6 6 9 7 9 0 % 8 0 6 1 .7 % 2 2 2 0 4 .8 % 3 0 1 3 6 .5 % 4 1 0 .0 9 % 3 0 2 6 6 .5 % 3 8 1 9 8 .2 % 6 0 8 0 1 3 % G am bi a 2 0 1 1 –2 0 1 3 5 5 2 8 1 5 5 8 3 2 4 3 1 0 0 7 7 % 1 0 2 6 2 .4 % 2 4 3 6 5 .7 % 2 3 8 4 5 .5 % 1 0 2 0 .2 4 % 3 4 6 2 8 .0 % 3 4 1 0 7 .9 % 5 9 4 8 1 4 % G ha na 2 0 1 3 6 3 9 0 5 6 7 7 5 7 6 1 7 2 6 9 1 % 7 7 1 1 .2 % 1 1 9 8 1 .9 % 4 3 8 7 7 .1 % 1 9 4 2 3 .1 % 1 9 6 9 3 .2 % 5 1 5 8 8 .4 % 8 2 9 8 1 3 % K en ya 2 0 1 5 –2 0 1 6 7 2 0 0 0 7 6 2 9 1 6 3 0 5 0 8 3 % 1 2 4 1 2 .0 % 2 8 9 6 4 .6 % 5 1 8 4 8 .2 % 3 9 4 0 .6 2 % 4 1 3 7 6 .6 % 6 4 2 5 1 0 % 9 7 1 5 1 5 % M al aw i 2 0 1 3 –2 0 1 4 3 7 2 0 0 3 9 0 2 6 3 1 5 7 9 8 1 % 3 8 1 1 .2 % 2 3 3 4 7 .4 % 7 1 7 2 .3 % N /A N /A 2 7 1 5 8 .6 % 1 0 9 8 3 .5 % 3 4 3 2 1 1 % N ig er ia 2 0 1 2 4 9 0 0 0 7 7 7 9 7 4 4 1 8 6 5 7 % 7 4 6 1 .7 % 1 7 2 0 3 .9 % 2 2 2 2 5 .0 % N /A N /A 2 4 6 6 5 .6 % 2 9 6 8 6 .7 % 4 6 8 8 1 1 % R w an da 2 0 1 2 4 4 5 0 0 4 5 0 5 8 4 3 1 2 8 9 6 % 5 4 5 1 .3 % 2 0 9 2 4 .9 % 2 1 0 7 4 .9 % 3 0 .0 1 % 2 6 3 7 6 .1 % 2 6 5 2 6 .1 % 4 7 4 7 1 1 % S ud an 2 0 1 3 –2 0 1 4 9 1 1 3 1 9 6 9 7 9 8 3 2 0 2 8 6 % 1 8 2 3 2 .2 % 8 4 0 1 .0 % 9 8 3 8 1 2 % 5 0 4 0 6 .1 % 2 6 6 3 3 .2 % 1 1 6 6 1 1 4 % 1 7 5 4 1 2 1 % U ga nd a 2 0 1 4 –2 0 1 5 4 0 1 8 0 4 5 2 9 3 4 1 1 5 4 9 1 % 5 5 2 1 .3 % 2 1 6 2 5 .3 % 2 2 9 8 5 .6 % 1 3 0 0 .3 2 % 2 7 1 4 6 .6 % 2 8 5 0 6 .9 % 5 1 4 2 1 2 % U R T an za ni a 2 0 1 1 –2 0 1 2 4 6 7 9 2 6 5 6 6 4 5 0 4 4 7 7 7 % 8 0 4 1 .6 % 3 4 5 9 6 .9 % 2 0 3 9 4 .0 % N /A N /A 4 2 6 3 8 .5 % 2 8 4 3 5 .6 % 6 3 0 2 1 2 % Za m bi a 2 0 1 3 –2 0 1 4 5 4 4 0 0 5 4 8 3 0 4 6 0 9 9 8 4 % 1 5 0 5 3 .3 % 2 9 4 8 6 .4 % 2 2 5 5 4 .9 % N /A N /A 4 4 5 3 1 0 % 3 7 6 0 8 .2 % 6 7 0 8 1 5 % Zi m ba bw e 2 0 1 4 4 4 9 5 1 4 3 4 7 8 3 3 7 3 6 7 8 % 6 2 8 1 .9 % 1 2 0 5 3 .6 % 2 8 0 3 8 .3 % 1 1 8 4 3 .5 % 1 8 3 3 5 .4 % 3 4 3 1 1 0 % 5 8 2 0 1 7 % To ta l 83 % 1. 8% 4. 7% 6. 2% 1. 7% 6. 5% 8. 0% 14 % As ia B an gl ad es h 2 0 1 5 –2 0 1 6 1 0 0 0 0 0 1 0 8 8 3 4 9 8 7 1 0 9 1 % 3 0 7 7 3 .1 % 4 2 1 7 4 .3 % 1 3 3 0 0 1 3 % N /A N /A 7 2 9 4 7 .4 % 1 6 3 7 7 1 7 % 2 0 5 9 4 2 1 % C am bo di a 2 0 1 0 –2 0 1 1 3 9 6 8 0 4 0 4 2 3 3 7 4 1 7 9 3 % 7 1 0 1 .9 % 1 2 0 6 3 .2 % 2 6 9 9 7 .2 % 1 6 5 0 .4 4 % 1 9 1 6 5 .1 % 3 4 0 9 9 .1 % 4 7 8 0 1 3 % C hi na 2 0 1 0 2 6 4 0 0 0 2 6 3 2 8 1 25 2 94 0 9 6 % 7 9 7 0. 32 % 4 6 6 5 1 .8 % 2 1 8 9 0. 87 % 2 1 7 4 0 .8 6 % 5 4 6 2 2 .2 % 2 9 8 6 1 .2 % 9 8 2 5 3 ,9 % D P R K or ea 2 0 1 5 –2 0 1 6 7 0 0 0 0 7 1 8 7 7 6 0 6 8 3 8 4 % 1 0 2 8 1 .7 % 1 9 1 6 3 .2 % 1 8 5 8 3 .1 % N /A N /A 2 9 4 4 4 .9 % 2 8 8 6 4 .8 % 4 8 0 2 7 ,9 % In do ne si a 2 0 1 3 –2 0 1 4 7 8 0 0 0 7 6 5 7 6 6 7 9 4 4 8 9 % 4 4 5 9 6 .6 % 4 0 9 3 6 .0 % 6 7 4 3 1 0 % 1 5 1 0 .2 2 % 8 5 5 2 1 3 % 1 1 2 0 2 1 6 % 1 5 4 4 6 2 3 % La o P D R 2 0 1 0 –2 0 1 2 4 0 0 0 0 4 6 0 7 9 3 9 2 1 2 8 5 % 1 3 1 2 3 .3 % 1 9 2 7 4 .9 % 3 1 0 7 7 .9 % N /A N /A 3 2 3 9 8 .3 % 4 4 1 9 1 1 % 6 3 4 6 1 6 % M on go lia 2 0 1 4 –2 0 1 5 4 9 0 0 0 6 0 0 3 1 5 0 3 0 9 8 4 % 8 1 7 1 .6 % 1 7 2 9 3 .4 % 7 0 6 4 1 4 % 7 4 9 1 .5 % 2 5 4 6 5 .1 % 7 8 8 1 1 6 % 1 0 3 5 9 2 1 % M ya nm ar 2 0 0 9 –2 0 1 0 4 9 6 9 0 5 7 6 0 7 5 1 3 6 7 8 9 % 1 2 5 8 2 .4 % 4 3 3 0. 84 % 9 3 6 4 1 8 % 1 1 8 0 2 .3 % 1 6 9 1 3 .3 % 1 0 6 2 2 2 1 % 1 2 2 3 5 2 4 % P ak is ta n 2 0 1 0 –2 0 1 1 1 3 3 0 0 0 1 3 1 3 2 9 10 5 91 3 8 1 % 2 8 1 9 2 .7 % 2 5 9 8 2 .5 % 5 0 4 2 4 .8 % 1 2 0 .0 1 % 5 4 1 7 5 .1 % 7 8 6 1 7 .4 % 1 0 4 7 1 1 0 % P hi lip pi ne s 2 0 0 7 3 0 0 0 0 2 2 8 6 7 2 0 6 4 3 9 0 % N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A 5 3 7 8 2 6 % P hi lip pi ne s 2 0 1 6 5 4 0 0 0 6 1 4 6 6 4 6 6 8 9 7 6 % 1 4 4 4 3 .1 % 1 3 7 1 2 .9 % 1 0 7 0 2 2 3 % 5 0 8 0 1 1 % 2 8 1 5 6 .0 % 1 2 1 4 6 2 6 % 1 8 5 9 7 4 0 % Th ai la nd 2 0 1 2 –2 0 1 3 9 0 0 0 0 7 8 8 3 9 6 2 5 3 6 7 9 % 5 2 6 0. 84 % 1 7 5 7 2 .8 % 3 7 6 7 6 .0 % N /A N /A 2 2 8 3 3 .7 % 4 2 9 3 6 .9 % 6 0 5 0 1 0 % Vi et N am 2 0 0 6 –2 0 0 7 1 0 5 0 0 0 1 0 3 9 2 4 9 4 1 7 9 9 1 % 5 1 8 0. 55 % 3 5 2 2 3 .7 % 2 9 7 2 3 .2 % 9 9 3 1 .1 % 4 0 4 0 4 .3 % 3 4 9 0 3 .7 % 8 0 0 5 8 ,5 % To ta l 87 % 2. 3% 3. 3% 9. 3% 2. 2% 5. 7% 12 % 17 % PART I: An overview of the 25 surveys implemented 2007–2016 35 among countries and between Africa and Asia (Fig. 3.2 and Table 3.2). In African countries, the prevalence of smear-positive pulmonary TB per 100 000 population aged 15 years or above ranged from 74 (95% confidence interval [CI]: 48–99) in Rwanda to 319 (95% CI 232–406) in Zambia. Similarly, the prevalence of bacteriologically confirmed pulmonary TB per 100  000 population aged 15  years or above ranged from 119 (95% CI: 79–160) in Rwanda to 638 (95% CI: 502–774) in Zambia. There was great variation in the proportion of bacteriologically confirmed pulmonary TB cases that were smear-positive in Africa, from a low of 24% in Zimbabwe to a high of 62% in Rwanda (Table 3.2). In Asian countries, the prevalence of smear-positive pulmonary TB per 100 000 population aged 15 years or above ranged from 66 (95% CI: 53–79) in China to 434 (95% CI: 350–518) in the Philippines (in 2016). Similarly, the prevalence of bacteriologically confirmed TB per 100 000 population aged 15 years or above ranged from Fig. 3.1 Participation rate in 25 surveys (24 countries) implemented in 2007−2016 DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. 1 Bangladesh (2015), Kenya (2015) and the Philippines (2016). 119 (95% CI: 103–135) in China to 1159 (95% CI: 1016– 1301) in the Philippines (in 2016). As in the surveys in African countries, the proportion of bacteriologically confirmed pulmonary TB cases that were smear-positive varied widely, from 33% in Cambodia to 68% in Pakistan (Table 3.2). The systematic use of culture (as well as Xpert® MTB/ RIF in three of the later surveys)1 identified more smear- negative than smear-positive pulmonary TB cases in all but the following eight surveys: China, the Democratic People’s Republic of Korea, Nigeria, Pakistan, Rwanda, Sudan, Viet Nam and Zambia (Table 3.2). 3.2.2 Prevalence of pulmonary TB disease disaggregated by age and sex The distribution of prevalent cases by age is shown in Fig. 3.3a for surveys in African countries and Fig. 3.3b for surveys in Asian countries. In the latter, there was National TB prevalence surveys 2007–201636 1 The estimated absolute number of TB cases in each age group is shown in Fig. 3 of the country profiles in Part II. a clear pattern in which prevalence increased with age (an exception was the Democratic People’s Republic of Korea). In African countries, this pattern was only observed in Ghana and Rwanda, although in Malawi and the United Republic of Tanzania there was an increase between the age groups of 45–54 years and of 65 years or more, after an earlier peak in the age group 35–44 years. In most African countries, there was a peak in prevalence per 100 000 population in the age groups 35–44 or 45– 54 years,1 which could be explained at least in part by the impact of the HIV epidemic. As transmission declines, more incident cases arise from past rather than recent infection. Therefore, a pattern in which prevalence increases with age suggests that transmission is falling. It is encouraging that prevalence surveys indicated that transmission is potentially declining in many Asian countries as well as in Ghana, Malawi, Rwanda and the United Republic of Tanzania. In other countries, surveys suggested considerable community transmission. A striking finding across all surveys was the much higher burden of TB disease in men compared with women (Fig. 3.4). The male to female ratio of the prevalence of bacteriologically confirmed TB was 2.3 (95% CI: 2.0–2.6) overall, but ranged from 1.2 in Ethiopia to more than 4 in Uganda and Viet Nam. It was higher in Asia (2.6) than in Africa (2.0). These results mean that men account for about 66–70% of the burden of TB disease among adults in Asia and Africa. 3.2.3 Estimates of the prevalence of TB, all ages and all forms Following surveys, estimates of the prevalence of TB for all ages (i.e. including those aged <15 years) and all forms (i.e. including extrapulmonary as well as pulmonary TB) were updated by WHO in consultation with national authorities. Fig 3.5 compares the updated estimates with the most recent estimates published before survey results became available. Fig. 3.2 Estimates of the prevalence of bacteriologically confirmed pulmonary TB in those aged ≥15 years in 25 surveys (24 countries) implemented in 2007−2016 DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. PART I: An overview of the 25 surveys implemented 2007–2016 37 Ta bl e 3. 2 Es tim at ed p re va le nc e of s m ea r- po si tiv e an d ba ct er io lo gi ca lly c on fir m ed p ul m on ar y TB Co un tr y Sm ea r- po si tiv e pu lm on ar y TB Ba ct er io lo gi ca lly c on fir m ed p ul m on ar y TB Pr op or tio n of ba ct er io lo gi ca lly co nfi rm ed c as es th at w er e sm ea r- po si tiv e N um be r of c as es Pr ev al en ce p er 10 0 00 0 po pu la tio n ag ed ≥1 5 ye ar s a 95 % co nfi de nc e in te rv al k b N um be r of c as es Pr ev al en ce p er 10 0 00 0 po pu la tio n ag ed ≥1 5 ye ar s a 95 % co nfi de nc e in te rv al kb Af ri ca E th io pi a 4 7 1 0 8 7 3 –1 4 3 0 .7 1 1 0 2 7 7 2 0 8 –3 4 7 0 .4 3 9 G am bi a 3 4 9 0 5 3 –1 2 7 1 .3 7 7 2 1 2 1 5 2 –2 7 2 0 .7 4 2 G ha na 6 4 1 1 1 7 6 –1 4 5 0 .9 2 0 2 3 5 6 2 8 8 –4 2 5 0 .7 3 1 K en ya 1 2 3 2 3 0 1 7 4 –2 8 6 0 .7 3 0 5 5 5 8 4 5 5 –6 6 2 0 .7 4 1 M al aw i 6 2 2 2 0 1 4 2 –2 9 7 1 .1 1 3 2 4 5 2 3 1 2 –5 9 3 1 .1 4 9 N ig er ia 1 0 7 3 1 8 2 2 5 –4 1 2 0 .9 1 4 4 5 2 4 3 7 8 –6 7 0 0 .7 6 1 R w an da 2 7 7 4 4 8 –9 9 N /A c 4 0 1 1 9 7 9 –1 6 0 0 .7 6 2 S ud an 5 7 8 7 5 2 –1 2 1 1 .3 1 1 2 1 8 3 1 2 8 –2 3 8 1 .3 4 8 U ga nd a 6 6 1 7 4 1 1 1 –2 3 8 0 .9 1 6 0 4 0 1 2 9 2 –5 0 9 0 .8 4 3 U R T an za ni a c 1 3 4 2 7 5 2 3 2 –3 2 6 0 .6 N /A N /A N /A N /A N /A Za m bi a 1 3 5 3 1 9 2 3 2 –4 0 6 0 .8 2 6 5 6 3 8 5 0 2 –7 7 4 0 .7 5 0 Zi m ba bw e 2 3 8 2 4 7 –1 1 8 N /A d 1 0 7 3 4 4 2 6 8 –4 2 0 0 .3 2 4 As ia B an gl ad es h 1 0 8 1 1 3 8 7 –1 3 9 0 .7 2 7 8 2 8 7 2 4 4 –3 3 0 0 .5 3 9 C am bo di a 1 0 3 2 7 1 2 1 2 –3 4 8 0 .6 3 1 4 8 3 1 7 0 7 –9 7 7 0 .5 3 3 C hi na 1 8 8 6 6 5 3 –7 9 0 .9 3 4 7 1 1 9 10 3– 13 5 0 .5 5 5 D P R K or ea 1 8 7 3 3 0 2 8 3 –3 7 7 2 .0 3 4 0 5 8 7 5 2 0 –6 5 5 0 .6 5 6 In do ne si a 1 6 5 2 5 7 2 1 0 –3 0 3 0 .7 4 2 6 7 5 9 5 9 0 –9 6 1 0 .5 3 4 La o P D R 1 0 7 2 7 8 1 9 9 –3 5 6 0 .7 2 3 7 5 9 5 4 5 7 –7 3 3 0 .7 4 7 M on go lia 8 8 2 0 4 1 4 3 –2 6 5 1 .5 2 4 8 5 6 0 4 5 5 –6 6 5 0 .9 3 6 M ya nm ar 1 2 3 2 4 2 1 8 6 –3 1 5 0 .8 3 1 1 6 1 3 5 0 2 –7 4 8 0 .7 39 P ak is ta n 2 3 3 2 7 0 2 1 7 –3 2 3 0 .6 3 4 1 3 9 8 3 3 3 –4 6 3 0 .6 6 8 P hi lip pi ne s (2 0 0 7 ) 5 5 2 8 0 1 9 0 –3 7 0 1 .0 1 3 6 6 6 0 5 3 0 –8 0 0 0 .6 4 2 P hi lip pi ne s (2 0 1 6 ) 1 7 3 4 3 4 3 5 0 –5 1 8 0 .6 4 4 6 1 1 5 9 1 0 1 6 –1 3 0 1 0 .6 3 7 Th ai la nd 5 8 1 0 4 5 5 –1 9 5 1 .0 1 4 2 2 4 2 1 7 6 –3 3 2 0 .5 4 3 Vi et N am 1 7 4 1 9 7 1 5 0 –2 4 4 0 .8 2 6 9 3 0 7 2 4 9 –3 6 6 0 .6 6 4 D P R K or ea , D em oc ra ti c P eo pl e’ s R ep ub lic o f K or ea ; La o P D R , La o P eo pl e’ s D em oc ra ti c R ep ub lic ; N /A , no t ap pl ic ab le ; TB , tu be rc ul os is ; U R T an za ni a, U ni te d R ep ub lic o f Ta nz an ia . a E st im at es a re b as ed o n th e us e of ro bu st s ta nd ar d er ro rs w it h m is si ng v al ue im pu ta ti on a nd in ve rs e pr ob ab ili ty w ei gh ti ng fo r a ll co un tr ie s ex ce pt th e U ni te d R ep ub lic o f T an za ni a, fo r w hi ch a c lu st er -l ev el m od el o f a na ly si s w as u se d. b k is t he c oe ffi ci en t of v ar ia ti on o f th e cl us te r- sp ec ifi c TB p re va le nc es . c Th e nu m be r of b ac te ri ol og ic al ly c on fir m ed p ul m on ar y TB c as es c ou ld n ot b e ve ri fie d fo r th e U ni te d R ep ub lic o f Ta nz an ia . d Fo r R w an da a nd Z im ba bw e, k c ou ld n ot b e ca lc ul at ed b ec au se t he d es ig n ef fe ct w as le ss t ha n on e. National TB prevalence surveys 2007–201638 UR Tanzania, United Republic of Tanzania. a Bacteriologically confirmed TB cases could not be verified for United Republic of Tanzania, so smear-positive TB prevalence rates are shown instead. Fig. 3.3a Estimated age-specific prevalence of bacteriologically confirmed pulmonary TB for surveys implemented in Africa in 2010−2016 The red line denotes the best estimate and the blue shaded areas are the 95% confidence intervals. In all countries, estimates of TB prevalence based on national surveys were much more precise than presurvey estimates (i.e. uncertainty intervals were much narrower). In most countries, best estimates based on surveys were also within the uncertainty interval of presurvey estimates. Best estimates of TB prevalence based on survey results were higher than presurvey estimates in 15 countries (most noticeably in Ghana, Indonesia, Lao People’s Democratic Republic, Malawi, Mongolia, the Philippines (2016) and the United Republic of Tanzania) and lower in 10 countries (most noticeably in Ethiopia, Gambia and Zimbabwe). 3.3 Trends in TB prevalence measured in repeat surveys Among countries that conducted prevalence surveys between 2007 and 2016, three countries had undertaken at least one survey in the preceding 20 years: Cambodia (2002), China (1990, 2000 and 2010) and the Philippines PART I: An overview of the 25 surveys implemented 2007–2016 39 DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; TB, tuberculosis. Fig. 3.3b Estimated age-specific prevalence of bacteriologically confirmed pulmonary TB for surveys implemented in Asia in 2007−2016 The red line denotes the best estimate and the blue shaded areas are the 95% confidence intervals. (1997 and 2007). Trends in TB prevalence based on surveys conducted since 2007 are shown in Fig. 3.6. The repeat surveys in Cambodia and China demonstrated that substantial reductions in TB prevalence can be achieved within 10  years. Observed reductions in the prevalence of smear-positive pulmonary TB in particular were consistent with the prioritization given to detection and cure of the most infectious cases within the framework of the DOTS strategy, which was recommended by WHO between the mid-1990s and 2006 (see also Chapter 1). The reduction in TB prevalence in China between 2000 and 2010 occurred during a period of nationwide expansion of DOTS (from half to all of the country). The reduction in Cambodia occurred during a period when DOTS services were expanded to health centres as well as hospitals, making TB diagnostic and treatment services much more accessible (2-4). However, the Philippines fourth national survey, in 2016, showed concerning results. Following a reduction in TB prevalence between 1997 and 2007, no decline occurred National TB prevalence surveys 2007–201640 Fig. 3.4 The sex ratio (male to female) of bacteriologically confirmed pulmonary TB cases detected in prevalence surveys implemented in 2007−2016 Fig. 3.5 Estimates of TB prevalence (all ages, all forms of TB) for 25 surveys (24 countries), before (in blue) and after (in red) results became available from national TB prevalence surveys implemented in 2007−2016 Countries are listed in decreasing order according to the before−after difference. DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; TB, tuberculosis; UR Tanzania, United Republic of Tanzania. a The sex ratio of smear−positive TB cases is shown for the United Republic of Tanzania. DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. PART I: An overview of the 25 surveys implemented 2007–2016 41 a The trend is for culture-confirmed cases. between 2007 and 2016. This may be linked to broader determinants of the TB epidemic, notably levels of poverty and undernutrition (5). Although not featured in this book, repeat surveys in Myanmar in 2018 and Viet Nam in 2017 showed large reductions in disease burden from 2009 to 2018 and from 2007 to 2017, respectively (6). 3.4 Proportion of survey cases reporting symptoms that met screening criteria A consistent finding in all surveys was that a high proportion of people with bacteriologically confirmed pulmonary TB did not report symptoms that met screening criteria. Although symptom screening criteria varied between countries (Table 3.2), only about half of the bacteriologically confirmed pulmonary TB cases Fig. 3.6 Trends in bacteriologically confirmed pulmonary TB measured in repeat surveys in Cambodia, China and the Philippines Shaded areas represent uncertainty intervals. (median 48%, range 21–70%) would have been identified if relying on symptom screening alone (Table 3.3). Other cases were identified due to chest X-ray screening. Among countries that used chronic cough alone as a symptom screening criterion, the proportion of people with bacteriologically confirmed pulmonary TB that did not report this symptom ranged from 36% in Nigeria to 79% in Mongolia. When chronic cough or haemoptysis (or both) were used, the proportion ranged from 43% in the Democratic People’s Republic of Korea and Indonesia to 79% in Myanmar. When cough and other TB-related symptoms were used, the proportion ranged from 30% in Malawi to 66% in Thailand (Fig. 3.7, Table 3.3). These findings can be explained by the fact that a prevalence survey identifies many people in the earlier stages of TB disease, before symptoms become more serious. These people will remain a source of transmission National TB prevalence surveys 2007–201642 Ta bl e 3. 3 Sc re en in g ou tc om es o f b ac te rio lo gi ca lly c on fir m ed p ul m on ar y TB c as es Co un tr y N um be r of ba ct er io lo gi ca lly co nfi rm ed T B ca se s Sy m pt om s cr ee ni ng d efi ni tio n a Sy m pt om po si tiv e, ch es t X -r ay po si tiv e Sy m pt om po si tiv e, c he st X- ra y ne ga tiv e/ N /A Sy m pt om ne ga tiv e, c he st X- ra y po si tiv e Ot he r sc re en in g ca te go ry a Pr op or tio n id en tifi ed by s ym pt om sc re en in g (% ) Pr op or tio n id en tifi ed b y ch es t X -r ay sc re en in g (% ) E th io pi a 1 1 0 C ou gh ≥ 2 w ee ks 4 5 1 2 5 3 0 5 2 % 8 9 % G ha na 2 0 2 C ou gh ≥ 2 w ee ks 6 7 1 5 8 5 3 5 4 1 % 7 5 % K en ya 3 0 5 C ou gh ≥ 2 w ee ks 1 1 5 3 2 1 5 4 4 4 8 % 8 8 % M on go lia 2 4 8 C ou gh ≥ 2 w ee ks 4 4 7 1 9 4 3 2 1 % 9 6 % N ig er ia 1 4 4 C ou gh ≥ 2 w ee ks 7 6 1 6 5 2 N /A 6 4 % 8 9 % S ud an 1 1 2 C ou gh ≥ 2 w ee ks 4 3 8 4 5 1 6 4 6 % 7 9 % U ga nd a 1 6 0 C ou gh ≥ 2 w ee ks 6 3 1 6 8 1 0 4 9 % 9 0 % Vi et N am 2 6 9 C ou gh ≥ 2 w ee ks 4 8 2 3 1 8 1 1 7 2 6 % 8 5 % C am bo di a 3 1 4 C ou gh ≥ 2 w ee ks o r ha em op ty si s 8 8 5 2 1 8 3 3 0 % 9 7 % C hi na 3 4 7 C ou gh ≥ 2 w ee ks o r ha em op ty si s 1 4 3 1 7 1 8 2 5 4 6 % 9 4 % D P R K or ea 3 4 0 C ou gh ≥ 2 w ee ks o r ha em op ty si s 1 8 7 7 1 4 6 N /A 5 7 % 9 8 % In do ne si a 4 2 6 C ou gh ≥ 2 w ee ks o r ha em op ty si s 2 1 7 2 5 1 8 4 0 5 7 % 9 4 % La o P D R 2 3 7 C ou gh ≥ 2 w ee ks o r ha em op ty si s 1 1 1 7 1 1 9 N /A 5 0 % 9 7 % M ya nm ar 3 1 1 C ou gh ≥ 3 w ee ks o r ha em op ty si s 6 5 1 2 3 1 1 4 2 1 % 9 5 % P hi lip pi ne s (2 0 1 6 ) 4 4 6 C ou gh ≥ 2 w ee ks o r ha em op ty si s 1 3 2 1 8 2 9 8 1 8 3 4 % 9 6 % B an gl ad es h 2 7 8 O th er 7 9 2 7 1 7 2 N /A 3 8 % 9 0 % G am bi a 7 7 O th er 3 2 1 2 3 3 0 5 7 % 8 4 % M al aw i 1 3 2 O th er 2 5 6 7 4 0 N /A 7 0 % 4 9 % P ak is ta n 3 4 1 O th er 1 5 7 4 1 1 4 2 1 5 8 % 8 8 % P hi lip pi ne s (2 0 0 7 ) b 1 3 6 O th er N /A N /A N /A N /A N /A N /A R w an da 4 0 O th er 1 5 4 2 1 0 N /A 9 0 % Th ai la nd 1 4 2 O th er 4 2 6 9 4 N /A N /A 9 6 % U R T an za ni a c 1 3 4 O th er 5 5 1 8 4 8 1 3 N /A 7 7 % Za m bi a 2 6 5 O th er 1 1 5 4 6 1 0 4 N /A N /A 8 3 % Zi m ba bw e 1 0 7 O th er 2 9 1 0 6 4 4 N /A 8 7 % D P R K or ea , D em oc ra ti c P eo pl e’ s R ep ub lic o f K or ea ; La o P D R , La o P eo pl e’ s D em oc ra ti c R ep ub lic ; N /A , no t ap pl ic ab le ; U R T an za ni a, U ni te d R ep ub lic o f Ta nz an ia . a O th er s cr ee ni ng c ri te ri a ge ne ra lly in cl ud ed m or e sy m pt om s or s ym pt om s of lo ng er d ur at io n. S ee c ou nt ry s pe ci fic c ha pt er s in P ar t I I f or d et ai ls . b In t he P hi lip pi ne s (2 0 0 7 ), a s ym pt om in te rv ie w w as n ot u se d to s cr ee n pa rt ic ip an ts . c In t he U ni te d R ep ub lic o f Ta nz an ia , on ly s m ea r- po si ti ve T B s ur ve y ca se s w er e re po rt ed . PART I: An overview of the 25 surveys implemented 2007–2016 43 until they experience symptoms that prompt them to seek health care. Even if they had sought care at an earlier stage, it is unlikely (with existing screening criteria) that they would have been referred for further laboratory testing on the basis of reported symptoms. As access to TB diagnostic and treatment services improve, the proportion of prevalent cases in the community that do report the ‘classic’ symptoms of pulmonary TB should fall. A prevalence survey in which a high proportion of cases do not report symptoms may indicate relatively good access to TB diagnosis and care, whereas a low proportion tends to suggest that access needs to be improved. An example of this was Nigeria, where many cases found in the survey already had symptoms that should have prompted care seeking and prompt diagnostic testing at health facilities. An increased proportion of cases not reporting symptoms in a repeat survey is consistent with improved health care services. This was a pattern found in the 2010–2011 survey in Cambodia, in which the prevalence of people with smear- positive pulmonary TB that reported symptoms fell by 56% compared with 2002. Fig. 3.7 Proportion of prevalent TB cases that were symptom−screen negative in surveys implemented in 2007−2016 DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The Philippines (2007) survey did not use symptom screening; however, symptom−related data were collected from all detected TB cases. Among those who do seek care, widening the use of chest X-ray screening in primary health care facilities and raising awareness among health care staff about the magnitude and characteristics of TB cases in the community could contribute to earlier diagnosis. 3.5 Detection and reporting gaps When measurements of prevalence are compared with official case notification data, prevalence surveys can identify gaps in detection and reporting. Overall ratios of prevalent (P) to notified (N) cases are shown in Fig. 3.8a, and ratios disaggregated by sex are shown in Fig. 3.8b–d.1 Ratios ranged from 0.62 in Gambia to 5.8 in Nigeria. For all countries except the Philippines in 2007 and Zimbabwe, the ratio was higher in men than women. Cross-country and male/female differences in the P:N ratio show that in several countries it should be possible 1 The P:N ratio is an approximate indicator (expressed in years) of case detection by the NTP (7). The higher the ratio, the longer the time taken for a prevalent case to be notified to the NTP. Some cases may exit the pool of prevalent cases without being notified, for example because they self-cure or die, or because they are detected and treated by providers not linked to official reporting systems. National TB prevalence surveys 2007–201644 Fig. 3.8a TB prevalence to TB notification (P:N) ratio in surveys implemented in 2007−2016 a Fig. 3.8b TB prevalence to TB notification (P:N) ratio (male) in surveys implemented in 2007−2016 a DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The comparison is for smear−positive pulmonary TB for all countries except for Bangladesh, DPR Korea, Kenya, Uganda and Zimbabwe, for which the comparison is for bacteriologically confirmed pulmonary TB. DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The comparison is for smear−positive pulmonary TB for all countries except for Bangladesh, DPR Korea, Kenya, Uganda and Zimbabwe, for which the comparison is for bacteriologically confirmed pulmonary TB. PART I: An overview of the 25 surveys implemented 2007–2016 45 Fig. 3.8c TB prevalence to TB notification (P:N) ratio (female) in surveys implemented in 2007−2016 a Fig. 3.8d Comparison of the TB prevalence to TB notification (P:N) ratio between men (green) and women (orange) in surveys implemented in 2007−2016 a DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The comparison is for smear−positive pulmonary TB for all countries except for Bangladesh, DPR Korea, Kenya, Uganda and Zimbabwe, for which the comparison is for bacteriologically confirmed pulmonary TB. DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The comparison is for smear-positive pulmonary TB for all countries except for Bangladesh, DPR Korea, Kenya, Uganda and Zimbabwe, for which the comparison is for bacteriologically confirmed pulmonary TB. National TB prevalence surveys 2007–201646 1 Cases detected before survey investigations are typically not as well documented as survey cases detected during investigations, particularly in countries where culture or Xpert MTB/RIF are not routinely used. 2 Results and lessons learned from this study were documented in the 2018 WHO global TB report (12). to achieve better (i.e. lower) ratios with strategies and technologies for TB diagnosis and treatment that are already available, and to close reporting and detection gaps for men. Although the burden of TB disease was consistently higher in men, P:N ratios were systematically lower among women, suggesting that women were accessing available diagnostic and treatment services more effectively (8). Development of strategies to improve care seeking and diagnosis among men are warranted in many countries. In some countries, P:N ratios also indicated that older people with TB were detected less effectively (Fig. 3.9). This may reflect financial and geographical accessibility barriers. Older people may also have greater tolerance of symptoms or associate symptoms with other chronic health conditions, leading to delayed care seeking and associated investigations. In Indonesia (9) and Viet Nam (10), the records of survey participants on treatment at the time of the survey were linked to the records of newly detected cases from routine TB surveillance, enabling the magnitude of underreporting of detected cases to be measured. In Indonesia, of the participants who reported that they were on TB treatment, only 19% (24/125) were identified in the national TB register, which helps to explain the high P:N ratio. In Viet Nam, 10% (37/353) of the participants that screened positive and were recently treated for TB had not been reported to the NTP. Whenever possible, future surveys should include comparison of the records of cases on treatment at the time of the survey with a national case-based electronic TB database, to assess the level of underreporting. Alternatively, or in addition,1 national inventory studies (11) can be used to measure levels of underreporting. A good example was the national inventory study in Indonesia, which was prompted by findings from the national TB prevalence survey.2 3.6 HIV testing and the prevalence of HIV Although HIV testing is a routine part of TB case management, collection of data about HIV status was not standardized in prevalence surveys implemented in 2007–2016. Reasons included variation in national HIV testing policies, the logistics of taking blood samples in the field, and the concern that survey participation might be negatively affected by refusing an HIV test. None of the surveys in Asia included HIV testing. HIV testing results or the HIV status of participants (or both) were obtained as part of the surveys in seven African countries: Kenya, Malawi, Rwanda, Uganda, the United Republic of Tanzania, Zambia and Zimbabwe (Table 3.4). HIV testing during field operations was done in only four countries: Rwanda, Uganda, the United Republic of Tanzania and Zambia. In Rwanda, Uganda and the United Republic of Tanzania, only those eligible for sputum examination were offered an HIV test. In Zambia, HIV testing was offered to all survey participants. In Zambia, 2063 (6.7%) of those tested were HIV-positive. In Rwanda, Uganda and the United Republic of Tanzania, the proportions of those tested who were HIV-positive were 4.9%, 9.6% and 5.0%, respectively. In Malawi, all participants were asked if they had ever been tested for HIV, and were invited to disclose their status; verbal acknowledgement of HIV status was provided for 19  703 (62%) participants, of which 1840 (9.3%) reported that they were HIV-positive. In Kenya and Zimbabwe, records of survey cases were linked to records from routine HIV treatment and care programmes. The proportion of survey cases who were HIV-positive was 13% in Kenya and 51% in Zimbabwe. HIV prevalence among prevalent TB cases was systematically lower than HIV prevalence among newly notified cases (Fig. 3.10), probably reflecting the faster progression of TB disease in people living with HIV, which prompts earlier care seeking. It is also plausible that the expansion of HIV care programmes since the early 2000s contributed to earlier detection and treatment of TB among people living with HIV. 3.7 Health care seeking behaviour Patterns of health care seeking behaviour can help to identify actions that could be taken to shorten the time to TB diagnosis and treatment. They may also indicate care providers that need to be better engaged with the NTP, including to ensure reporting of detected cases. Although there was limited standardization in the data on health care seeking behaviour that were collected during surveys implemented in 2007–2016, it was clear that a large proportion of symptomatic participants had not sought care before the survey (Table 3.5). The median proportion of those reporting symptoms that PART I: An overview of the 25 surveys implemented 2007–2016 47 Fig. 3.9 TB prevalence to TB notification (P:N) ratio by age group in surveys implemented in 2007−2016 a DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. a The comparison is for smear-positive pulmonary TB for all countries except for Bangladesh, DPR Korea, Kenya, Uganda and Zimbabwe, for which the comparison is for bacteriologically confirmed pulmonary TB. National TB prevalence surveys 2007–201648 Ta bl e 3. 4 HI V st at us o f p ar tic ip an ts a nd b ac te rio lo gi ca lly c on fir m ed p ul m on ar y TB c as es Pa rt ic ip an ts Pa rt ic ip an ts w ho s cr ee ne d po si tiv e Ba ct er io lo gi ca lly c on fir m ed p ul m on ar y TB c as es Co un tr y Nu m be r o f pa rt ic ip an ts wh o we re te st ed fo r H IV , o r h ad do cu m en te d HI V st at us % a HI V- po si tiv e % HI V- ne ga tiv e % To ta l sc re en ed po si tiv e Nu m be r o f pa rt ic ip an ts wh o we re te st ed fo r HI V, o r w ith do cu m en te d HI V st at us % HI V- po si tiv e % HI V- ne ga tiv e % To ta l TB ca se s Nu m be r wh o we re te st ed fo r HI V or w ith re po rt ed HI V st at us % HI V- po si tiv e % HI V- ne ga tiv e % K en ya b N /A N /A N /A N /A N /A N /A 9 7 1 5 N /A N /A N /A N /A N /A N /A 3 0 5 2 4 5 8 0 % 4 1 1 7 % 2 0 4 6 7 % M al aw i c 1 9 7 0 3 6 2 % 1 8 4 0 9 .3 % 1 7 8 6 3 9 1 % 3 4 3 2 2 0 6 6 6 0 % 3 3 9 1 6 % 1 7 0 8 8 3 % 1 3 2 7 8 a 5 9 % 2 2 2 8 % 5 2 6 7 % R w an da d N /A N /A N /A N /A N /A N /A 4 7 4 7 4 4 4 5 9 4 % 2 1 8 4 .9 % 4 2 2 7 9 5 % 4 0 3 6 9 0 % 1 2 .8 % 3 5 9 7 % U ga nd a d N /A N /A N /A N /A N /A N /A 5 1 4 2 4 3 8 6 8 5 % 4 2 2 9 .6 % 3 9 6 4 9 0 % 1 6 0 1 4 5 9 1 % 3 9 2 7 % 1 0 6 7 3 % U R T an za ni a d, e N /A N /A N /A N /A N /A N /A 6 3 0 2 6 3 0 2 1 0 0 3 1 8 5 .0 % 5 9 8 4 9 5 % N /A N /A N /A N /A N /A N /A N /A Za m bi a f 3 0 5 8 4 6 6 % 2 0 6 2 6 .7 % 2 8 5 2 2 9 3 % 6 7 0 8 N /A N /A N /A N /A N /A N /A 2 6 5 1 3 4 5 1 % 3 6 2 7 % 9 8 7 3 % Zi m ba bw e g N /A N /A N /A N /A N /A N /A 5 8 2 0 N /A N /A N /A N /A N /A N /A 1 0 7 8 3 7 8 % 4 2 5 1 % 4 1 4 9 % H IV , hu m an im m un od efi ci en cy v ir us ; N /A , no t ap pl ic ab le ; U R T an za ni a, U ni te d R ep ub lic o f Ta nz an ia . a Fo r M al aw i a nd Z am bi a, t he d en om in at or is t he t ot al n um be r of s ur ve y pa rt ic ip an ts a s sh ow n in T ab le 3 .1 . b In K en ya , H IV t es ti ng w as n ot d on e as p ar t of t he s ur ve y. H IV d at a w er e ob ta in ed f ro m t he n at io na l H IV r ep or ti ng s ys te m o f lin ke d TB c as es . c In M al aw i, th er e w as v er ba l r ep or ti ng o f H IV s ta tu s on ly . O f 2 0 6 6 p ar ti ci pa nt s w ho s cr ee ne d po si ti ve , 1 9 h ad u nk no w n st at us . Fo r 4 8 o ut o f 7 8 b ac te ri ol og ic al ly c on fir m ed T B c as es , th ei r H IV s ta tu s w as u nk no w n. d I n R w an da , U ga nd a an d th e U ni te d R ep ub lic o f Ta nz an ia , on ly t ho se w ho s cr ee ne d po si ti ve w er e te st ed f or H IV d ur in g fie ld o pe ra ti on s. e In t he U ni te d R ep ub lic o f Ta nz an ia , ba ct er io lo gi ca lly c on fir m ed T B c as es c ou ld n ot b e ve ri fie d. f In Z am bi a, a ll su rv ey p ar ti ci pa nt s w er e in vi te d to b e te st ed f or H IV d ur in g fie ld o pe ra ti on s. g In Z im ba bw e, a ll ba ct er io lo gi ca lly c on fir m ed T B c as es d et ec te d by t he s ur ve y w er e of fe re d H IV c ou ns el lin g an d te st in g as p ar t of r ou ti ne t re at m en t m an ag em en t an d w er e no t di re ct ly t es te d as p ar t of t he s ur ve y. PART I: An overview of the 25 surveys implemented 2007–2016 49 met screening criteria who had not yet sought care was 42% (range, 10–67%), suggesting that there are barriers to accessing health services. Among those that had sought health care, most did so within the public sector (Table 3.6). In a few countries (mostly in Asia), 30% or more of the symptomatic participants sought care in the private sector; examples included Bangladesh, Indonesia, Malawi, Myanmar and the Philippines (in 2016). Pharmacies were also an important point of care in a few countries, especially in Asia. The observed proportion of cases treated in the private health sector is a useful measure of the need for engagement of NTPs with the private sector. 3.8 Diagnostic performance of smear microscopy High proportions of false-positive results from direct microscopic examination of smears were observed in several surveys (Table 3.7). In these surveys, TB was ruled out based on results from culture and Xpert MTB/ RIF (or LPA), with false-positive results probably due to nontuberculous mycobacteria. These findings provide evidence that sputum smear microscopy is also likely to be an unreliable diagnostic test for TB in the context of active case finding, unless high positive predictive values can be demonstrated in the population group targeted by active case finding. Commonly used diagnostics – particularly direct microscopic examination of sputum smear samples – need to be upgraded with better technology, including WHO-approved rapid diagnostics that are more sensitive and more specific than sputum microscopy. 3.9 Conclusions The 25 national TB prevalence surveys implemented in Africa and Asia between 2007 and 2016 provided a better understanding of the national, regional and global burden of TB disease, and of gaps in TB detection and treatment. The surveys showed a much higher burden in men than women, an ageing epidemic in most of Asia and a peak in prevalence in the younger age groups in most African countries. They also showed that actions are needed to improve access to health care and to ensure prompt diagnosis when care is sought, especially among men. Repeat surveys in Asian countries have demonstrated that substantial reductions in the burden of TB disease can be achieved within 10 years, and all 25 surveys provide a valuable baseline for future assessment of trends. Fig. 3.10 HIV prevalence in TB survey cases compared with HIV prevalence in notified TB cases expressed as a ratio, in surveys implemented in 2007−2016 UR Tanzania, United Republic of Tanzania. National TB prevalence surveys 2007–201650 Ta bl e 3. 5 He al th c ar e se ek in g be ha vi ou r a m on g pa rt ic ip an ts w ho w er e sy m pt om -s cr ee n po si tiv e Co un try Pa rti ci pa nt s wh o we re sy m pt om - sc re en po sit ive No ac tio n ta ke n % Lo ca tio n of c ar e so ug ht Co ns ul te d m ed ic al fa ci lit y % Ty pe o f f ac ili ty Ph ar - m ac y % Tr ad i- tio na l % Ot he r % Un sp ec - ifi ed % Se lf- tre at ed % Un - kn ow n % Pu bl ic fa ci lit y % Pr iva te fa ci lit y % Ot he r fa ci lit y % Af ri ca E th io pi a 3 0 2 6 1 9 3 2 6 4 % 8 4 8 2 8 % 6 2 8 7 4 % 1 9 9 2 3 % 2 1 2 .5 % 4 0 1 .3 % 3 0 .1 0 % N /A N /A 5 5 1 .8 % N /A N /A 1 4 8 4 .8 % G am bi a 3 4 6 2 1 4 2 4 4 1 % 1 7 0 6 4 9 % 1 3 9 8 8 2 % 2 2 0 1 3 % 8 8 5 .2 % 1 7 0 .4 9 % 1 4 0 .4 0 % 2 4 0 .6 9 % N /A N /A N /A N /A 2 7 7 8 .0 % G ha na 1 9 6 9 2 6 4 1 3 % 7 9 3 4 0 % 6 9 5 8 8 % 6 1 7 .7 % 3 7 4 .7 % 3 2 4 1 7 % 2 0 1 .0 % N /A N /A N /A N /A 5 6 7 2 9 % 1 0 .1 0 % K en ya a 4 1 3 7 2 7 6 3 6 7 % 1 2 5 7 3 0 % 1 0 4 7 N /A 1 9 8 N /A 3 N /A 5 6 N /A 9 N /A N /A N /A N /A N /A N /A N /A 1 1 7 2 .8 % M al aw i 2 7 1 5 1 0 9 6 4 0 % 1 2 8 0 4 7 % 9 0 1 7 0 % 3 7 9 3 0 % N /A N /A 3 2 1 .2 % 4 1 1 .5 % 4 0 .1 5 % N /A N /A 2 3 6 8 .7 % 2 6 0 .9 6 % N ig er ia 2 4 6 6 6 0 4 2 4 % 8 0 0 3 2 % 6 2 8 7 9 % 1 7 2 2 1 % N /A N /A 3 1 9 1 3 % 1 1 0 .4 5 % 9 0 .3 6 % 3 0 .1 2 % 6 8 0 2 8 % 4 0 1 .6 % R w an da a 2 8 5 5 1 9 3 4 6 8 % 9 2 1 3 2 % 9 4 1 N /A 4 8 N /A 3 8 N /A 1 0 1 N /A 5 4 N /A N /A N /A N /A N /A 0 0 N /A N /A S ud an 2 6 6 3 5 7 5 2 2 % 1 3 0 8 4 9 % 1 0 7 7 8 2 % 9 0 6 .9 % 1 4 1 1 1 % 5 2 2 .0 % 4 9 1 .8 % N /A N /A 6 9 2 .6 % N /A N /A 6 1 0 2 3 % U ga nd a 2 7 1 4 1 0 5 9 3 9 % 1 2 0 1 4 4 % 1 0 3 8 8 6 % 1 4 6 1 2 % 1 7 1 .4 % 4 2 1 1 6 % 1 1 0 .4 1 % N /A N /A N /A N /A 2 2 0 .8 1 % 0 0 % U R T an za ni a 3 3 8 8 1 6 8 8 5 0 % 4 8 1 1 4 % 4 4 5 9 3 % 3 6 7 .5 % N /A N /A 1 4 7 4 .3 % 1 1 0 .3 2 % 2 5 7 7 .6 % 1 5 5 4 .6 % N /A N /A 6 4 9 1 9 % Za m bi a 4 4 5 3 2 5 3 4 5 7 % 1 8 2 9 4 1 % 1 6 8 0 9 2 % 7 5 4 .1 % 7 4 4 .0 % 1 6 0 .3 6 % 1 0 .0 2 % N /A N /A N /A N /A N /A N /A 7 3 1 .6 % Zi m ba bw ea 1 8 3 3 1 1 3 0 6 2 % 4 8 6 2 6 % 4 3 8 N /A 4 5 N /A N /A N /A 1 7 N /A 1 3 N /A N /A N /A N /A N /A N /A N /A 2 1 7 1 2 % As ia B an gl ad es hb 2 6 8 8 2 1 2 9 4 7 4 8 % 6 5 4 5 2 4 % 1 8 1 6 2 8 % 2 1 8 2 3 3 % 2 5 4 7 3 9 % 6 5 3 3 2 4 % 2 3 0 .1 0 % 1 9 1 0 .7 1 % N /A N /A 6 4 3 2 .4 % N /A N /A C am bo di a 1 9 1 6 1 9 7 1 0 % 1 2 6 1 6 6 % 9 4 7 7 5 % 3 0 5 2 4 % 9 0 .7 1 % 4 0 1 2 1 % 2 1 1 .1 0 % 6 0 .3 1 % N /A N /A 2 8 1 .5 % 2 0 .1 0 % C hi na 5 4 6 2 N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A D P R K or ea 2 9 4 4 1 1 9 2 4 1 % 1 7 4 3 N /A 1 7 4 3 1 0 0 % N /A N /A N /A N /A N /A N /A 3 0 .1 0 % N /A N /A N /A N /A 0 0 % 6 0 .2 0 % In do ne si a 8 5 5 2 3 6 8 5 4 3 % 2 2 3 1 2 6 % 1 1 7 8 5 3 % 6 7 2 3 0 % 3 8 1 1 7 % 2 6 3 6 3 1 % N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A La o P D R 3 2 3 9 1 2 1 0 3 7 % 1 1 4 8 3 5 % 9 9 0 8 6 % 1 0 6 9 .2 % 5 2 4 .5 % 6 9 0 2 1 % 2 6 0 .8 0 % N /A N /A N /A N /A N /A N /A 1 6 5 5 .1 % M on go lia 2 5 4 6 1 1 7 9 4 6 % 9 5 0 3 7 % 9 2 0 9 7 % 3 0 3 .1 % N /A N /A 2 2 2 8 .7 % 2 0 .0 8 % 5 9 2 .3 % N /A N /A N /A N /A 3 0 1 .2 % M ya nm ar 1 6 9 1 4 4 0 2 6 % 3 6 3 2 2 % 1 9 7 5 4 % 1 6 6 4 6 % N /A N /A 2 7 1 1 6 % 2 4 3 1 4 % 3 9 2 .3 % N /A N /A 3 0 7 1 8 % 2 8 1 .7 % P ak is ta n 5 4 1 7 N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A P hi lip pi ne s (2 0 0 7 )c N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A P hi lip pi ne s (2 0 1 6 )a 2 8 1 5 1 1 4 2 4 1 % 5 3 0 N /A 3 5 9 6 7 % 1 6 2 3 1 % 9 1 .7 % 4 N /A 1 0 N /A N /A N /A N /A N /A 1 1 3 0 N /A 1 8 0 .6 4 % Th ai la nd d 2 2 8 3 N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A N /A Vi et N am 4 1 7 2 2 2 4 8 5 4 % 1 2 2 8 2 9 % 1 0 2 9 8 4 % 1 9 9 1 6 % N /A N /A 6 7 1 1 6 % 2 5 0 .6 0 % N /A N /A N /A N /A N /A N /A N /A N /A D P R K or ea , D em oc ra ti c P eo pl e’ s R ep ub lic o f K or ea ; La o P D R , La o P eo pl e’ s D em oc ra ti c R ep ub lic ; N /A , no t ap pl ic ab le ; U R T an za ni a, U ni te d R ep ub lic o f Ta nz an ia . a a In K en ya , th e P hi lip pi ne s (2 0 1 6 ), R w an da a nd Z im ba bw e pa rt ic ip an ts c ou ld s el ec t m or e th an o ne c at eg or y. b In B an gl ad es h, d at a on h ea lt h ca re s ee ki ng b eh av io ur w er e av ai la bl e fo r pa rt ic ip an ts w ho r ep or te d at le as t on e TB s ym pt om ( i.e . co ug h, h ae m op ty si s, w ei gh t lo ss , fe ve r, ni gh t sw ea ts ). c In P hi lip pi ne s (2 0 0 7 ), t he re w as n o sy m pt om s cr ee ni ng a nd t he re fo re n o he al th s ee ki ng b eh av io ur d at a ob ta in ed . d In T ha ila nd , lim it ed d at a w er e av ai la bl e an d w er e no t re po rt ed . PART I: An overview of the 25 surveys implemented 2007–2016 51 Table 3.6 Location of treatment for participants who were on treatment at the time of the survey Country Number of participants who were on treatment at the time of the survey Public sector % Private sector % Other sector % Unknown sector Location of treatment for participants who were on treatment at the time of the survey Africa Ethiopia 75 54 72% 7 9.3% 3 4.0% 11 15% Gambia 38 38 100% 0 0% 0 0% 0 0% Ghana 48 42 88% 1 2.1% 5 10% 0 0% Kenyaa 62 23 37% 0 0% 1 1.6% 38 61% Malawia 12 10 83% 2 17% 0 0% 0 0% Nigeria 82 56 68% 14 17% 5 6.1% 7 8.5% Rwanda 21 – – – – – – – – Sudan 104 69 66% 1 1.0% 4 3.8% 30 29% Uganda 61 57 93% 4 6.6% 0 0% 0 0% UR Tanzania 88 – – – – – – – – Zambia 114 61 54% 1 0.9% 0 0% 52 46% Zimbabwe 84 – – – – – – – – Asia Bangladesh 57 16 28% 10 18% 18 32% 13 23% Cambodia 80 72 90% 6 7.5% 0 0% 2 2.5% China 73 72 99% 0 0% 1 1.4% 0 0% DPR Korea 106 101 95% 0 0% 0 0% 5 4.7% Indonesia 125 68 54% 52 42% 5 4.0% 0 0% Lao PDR 42 21 50% 0 0% 0 0% 21 50% Mongolia 129 126 98% 0 0% 3 2.3% 0 0% Myanmar 79 63 80% 14 18% 0 0% 2 2.5% Pakistan 146 – – – – – – – – Philippines (2007) – – – – – – – – – Philippines (2016)b 170 134 79% 15 8.8% 24 14% 1 0.59% Thailand 66 53 80% 3 4.5% 3 4.5% 7 11% Viet Nam 64 46 72% 2 3.1% 0 0% 16 25% DPR Korea, Democratic People’s Republic of Korea; Lao PDR, Lao People’s Democratic Republic; UR Tanzania, United Republic of Tanzania. – no data were available. a In Kenya and Malawi, data were available only for participants who were eligible for sputum submission. b In the Philippines (2016), some participants identified more than one location. National TB prevalence surveys 2007–201652 a Results are shown for surveys in which specimens were tested using smear microscopy and the systematic use of rapid molecular tests. All surveys used Xpert MTB/RIF except Sudan which used line probe assay (LPA). Bangladesh and Kenya used both culture and Xpert MTB/RIF whereas other surveys used Xpert (or LPA) to confirm smear-positive specimens only. References 1. Hong YP, Kim SJ, Lew WJ, Lee EK, Han YC. The seventh nationwide tuberculosis prevalence survey in Korea, 1995. Int J Tuberc Lung Dis. 1998;2(1):27–36 (https://www.ncbi.nlm.nih. gov/pubmed/9562108, accessed 22 November 2019). 2. Mao TE, Okada K, Yamada N, Peou S, Ota M, Saint S et al. Cross- sectional studies of tuberculosis prevalence in Cambodia between 2002 and 2011. Bull World Health Organ. 2014;92(8):573–81 (https://www.ncbi.nlm.nih.gov/pubmed/25177072, accessed 28 November 2019). 3. Tupasi TE, Radhakrishna S, Chua JA, Mangubat NV, Guilatco R, Galipot M et al. Significant decline in the tuberculosis burden in the Philippines ten years after initiating DOTS. Int J Tuberc Lung Dis. 2009;13(10):1224–30 (https://www.ncbi.nlm.nih.gov/ pubmed/19793426, accessed 28 November 2019). 4. Wang L, Zhang H, Ruan Y, Chin DP, Xia Y, Cheng S et al. Tuberculosis prevalence in China, 1990–2010; a longitudinal analysis of national survey data. Lancet. 2014;383(9934):2057–64 (https://www.ncbi.nlm.nih.gov/pubmed/24650955, accessed 28 November 2019). 5. Global tuberculosis report 2017. Geneva: World Health Organization; 2017 (http://apps.who.int/iris/bitstre am/10665/259366/1/9789241565516-eng.pdf?ua=1, accessed 1 February 2020). 6. Global tuberculosis report 2019. Geneva: World Health Organization; 2019 (https://apps.who.int/iris/bitstream/hand le/10665/329368/9789241565714-eng.pdf, accessed 1 February 2020). 7. Borgdorff MW. New measurable indicator for tuberculosis case detection. Emerg Infect Dis 2004: 10: 1523–1528. Table 3.7 Percentage of smear-positive results that were not confirmed TB. Results shown for surveys in which specimens were tested using smear microscopy, rapid molecular tests and culture a Country Number of participants with at least one smear-positive specimen Participants with smear-positive specimens excluded as a TB case Number % Bangladesh 125 17 14% Ghana 198 138 70% Indonesia 291 126 43% Kenya 141 18 13% Malawi 163 101 62% Mongolia 92 5 5.4% Pakistan 236 29 12% Philippines (2016) 183 10 5.5% Sudan 61 4 6.6% Uganda 91 25 27% Zambia 356 221 62% Zimbabwe 206 183 89% 8. Horton KC, MacPherson P, Houben RM, White RG, Corbett EL. Sex Differences in tuberculosis burden and notifications in low- and middle-income countries: a systematic review and meta- analysis. PLoS Med 2016: 13: e1002119. 9. Ministry of Health – Republic of Indonesia, National Institute of Health Research and Development. Indonesia tuberculosis prevalence survey 2013–2014. Jakarta, Indonesia: 2015. 10. Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, EW T. Diagnosis and treatment of tuberculosis in the private sector, Vietnam. Emerging infectious diseases. 2011;17(3):562–4 (https://www.ncbi.nlm.nih.gov/pubmed/21392464, accessed 28 November 2019). 11. World Health Organization. Assessing tuberculosis under- reporting through inventory studies. Geneva, Switzerland 2012 (https://www.who.int/tb/publications/inventory_studies/en/, accessed 20 November 2019). 12. Global tuberculosis report 2018. Geneva: World Health Organization; 2018 (https://apps.who.int/iris/bitstream/han dle/10665/274453/9789241565646-eng.pdf?ua=1, accessed 1 February 2020). 53 Chapter 4 Successes, challenges and lessons learned Chapter 3 provided an overview of the main results from the 25 national TB prevalence surveys completed between 2007 and the end of 2016, including what they showed about the distribution of TB disease by age and sex. The overview also showed trends over time (for any countries that completed repeat surveys) and discussed the implications of all results for policy, planning and programmatic action. In addition to the major success of producing valuable new data, this chapter highlights other aspects of survey success. It also identifies the main challenges that were faced during the process from deciding to implement a survey through to finalizing and disseminating the results. Lessons learned from both survey successes and challenges, which should be useful for informing future surveys, are then summarized. For all three topics, this chapter synthesizes the more detailed assessments of successes, challenges and lessons learned that are reported by those who led or contributed to each survey in the country-specific chapters (Part II) of this book. 4.1 Successes Survey successes are summarized in Table 4.1. All surveys provided an up-to-date direct measurement of the burden of TB disease, and other valuable information about the status of the TB epidemic and access to care. This information was used to inform national policy, national strategic plans, advocacy and resource mobilization. Of the 25 surveys, 21 were in countries that completed either their first-ever national TB prevalence survey (n=18) or the first survey to include culture testing (n=3) according to the screening and diagnostic algorithm recommended in the lime book (1). In 2011, Ethiopia became the first African country in decades to implement a national survey using this algorithm; also impressive was the short time (about 1 year) between the decision by Ethiopia’s Ministry of Health to conduct a survey and the start of field operations. Three countries completed repeat surveys that enabled assessment of trends in TB disease burden: Cambodia, China and the Philippines. Most countries (19 of 25) also succeeded in achieving a high participation rate (more than 80%). Nine countries (Bangladesh, Cambodia, China, Ethiopia, Ghana, the Philippines in 2007, Rwanda, Uganda and Viet  Nam) managed to achieve participation rates of 90% or more, with an exceptionally high participation rate (96%) in China and Rwanda. Other survey successes identified by multiple countries included good data management (n=6), a strong laboratory (n=5), and timely finalization and dissemination of results (n=8). Surveys that described laboratory work as a “survey success” included those in which the laboratory used was either part of a long- established research unit (e.g. Gambia) or a national reference laboratory. In the survey in Uganda, the national reference laboratory was also a supranational reference laboratory. Survey successes mentioned by a single country were: • the ability to generate subnational (provincial) as well as national estimates of prevalence (China, reflecting the survey’s very large sample size); • full domestic funding for the survey (China); • capacity development for health care workers during the survey (Cambodia); • smooth field operations (Cambodia); • the enhancement of laboratory and operational research capacity (Ghana); and • the opportunity to see challenges in case management and surveillance in the most remote areas of the country, often for the first time (Lao People’s Democratic Republic). 4.2 Challenges The major challenges faced in surveys are summarized in Table 4.2, with the top five challenges shown in Fig. 4.1. National TB prevalence surveys 2007–201654 Ta bl e 4. 1 Su rv ey s uc ce ss es a s re po rt ed b y co un tr ie s (s ee P ar t I I f or d et ai ls ) Co un tr y Fi rs t n at io na l su rv ey c om pl et ed (e ve r or fo r m an y ye ar s) a Re pe at na tio na l su rv ey co m pl et ed U p- to -d at e di re ct m ea su re m en t o f T B di se as e bu rd en a nd o th er v al ua bl e in fo rm at io n ab ou t t he s ta tu s of th e TB ep id em ic a nd a cc es s to c ar e pr ov id ed Di re ct m ea su re m en t of tr en ds in T B di se as e bu rd en H ig h pa rt ic ip at io n ra te ( > 80 % ) Go od d at a m an ag em en t b St ro ng la bo ra to ry c Ti m el y fin al iz at io n an d di ss em in at io n of r es ul ts d B an gl ad es h ● ● ● ● C am bo di a ● ● ● ● ● C hi na ● ● ● ● ● D P R K or ea ● ● ● E th io pi a ● ● ● G am bi a ● ● ● G ha na ● ● ● ● In do ne si a ● ● ● K en ya ● ● ● La o P D R ● ● ● M al aw i ● ● ● ● ● M on go lia ● ● ● ● M ya nm ar ● ● ● ● ● N ig er ia ● ● P ak is ta n ● ● ● P hi lip pi ne s (2 0 0 7 ) ● ● ● ● P hi lip pi ne s (2 0 1 6 ) ● ● ● ● R w an da ● ● ● S ud an ● ● ● Th ai la nd ● ● U ga nd a ● ● ● ● ● ● U R T an za ni a ● ● Vi et N am ● ● ● ● Za m bi a ● ● ● ● ● Zi m ba bw e ● ● ● ● a Th e su rv ey i n B an gl ad es h w as t he fi rs t na ti on al s ur ve y th at u se d th e m et ho ds r ec om m en de d in t he li m e bo ok ( 1) ; th e su rv ey i n E th io pi a w as t he fi rs t in t he c ou nt ry a s w el l as t he fi rs t in d ec ad es i n A fr ic a th at u se d cu lt ur e; t he su rv ey in I nd on es ia w as t he fi rs t fo r de ca de s us in g bo th X -r ay a nd c ul tu re ; an d th e su rv ey in M ya nm ar w as t he fi rs t in t he c ou nt ry t o us e cu lt ur e. b C ou nt ri es t ha t ha d a da ta m an ag em en t pl an , ha d no m aj or d at a is su es in t he fi el d, a nd t oo k < 1 y ea r to c le an d at a af te r fie ld o pe ra ti on s w er e co m pl et ed . c C ou nt ri es t ha t di d no t ha ve h av e la bo ra to ry p ro to co l v io la ti on s an d ha d hi gh c ul tu re c on fir m at io n of s m ea r- po si ti ve c as es ( ≥ 8 5 % ). d C ou nt ri es t ha t di d no t ha ve lo ng d el ay s be fo re r es ul ts w er e ac ce pt ed b y pu bl ic h ea lt h au th or it ie s, a nd p ro vi de d an o ffi ci al r ep or t an d/ or p ap er w it hi n a fe w y ea rs o f co m pl et in g fie ld o pe ra ti on s. PART I: An overview of the 25 surveys implemented 2007–2016 55 Ta bl e 4. 2 M aj or c ha lle ng es fa ce d in s ur ve ys a s re po rt ed b y co un tr ie s (s ee P ar t I I f or d et ai ls ) Co un tr y Ti m e to s ec ur e fu nd in g or fu nd in g in te rr up tio ns du ri ng s ur ve y Le ng th y pr oc es s to p ro cu re X- ra y eq ui pm en t Se cu ri ty is su es Ga ps be tw ee n po pu la tio n in n at io na l an d su rv ey ce ns us In te rn al m ig ra tio n af fe ct in g re si de nt ia l el ig ib ili ty cr ite ri a Pa rt ic ip at io n (≤ 80 % )a Ov er he at in g or b re ak do w n of X -r ay m ac hi ne s du ri ng fi el d op er at io ns Da ta m an ag em en t La bo ra to ry w or k (p ri m ar ily is su es re la te d to c ul tu re te st in g) De la ys in ce nt ra l r ea di ng of X -r ay s or di ffi cu lti es in r et ai ni ng ra di ol og is ts De la ys in w ri tin g th e su rv ey re po rt B an gl ad es h ● ● ● C am bo di a ● ● ● C hi na ● ● ● D P R K or ea ● ● ● E th io pi a ● ● ● ● G am bi a ● ● ● G ha na ● ● ● ● ● In do ne si a ● ● ● K en ya ● ● ● La o P D R ● ● ● M al aw i ● ● ● ● ● M on go lia ● M ya nm ar ● ● ● ● ● N ig er ia ● ● ● ● ● P ak is ta n ● ● ● P hi lip pi ne s (2 0 0 7 ) ● ● P hi lip pi ne s (2 0 1 6 ) ● ● ● R w an da ● ● ● S ud an ● ● ● ● Th ai la nd ● ● ● U ga nd a ● ● U R T an za ni a ● ● ● ● Vi et N am ● Za m bi a Zi m ba bw e ● ● ● a In a dd it io n, m an y co un tr ie s re po rt ed c ha lle ng es w it h pa rt ic ip at io n in a t le as t on e of t he f ol lo w in g su bc at eg or ie s: t he fi rs t su rv ey c lu st er s, y ou ng er a ge g ro up s, m en a nd u rb an ( es pe ci al ly w ea lt hi er ) ar ea s. ● y es National TB prevalence surveys 2007–201656 The top challenge, identified by 16 countries, was laboratory-related work (primarily issues related to culture testing). Examples of such challenges included: • potential cross-contamination of samples during transportation from the field to the laboratory (e.g. Bangladesh and Malawi); • a need to rely on only two laboratories owing to difficulties in standardizing laboratory work (Cambodia); • the difficulty of standardizing techniques when multiple laboratories were used (e.g. China and the Philippines); • a lower yield than expected from culture specimens (e.g. China, Pakistan, Rwanda and the United Republic of Tanzania); • backlogs and delays in culture inoculation, linked to the high volume of specimens (e.g. Sudan and the United Republic of Tanzania); • testing of only one specimen (instead of the recommended number of two) using culture in some (e.g. Indonesia) or all (e.g. Ethiopia) clusters owing to limited laboratory capacity; • the time required to establish the laboratory capacity needed for culture testing (e.g. this took 2 years in Lao People’s Democratic Republic); Fig. 4.1 Top five challenges in 25 surveys as reported by countries • use of sputum cups that were suboptimal for culture testing (Myanmar); • security issues in the part of the country where the national reference laboratory was located, which limited monitoring and technical assistance (Nigeria); and • difficulties maintaining a cold chain, especially in hot or heatwave conditions (e.g. Pakistan and the Philippines). Despite these issues, in 15 of 16 countries, the number of culture-confirmed survey cases was considerably higher than the number of survey cases that were smear positive, as expected. The exception was the United Republic of Tanzania, for which it was concluded that culture results could not be used (and hence the results in this book are restricted to smear-positive cases). The second most frequent challenge, identified by 11 countries, was data management. Examples of such challenges included: • slow data entry (e.g. Cambodia, the Democratic People’s Republic of Korea and Nigeria); • use of software designed for a national census that was not suited to a prevalence survey (e.g. Ethiopia); • overreliance on internet connectivity in the field for electronic data entry (Kenya and Sudan), which was later resolved through use of a local area PART I: An overview of the 25 surveys implemented 2007–2016 57 network (Kenya) or paper-based recording of data as a backup (Sudan); • linking data from the field with laboratory results when different data management systems were used (e.g. Kenya); • transcription errors and serious difficulties in matching records for the same individual when multiple paper-based forms were used to collect data (e.g. Pakistan, Rwanda and the United Republic of Tanzania) – this caused long delays in the production of a final, clean dataset in Pakistan (>1 year) and the United Republic of Tanzania, while intensive efforts were needed by the survey team in Rwanda to successfully ameliorate the problem; • difficulties with the data management system that were hard to resolve until the survey implementing agency (rather than an externally contracted separate agency) assumed direct responsibility for it (e.g. the Philippines); and • delayed sharing of datasets and different data management processes between the survey team and the national statistics agency (e.g. Zimbabwe). Challenges related to X-ray equipment – either the initial process to procure it (eight countries) or breakdowns or overheating in the field (eight countries) – were also common. One or both of these two challenges affected 10 countries in total. Producing a final survey report was a considerable challenge in 10 countries. The main reasons were the lack of a budget specifically for this activity, that the contracts of members of the survey team who could have worked on the report expired before they were able to spend time on writing the report, and no funding was available for the option of employing someone to help the survey team to produce it. The other challenges identified by at least three countries were: • the time taken to secure funding, or interruptions to funding during the survey; • security issues; • discrepancies between the national census data and the survey census; • internal migration, which affected the proportion of the population eligible to participate according to residential criteria; • participation; and • delays in central reading of X-rays. Even in countries that achieved high participation rates, many countries experienced challenges with participation in at least one of the following subcategories: the first survey clusters, younger age groups, men, and urban (especially wealthier) areas. Other challenges mentioned by a single country were: • interruption to field operations during the long winter season (Mongolia); • expiry of X-ray software licences due to delays in starting the survey, with the software then having to be repurchased (Nigeria); • lack of access to national census data by the survey team responsible for the prevalence survey, and the changing of bureau of statistics staff for each cluster (Nigeria); • extreme rainfall that forced field operations to be suspended for 1 month (Nigeria); • a natural disaster (a flood) that delayed field operations (Pakistan); • high staff turnover (Sudan); • some recommendations from external monitoring missions not being implemented in a timely manner (United Republic of Tanzania); • some myths and misconceptions about TB in the community, which had an effect on the participation rate (Zambia); • the need for field staff to work long hours when participants arrived at the main survey camp site relatively late in the day, especially in rural areas (Zambia); • hot weather conditions that affected participation (Zimbabwe); • religious groups that were opposed to modern medical interventions (Zimbabwe); and • issues with retrieval of X-ray images because the archiving and communications system was controlled by an X-ray supplier in the Netherlands (Zimbabwe). National TB prevalence surveys 2007–201658 Table 4.3 Lessons learned for future surveys as reported by countries (see Part II for details) Topic Lessons learned First-ever surveys • These are strongly facilitated by the use of experts from other countries that have recently completed a survey successfully and of international experts that have already supported multiple surveys; it is important to have continuity of support from these experts throughout the survey. • Many surveys benefited from “AA collaboration” (Asia–Asia, Asia–Africa and Africa–Africa) and technical assistance from international agencies. Repeat surveys • These are facilitated if at least some of the same national staff and international experts involved in the previous survey are involved in the repeat survey. Stakeholder commitment and involvement, and regular communication among stakeholders • Involvement of and ownership by the national TB programme and, more broadly, the Ministry of Health are especially important, even if another agency is contracted to implement the survey. • The roles and responsibilities of each stakeholder should be clearly defined. • Good collaboration with the national bureau of statistics is essential for proper sampling design. Survey team leadership, management capacity and monitoring • A high level of leadership and management capacity in the team responsible for implementing the survey is a major contributor to a successful survey. • All survey procedures should be carefully monitored to prevent protocol violations, or to ensure prompt remedy if violations occur. Community engagement and survey participation • Involvement of stakeholders and community leaders at local level is essential; use of the media to inform people about the survey can also be helpful. Participation can be increased by extended hours of field operations (including in the evenings and on weekends), provision of transport to those living far from the survey field site, and high levels of motivation of the field and survey teams. X-ray equipment procurement and breakdowns • Procurement needs to be planned well in advance, and national regulations checked to ensure that what is ordered complies with the regulations. • If equipment is procured from an international supplier, it is important to ensure that there is a contract to provide local support in the event of breakdowns. The availability of in-country servicing of equipment is essential to ensure timely repairs and troubleshooting. • Back-up machines should be available in case of breakdowns. Data management • A competent and responsive data management team that is involved from survey design to completion is essential. • Use of multiple paper-based forms for the same individual should be avoided. • Electronic data management facilitates timely entry, validation and analysis of survey data. • Internet connectivity may be a challenge in some parts of a country; solutions identified included use of a local area network in the field with later uploading of data to a central server and use of paper forms as a back-up. • Use of bar codes (as opposed to writing individual identifiers by hand) reduces errors in data entry and matching of records. Laboratory issues • It is essential to ensure that good laboratory practices are maintained and standardized in all involved laboratories, and are properly monitored throughout the survey, including during periods of high volume and throughput of specimens to be tested. • Strong leadership from the principal investigator and survey team can help to identify and resolve problems in a timely way, as can following the advice of an expert technical advisory group for the survey. • Xpert MTB/RIF is useful for checking smear-positive results that are negative on culture, or for which the culture result is missing. • Surveys that described the quality of laboratory work as a “survey success” included countries where the laboratory that was used was either part of a long-established research unit (Gambia) or a national reference laboratory (including one, in Uganda, that was a supranational reference laboratory). Security issues • Survey protocols should clearly define how clusters will be replaced in the event of security or other issues that require cluster replacement. • How clusters were replaced should also be documented in the final survey report. Production of final survey report • A budget should be allocated specifically for the writing of the survey report. In several countries, the report was delayed because no funding had been allocated to prepare and write the survey report. Delays in reading X-rays • It may be necessary to allocate a budget specifically for reading X-rays. In some surveys, additional funds had to be mobilized at the end of the survey (including from WHO) to enable review by qualified radiographers. A local supplier for software and for archiving or communication of images should be used if possible. WHO: World Health Organization. PART I: An overview of the 25 surveys implemented 2007–2016 59 4.3 Lessons learned Lessons were learned from both successes and challenges. These lessons, which are important for guiding and informing future surveys, are summarized in Table 4.3. They included the following: • There is much value in cross-country collaboration and international technical assistance from experts with experience of supporting multiple surveys, especially for countries implementing their first- ever (or first for many years) survey. Asia–Asia, Asia–Africa and Africa–Africa collaborations (collectively referred to as “AA collaboration”) were all strongly promoted and facilitated by the WHO Global Task Force on TB Impact Measurement’s subgroup on national TB prevalence surveys. • There is high value in having at least some continuity in the national staff and international experts involved in repeat surveys. • It is important to have stakeholder commitment and involvement, and regular communication among stakeholders, throughout a survey. • Strong leadership and management of the survey team are major contributors to survey success. • Procurement needs to be planned well in advance, and national regulations checked, to ensure that the ordered equipment complies with national regulations. • The availability of in-country servicing for X-ray equipment, and the availability of back-up machines, help to ensure that issues during field operations (e.g. overheating or breakdown) can be resolved quickly. • Laboratory work must be carefully planned, maintained and closely monitored throughout a survey; advice from laboratory experts or an expert technical advisory group must be promptly acted upon. • Xpert MTB/RIF can be helpful to check smear- positive results when culture results are missing or negative.1 • A competent and responsive data management team is essential; this team should be involved from the initial stages of survey preparations through to completion of data analysis and report writing. 1 Further discussion of the role of molecular tests in addressing challenges with culture testing in prevalence surveys is included in Chapter 5. • Electronic data capture systems can significantly facilitate and increase the efficiency of data collection, validation and analysis. • Use of multiple paper-based forms for the same individual should be avoided. • A budget should be allocated specifically for the writing of the final survey report. • Involvement of stakeholders and community leaders at local level is essential; also, use of the media can facilitate community engagement and participation. These lessons learned echo and reinforce the 11 factors that were identified in the lime book as prerequisites for the successful implementation of a national TB prevalence survey (1). The 11 prerequisites were: strong commitment and leadership from the NTP, the ministry of health and a core group of professionals; identification of a suitable institute, organization or agency to lead and manage the survey; adequate laboratory capacity, especially for culture; compliance with the regulations of the national radiation authority; reliable and timely procurement and logistics; funding; assurance of security in the field for survey teams and participants; professional data managers and associated data management practices; community participation; expert review and clearance of protocols, including ethical clearance; external support and technical assistance. Reference 1. Tuberculosis prevalence surveys: a handbook. Geneva: World Health Organization; 2011 (https://apps.who.int/iris/bitstream/ handle/10665/44481/9789241548168_eng.pdf, accessed 22 November 2019). 60 Photo credit: Raldy Benavente / FACE Inc (Philippines) Transporting chest X-ray equipment during the 2016 national TB prevalence survey of the Philippines. 61 Chapter 5 Future direction The introductory chapter of this book highlighted that national notification and vital registration systems can be used to reliably monitor the burden of TB disease (in terms of numbers of cases and deaths each year) in many high-income countries, with a few countries having time series of data that cover a span of more than 100 years. It also highlighted that while the ultimate goal is that all countries can reliably track their TB epidemics using such systems, in the early 2000s this goal had not been achieved in many countries with a high burden of TB. Although all countries (including high TB burden countries) had national notification systems for TB and were reporting notification data to WHO on an annual basis, in most countries the number of notified cases was not a good proxy for the actual number of new cases. This was due to a mixture of underreporting of detected cases, duplicated case reporting, some level of overdiagnosis of bacteriologically unconfirmed cases and underdiagnosis. National VR systems of high quality and coverage had yet to be established in many parts of the world. This situation was the reason for the establishment of the WHO Global Task Force on TB Impact Measurement in 2006. The task force had the aim of ensuring a robust, rigorous and consensus-based assessment of whether TB targets set for 2015 in the UN MDGs and the WHO Stop TB Strategy were achieved. It included national surveys of the prevalence of TB disease in 22 global focus countries (GFCs) as one of its three strategic areas of work during the period 2007–2015 (1). Such surveys were recognized as providing an alternative way of directly and reliably measuring the burden of TB disease, with repeat surveys allowing assessment of trends. Other recognized benefits of surveys were that they could be used to document the distribution of disease by age and sex; to better understand health care seeking behaviour in the public and private sectors; to identify reasons why people with TB were not diagnosed before the survey or officially reported to national authorities (or both); and to inform the development or improvement of strategies and interventions for TB case finding, diagnosis and treatment. As illustrated in Chapter 3 and in the country-specific chapters that form Part II of this book, a substantial new body of knowledge was generated by the 25 surveys completed in 24 countries1 (including 18 of the 22 GFCs) between 2007 and 2016. Data were used to update estimates of TB disease burden, including time trends in the three countries that conducted repeat surveys, and to inform national policy, national strategic plans, advocacy and resource mobilization. Chapter 4 then synthesized survey successes (including and beyond the generation and use of survey data), challenges and lessons learned during the time between the initial decision to implement a survey and dissemination of results, based on the more detailed descriptions provided in Part II. Looking forward, and building on Chapter 3 and Chapter 4, this final chapter of Part I addresses three important questions: • Are national TB prevalence surveys still relevant? • Where do national TB prevalence surveys remain relevant? • Should national TB prevalence surveys be done differently in future? 5.1 Are national TB prevalence surveys still relevant? In 2013, WHO published a TB surveillance checklist of standards and benchmarks that can be used to assess the quality and coverage of national notification and VR systems (2). Although much progress in strengthening national TB notification systems was made between 2007 and 2016, at the end of this period, most countries with a high burden of TB still lacked systems that met the levels of quality and coverage necessary for notification data to provide a direct measure of TB incidence (3). In WHO’s Global tuberculosis report 2019, the data used to estimate TB incidence in high TB burden countries were sourced mainly from national TB prevalence surveys (4). In the 1 Two surveys were implemented in the Philippines (2007 and 2016). National TB prevalence surveys 2007–201662 same report, estimates of TB mortality were based on national VR data for 123 countries (including nine of the 30 included in WHO’s list of high TB burden countries). Given this situation, the rationale for using national TB prevalence surveys as an alternative way to directly measure the burden of TB disease and trends remained as valid at the end of 2016 as it was in 2007. 5.2 Where do national TB prevalence surveys remain relevant? In April 2016, the WHO Global Task Force on TB Impact Measurement held a meeting to discuss progress achieved during the period 2007–2015, and its work in the post- 2015 era of the UN Sustainable Development Goals (SDGs, which succeeded the MDGs) and WHO’s End TB Strategy (which succeeded the Stop TB Strategy) (5). The SDGs, set for 2030, were adopted by all UN Member States in September 2015 (6). The End TB Strategy was adopted by all WHO Member States at the World Health Assembly in 2014; it covers the period 2016–2035, with milestones for 2020 and 2025 and targets for 2030 and 2035 (7). During its April 2016 meeting, the task force agreed on an updated mission and five strategic areas of work, initially for the period 2016–2020 (likely to apply and be extended to 2021–2025). National TB prevalence surveys were retained under the new third strategic area of work, Table 5.1 Suggested epidemiological criteria for assessing whether a country should consider implementing a prevalence survey post-2016 for two major groups of countries, as discussed by the WHO Global Task Force on TB Impact Measurement in April 2016 which was defined as “Priority studies to periodically measure TB disease burden”. The task force meeting was also used to discuss the countries in which national TB prevalence surveys remained relevant. The suggested epidemiological criteria for assessing whether a country should consider implementing a survey are shown in Table 5.1, and the countries in each of the two groups defined in Table 5.1 (based on data available at the end of 2019) are shown in Fig. 5.1 and Fig. 5.2. Among the 24 countries in Group 1 (i.e. those that implemented a survey in 2007–2016 and that met the criteria shown in Table 5.1), it is worth highlighting that five did not meet the criteria for a further survey because of their relatively low measured level of TB disease burden; these countries were China, Gambia, Rwanda, Sudan and Thailand. In these countries, the focus should be on maintaining or strengthening national notification and VR systems. Of the 29 countries in Group 2, four stood out in terms of their share of estimated TB disease burden from a global perspective: Democratic Republic of the Congo, India, Mozambique and South Africa. Of these, South Africa completed a survey in 2019, Mozambique completed one in 2020 and India started a survey in 2019. In addition, surveys were completed in Namibia (2018), Nepal (2019) and Lesotho (2019), and planning for a survey in Botswana was initiated in 2018. Criteria Explanation Group 1 → Countries that conducted a national prevalence survey in 2007–2016a (Fig. 5.1) 1. Estimated prevalence of bacteriologically confirmed TB ≥250 per 100 000 population aged ≥15 years during the previous survey. and 2. More than 7 years since the last survey.a • Sample size small enough (<70 000 individuals) to make surveys feasible in terms of cost and logistics. • Time between surveys sufficient to allow a statistically meaningful comparison of prevalence. Group 2 → Countries that did not implement a national prevalence survey in 2007–2016 (Fig. 5.2) 1. Estimated TB incidenceb ≥150 per 100 000 population per year (all forms, all ages). and 2. No nationwide VR system with standard coding of causes of deaths. and 3. Infant mortality rate >10/1000 live births. • Sample sizeb small enough (<70 000 individuals) to make surveys feasible in terms of cost and logistics, taking into account added uncertainty due to the use of rapid molecular tests with performance that may be inferior to culture. • No reliable direct measurement of TB disease burden. • Indirect indicator of low access to quality health services, as defined in the WHO checklist of standards and benchmarks for TB surveillance and VR. VR: vital registration; WHO: World Health Organization. a Surveys conducted before 2000 may lack comparability with surveys implemented according to the screening and diagnostic algorithm recommended in the lime book (8). An interval of about 7–10 years between two surveys is recommended. b Country-specific prevalence estimates have not been published by WHO post-2016 because prevalence is not a high-level indicator of the End TB Strategy. For sample size calculations, prevalence in the age group 15 years or more may be predicted from incidence. PART I: An overview of the 25 surveys implemented 2007–2016 63 a These 19 countries are Bangladesh, Cambodia, Democratic People’s Republic of Korea, Ethiopia, Ghana, Indonesia, Kenya, Lao People’s Democratic Republic, Malawi, Mongolia, Myanmar, Nigeria, Pakistan, Philippines, Uganda, United Republic of Tanzania, Viet Nam, Zambia and Zimbabwe. Fig. 5.1 Countries that conducted a national TB prevalence survey in 2007–2016 and that met the Group 1 criteria based on data available at the end of 2019 (N=19, red) a Fig. 5.2 Countries that met the Group 2 criteria for implementing a national TB prevalence survey based on data available at the end of 2019 a Countries that had already completed or started implementation of a survey by the end of 2019 are shown in blue and remaining countries are shown in red. Grey, not applicable. Grey, not applicable. a The 8 countries in blue are Botswana, Eswatini, India, Lesotho, Mozambique, Namibia, Nepal and South Africa. The 21 other countries in red are Afghanistan, Angola, Cameroon, Central African Republic, Congo, Democratic Republic of the Congo, Djibouti, Equatorial Guinea, Gabon, Guinea, Guinea-Bissau, Haiti, Kiribati, Liberia, Madagascar, Marshall Islands, Papua New Guinea, Sierra Leone, Somalia, Timor-Leste and Tuvalu. National TB prevalence surveys 2007–201664 For any country meeting the epidemiological criteria shown in Table 5.1, it was stressed that survey feasibility must also be carefully assessed. As set out in the lime book (8), there are 11 prerequisites for a survey to be feasible: • there is strong commitment and leadership from the NTP, ministry of health and a core group of professionals; • a suitable institute, organization or agency to lead and manage the survey can be identified; • there is adequate laboratory capacity; • X-ray equipment can comply with the regulations of the national regulatory authority; • reliable and timely procurement and logistics is possible; • funding is available; • security in the field for survey teams and participants can be assured; • data management can be done according to recommended standards; • community participation is likely to be sufficiently high, including in urban areas; • expert review and clearance of protocols, including ethical clearance, can be undertaken; and • external support and technical assistance are available if needed.1 These prerequisites remain valid post-2016. Among the countries in Group 2, several are likely to face challenges in meeting the feasibility criteria; examples include Afghanistan, Democratic Republic of the Congo and Papua New Guinea. 5.3 Should national TB prevalence surveys be done differently in future? The successes, challenges and lessons learned during surveys completed in 2007–2016 (Chapter 4) are useful for informing surveys implemented after 2016. They clearly show what challenges are likely to be encountered and how these can be prevented or mitigated. For the top three challenges that affected by far the largest number of countries (i.e. between 10 and 16), it is worth considering what could be done differently to avoid or mitigate them in future surveys. The top three challenges were: • laboratory work, notably issues related to culture 1 This is likely to be especially important for countries implementing a survey for the first time. 2 Use of rapid molecular tests – notably the Xpert® MTB/RIF and Xpert Ultra cartridges – started following WHO’s endorsement of Xpert MTB/ RIF in 2010. testing; • data management; and • delays in producing the final survey report. 5.3.1 Are there alternatives to relying on culture testing of samples from all survey participants who meet survey screening criteria? The reference standard test for diagnosis of active pulmonary TB disease is culture of M. tuberculosis from sputum samples. In many countries with a high burden of TB, sputum smear microscopy remained the most commonly used diagnostic test for TB in the period 2007–2016.2 For these two reasons, testing of sputum samples using both smear microscopy and culture to diagnose TB was the method recommended for national TB prevalence surveys in the lime book (8) (see also Chapter 2). As stated in the lime book: Surveys of the prevalence of TB disease aim to measure the burden of bacteriologically confirmed pulmonary TB in the community… as such, laboratory tests of sputum samples (using sputum smear microscopy and culture) are a fundamental component of a prevalence survey. Nonetheless, the challenges of culture testing in the context of a national TB prevalence survey were always well recognized. Challenges included the following: • Samples taken among the general population in a field-site setting can be of poorer quality and lower volume than those taken in clinical settings. They are also likely to be more paucibacillary in nature, since on average those with TB disease will be at an earlier stage of disease progression compared with those diagnosed when seeking care at a health facility. • There can be long transportation times between a survey cluster and the laboratory or laboratories being used for testing, and a cold chain needs to be maintained during these times. The recommended time between obtaining a sample and its arrival at the laboratory is 3 days or less, and no more than 5 days. If these times are exceeded, contaminated tubes or false-negative test results become likely. • There is a risk of cross-contamination from positive to negative specimens. PART I: An overview of the 25 surveys implemented 2007–2016 65 3 The main explanation for false-positive Xpert results is that Xpert detects dead TB bacilli, whereas a culture-positive result requires live TB bacilli to be present. This means that Xpert may detect people who had TB in the past as well as those who have been infected by M. tuberculosis but contained the infection. 1 A recent systematic review found that 2% (95% confidence interval [CI]: 1–2%) of all positive cultures were false-positive results due to laboratory cross-contamination. See Barac et al. (2019) (9). 2 For the detection of M. tuberculosis, Ultra incorporates two new multicopy amplification targets (IS6110 and IS1081) and a larger DNA amplification reaction chamber than Xpert MTB/RIF. The semiquantitative scale for Xpert Ultra results is as follows: trace, very low, low, medium or high. Trace corresponds to the lowest bacillary burden for detection of M. tuberculosis and indicates that only the multicopy targets were detected, as opposed to the TB-specific regions in the rpoB gene. • The workload of culture testing generated by a prevalence survey may be challenging for laboratories to manage. Without careful planning it is possible for laboratories to become overloaded, affecting testing timeliness and quality. For these reasons, in all surveys, culture results have been missing for some survey participants, and some results may have been false-negative or (if there was cross-contamination) false-positive.1 Surveys have made use of expert panel reviews using all sources of evidence (symptom screen, X-ray, smear microscopy, sometimes a molecular test result) to make a final determination of whether someone with a missing culture result, or with a culture-negative but smear-positive or Xpert-positive result, should be classified as a survey case. The challenges of culture testing not surprisingly led to growing interest in the role of Xpert (both the MTB/RIF® and more recent Xpert Ultra® assays) in a national TB prevalence survey, following WHO’s endorsement of the Xpert MTB/RIF assay in December 2010, and publication of a policy update (10) and an implementation manual (11). Compared with culture, the advantages of this molecular test include that it is rapid (results available within hours), is automated, does not require fresh samples to perform optimally and does not require stringent laboratory containment. Direct testing of sputa without centrifugation has the added advantage of minimizing cross-contamination. Nonetheless, both Xpert assays also have disadvantages compared with culture. In particular, they have lower sensitivity and specificity. Results from evaluations by the Foundation for Innovative New Diagnostics (FIND) in which Xpert was compared with the reference standard of culture using sputum samples collected in clinical settings in a variety of countries are shown in Table 5.2. The best estimate of sensitivity (the percentage of culture-positives identified by Xpert) was 83% for Xpert MTB/RIF, 88% for Xpert Ultra if trace results were used, and 85% for Xpert Ultra if trace results were not used.2 These findings mean that if Xpert alone is used to Table 5.2 Sensitivity and specificity of the Xpert® MTB/RIF and Ultra assays as measured in an evaluation by FIND Assay Sensitivity (%) compared with the reference standard of culture Specificity (%) compared with the reference standard of culture Xpert MTB/RIF 83 (78–87) 98 (96–99) Xpert Ultra, if trace results are used 88 (84–91) 95 (93–97) Xpert Ultra, if trace results are excludeda 85 97 FIND: Foundation for Innovative New Diagnostics. Sources: Dorman et al. (2018) (12) and WHO (2017) (13). a Uncertainty bounds could not be calculated. test samples from survey participants that meet survey screening criteria (i.e. reported symptoms suggestive of TB or an abnormal chest X-ray), true cases of TB (that could be identified by culture) would be missed. Sensitivity may be improved by repeating Xpert testing on another sample (interpreting the test combination as positive if at least one of the two tests is found to be positive). The best estimates of specificity (the percentage of culture-negatives found to be negative by Xpert) were 98% for Xpert MTB/RIF, 95% for Xpert Ultra if trace results were considered positive and 97% for Xpert Ultra if trace results were considered negative.3 In the setting of a population-based national TB prevalence survey, the proportion of screen-positive individuals (in terms of reported symptoms or an abnormal X-ray) with culture-positive TB disease will be low. In surveys implemented in 2007–2016, the proportion was typically in the range 1–5% (Fig. 5.3). This means that of those tested, typically 95–99% will be culture negative; if tested using an Xpert assay, given the specificity of Xpert MTB/RIF and Ultra (excluding trace results), 2–3% of this 95–99% would have a false- positive Xpert result. In other words, if 100 individuals who screen positive in a prevalence survey are tested with Xpert, about 1–5 people with TB will be correctly identified (in reality a bit less given that Xpert is less sensitive than culture) and about 2–3 people will have a false-positive result. National TB prevalence surveys 2007–201666 Fig. 5.3 Percentage of people who were eligible for sputum testing (i.e. they reported symptoms suggestive of TB or had an abnormal chest X-ray) that had culture confirmed pulmonary TB, for surveys implemented in 2007–2016 Grey, not applicable. Table 5.3 Estimated percentage of Xpert-positive results that would be false-positive in a national TB prevalence survey, based on the specificity of Xpert estimated in the FIND evaluation Percentage of people screened positive in a national TB prevalence survey who have bacteriologically confirmed TB (culture positive) Estimated percentage of Xpert MTB/RIF positive results that will be false-positive Estimated percentage of Xpert Ultra positive results that will be false-positive (assuming trace results are excluded)a 2 54 (49–59) 63 3 44 (39–49) 53 4 37 (32–42) 46 5 31 (20–29) 40 FIND: Foundation for Innovative New Diagnostics. a There were insufficient data to estimate uncertainty intervals. This means that between around one third and two thirds of Xpert-positive results would be expected to be false-positive results in the context of a national TB prevalence survey (Table 5.3). This is an unacceptable level of error when the main objectives of a survey are to reliably measure the level of pulmonary TB disease in the community and (in a repeat survey) trends in that level of TB disease. These expectations were borne out in six national and two subnational TB prevalence surveys completed between 2015 and 2019, in which Xpert MTB/RIF or Xpert Ultra were used alongside culture for testing of all survey participants that screened positive.1 These surveys showed a high proportion of discordant results. The discordance was higher for those reporting a previous history of TB disease compared with those reporting no history; this finding is as expected, given that in those with a treatment history, Xpert is more likely to detect dead TB bacilli. The estimated pooled sensitivity of Xpert MTB/RIF compared with culture was 73% (62–82%) and for Xpert Ultra (excluding trace results) it was 68% (55– 79%);2 the estimated pooled specificity of Xpert MTB/ RIF was 98% (98–99%) and for Xpert Ultra it was 98% (97–99%). 1 The national surveys were those in Bangladesh (2015), Kenya (2015), Myanmar (2017–2018), the Philippines (2016), South Africa (2018–2019) and Viet Nam (2017). The subnational surveys were implemented in 2019 as part of community randomized trials in South Africa and Zambia (the TREATS study). 2 The pooled estimates of sensitivity for Xpert MTB/RIF and Xpert Ultra were based on a small number of prevalence surveys, with wide credibility intervals. There was no demonstrated statistical difference in the sensitivity of the two tests. PART I: An overview of the 25 surveys implemented 2007–2016 67 Recognizing the limitations of both culture and Xpert testing in the context of a national TB prevalence survey, WHO organized meetings between 2018 and 2020 to discuss the best way forward, based on accumulating evidence from surveys in which Xpert and culture were used alongside each other. As of early 2020, one option under consideration for countries without the capacity to conduct high-quality culture testing in the context of a national prevalence survey was as follows: the use of two Xpert Ultra tests on two separate sputum samples for all participants who screen positive (to maximize sensitivity), followed by culture testing for any participant with an Xpert Ultra positive test result (thus addressing the suboptimal specificity of Xpert Ultra by using the reference standard as a confirmatory test to eliminate false-positive Xpert Ultra results). Prevalence estimates would then need to be adjusted to account for the lower sensitivity of the Xpert Ultra test (i.e. adjustment for false-negative Xpert Ultra results). A final set of recommendations related to the diagnostic algorithm to be used in future surveys, designed to make optimal use of both culture and Xpert, is planned for publication in a new edition of the WHO handbook on prevalence surveys that will succeed the lime book.1 5.3.2 Adapting and using the principles of good clinical practice that have been established for clinical trials in the context of national TB prevalence surveys Good clinical practice (GCP) is a set of internationally recognized ethical and scientific quality requirements that must be followed when designing, conducting, recording and reporting clinical trials that involve people (14). They have been used in the context of drug development in particular. Adapting GCP principles to the context of a national TB prevalence survey could help to prevent or mitigate challenges related to data management. They could also contribute to enhancing survey quality more broadly, by strengthening oversight, monitoring processes and ensuring that any recommendations are implemented in a timely way. An independent evaluation of national TB prevalence surveys conducted in 2015 included a recommendation to explore the relevance of GCP to future national TB prevalence surveys (5). GCP requirements are designed to ensure two things: the protection of the rights, safety and well-being of all participants; and that data are comprehensive and accurate. To facilitate their use, roles and responsibilities are defined as follows: • The sponsor or sponsors provide the financing for a survey. Examples include external agencies (e.g. the Global Fund to Fight AIDS, Tuberculosis and Malaria, development agencies or the national government) and may include a mixture of agencies. Sponsors can request regular reports from survey implementing agencies, and reports may be linked to periodic release of funds. • The principal investigator represents all survey investigators. That person is responsible for leading the development of the protocol and standard operating procedures (SOPs) and for ensuring review. The principal investigator is also responsible for the recruitment of competent staff, and leads the writing of the final report and scientific papers. • Investigators contribute to survey design (including the development of a protocol and SOPs, and ethics review and approval), implementation of field operations including quality control, analysis of results and preparation of a survey report. During field operations, this includes ensuring the accuracy, completeness, legibility and timeliness of the data reported in data collection tools. Data that are derived from source documents should be consistent with the source documents; if this is not the case, discrepancies should be explained. To achieve maximum data quality, a standard set of quality assurance procedures2 should be in place. These include checking that batches of newly entered records are consistent with defined standards. • Survey monitors assess the implementation of survey operations, including checking protocol modifications and checking for protocol violations. They may conduct batch checks of data. They advise investigators about their findings and provide recommendations for corrective actions if needed. They also report to an independent data monitoring committee (or board) and may assist the principal investigator to prepare the final 1 At the time of writing, this was planned for publication in 2021. 2 Quality assurance is a process of systematic activities designed to ensure, assess and confirm the quality of the data collected during a survey. Quality-assured data are those that are suitable for their intended purpose in terms of their accuracy, timeliness, accessibility and comparability between database and source documents. National TB prevalence surveys 2007–201668 report. In the context of GCP, study monitors represent the sponsor. • An independent data monitoring committee (or board) may be established by the sponsor to assess the progress of the survey at regular intervals (based on reports from survey monitors) and to provide recommendations to the sponsor about whether to continue, modify or stop the survey. The first three of these elements were present in all national TB prevalence surveys implemented in 2007–2016. Survey monitoring by external experts (the fourth element) was also commonly in place, provided by staff of international agencies or by people who had held senior roles in previous surveys in other countries (via the Asia–Asia, Asia–Africa and Africa–Africa collaboration highlighted in Chapter 4). However, a formal and independent data monitoring committee was not established for any of the surveys (although many had oversight from a survey committee, expert advisory group or equivalent). It was also the case that there was not necessarily any obligation for investigators to implement all of the recommendations made by external experts. WHO initiated the development of guidance on the adaptation and use of GCP and good data management practices (GDMP) within the context of national population-based surveys of TB disease (including national TB prevalence surveys) and health facility based surveys in 2019, in collaboration with WHO/TDR – the WHO Special Programme for Research and Training in Tropical Diseases, which has conducted extensive training in the application of GCP in clinical trials. The final document will provide guidance on how to implement the key GCP/GDMP principles to maximize data credibility (i.e. comprehensive and accurate data collected in an ethical manner) within the scope of population- based surveys and health facility based surveys. Other challenges related to data management can be addressed using the lessons learned from previous surveys (documented in Chapter 4). Examples of lessons learned are that: • a competent and responsive data management team is essential, and this team should be involved from the initial stages of survey preparations through to completion of data analysis and report writing; • electronic data capture systems in the field and laboratories can significantly facilitate and increase the efficiency of data collection, validation and analysis; and • use of multiple paper-based forms for the same individual should be avoided. 5.3.3 Invest more resources in the work required once results are finalized, especially to ensure the timely production of survey reports and effective communication of findings and their implications In 10 of the 25 surveys implemented in 2007–2016, producing the final survey report took a considerable amount of time (more than 1 year in 8 countries). The presence of a permanent full-time survey monitor (in line with GCP) could help to address this challenge, since one of that person’s responsibilities would be to provide regular reports with material that could subsequently be used in the final survey report. More generally, more resources for report writing (people with the right skills and time, and funding for production costs including editing and printing) need to be committed when a survey budget is first developed and approved. Experience in several countries also highlighted the importance of good communication of results to key decision-makers (e.g. planners, policy-makers and those with responsibility for communicable diseases in the ministry of health). During discussions, emphasis should be given to survey validity; quality assurance procedures; monitoring (including external monitoring); and how survey findings provide valuable information for decision-making on policies, prioritization and future budgeting for TB control. When to engage with national and local media also needs careful thought. The last chapter of the lime book (8), on “Analysis and reporting”, focused on best-practice methods for the analysis of survey data and how to present results.1 The book did not include a subsequent chapter on the production of a survey report and communication of results. Such additional guidance will be part of the next WHO edition of this handbook. References 1. Global tuberculosis report 2015. Geneva: World Health Organization; 2015 (https://apps.who.int/iris/bitstream/ handle/10665/191102/9789241565059_eng.pdf, accessed 8 January 2020). 2. WHO. Standards and Benchmarks for tuberculosis surveillance and vital registration systems. WHO, 2014 http://apps.who.int/ iris/bitstream/10665/112673/1/9789241506724_eng.pdf?ua=1. 1 This guidance was subsequently updated and published in a journal article by Floyd et al. (2013) (15). PART I: An overview of the 25 surveys implemented 2007–2016 69 3. Anderson L, Floyd K, Sismanidis B. Strengthening national notification systems for direct measurement of TB cases: an overview of progress. Geneva: World Health Organization; 2018 (https://www.who.int/tb/advisory_bodies/impact_measurement_ taskforce/meetings/tf7_background_1_strengthen_notification. pdf, accessed 22 January 2020). 4. Global tuberculosis report 2019. Geneva: World Health Organization; 2019 (https://apps.who.int/iris/bitstream/hand le/10665/329368/9789241565714-eng.pdf, accessed 1 February 2020). 5. WHO Global Task Force on TB Impact Measurement. Report of the sixth meeting of the full Task Force (19–21 April 2016, Glion-sur-Montreux, Switzerland). Geneva: World Health Organization; 2016 (https://www.who.int/tb/advisory_bodies/ impact_measurement_taskforce/meetings/tf6_report.pdf?ua=1, accessed 8 January 2020). 6. Sustainable development goals [website]. New York: United Nations; 2016 (https://sustainabledevelopment.un.org/topics/ sustainabledevelopmentgoals, accessed 22 January 2020. 7. Uplekar M, Weil D, Lonnroth K, Jaramillo E, Lienhardt C, Dias HM et al. WHO’s new End TB Strategy. Lancet. 2015;385(9979):1799– 801 (https://www.ncbi.nlm.nih.gov/pubmed/25814376, accessed 22 January 2020). 8. Tuberculosis prevalence surveys: a handbook. Geneva: World Health Organization; 2011 (https://apps.who.int/iris/bitstream/ handle/10665/44481/9789241548168_eng.pdf, accessed 22 November 2019). 9. Barac A, Karimzadeh-Esfahani H, Pourostadi M, Rahimi MT, Ahmadpour E, Rashedi J et al. Laboratory cross-contamination of Mycobacterium tuberculosis: a systematic review and meta- analysis. Lung. 2019;197(5):651–61. 10. Policy update: Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Geneva: World Health Organization; 2013 (https://apps.who.int/ iris/bitstream/handle/10665/112472/9789241506335_eng.pdf, accessed 14 February 2020). 11. Xpert MTB/RIF implementation manual: technical and operational ‘how-to’; practical considerations. Geneva: World Health Organization; 2014 (https://apps.who.int/iris/bitstream/ handle/10665/112469/9789241506700_eng.pdf, accessed 14 February 2020). 12. Dorman SE, Schumacher SG, Alland D, Nabeta P, Armstrong DT, King B et al. Xpert MTB/RIF Ultra for detection of Mycobacterium tuberculosis and rifampicin resistance: a prospective multicentre diagnostic accuracy study. Lancet Infect Dis. 2018;18(1):76–84. 13. WHO meeting report of a technical expert consultation: non- inferiority analysis of Xpert MTB/RIF Ultra compared to Xpert MTB/RIF. Geneva: World Health Organization; 2017 (https:// apps.who.int/iris/bitstream/handle/10665/254792/WHO-HTM- TB-2017.04-eng.pdf?sequence=1, accessed 22 January 2020). 14. International Conference on Harmonisation Working Group. ICH harmonised tripartite guideline: guideline for good clinical practice E6 (R1). International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Washington, DC. 1996. 15. Floyd S, Sismanidis C, Yamada N, Daniel R, Lagahid J, Mecatti F et al. Analysis of tuberculosis prevalence surveys: new guidance on best-practice methods. Emerg Themes Epidemiol. 2013;10(1):10 (https://ete-online.biomedcentral.com/ articles/10.1186/1742-7622-10-10, accessed 14 February 2020). 70 71 PART II Country-by-country survey profiles 72 73 Name Role Organization Mahmudur Rahman Principal investigator (PI) Institute of Epidemiology, Disease Control and Research (IEDCR) Meerjady Sabrina Flora Co-investigator IEDCR Mohammad Mushtuq Husain Co-investigator and chief coordinator IEDCR S.M. Mostofa Kamal Co-investigator and laboratory manager National TB Reference Laboratory (NTRL), National Institute of Diseases of the Chest and Hospital (NIDCH) Asif Mujtoba Mahmud Co-investigator IEDCR Iqbal Ansary Khan Co-investigator IEDCR Akter Hossain Co-investigator and central radiologist IEDCR Ahmad Raihan Sharif Co-investigator and data manager IEDCR Mahbubur Rahman Co-investigator and assistant data manager IEDCR Vikarunnessa Begum Co-investigator WHO Bangladesh Mohammed Sayeedur Rahman Co-investigator and survey coordinator WHO Bangladesh Ashaque Husain Chairperson, Executive Committee TB-Leprosy, Directorate General of Health Services (DGHS) Ahmed Hussain Khan Chairperson, Executive Committee TB-Leprosy, DGHS Md Mozammel Haque Chairperson, Executive Committee TB-Leprosy, DGHS Md Quamrul Islam Chairperson, Executive Committee TB-Leprosy, DGHS Shahid Md Sadiqul Islam Chairperson, Executive Committee TB-Leprosy, DGHS Rouseli Haq Chairperson, Executive Committee TB-Leprosy, DGHS Md Ehteshamul Huq Choudhury Chairperson, Executive Committee TB-Leprosy, DGHS Md Jahangir Alam Sarker Member secretary, Executive Committee National TB Control Programme Md Ashraf Uddin Member secretary, Executive Committee National TB Control Programme Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Irwin Law Technical assistance (design and analysis) WHO headquarters Sayori Kobayashi Technical assistance (data management) WHO headquarters J. Sean Cavanaugh Technical assistance (design and analysis) US Centers for Disease Control and Prevention (CDC) Shua Chai Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Mourad Gumusboga Technical assistance (laboratory) Supranational Reference Laboratory (SRL), Antwerp, Belgium Susumu Hirao Technical assistance (X-ray interpretation) Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) Key people Surveyed clusters (N=125)a Data sources ■ National Tuberculosis Prevalence Survey, Bangladesh 2015– 2016. Institute of Epidemiology, Disease Control & Research (IEDCR), Directorate General of Health Services, Ministry of Health & Family Welfare, Government of the People’s Republic of Bangladesh; 2017. ■ Survey dataset. Summary statistics Participation rate 91% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 287 3.2 Prevalence:notification ratio (Bacteriologically confirmed TB, ≥15 years) 2.8 Finance Amount (US$) USAID 1 689 004 The Global Fund 1 849 334 TB CARE II   56 440 Total budget 3 594 778 Survey organization and financing Implementing agency: Institute of Epidemiology, Disease Control and Research (IEDCR) BANGLADESH a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2015–2016 National TB prevalence surveys 2007–201674 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Urban: ward/mohalla or para Rural: union/mouza or village Sample size assumptions • Smear-positive prevalence 100 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.3 • k 0.6 • Response rate 80% • Sample size (estimated) 100 000 Number of clusters 125a Cluster size 800 Eligibility criteria • Age ≥15 years • Residency Lived in the cluster for at least 2 weeks before the census Survey design and methodology a One cluster was replaced by another in the same district, due to a security issue (the planned survey site was set on fire by people from a neighbouring village). Laboratory methodology Smear Two samples (spot, morning): direct preparation FM (LED, auramine stain) Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB Capilia TB drug susceptibility test Done Xpert® MTB/RIF One sample (morning). A spot sample was used if the following conditions were met: morning sample was not available; smear-positive but Xpert-negative in a morning sample; smear-negative and Xpert-negative in a morning sample, but a spot sample was smear-positive. HIV test Not done Screening criteria Interviewa Cough ≥2 weeks (3 points) Cough <2 weeks (1 point) Haemoptysis in the past month (3 points) Weight loss in the past month (1 point) Fever ≥1 week in the past month (1 point) Night sweats in the past month (1 point) Symptom-screen positive: Total clinical score ≥3 points Clinical score 1 or 2 with chest X-ray exempted Chest X-rayb Any lung abnormality Other N/A a An in-depth interview on health-care seeking behaviour was done only for participants who reported any TB symptoms (cough, haemoptysis, weight loss, fever, night sweats). b Portable digital direct radiography. Field data collection Paper (interview)/ electronic (census) Database SQL Method of analysis MI+IPW Results first published in a report/paper August 2017 Official dissemination event August 2018 Analysis and reporting Health-care seeking behaviour among participants who reported symptoms Number % Participants who reported symptomsa 26 882 – Location of care sought • Consulted medical facility 6 545 24 Public facility 1 816 28 Private facility 2 182 33 Other (NGO, village doctor) 2 547 39 • Pharmacy 6 533 24 • Traditional healer 23 0.1 • Otherb 191 0.7 Self-treated 643 2.4 No action taken 12 947 48 Key survey results a Data on health-care seeking behaviour were available for participants who reported at least one of TB symptoms (cough, haemoptysis, weight loss, fever, night sweats). b Ayurvedic/homeo/unani (177), not specified (14). a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b DST was done for 157 subjects. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 57 – • Treated in the public sector 16 28 • Treated in the private sector 10 17 • Treated in other sector 18 32 • Treated in unknown sector 13 23 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 9 3.2 Design effect k Smear-positive TB 1.5 0.8 Bacteriologically confirmed TB 1.6 0.5 Other sputum results Number % Total smear-positive participants 125 – Smear-positive participants without MTB confirmationa 17 14 Isolates with MDR-TB detectedb 1 0.6 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 113 87–139 287 244–330 Male 187 141–234 452 379–526 Female 48 30–67 143 109–178 15–24 years 45 21–69 103 65–152 25–34 years 77 39–116 183 122–244 35–44 years 138 74–202 302 215–389 45–54 years 137 72–202 338 235–441 55–64 years 147 64–229 462 317–607 ≥65 years 333 185–480 954 715–1 194 Urban 131 78–185 316 239–392 Rural 103 77–129 270 220–321 a Age ≥15 years unless otherwise specified. BANGLADESH 75 Survey flow: census to final outcomes Field operations: March 2015 to April 2016 Individuals enumerated in census 148 126 Eligible for sputum examination 20 594 (21%) Total bacteriologically confirmed cases 278 Eligible study population 108 834 (73%) Total participants 98 710 (91%) Interview and chest X-ray 98 559 (99.8%) Interview only 151 (0.2%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 3 077 (15%) Symptom positive, chest X-ray negative or N/A 4 217 (20%) Symptom negative, chest X-ray positive 13 300 (65%) Other N/A Symptom positive, chest X-ray positive 79 (28%) Symptom positive, chest X-ray negative or N/A 27 (9.7%) Symptom negative, chest X-ray positive 172 (62%) Other N/A Submitted specimens At least one specimen 20 463 (99%) Both specimens 20 010 (97%) Smear-positive casesc 108 (39%) Definite 108 Probable N/A Smear-negative casesc 170 (61%) Definite 170 Probable N/A Laboratory result At least one culture result availableb 20 378 (99%) At least one Xpert result available 20 425 (99%) Symptom screening Cough ≥2 weeks 6 467 (6.6%) Cough <2 weeks 12 909 (13%) Haemoptysis in the past month 472 (0.5%) Weight loss in the past month 3 615 (3.7%) Fever ≥1 week in the past month 7 546 (7.6%) Night sweats in the past month 3 819 (3.9%) Score ≥3 pointsa 7 260 (7.4%) Score 1 or 2 with chest X-ray exempteda 34 (0.03%) Total symptom-screen positivea 7 294 (7.4%) Chest X-ray screening Normal 79 865 (81%) Abnormala 16 377 (17%) Other abnormality 2 317 (2.4%) Result not available 0 (0%) Total chest X-rays taken 98 559 a Eligible for sputum collection. b Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. c Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. Ineligible individuals 39 292 (27%) Children <15 years 35 770 (24%) Did not meet residency criteria 3 522 (2.4%) National TB prevalence surveys 2007–201676 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of bacteriologically confirmed TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 3.0 2.5 3.5 4.5 4.0 5.0 0 15–24 25–34 35–44 45–54 55–64 Male Female Total (≥15) ≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 15 5 20 25 30 0 0 1 2 3 4 5 6 7 8 9 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 300 400 500 600 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 200 400 600 800 1000 1200 0 Prevalence per 10 0 0 0 0 population10 000 30 000 40 000 50 000 60 000 80 000 70 000 90 000 20 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1200 600 800 1000 400 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence per 100 000 population and population estimates from the UN Population Division (2015 revision). b The data suggest that the distribution of cases by cluster (blue bars) is significantly different from the theoretical distribution (red line) (mean 2.22, variance 3.63, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications of bacteriologically confirmed pulmonary TB (2015) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb BANGLADESH 77 Background Bangladesh’s population was 161 million in 2015. It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016– 2020. In 2015, Bangladesh was a lower-middle-income country with an average gross national income (GNI) per person of US$ 1190 per year (1). The prevalence of HIV in the general population aged 15–49 years was <0.1% (95% confidence interval [CI]: <0.1–<0.1%) (2), and it was estimated that 0.1% (95% CI: 0.08–0.2%) of TB patients were coinfected with HIV (3). In Bangladesh in the 1960s and 1970s, TB services were based in TB clinics or hospitals, and then expanded to 124 upazila health complexes (UHCs) between 1980 and 1986 (the period of the second health and population plan). During the third health and population plan (1986– 1991), TB services were integrated with leprosy under the Mycobacterial Disease Control unit of the Directorate General of Health Services (DGHS). The National TB Control Programme (NTP) adopted the WHO- recommended DOTS strategy during the fourth health and population plan (1992–1998); it was implemented in four upazilas in November 1993 and expanded to cover all upazilas by mid-1998 (4-6). The notification rate increased from 45 per 100 000 popu- lation in 1990 to 102 per 100 000 population in 2010 (7). WHO estimated incidence to be 227 (95% CI: 200–256) per 100 000 population and prevalence to be 404 (95% CI: 211–659) per 100 000 population in 2014. The case detection rate was 53% (95% CI: 47–60) in 2014 (8). Bangladesh carried out national TB prevalence surveys in 1964–1966, in 1987–1988 and in 2007–2009 (9–11). In contrast to the methodology recommended by the WHO Global Task Force on TB Impact Measurement, the 2007– 2009 survey was based on “smear from everybody”; that is, without screening, sputum samples were collected from every eligible participant for smear examinations (and a subsequent chest X-ray was taken if a smear was positive). Adjusted smear-positive TB prevalence in those aged 15 years or more in the 2007–2009 survey was 79 (95% CI: 47–134) per 100 000 population. In December 2007, Bangladesh was one of the 22 global focus countries for a national TB prevalence survey selected by the WHO Global Task Force on TB Impact Measurement. Recognizing that a new prevalence survey – carried out in accordance with recommended methods – was needed to understand the current TB burden (12) and to measure the impact of the NTP, the Ministry of Health decided in 2012 to implement a fourth national TB prevalence survey. Field operations were conducted from March 2015 to April 2016. Photo credit: Irwin Law National TB prevalence surveys 2007–201678 Key methods and results There were 125 survey clusters in two strata (urban and rural), with a target cluster size of 800 individuals. A total of 148 126 individuals from 9594 households were enumerated in the survey census, of whom 108 834 (73%) were eligible and invited to participate. Of these, 98 710 (91%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using a chest X-ray and an interview about symptoms (12). A total of 20 594 participants (21%) were eligible for sputum examination; of these, 20 463 (99%) submitted at least one sputum specimen and 20 010 (97%) submitted two sputum specimens. Sputum from 20 425 participants was tested with Xpert® MTB/RIF. Of these participants, 269 (1.3%) were positive for Mycobacterium tuberculosis (MTB); of these, 12 (4.4%) were also rifampicin (RIF) resistant, 231 (86%) were RIF sensitive and 26 (9.6%) were indeterminate. Due to potential cross-contamination, 13 Xpert-positive results were annulled. A total of 278 bacteriologically-confirmed pulmonary TB cases were identified, including 108 (39%) cases of smear-positive TB. Of these 278 cases, 132 (47%) were confirmed by both culture and Xpert MTB/RIF, 22 (7.9%) only by culture, and 124 (45%) only by Xpert MTB/RIF (the accompanying culture result was either culture MTB negative, nontuberculous mycobacteria or contaminated). Among 124 cases that were diagnosed only by Xpert MTB/RIF, 103 were smear-negative. The prevalence of smear-positive TB was 113 (95% CI: 87–139) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 287 (95% CI: 244–330) per 100 000 population. The prevalence of smear-positive and bacteriologically confirmed TB per 100 000 population did not vary by strata. Other key results were: • the male to female ratio was 3.9 for smear- positive TB and 3.2 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age and was especially high in those aged 55 years or more; the absolute number of TB cases was consistently high in all age groups, with two peaks in those aged 35–44 years and 65 years or more; • among bacteriologically confirmed TB cases, 38% were symptom-screen positive, and among the smear-positive cases, 48% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.8 overall, but varied from 1.9 in those aged 55–64 years to 4.3 in those aged 65 years or more, and was higher for men than women (3.6 versus 1.9); • among the bacteriologically confirmed TB cases, 90% had no previous history of anti- TB treatment and only 3.2% were on anti-TB treatment at the time of the survey; and • of the 101 bacteriologically confirmed and 48 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 32 (32%) and 15 (31%), respectively, had previously sought care in a public or private health facility for their symptoms. Photo credit: Irwin Law BANGLADESH 79 Implications of results Based on the results from the national TB prevalence survey, the overall prevalence (for all forms and all ages) was estimated at 260 (95% CI: 220–301) per 100  000 population. This was lower than the pre-survey estimate of 404 (95% CI: 211–659) per 100 000 population (6, 8). However, it was higher than had been anticipated by national authorities based on the country’s notification data and the results from the 2007–2009 survey. Possible explanations for the lower-than-expected burden include improved access and use of TB diagnostic services, better case detection and treatment of TB cases in the community (especially in urban areas), and reductions in the level of poverty and undernourishment in the decade prior to the survey (1, 13). The estimated incidence was 221 (95% CI: 160–290) per 100  000 population, which was similar to the pre-survey estimate of 227 (95% CI: 200–256) per 100 000 population (6, 8). Other implications included: • a need for case detection to be improved by including chest X-ray examination in the diagnostic algorithm, given that only one third of survey TB cases met the symptom screening criteria according to the scoring system used, and the remaining cases were identified as eligible for diagnostic testing only by chest X-ray; • a need for strategies to improve access to diagnosis and treatment for men and those aged 55 years or more, given the higher prevalence and higher ratio of prevalence to notifications in these groups; • a need for improved diagnostic capacity to detect the large pool of smear-negative disease; • a need for strengthened community awareness about TB and efforts to reduce stigma associated with the disease, given that about half of participants who reported at least one TB symptom had not sought care for their symptoms at the time of the survey; and • a need for informal private providers to be integrated into public-private networks given that among participants who reported at least one TB symptom, 24% sought care in pharmacies as the first point of care. Major successes, challenges and lessons learned The 2015–2016 national TB prevalence survey in Bangladesh was carried out successfully. As with the survey in Kenya, this was one of the first national TB prevalence surveys that used both culture and Xpert MTB/RIF for all participants who screened positive. A key factor in the success of the survey was the strong leadership, strong technical capacity and collaborative culture of the implementing agency. Together with extensive experience with health research, the team had good channels of communication with the Ministry of Health, the NTP, the National TB Reference Laboratory, nongovernmental organizations and external partners such as WHO and the United States Centers for Disease Control and Prevention. Ensuring that only one agency was responsible for the survey helped to streamline funding mechanisms, procurement and human resource management. Another reason for success was the ingenuity and responsiveness of the information technology (IT) team. This included medical and epidemiological officers, a data manager and software engineers. Practical understanding of the survey in combination with technical capacity ensured a system that was fit for purpose, and which could provide a high-quality dataset shortly after field operations were completed. The use of barcodes and real-time data entry using tablets helped to minimise transcription errors and the overall workload of the survey team. The use of paper as a backup for key variables assisted with validation of data and for tracking of individuals during field operations. Photo credit: Sayori Kobayashi National TB prevalence surveys 2007–201680 Challenges faced during the survey, and associated lessons learned, are listed below. • A lengthy procurement process that took more than 12 months. Some laboratory equipment was only received after the pilot survey had been completed. • Field operations had to be rescheduled several times due to the failure (due to overheating) of some digital chest X-ray machines. None of the five machines were simultaneously functional, which slowed survey implementation and increased overall costs (e.g. for human resources). There was no local vendor of the equipment to provide service support. Procurement of major capital from international suppliers should always include local support. • Potential cross-contamination of specimens from the field to the laboratory. In defining a survey TB case, laboratory source documents were examined to identify any potential clustering of positive Xpert MTB/RIF results. All available data from those with any results that were consecutively positive (because they were processed in the same numerical order) were reviewed, and 13 results from three clusters were excluded. It is essential to ensure that good laboratory practices are maintained in situations of high volume and throughput. • One cluster had to be replaced by another due to local conflict – the planned survey site was set on fire by people from a neighbouring village. While such situations are likely to be infrequent, survey protocols should clearly define how clusters will be replaced in such circumstances, and cluster replacement should be documented including in the final survey report. References 1. The World Bank. (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS. (http://aidsinfo.unaids.org/, accessed May 2017). 3. World Health Organization. Global tuberculosis database. Geneva: WHO; 2017 (http://www.who.int/tb/data/en/, accessed April 2017). 4. National tuberculosis prevalence survey, Bangladesh 2015– 2016. Bangladesh: Institute of Epidemiology, Disease Control & Research (IEDCR), Directorate General of Health Services, Ministry of Health & Family Welfare, Government of the People’s Republic of Bangladesh; 2017. 5. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 6. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 7. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://reliefweb.int/sites/reliefweb.int/files/ resources/9789241564656_eng.pdf, accessed May 2017). 8. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015 (http://apps.who.int/iris/ bitstream/10665/191102/1/9789241565059_eng.pdf, accessed July 2017). 9. Government of Bangladesh. The report of the tuberculosis survey of Bangladesh. National TB prevalence survey in Bangladesh. 1973. National Tuberculosis Control and Research Project. Ministry of Health and Family Planning, Government of Bangladesh; 1973. 10. Zaman K, Hossain S, Banu S, Quaiyum MA, Barua PC, Salim MA et al. Prevalence of smear-positive tuberculosis in persons aged ≥ 15 years in Bangladesh: results from a national survey, 2007–2009. Epidemiol Infect. 2012;140(6):1018–1027 (https://www.ncbi.nlm. nih.gov/pubmed/21880168, accessed March 2018). 11. Directorate General of Health Services. Report on the national prevalence survey on tuberculosis in Bangladesh, 1987–88. Dhaka: Ministry of Health and Family Welfare, Government of Bangladesh; 1989. 12. World Health Organization. Tuberculosis prevalence surveys: a handbook. Geneva: WHO; 2011 (https://apps.who.int/iris/ bitstream/handle/10665/44481/9789241548168_eng.pdf, accessed August 2017). 13. Food and Agriculture Organization of the United Nations (FAO). FAOSTAT, Bangladesh. (http://www.fao.org/faostat/ en/#country/16, accessed March 2018). 81 Name Role Organization Mao Tan Eang Chairman National Centre for TB and Leprosy Control (CENAT) Peou Satha Survey coordinator/chief of radiology CENAT Pheng Sok Heng Chief of laboratory CENAT Koy Bonamy Chief of census CENAT Tieng Sivanna Chief of statistics CENAT Kouet Pichenda Field team leader CENAT Keo Sokonth Field team leader CENAT Saint Saly Field team leader CENAT Chea Manith Field team leader CENAT Kosuke Okada Supervisor (project leader) Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) Norio Yamada Supervisor (epidemiology/statistics) RIT/JATA Masaki Ota Supervisor (epidemiology/data management) RIT/JATA Takashi Yoshiyama Supervisor (chest X-ray examination (diagnosis)) RIT/JATA Kunihiko Ito Supervisor (chest X-ray examination (diagnosis)) RIT/JATA Hiroyuki Nishiyama Supervisor (chest X-ray examination (diagnosis)) RIT/JATA Yutaka Hoshino Supervisor (chest X-ray examination (film shooting)) RIT/JATA Hiroko Matsumoto Supervisor (bacteriological examination (quality assurance)) RIT/JATA Tetsuhito Sugamoto Supervisor (bacteriological examination (culture, identification and DST)) RIT/JATA Kiyomi Yamamoto Coordinator/data management RIT/JATA Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Rajendra Yadav Technical assistance (survey advisor) WHO Cambodia Emily Bloss Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Sara Whitehead Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Philippe Glaziou Technical assistance (statistics) WHO headquarters Charalampos Sismanidis Technical assistance (analysis) WHO headquarters Sian Floyd Technical assistance (analysis) London School of Hygiene & Tropical Medicine Key people Surveyed clusters (N=62)a Data sources ■ Report of the second national TB prevalence survey, 2011. Phnom Penh: Cambodia Ministry of Health; 2012. ■ Survey dataset. Summary statistics Participation rate 93% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 831 1.8 Prevalence:notification ratio (smear-positive TB, ≥15 years) 1.7 Finance Amount (US$) The Global Fund 203 650 JICA 760 300 USAID 53 600 Total budget 1 017 550 Survey organization and financing Implementing agency: The National Centre for TB and Leprosy Control (CENAT) CAMBODIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2010–2011 National TB prevalence surveys 2007–201682 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural/othersa Sampling unit District/commune/village Sample size assumptions • Smear-positive prevalence 256 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.4 • k 0.5 • Response rate 90% • Sample size (estimated) 39 680 Number of clusters 62 Cluster size 640 Eligibility criteria • Age ≥15 years • Residency Resided ≥2 weeks in the household prior to the census Survey design and methodology a Mondulkiri, Rattanakiri, Preah Vihear and Steung Treng. Laboratory methodology Smear Two samples (spot, morning); direct preparation FM (LED, auramine stain), cross-examination by ZN for specific slidesa Culture Two specimens (spot, morning); direct preparation, Ogawa media Identification of MTB Capilia TB drug susceptibility test Doneb Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interview Cough ≥2 weeks and/or haemoptysis Chest X-raya Any lung abnormalityb Other Chest X-ray exempted a Conventional radiography. b Other than a single small calcification nodule less than 10mm or pleural adhesion at costophrenic angles. a ZN was used on smears that were FM positive; those with positive cultures; those with negative smears and negative cultures but chest X-ray suggestive of active TB; and 5% of those smears that were FM negative as negative controls. b 278 MTB strains were sent to RIT/JATA. Field data collection Paper Database Microsoft® Access Method of analysis Survey analysis based on participants without imputation Results first published in a report/paper December 2012 Official dissemination event February 2012 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 1 916 – Location of care sought • Consulted medical facility 1 261 66 Public facility 947 75 Private facility 305 24 Unspecified 9 0.7 • Pharmacy 401 21 • Traditional healer 21 1.1 Self-treated 28 1.5 Otherb 6 0.3 No action taken 197 10 Unknown 2 0.1 Key survey results a Cough ≥2 weeks and/or haemoptysis. b Family member. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). Survey participants currently on TB treatment Number % Total participants currently on TB treatment 80 – • Treated in the public sector 72 90 • Treated in the private sector 6 8.0 • Treated in unknown sector 2 2.0 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 6 1.9 Design effect k Smear-positive TB 1.6 0.6 Bacteriologically confirmed TB 2.5 0.6 Other sputum results Number % Total smear-positive participants 114 – Smear-positive participants without MTB confirmationa 24 21 Isolates with MDR-TB detected 0 0 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 271 212–348 831 707–977 Male 361 265–493 1 097 895–1 344 Female 197 127–303 609 486–763 15–24 years 18 4.3–71 130 74–227 25–34 years 87 41–185 427 304–598 35–44 years 266 169–420 881 667–1 163 45–54 years 364 218–607 1 029 780–1 358 55–64 years 799 534–1 194 1 844 1 388–2 446 ≥65 years 1 007 653–1 550 3 046 2 353–3 936 Urban 134 61–292 593 357–983 Rural 310 236–408 882 738–1 055 Other 249 4.4–12 273 1 175 24–36 964 a Age ≥15 years unless otherwise specified. CAMBODIA 83 Survey flow: census to final outcomes Field operations: December 2010 to September 2011 Individuals enumerated in census 68 087 Eligible for sputum examination 4 780 (13%) Total bacteriologically confirmed cases 314 Eligible study population 40 423 (59%) Total participants 37 417 (93%) Interview and chest X-ray 37 221 (99.5%) Interview only 196 (0.5%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 710 (15%) Symptom positive, chest X-ray negative or N/A 1 206 (25%) Symptom negative, chest X-ray positive 2 699 (57%) Otherb 165 (3.5%) Symptom positive, chest X-ray positive 88 (28%) Symptom positive, chest X-ray negative or N/A 5 (1.6%) Symptom negative, chest X-ray positive 218 (69%) Otherb 3 (1.0%) Submitted specimens At least one specimen 4 612 (97%) Both specimens 4 598 (96%) Smear-positive casesd 103 (33%) Definite 90 Probable 13 Smear-negative casese 211 (67%) Definite 211 Probable 0 Laboratory result At least one culture result availablec 4 602 (98%) Symptom screening Cough ≥2 weeksa 1 804 (4.8%) Haemoptysisa 319 (0.9%) Sputum production 15 698 (42%) Chest pain 11 405 (31%) Fever 17 811 (48%) Total symptom-screen positivea 1 916 (5.1%) Chest X-ray screening Normal 33 502 (90%) Abnormala 3 409 (9.2%) Other abnormality 310 (0.8%) Result not available N/A Total chest X-rays taken 37 221 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative, and other. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but two smear-positive slides, or one smear-positive slide with chest X-ray suggestive of TB. e Definite: MTB confirmed by culture. Probable: culture-positive (but MTB not confirmed) and chest X-ray suggestive of TB. Ineligible individuals 27 664 (41%) Children <15 years 19 681 (29%) Did not meet residency criteria 7 983 (12%) National TB prevalence surveys 2007–201684 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 6 2 4 10 8 12 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 500 700 900 1100 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 500 1000 1500 2000 2500 3000 3500 0 Prevalence per 10 0 0 0 0 population2 000 6 000 8 000 10 000 12 000 14 000 16 000 18 000 4 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 4000 2500 3000 3500 2000 1500 1000 500 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 5.06, variance 11.0, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimate of using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimated prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Total (≥15) CAMBODIA 85 Background Cambodia’s population was 15 million in 2011, and the average gross national income (GNI) per person was US$ 810 per year, making it a low-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000– 2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2011, the prevalence of HIV in the general population aged 15–49 years was 0.8% (95% confidence interval [CI]: 0.7–0.9%) (2), and it was estimated that 5.1% (95% CI: 4.6–5.6%) of TB patients were coinfected with HIV (3). Cambodia experienced a long period of political, economic and social turmoil following the regime of the Khmer Rouge (1975–1979), during which about three million people died (4) and many people left the country.1 In the 1980s, there were few health personnel per capita. In 1992, the country began to be rebuilt with United Nations support. In 1994, the National TB Programme (NTP) introduced the WHO-recommended DOTS strategy in hospitals (5,6). Further decentralization to primary care health centres was implemented between 1999 and 2004, with technical support from WHO and the Japanese International Cooperation Agency (JICA). By 2001, DOTS had been introduced in 268 (31%) health centres, and by 2005 all 853 health centres had been covered. Subsequently, the NTP strengthened the community DOTS programme with support from USAID, the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) and other partners. Treatment success rates were consistently maintained at above 90%. The NTP implemented the country’s first national TB prevalence survey in 2002, during the early stages of DOTS decentralization (7). The results showed a prevalence of smear-positive pulmonary TB of 362 (95% CI: 284–461) per 100  000 population aged 10 years or older; the prevalence of bacteriologically confirmed TB was 1208 (95% CI: 997–1463) per 100  000 population. The notification rate of new smear-positive TB cases peaked in 2005, and subsequently stagnated for 3 years. In December 2007, Cambodia was selected by the WHO Global Task Force on TB Impact Measurement as one of 22 global focus countries to undertake a national TB prevalence survey. The aim was to better understand the burden of TB disease at national and global levels, and to assess trends in countries with a baseline survey. A second national TB prevalence survey was needed to obtain an up-to-date measurement of the burden of TB disease and to assess trends since the 2002 survey. Planning for this second survey started in September 2009, and the survey was implemented in 2010–2011 (8). Key methods and results There were 62 survey clusters in three strata (urban, rural and other2), with a target cluster size of 640 individuals. A total of 68  087 individuals from 12  651 households were enumerated in the survey census, of whom 40 423 (59%) were eligible and invited to participate. Of these, 37  417 (93%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (9). A total of 4780 participants (13%) were eligible for sputum examination, of whom 4612 (97%) submitted at least one sputum specimen and 4598 (96%) submitted two sputum specimens. A total of 314 bacteriologically confirmed pulmonary TB cases were identified, including 103 cases of smear- positive TB. The prevalence of smear-positive TB was 271 (95% CI: 212–348) per 100  000 population (among those aged ≥15 years) and for bacteriologically confirmed TB it was 831 (95% CI: 707–977) per 100 000 population. When extrapolated to all forms of TB and to all ages, prevalence was 817 (95% CI: 690–954) per 100  000 population. There was variation between the three geographical strata, with a significantly lower prevalence per 100 000 population in urban areas than in rural and other regions. 1 The original report is in the Khmer language; excerpts have been translated by the Documentation Center of Cambodia for the Cambodian Genocide Program. 2 In the 2002 survey, four provinces (Mondulkiri, Rattanakiri, Preah Vihear and Steung Treng) were excluded because of serious difficulties in accessing these provinces and their relatively small population (<3% at that time). In the 2010 survey, for the purposes of comparisons between the two surveys, these four provinces were grouped into a stratum separate from the areas covered in the 2002 survey. National TB prevalence surveys 2007–201686 Other key results were: • the male to female ratio was 1.8 for smear- positive TB and 1.8 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age and was especially high in those aged 55 years and above; however, given the population distribution, the absolute number of bacteriologically confirmed TB cases was consistently high in most age groups (from those aged 25 years and above); • among bacteriologically confirmed TB cases, 30% were symptom-screen positive, and among the smear-positive cases, 44% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 1.7 overall, but varied from 0.4 in those aged 15–24 years to 2.2 in those aged 65 years or more, and was slightly higher for men than for women (2.0 versus 1.4); • among bacteriologically confirmed TB cases, 90% had no previous history of anti-TB treatment and only 1.9% were on anti-TB treatment at the time of the survey; and • of the 88 bacteriologically confirmed and 41 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 62 (70%) and 27 (66%), respectively, had previously sought care in a public or private health facility for their symptoms. Comparing the results between the 2002 and 2011 surveys (for the population aged ≥15 years and for the same or equivalent provinces): • there was a statistically significant decline of 38% and 46% in smear-positive and bacteriologically confirmed TB prevalence per 100  000 population, respectively; prevalence per 100 000 population was reduced in all age groups, although not all of the reductions were statistically significant; • the prevalence of those with smear-positive TB who reported symptoms decreased by 56%, while the prevalence of smear-positive TB among those who did not report symptoms decreased by only 8%; and • the P:N ratio declined from 2.0 to 1.7 between the two surveys, with an especially large change in those aged 15–24 years; and the P:N ratio in elderly men remained high. Implications of results The 2011 survey showed that the prevalence of TB fell significantly in the nine years between 2002 and 2011. A key factor in this reduction was the expansion of DOTS from hospitals to health centres, which was achieved with technical support from a JICA project. As part of DOTS expansion, enormous efforts were made by the NTP and its development partners to detect and treat the most infectious cases, and to increase their treatment success rate to more than 90%. The NTP in Cambodia maintained facility-level DOTS services at hospital and health-centre level as the core of TB control, while also expanding efforts to encompass community-level DOTS and public–private mix DOTS. Other factors that could have contributed to a reduction in TB prevalence included a reduction in the prevalence of HIV coinfection and a more than a doubling of GNI per capita between 2002 and 2011 (US$ 320 to US$ 810) (10). There were clear differences in the extent to which the prevalence of TB fell in those screening symptom- positive (56% decline, 2002–2011) compared with those screening symptom-negative (8% decline, 2002–2011). These differences are consistent with the emphasis on passive detection of self-referring symptomatic TB cases under the DOTS strategy. In 2002, symptomatic smear-positive TB cases with a cough of 2 weeks or longer or haemoptysis (62%) were more common than Photo credit: Kosuke Okada CAMBODIA 87 Major successes, challenges and lessons learned Major successes included smooth survey operations conducted in a highly transparent manner, a high participation rate, capacity development of health workers at the central and local level, and rapid dissemination of key results at a large dissemination event (in February 2012, within 5 months of the completion of field operations). Funding was mobilized from several sources (JICA, the Global Fund and USAID) and was efficiently managed. After the survey, staff and equipment (e.g. chest X-ray machines) deployed for the survey were used to undertake active case finding in specific geographical hotspots identified by the survey and in specific subpopulations (e.g. the elderly). Challenges were limited, but included a need to rely on two laboratories, given issues with standardizing laboratory work in other parts of country; slow data entry; some gaps between the population identified in the survey census and the national census data due to seasonal migration; and rescheduling of one cluster operation due to border security issues. Important lessons learned for future surveys were that: • institutional memory from a previous survey substantially facilitates a subsequent survey; the core staff of the 2002 survey led the 2011 survey, and the same international experts (from WHO and JICA) provided technical assistance; and • the availability of trained staff and survey equipment previously mobilized for active case detection in high-risk populations (in the case of Cambodia, since 2006) can help to ensure smooth survey operations. Photo credit: Kosuke Okada Photo credit: Kosuke Okada asymptomatic TB cases (38%). By 2011, symptomatic smear-positive TB cases accounted for 44% of all cases. Only 23% of people with smear-negative, culture-positive TB met the 2011 NTP definition of an individual with presumptive TB. This evolution in the TB epidemic had two major programmatic implications. The first was a need to strengthen diagnostic capacity for outpatients with respiratory symptoms, by reviewing and updating the diagnostic algorithm which had previously relied heavily on smear microscopy. Suggested updates included more extensive use of chest X-ray for people with any respiratory symptom, including a referral system for people with smear-negative presumptive TB to a health facility equipped to carry out chest X-rays, and the replacement of smear microscopy with more sensitive diagnostic tools, such as Xpert® MTB/RIF. The second implication was that active case detection activities should be expanded to specific groups with a high prevalence of TB, such as the elderly, household contacts of people with smear-positive TB and people coinfected with HIV. Other implications included: • a need to improve the capacity of health-care workers to clinically recognize TB disease, given that 55% of those with smear-positive TB and cough of any duration had already sought care (and 45% of these cases had consulted a public health facility); more than half (55%) of those with smear-negative, culture-positive TB and a cough of any duration had also previously sought care; and • a need to consider the wider use of TB preventive therapy, especially among older people with a chest X-ray suggestive of inactive TB and negative bacteriological test results. National TB prevalence surveys 2007–201688 The expertise and experience of those involved in leading and managing the 2002 and 2011 surveys in Cambodia proved to be an invaluable source of assistance to surveys in other countries. Survey staff from Cambodia provided direct technical assistance to the surveys in Ethiopia, Kenya, Lao People’s Democratic Republic, Malawi, Rwanda and Uganda. In addition, two training courses were held in Cambodia during the 2011 survey, which provided survey coordinators and their technical partners with an opportunity to witness and learn from a model survey operation at first-hand. Staff from the Cambodia survey played a crucial role in Asia–Asia and Asia–Africa collaborations that were strongly promoted by WHO to support surveys implemented from 2009–2015. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. Report of the Research Committee on Pol Pot’s Genocidal Regime Phnom Penh, Cambodia: 1983 (http://www.dccam.org/, accessed May 2017). The original report is in the Khmer language; excerpts have been translated by the Documentation Center of Cambodia for the Cambodian Genocide Program. 5. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 6. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 7. Report of the national TB prevalence survey, 2002. Phnom Penh: Cambodia Ministry of Health; 2005. 8. Report of the second national TB prevalence survey, 2011. Phnom Penh: Cambodian Ministry of Health; 2012 (http://open_ jicareport.jica.go.jp/pdf/12120325.pdf, accessed January 2018). 9. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 10. Mao TE, Okada K, Yamada N, Peou S, Ota M, Saint S et al. Cross- sectional studies of tuberculosis prevalence in Cambodia between 2002 and 2011. Bull World Health Organ. 2014;92(8):573–581 (https://www.ncbi.nlm.nih.gov/pubmed/25177072, accessed May 2017). 89 Surveyed clusters (N=176)a CHINA Name Role Organization Wang Lixia Principal investigator National Center for Tuberculosis Control and Prevention, Chinese Center for Disease Control and Prevention (NCTB, China CDC) Zhang Hui Survey coordinator NCTB, China CDC Cheng Shiming, Chen Mingting, He Guangxue Survey design NCTB, China CDC Jiang Shiwen Survey design, data collection NCTB, China CDC Zhao Yanlin Survey design, laboratory manager NCTB, China CDC Ruan Yunzhou Survey design, data collection and analysis NCTB, China CDC Du Xin, Chen Wei Sampling NCTB, China CDC Zhou Lin Diagnosis NCTB, China CDC Zhou Xinhua Radiology coordinator Beijing Tuberculosis and Thoracic Tumor Research Institute Li Renzhong Data collection and analysis NCTB, China CDC Xia Yinyin Data manager, data analysis NCTB, China CDC Xu Caihong, Li Jun Data manager NCTB, China CDC Wang Shengfen Data analysis, laboratory manager NCTB, China CDC Chen Yude Technical assistance (survey design, quality control, data analysis) Peking University Health Science Center Wang Xiexiu Technical assistance (survey design, quality control, data analysis) Tianjin Center for Disease Control and Prevention Jin Shuigao Technical assistance (survey design, quality control, data analysis) China CDC Tang Danlin Technical assistance (survey design) China-Japan friendship hospital Qian Yuanfu, Wang Zhongren, Duanmu Hongjin, Zhao Fengzeng Technical assistance (survey design) Beijing Tuberculosis and Thoracic Tumor Research Institute Wu Zhenglai Technical assistance (survey design) Peking Union Medical College Zhu Guilin Technical assistance (survey design) Chinese Anti-tuberculosis Association Tu Dehua Technical assistance (radiology) Beijing Research Institute for Tuberculosis Control Pan Yuxuan, Zou Jiqian, Zhu Lizhen Technical assistance (radiology) Beijing Tuberculosis and Thoracic Tumor Research Institute Shi Hongsheng Technical assistance (data analysis) Beijing Tuberculosis and Thoracic Tumor Research Institute Cao Jiping Technical assistance (data analysis) Hebei Center for Disease Control and Prevention Xu Weiguo Technical assistance (data analysis) Jiangsu Center for Disease Control and Prevention Zheng Suhua, Zhang Zongde Technical assistance (data analysis) Beijing Tuberculosis and Thoracic Tumor Research Institute Sang Jae Kim Technical assistance (laboratory advisor) Korean Institute of Tuberculosis, Republic of Korea Philippe Glaziou, Ikushi Onozaki, Charalampos Sismanidis. Technical assistance (survey advisor, analysis) WHO headquarters Key people Data sources ■ Disease Control Bureau of the Ministry of Health – Chinese Center for Disease Control and Prevention. Report on the 5th national tuberculosis epidemiological survey in China – 2010. Beijing, China: Military Medical Science Press; 2011. ■ Wang L, Zhang H, Ruan Y, Chin DP, Xia Y, Cheng S et al. Tuberculosis prevalence in China, 1990–2010; a longitudinal analysis of national survey data. Lancet. 2014;383(9934):2057–2064. Summary statistics Participation rate 96% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 119 3.0 Prevalence:notification ratio (smear-positive TB, ≥15 years) 1.7 Finance Amount (US$) Ministry of Health, China 5 620 520 Total budget 5 620 520 Survey organization and financing Implementing agency: National TB Control Programme NCTB* * A leading group, technical advisory group and survey office were set up at all administrative levels (national, provincial, prefectural, county/district) to support survey implementation. There were 160 field survey teams in 31 provinces (autonomous regions and municipalities). a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2010 National TB prevalence surveys 2007–201690 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Province/prefecture/county/township/ village Sample size assumptions • Smear-positive prevalence 116 per 100 000 (≥15 years) • Precision 0.15 • Design effect 1.8 • k 0.7 • Response rate 95% • Sample size (estimated) 264 000 Number of clusters 176 Cluster size 1 500 Eligibility criteria • Age ≥15 years • Residency Individuals who lived for ≥6 months in the household Survey design and methodology Laboratory methodology Smear Three samples (spot, night and morning): direct preparation, ZN Culture Two samples out of three (spot, night and morning) selecteda: direct preparation, LJ media Identification of MTB PNB TB drug susceptibility test Done Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interviewa Cough ≥2 weeks and/or haemoptysis for any duration Chest X-rayb Any lung abnormality Other Participants with known active pulmonary TB with normal chest X-ray, and chest X-ray exempted a The questionnaire on the socioeconomic conditions was done only for active pulmonary TB patients. b Conventional radiography. a Two samples were selected based on their smear result (-, +, ++, +++, ++++) and appearance (bloody, mucopurulent or salivary). Samples with higher smear grades and better appearance were selected for culture. Field data collection Paper Database SPV Method of analysis Complex sampling method-based weighted adjustment Results first published in a report/paper December 2011 Official dissemination event March 2011 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 5 462 – Location of care sought • Consulted medical facility N/A N/A Public facility N/A N/A Private facility N/A N/A Other N/A N/A • Pharmacy N/A N/A • Traditional healer N/A N/A No action taken N/A N/A Unknown N/A N/A Key survey results a Cough ≥2 weeks or haemoptysis for any duration. a Among 1 310 participants (active pulmonary TB patients), 1 301 were interviewed about their health-care seeking behaviour. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). b A total of 280 MTB strains were examined. Survey participants currently on TB treatment Number % Total participants currently on TB treatmenta 73 – • Treated in the public sector 72 99 • Treated in the village/community clinic 1 1.4 • Treated in unknown sector 0 0 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 9 2.6 Design effect k Smear-positive TB 1.7 0.9 Bacteriologically confirmed TB 1.4 0.5 Other sputum results Number % Total smear-positive participants 207 – Smear-positive participants without MTB confirmationa 61 30 Isolates with MDR-TB detectedb 19 6.8 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 66 53–79 119 103–135 Male 99 74–123 177 149–204 Female 32 23–42 59 46–72 15–24 years 20 0.8–39 45 16–75 25–34 years 35 15–55 69 38–100 35–44 years 38 21–55 79 55–104 45–54 years 67 38–95 103 71–135 55–64 years 136 90–181 200 151–249 ≥65 years 188 138–238 369 303–435 Urban 49 25–74 73 46–99 Rural 78 64–93 153 133–172 a Age ≥15 years unless otherwise specified. CHINA 91 Survey flow: census to final outcomes Field operations: April 2010 to July 2010 Individuals enumerated in census 447 563 Eligible for sputum examination 9 825 (3.9%) Total bacteriologically confirmed cases 347 Eligible study population 263 281 (59%) Total participants 252 940 (96%) Interview and chest X-ray 250 716 (99%) Interview only 2 224 (0.9%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 797 (8.1%) Symptom positive, chest X-ray negative or N/A 4 665 (48%) Symptom negative, chest X-ray positive 2 189 (22%) Otherb 2 174 (22%) Symptom positive, chest X-ray positive 143 (41%) Symptom positive, chest X-ray negative or N/A 17 (5.0%) Symptom negative, chest X-ray positive 182 (52%) Otherf 5 (1.4%) Submitted specimens At least one specimen N/A Both specimens N/A Smear-positive casesd 188 (54%) Definite 146 Probable 42 Smear-negative casese 159 (46%) Definite 159 Probable N/A Laboratory result At least one culture result availablec 9 684 (99%) Symptom screening Cough ≥2 weeksa 5 364 (2.1%) Haemoptysisa 293 (0.1%) Sputum production N/A Chest pain N/A Fever N/A Total symptom-screen positivea 5 462 (2.2%) Chest X-ray screening Normal 242 152 (97%) Abnormala 2 986 (1.2%) Other abnormality 5 578 (2.3%) Result not available 0 (0%) Total chest X-rays taken 250 716 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative (2 167), other (not specified) (7). c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture, and non NTM (all 42 participants had culture-negative results). e Definite: MTB confirmed by culture. Probable: no definition. f Chest X-ray exempted and symptom-screen negative (4), other (not specified) (1). Ineligible individuals 184 282 (41%) Children <15 years 58 940 (13%) Did not meet residency criteria 125 342 (28%) National TB prevalence surveys 2007–201692 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 60 90 120 150 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 50 100 150 200 250 300 350 400 0 Prevalence per 10 0 0 0 0 population 100 000 50 000 350 000 300 000 400 000 450 000 150 000 200 000 250 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 500 300 350 400 450 200 250 100 150 50 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 1.97, variance 5.55, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 25 10 5 15 20 45 40 35 30 50 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Total (≥15) CHINA 93 Background China’s population was 1.3 billion in 2010, making it the most populous country in the world. China experienced rapid economic growth throughout the 1990s and 2000s, and by 2010 it was an upper-middle-income country with an average gross national income (GNI) per person of US$ 4340 (1). For most of the Millennium Development Goal (MDG) era (2000–2015), China ranked second (after India) in terms of the estimated number of new tuberculosis (TB) cases occurring each year. It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2010, it was estimated that 1.5% (95% confidence interval [CI]: 1.2–1.8%) of TB patients in China were coinfected with HIV (2). The TB epidemic began to be addressed as a high priority and on a large scale in the 1990s. At that time, a TB control project funded by the World Bank and China’s domestic resources was used to implement the WHO-recommended DOTS strategy in 13 provinces, which accounted for half the country’s population (3,4). National TB prevalence surveys implemented in 1990 and 2000 showed a 30% reduction in TB prevalence in areas where the project was implemented. In contrast, overall TB prevalence fell by less than 20% during this period (5). To accelerate progress in TB control, in 2001 the State Council of China launched a new 10-year TB control plan, which resulted in national coverage of DOTS by 2005. In the same year, China achieved the global TB control targets of detecting at least 70% of all estimated new smear-positive TB cases and successfully treating more than 85% of detected cases. To assess progress made by 2010 in reducing the burden of TB, a national TB prevalence survey was implemented. This was the fifth such survey in China, following previous surveys in 1979, 1984–1985, 1990 and 2000. The 2010 survey was by far the largest national TB prevalence survey undertaken worldwide in the period 2000–2015, having a sample size of 260 000 people. The survey started in April 2010 and was completed in July 2010 (6). Key methods and results There were 176 survey clusters in two strata (urban and rural); the target cluster size was 1500 individuals. A total of 447  563 individuals from 130  655 households were enumerated in the survey census, 263  281 (59%) of whom were eligible to participate. Of these, 252  940 (96%) did so. All participants were screened in line with the 2011 algorithm recommended by WHO; that is, using chest X-ray and a symptom-based questionnaire (7). A total of 9825 participants (3.9%) were eligible for sputum examination. Field operations were conducted by provincial teams, in contrast to the use of centrally managed cluster operations in other countries. Local reference laboratories were used for smear and culture. The central survey unit provided an operation manual, training of provincial staff, monitoring and supervision, and external quality assurance on diagnostic tools. Identification of isolated colonies and TB drug susceptibility testing were performed by the National TB Reference Laboratory in Beijing. A total of 347 bacteriologically confirmed pulmonary TB cases was identified, including 188 cases of smear-positive TB. The prevalence of smear-positive TB was 66 (95% CI: 53–79) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 119 (95% CI: 103–135) per 100  000 population. When extrapolated to all forms of TB and to all ages, prevalence was 108 (95% CI: 94–123) per 100 000 population. The prevalence of bacteriologically confirmed TB was higher in rural than in urban areas, and higher in the western region (198 per 100  000 population; 95% CI: 167–229) than the central region (118 per 100 000 population; 95% CI: 81–154) and eastern region (65 per 100  000 population; 95% CI: 50–81). These three regions were Photo credit: Yin Yin Xia National TB prevalence surveys 2007–201694 defined based on geography and economic status, with wealth generally declining from east to west. Other key results were: • the male to female ratio was 3.1 for smear- positive TB and 3.0 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age, as did the absolute number of bacteriologically confirmed TB cases; over 60% (219/347) of prevalent TB cases were aged 55 or more; • among the bacteriologically confirmed TB cases, 46% were symptom-screen positive, and of the smear-positive cases, 49% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 1.7 overall, but varied from 0.7 in those aged 15–24 years to 2.5 in the age groups 55–64 years and 65 years or more; also, it was higher for men than for women (1.8 versus 1.4); • among the bacteriologically confirmed TB cases, 85% had no previous history of anti-TB treatment and 2.6% were on anti-TB treatment at the time of the survey; and • of the 153 bacteriologically confirmed survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 48% (73 of the 151 for whom there were results about health-care seeking behaviour) had previously sought care in a public or private health facility for their symptoms; and of the 87 smear-positive TB cases who reported symptoms but were not on anti-TB treatment at the time of the survey, 57% (49 of the 86 for whom there were results about health-care seeking behaviour) had previously sought care in a public or private health facility. The survey showed that TB prevalence declined substantially between 1990 and 2010. Based on analysis of results according to the diagnostic protocol used in 1990 to allow for a fair comparison, the prevalence of smear- positive TB fell from 170 (95% CI: 166–174) per 100 000 population in 1990 to 59 (95% CI: 49–72) per 100  000 population in 2010. In the 1990s, the prevalence of smear-positive TB fell only in the provinces where DOTS was implemented. After 2000, declines were observed in all provinces. Of the total reduction in the prevalence of smear-positive TB from 1990–2010, 70% occurred after 2000. Implications of results The halving of TB prevalence in 20 years was assisted by a nationwide DOTS programme being implemented throughout the country’s network of local centres for disease control, improved reporting and referral hospital systems, and a policy of free treatment for all patients with active pulmonary TB, alongside rapid socioeconomic development. Specifically, there were tremendous increases in GNI per capita (from US$  330 in 1990 to US$ 4340 in 2010) and in living conditions overall (the human development index improved from 0.501 in 1990 to 0.699 in 2010) (1,8). The overall fall in the prevalence of TB, in combination with the reduction in the proportion of prevalent cases with a previous history of TB, also had Photo credit: Yin Yin Xia Photo credit: Yin Yin Xia CHINA 95 a major impact on reducing the burden of multidrug- resistant TB (MDR-TB). Clear differences in TB prevalence between men and women, and across age groups and geographic regions also showed the need for considerable further efforts, such as policy or programmatic measures targeted to particular population groups and regions. TB control and policy should prioritize western and central China, rural areas, the elderly, ethnic minorities and those who are poor. It was also recognized that central and provincial governments should strengthen funding support and input for infrastructure, facilities and human resources for these areas and population groups. The survey also showed that there was a need to improve TB notification and treatment of patients with TB within the hospital sector. Major successes, challenges and lessons learned The 2010 survey followed methods recommended in the first (2007) edition of WHO’s handbook on prevalence surveys (9). This included three modifications compared with the fourth (2000) survey: • inclusion of adults (aged ≥15 years) only;1 • no tuberculin skin testing; and • use of direct chest X-ray (posteroanterior) film images instead of fluoroscopy screening. Major successes included implementation of the survey within two years of initiating planning; full mobilization of funding required for field operations at provincial level, which enabled field operations to be completed within four months; a high participation rate; a sample large enough to produce precise provincial as well as national estimates of TB prevalence; and prompt finalization of results and production of a survey report. There were two major challenges. The first was the level of internal migration in China. The technical expert group established to provide advice on the survey suggested that the residential criteria for determining whether people were eligible to participate in the survey should be defined as “resident for one month at the time of the survey census”. In practice, the definition of “six months residency in the household” was used since this was the official government definition. Using this more restrictive criterion, 10% of otherwise-eligible invitees were defined as non-permanent residents, and just over 20% of people identified by the survey census were not included in the survey because they had moved in the past six months. In addition, the survey team could not find 30% of the registered population in the survey clusters; this probably also reflected internal migration, especially of young men to urban areas. The second challenge involved culture testing. Although the central team and the National TB Reference Laboratory made extensive efforts to standardize survey operations in all provinces, the yield from cultured sputum specimens was low or non-existent in some provinces. In other TB prevalence surveys in Asia, the number of smear-negative culture-positive TB cases was 1.2–2.0 times higher than the number of smear-positive TB cases.2 Among 31 provinces in the prevalence survey in China, only six (19%) had a ratio of smear-negative culture-positive to smear-positive TB cases of 1.5 and above. Five other provinces had no yield from cultured sputum specimens. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 3. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 4. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 5. Wang L, Zhang H, Ruan Y, Chin DP, Xia Y, Cheng S et al. Tuberculosis prevalence in China, 1990–2010; a longitudinal analysis of national survey data. Lancet. 2014;383(9934):2057– 2064 (https://www.ncbi.nlm.nih.gov/pubmed/24650955, accessed May 2017). 6. Disease Control Bureau of the Ministry of Health – Chinese Center for Disease Control and Prevention. Report on the 5th national tuberculosis epidemiological survey in China – 2010. Beijing, China: Military Medical Science Press; 2011. 1 The 2000 survey included children as well as adults; for every case found among children, 8000 children were screened. 2 Cambodia: 2.0 (smear-positive TB cases: smear-negative culture-positive TB cases = 103:211), Indonesia: 1.6 (165:261), Lao PDR: 1.2 (107:130), Mongolia: 1.8 (88:160), Myanmar: 1.5 (123:188) and Thailand: 1.5 (58:84). National TB prevalence surveys 2007–201696 7. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481 /9789241548168_eng.pdf, accessed August 2017). 8. UNDP. Human development report – China, human development indicators. United Nations Development Programme (UNDP);(http://hdr.undp.org/en/countries/profiles/CHN, accessed May 2017). 9. World Health Organization. Assessing tuberculosis prevalence through population-based surveys. Geneva: WHO; 2007 (https:// apps.who.int/iris/handle/10665/206962, accessed January 2018). 97 Finance Amount (US$) Ministry of Public Health 481 963 Global Fund 896 026 Total budget 1 377 989 Name Role Organization Kim Hyong Hun Chair of steering committee Ministry of Public Health (MoPH) Ri Chan Hyok Member of steering committee, principal investigator MoPH Choe Tong Chol Member of steering committee MoPH Jo Won Ryong Leader of central survey data management team MoPH Rim Gye Tong Leader of central survey management team MoPH Choe Tal Bom Leader of central survey interview team Pyongyang Medical college under Kim Il Sung university Ri Jong Chan Leader of central survey chest X-ray team Central TB Preventive Institute (CTPI) Yun Jong Chol Leader of central survey laboratory team CTPI Ko Jin Hyok Survey coordinator TB Programme Management Unit (PMU), MoPH Partha Pratim Mandal TB medical officer WHO South-East Asia Regional Office (SEARO) Mubeen Aslam Global Fund programme coordinator UNICEF, Democratic People's Republic of Korea M. Bintari Dwihardiani Technical assistance (survey advisor) WHO Indonesia Philippe Glaziou Technical assistance (analysis) WHO headquarters Charalambos Sismanidis Technical assistance (analysis) WHO headquarters Key people Data sources ■ Report of DPRK National TB Prevalence Survey (2015–2016), Department of TB and Hepatitis, Ministry of Public Health DPR Korea; 2017. Summary statistics Participation rate 84% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 587 2.9 Prevalence:notification ratio (Bacteriologically confirmed TB, ≥15 years) 1.2 Survey organization and financing Implementing agency: National TB Control Programme Surveyed clusters (N=100). Cluster location not provided.a a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. DEMOCRATIC PEOPLE’S REPUBLIC OF KOREA 2015–2016 National TB prevalence surveys 2007–201698 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural/construction area Region (east, west, in-land) Sampling unit Province/city, county or district/Ri or Dong Sample size assumptions • Smear-positive prevalence 220 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.4 • k 0.5 • Response rate 85% • Sample size (estimated) 70 000 Number of clusters 100 Cluster size 700 Eligibility criteria • Age ≥15 years • Residency Resident in the household for the last 2 weeks Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): concentrated preparation, FM (LED, auramine stain) Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB SD Bioline TB Ag MPT64 rapid test TB drug susceptibility test Not done as per protocol Xpert® MTB/RIF Not done as per protocol HIV test Not done as per protocol Screening criteria Interviewa Cough more than 2 weeks and/or haemoptysis Chest X-rayb Any lung abnormality Other N/A a An in-depth interview on health-care seeking behavoiur was done for those who screened positive (symptom) and/or who had TB history. b Conventional radiography. Field data collection Paper Database Microsoft Access® Method of analysis MI+IPW Results first published in a report/paper April 2018 Official dissemination event October 2017 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 330 283–377 587 520–655 Male 535 443–627 917 783–1 052 Female 164 118–210 319 256–382 15–24 years 42 0–84 155 70–240 25–34 years 333 218–448 579 410–748 35–44 years 417 302–531 764 611–916 45–54 years 525 400–651 877 705–1 049 55–64 years 341 207–474 595 410–781 ≥65 years 265 137–393 444 264–624 Urban 330 267–394 577 489–665 Rural 361 288–434 659 555–764 Construction unit 66 0–138 102 0–219 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 944 – Location of care sought • Consulted medical facility 1 743 N/A Public facility 1 743 100 Private facility 0 0 Other (NGO, village doctor) 0 0 • Pharmacy 0 0 • Traditional healer 3 0.1 Self-treated 0 0 No action taken 1 192 41 Unknown 6 0.2 a Cough more than 2 weeks and/or haemoptysis. a All were NTM. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 106 – • Treated in the public sector 101 96 • Treated in the private sector 0 0 • Treated in unknown sector 5 4.7 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 87 26 Design effect k Smear-positive TB 1.5 0.5 Bacteriologically confirmed TB 2.0 0.5 Other sputum results Number % Total smear-positive participants 203 – Smear-positive participants without MTB confirmationa 10 4.9 Isolates with MDR-TB detected NA NA a Age ≥15 years unless otherwise specified. 99DEMOCRATIC PEOPLE’S REPUBLIC OF KOREA Survey flow: census to final outcomes Field operations: October 2015 to May 2016 Individuals enumerated in census 90 466 Eligible for sputum examination 4 802 (7.9%) Total bacteriologically confirmed cases 340 Eligible study population 71 877 (80%) Total participants 60 683 (84%) Interview and chest X-ray 59 649 (98%) Interview only 1 034 (1.7%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 1 028 (21%) Symptom positive, chest X-ray negative or N/A 1 916 (40%) Symptom negative, chest X-ray positive 1 858 (39%) Other N/A Symptom positive, chest X-ray positive 187 (55%) Symptom positive, chest X-ray negative or N/A 7 (2.1%) Symptom negative, chest X-ray positive 146 (43%) Other N/A Submitted specimens At least one specimen 4 586 (96%) Both specimens 4 462 (93%) Smear-positive casesc 187 (55%) Definite 158 Probable 29 Smear-negative casesd 153 (45%) Definite 153 Probable N/A Laboratory result At least one culture result availableb 4 583 (95%) Symptom screening Cough ≥2 weeksa 2 805 (4.6%) Haemoptysisa 267 (0.4%) Weight loss 1 032 (1.7%) Fever 1 335 (2.2%) Night sweats 1 214 (2.0%) Total symptom-screen positivea 2 944 (4.9%) Chest X-ray screening Normal 56 763 (95%) Abnormala 2 886 (4.8%) Other abnormality N/A Result not available N/A Total chest X-rays taken 59 649 a Eligible for sputum collection. b Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. c Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but chest X-ray abnormal findings at central reading. d Definite: MTB confirmed by culture, with having either chest X-ray abnormal findings at central reading or follow-up evidence. Probable: no definition. Ineligible individuals 18 589 (21%) Children <15 years 18 589 (21%) Did not meet residency criteria 0 (0%) National TB prevalence surveys 2007–2016100 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 80 70 90 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of bacteriologically confirmed TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 400 800 1200 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 800 900 1000 0 Prevalence per 10 0 0 0 0 population 5 000 25 000 30 000 35 000 10 000 15 000 20 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1200 800 1000 400 600 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2019 revision). b The data did not suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 3.4, variance 4.22, p=0.12). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. C Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2019 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 12 6 4 2 8 10 20 18 16 14 22 0 0 1 2 3 4 5 6 7 8 9 10 11 Total (≥15) 101DEMOCRATIC PEOPLE’S REPUBLIC OF KOREA Background The Democratic People’s Republic of Korea had a population of 25 million in 2015 and was classified as a low-income country (1). The country was one of the top 30 high tuberculosis (TB) burden countries (HBCs) defined by the World Health Organization (WHO) for the period 2016–2020. Although no data about HIV prevalence were available (2), it was estimated that 0.32% (95% confidence interval [CI]: 0.26–0.38) of TB patients were coinfected with HIV in 2015 (3). The country’s National TB Control Programme (NTP) was established in 1968. By 2003, the NTP had adopted and expanded the WHO DOTS strategy nationally, including the establishment of a unified surveillance system for TB case registration. To estimate the burden of disease, the Democratic People’s Republic of Korea conducted a national survey of the annual risk of TB infection (ARTI) in 2007 (4). The estimated ARTI was 3.1% (95% CI: 2.8–3.3%); based on this result, the burden of new smear-positive pulmonary TB disease was estimated as 155±34 cases per 100 000 population per year. Thereafter, the notification rate for smear-positive pulmonary TB increased. Since 2010, reported treatment success rates were consistently 90% or higher. In mid-2013, the NTP decided to conduct a national TB prevalence survey to improve estimates of TB disease burden. Following 2 years of preparations, the NTP undertook a national TB prevalence survey between October 2015 and May 2016. Key methods and results There were 100 clusters across three population strata (defined as urban, rural and construction areas) and three geographical strata (defined as east, west and inland regions), with a target cluster size of 700 individuals. A total of 90 466 people were enumerated in the survey census, of whom 71 877 (80%) were eligible and invited to participate. Of these, 60 683 (84%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (5). A total of 4802 (7.9%) participants were eligible for sputum examination, of whom 4586 (96%) submitted at least one sputum sample and 4462 (93%) submitted two sputum samples. A total of 340 bacteriologically confirmed pulmonary TB cases were identified, including 187 cases of smear- positive TB. The prevalence of smear-positive TB was 330 (95% CI: 283–377) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 587 (95% CI: 520–655) per 100 000 population. The prevalence of bacteriologically confirmed TB varied by strata: 577 (95% CI: 489–665) per 100 000 population in urban areas, 659 (95% CI: 555–764) per 100 000 population in rural areas and 102 (95% CI: 0–219) per 100 000 population in construction areas. Photo credit: National TB Programme of the Democratic People’s Republic of Korea National TB prevalence surveys 2007–2016102 Other key results were as follows: • the male to female ratio was 3.3 for smear- positive TB and 2.9 for bacteriologically confirmed TB; • the prevalence per 100 000 population increased with age, with a peak in those aged 45–54 years, before declining with age; the absolute number of TB cases was high in those aged 25–54 years; • among the bacteriologically confirmed TB cases, 57% were symptom-screen positive, and among the smear-positive TB cases, 66% were symptom-screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 1.2 overall, but varied from 0.5 in those aged 15–24 years to 2.2 in those aged 65 years or more, and was higher for men than for women (1.5 versus 0.9); • among the bacteriologically confirmed TB cases, 91% (310/340) had no previous history of TB treatment and 26% (87/340) were on treatment at the time of the survey; and • of the 107 bacteriologically confirmed cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 96 (90%) had previously sought care in a public health facility for their symptoms; and of 123 smear-positive TB cases with TB symptoms, only 69 (56%) had sought care. Implications of results Based on the results from the national TB prevalence survey, the overall prevalence (for all forms and all ages) was estimated at 600 (95%  CI:  527–676) per 100 000 population. This was higher than the pre-survey estimate of 490 per 100 000 population used in the initial design in 2012 (4). Based on the survey, TB incidence was re- estimated at 513 per 100 000 population per year (95% CI: 446–584), equivalent to 131 000 new cases per year in 2017 (6). This was also 1.2 times higher than the previous estimate of 442 (95% CI: 412–473) per 100 000 population in 2014 (7). The survey findings were adopted by relevant stakeholders for the development of a new national TB control strategy (2019–2022) and a proposal to the Global Fund to Fight AIDS, Tuberculosis and Malaria (the Global Fund) (2018–2021). Clear differences in TB prevalence between men and women and across age groups showed the need for considerable further efforts, such as policy or programmatic measures targeted to specific population groups. TB control and policy should give particular attention to men, given the high burden in men (especially in those of working age) and the large gap between prevalence and notifications (especially in those aged 45–54 and ≥65 years). It was also clear that central and provincial governments needed to strengthen funding support and other inputs for infrastructure, facilities and human resources in specific areas and population groups. Photo credit: National TB Programme of the Democratic People’s Republic of Korea Photo credit: National TB Programme of the Democratic People’s Republic of Korea 103DEMOCRATIC PEOPLE’S REPUBLIC OF KOREA Other implications included the need to: • strengthen community screening for TB to ensure earlier detection, treatment and notification of cases; • review the surveillance system, given that a large proportion of TB cases (26%) on treatment were not notified to the NTP; • review the use of and access to chest X-ray screening in the early detection of cases, given that 43% (146/340) of bacteriologically confirmed survey cases were only identified by chest X-ray; • expand the range of laboratory tests to diagnose TB beyond smear and culture, recognizing that high-level negotiations would be required to ensure the sustainable use and expansion of Xpert MTB/RIF; • strengthen health services, especially at the peripheral level (Ri/Dong clinic and city/ county hospital) – for example, by raising health worker awareness of TB symptoms and making diagnostics more widely available; many bacteriologically confirmed cases had sought care before diagnosis, including nearly half of all symptomatic smear-positive TB cases, but not been diagnosed; and • strengthen community awareness of TB, since more than 40% (1193/2944) of participants with chronic cough or haemoptysis (or both) at the time of the survey had not sought care for their symptoms – men accounted for the vast majority of symptomatic participants. Major successes, challenges and lessons learned Despite financial and technological constraints, the first national TB prevalence survey of the Democratic People’s Republic of Korea managed to achieve its primary objective and field operations were successfully completed within a year. Almost all national TB prevalence surveys since 2015 used Xpert MTB/RIF or Xpert Ultra, in addition to culture, as part of the diagnostic algorithm for all participants who screened positive. However, given important limitations, the Democratic People’s Republic of Korea’s survey was only able to use smear and culture, as was originally recommended in the lime book (5). Other constraints included reliance on paper instead of electronic data collection systems, and conventional instead of digital mobile chest X-ray machines. The survey was among those that cost the least to implement, at US$ 1.4 million. The Global Fund contributed about US$  900 000, with the remainder being supplied by the Ministry of Public Health. The survey’s high level of participation (84%) was probably due to strong leadership and extensive community engagement by the large survey teams (six teams of 25 people) and central survey team (>100 people). Photo credit: National TB Programme of the Democratic People’s Republic of Korea National TB prevalence surveys 2007–2016104 There were few regular international missions to provide technical assistance, due to administrative challenges and access restrictions. A technical consultant from the national TB prevalence survey team of Indonesia provided some in-country advice (June 2016), and two laboratory experts from the supranational reference laboratory in Hong Kong Special Administrative Region reviewed laboratory progress and results (July 2016). There was good engagement with the WHO country office and WHO headquarters to support data review, final analysis and report writing. Collaboration with multiple international stakeholders, from procurement to dissemination, also helped to ensure that the survey was a success. Major challenges included interruptions of funding in 2014 that led to a 1-year delay before field operations could be started. In addition, there were long delays in the procurement of mobile conventional chest X-ray machines. This meant that the survey could only start with two instead of four field teams during phase one of the survey (October to November 2015). Extended delays from the end of field operations to the final dissemination of results were due to insufficient human resources for data entry, analysis and report writing. Other challenges included the replacement of five clusters due to poor road conditions. An important lesson learned for future surveys was the importance of good planning and collaboration for smooth implementation. Specifically, the procurement of laboratory and chest X-ray equipment should be completed before starting field operations, and sufficient lead times allowed for this purpose. References 1. The World Bank (https://data.worldbank.org, accessed July 2017). 2. UNAIDS (https://aidsinfo.unaids.org/, accessed July 2017). 3. World Health Organization. Global Tuberculosis Database. 2017 (https://www.who.int/tb/data/en/, accessed July 2017). 4. Report of DPRK National TB Prevalence Survey (2015–2016), Department of TB and Hepatitis, Ministry of Public Health, Democratic People’s Republic of Korea; 2017. 5. World Health Organization. Tuberculosis prevalence surveys: a handbook. Geneva: WHO; 2011 (https://apps.who.int/iris/ bitstream/handle/10665/44481/9789241548168_eng.pdf, accessed August 2017). 6. World Health Organization. Global tuberculosis report 2017. Geneva: WHO; 2017 (https://www.who.int/tb/data/en/, accessed July 2017). 7. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015 (https://www.who.int/tb/data/en/, accessed July 2017). Photo credit: National TB Programme of the Democratic People’s Republic of Korea 105 Name Role Organization Amha Kebede Director general, principal investigator Ethiopian Public Health Institute Zeleke Alebachew Survey coordinator Ethiopian Public Health Institute Fasil Tsegaye Deputy survey coordinator Ethiopian Public Health Institute Almaz Abebe Directorate director, Infectious and Non Infectious Disease Research Ethiopian Public Health Institute Eshetu Lema Senior laboratory advisor Ethiopian Public Health Institute Mulualem Agonafer Laboratory manager Ethiopian Public Health Institute Gashawtena Fantu Central X-ray radiologist Saint Paul’s Hospital Molla Endale Central X-ray radiologist Saint Paul’s Hospital Shewalem Negash Central X-ray radiologist Saint Paul’s Hospital Feleke Dana Data manager Ethiopian Public Health Institute Menelik Balcha Field team leader Ethiopian Public Health Institute Sale Workneh Field team leader Ethiopian Public Health Institute Tedla Fiseha Field team leader Ethiopian Public Health Institute Tibebu Biniam Field team leader Ethiopian Public Health Institute Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Peou Satha Technical assistance (survey advisor) Consultant, Cambodia Hazim Timimi Technical assistance (data management) WHO headquarters Charalampos Sismanidis Technical assistance (design and analysis) WHO headquarters Key people Surveyed clusters (N=85)a Data sources ■ First Ethiopian national population-based tuberculosis prevalence survey. Addis Ababa: Ministry of Health, Federal Democratic Republic of Ethiopia; Ethiopian Health and Nutrition Research Institute; 2011. ■ Kebede AH, Alebachew Z, Tsegaye F, Lemma E, Abebe A, Agonafir M et al. The first population-based national tubercu- losis prevalence survey in Ethiopia, 2010–2011. Int J Tuberc Lung Dis. 2014;18(6):635–639. ■ Survey dataset. Summary statistics Participation rate 90% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 277 1.2 Prevalence:notification ratio (smear-positive TB, ≥15 years) 1.2 Finance Amount (US$) The Global Fund/Ministry of Health, Ethiopia 2 625 520 WHO 106 900 TB CAP Ethiopia 100 000 Total budget 2 832 420 Survey organization and financing Implementing agency: Ethiopian Public Health Institute, Ethiopian Health and Nutrition Research Institute ETHIOPIA 2010–2011 a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. National TB prevalence surveys 2007–2016106 Sampling design Sampling frame 773 of 810 woredas (districts)a Sampling design Multistage cluster sampling using PPS Strata Rural/urban/pastoralist Sampling unit Strata (urban, rural, pastoralist)/woreda/ kebele Sample size assumptions • Smear-positive prevalence 364 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.5 • k 0.5 • Response rate 85% • Sample size (estimated) 46 514 Number of clusters 85 Cluster size 550 Eligibility criteria • Age ≥15 years • Residency Permanent residents who stayed in the household at least one night in the 14 days prior to the census, and temporary visitors who stayed in the household at least 14 days prior to the census Survey design and methodology a 37 woredas (3% of the total population) were excluded due to security and logistical challenges. Two clusters (kebele) were replaced before field operations started due to logistical challenges. Laboratory methodology Smear Two samples (spot, morning): direct preparation, FM (LED, auramine stain) Culture One sample (morning; if unavailable then spot): concentrated preparation, LJ media Identification of MTB Capilia TB drug susceptibility test Conducted as post-survey activity Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interviewa Cough ≥2 weeks Chest X-rayb Any lung abnormality Other Those exempt from chest X-ray but with one of the following criteria were also requested to submit sputum specimens: weight loss ≥3 kg in the past month, night sweats ≥2 weeks, fever ≥2 weeks or contact with a TB patient in the past year. a An in-depth interview about health-care seeking behaviour was done for participants who had a cough ≥2 weeks and for those with a history of TB treatment. b Conventional radiography. Field data collection Paper Database CSPro Method of analysis MI+IPW Results first published in a paper December 2012 Official dissemination event December 2011 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 3 026 – Location of care sought • Consulted medical facility 848 28 Public facility 628 74 Private facility 199 23 Other 21 2.5 • Pharmacy 40 1.3 • Traditional healer 3 0.1 • Unspecified 55 1.8 No action taken 1 932 64 Unknown 148 4.8 Key survey results a Cough ≥2 weeks. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). b 90 culture MTB-positive specimens were tested for drug susceptibility and 4 were identified as MDR-TB. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 75 – • Treated in the public sector 54 72 • Treated in the private sector 7 9.3 • Treated in other sector 3 4.0 • Treated in unknown sector 11 15 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 3 2.7 Design effect k Smear-positive TB 1.3 0.7 Bacteriologically confirmed TB 1.3 0.4 Other sputum results Number % Total smear-positive participants 61 – Smear-positive participants without MTB confirmationa 28 46 Isolates with MDR-TB detectedb 4 4.4 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 108 73–143 277 208–347 Male 133 80–185 304 219–388 Female 87 47–127 246 176–315 15–24 years 113 53–173 292 194–390 25–34 years 86 30–143 216 129–303 35–44 years 117 43–191 259 135–382 45–54 years 138 23–253 337 161–513 55–64 years 159 19–300 367 153–582 ≥65 years 41 1–123 227 33–421 Urban 70 6–135 273 130–416 Rural 109 67–151 273 189–356 Pastoralist 170 60–280 316 163–468 a Age ≥15 years unless otherwise specified. ETHIOPIA 107 Survey flow: census to final outcomes Field operations: October 2010 to June 2011 Individuals enumerated in census 95 092 Eligible for sputum examination 6 080 (13%) Total bacteriologically confirmed cases 110 Eligible study population 51 667 (54%) Total participants 46 697 (90%) Interview and chest X-ray 46 548 (99.7%) Interview only 149 (0.3%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 806 (13%) Symptom positive, chest X-ray negative or N/A 2 220 (36%) Symptom negative, chest X-ray positive 3 013 (50%) Otherb 41 (1.0%) Symptom positive, chest X-ray positive 45 (41%) Symptom positive, chest X-ray negative or N/A 12 (11%) Symptom negative, chest X-ray positive 53 (48%) Other 0 (0%) Submitted specimens At least one specimen 5 868 (97%) Both specimens 5 606 (92%) Smear-positive casesd 47 (43%) Definite 33 Probable 14 Smear-negative casese 63 (57%) Definite 59 Probable 4 Laboratory result At least one culture result availablec 5 503 (91%) Symptom screening Cough ≥2 weeksa 3 026 (6.5%) Cough (any duration) 5 930 (13%) Night sweats ≥2 weeks 8 177 (18%) Fever ≥2 weeks 5 920 (13%) Weight loss ≥3 kg in past month 10 014 (21%) Total symptom-screen positivea 3 026 (6.5%) Chest X-ray screening Normal 42 490 (91%) Abnormala 3 819 (8.0%) Other abnormality 239 (1.0%) Result not available N/A Total chest X-rays taken 46 548 a Eligible for sputum collection. b One of the following: weight loss ≥3 kg in the past month, night sweats ≥2 weeks, fever ≥2 weeks or contact with a TB patient in the past year while chest X-ray exempted. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but chest X-ray consistent with TB. e Definite: MTB confirmed by culture with chest X-ray consistent with TB. Probable: MTB confirmed by culture but without chest X-ray consistent with TB. Ineligible individuals 43 425 (46%) Children <15 years 41 125 (43%) Did not meet residency criteria 2 300 (2.4%) National TB prevalence surveys 2007–2016108 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 300 400 500 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 50 100 150 200 250 300 350 400 0 Prevalence per 10 0 0 0 0 population 10 000 30 000 40 000 50 000 60 000 20 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 600 300 400 500 200 100 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data did not suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 1.29, variance 1.59, p=0.12). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 15 5 10 25 20 30 0 0 1 2 3 4 5 Total (≥15) ETHIOPIA 109 Background Ethiopia’s population was 90 million in 2011, making it the second most populous country in Africa (after Nigeria) (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2011, Ethiopia was a low-income country with an average gross national income (GNI) per person of US$ 390 per year (1). In 2011, the prevalence of HIV in the general population aged 15–49 years was 1.3% (95% confidence interval [CI]: 1.1–1.5%) (2), and it was estimated that 14% (95% CI: 13–16%) of TB patients were coinfected with HIV (3). Although WHO launched the DOTS strategy in the mid-1990s, the Government of Ethiopia began to implement the key components of the strategy earlier, in 1992. Nationwide coverage was reached in 2009. The Ministry of Health (MoH) reported that, in 2010, TB was the eighth leading cause of hospital admissions and the third leading cause of hospital deaths in Ethiopia (4). Based on WHO estimates for the same year, Ethiopia had the seventh highest burden of TB globally in terms of estimated incident cases, and ranked third in Africa. Nonetheless, there was considerable uncertainty about the true level of the burden of TB disease. No national TB prevalence survey had been done, no direct measurements of TB mortality were available from vital registration, and the gap between notifications and incidence (due to underreporting or underdiagnosis of cases) was unquantified and hard to estimate. The national authorities in Ethiopia considered that the WHO estimate of TB incidence was too high. To better understand the burden of TB disease, in December 2008 the MoH decided to implement a national TB prevalence survey. From this point onwards, the WHO Global Task Force on TB Impact Measurement considered Ethiopia as one of 22 global focus countries for a national TB prevalence survey. The survey was implemented in 2010–2011 (4, 5). Key methods and results There were 85 survey clusters in three strata (urban, rural and pastoralist), with a target cluster size of 550 individuals. A total of 95  092 individuals from 19 267 households were enumerated in the survey census, of whom 51  667 (54%) were eligible and invited to participate. Of these, 46 697 (90%) did so. All participants were screened in line with the 2011 algorithm as recommended by WHO; that is, using chest X-ray and an interview about symptoms (6). A total of 6080 participants (13%) were eligible for sputum examination, of whom 5868 (97%) submitted at least one sputum specimen and 5606 (92%) submitted two sputum specimens. A total of 110 bacteriologically confirmed pulmonary TB cases was identified, including 47 cases of smear-positive TB. The prevalence of bacteriologically confirmed TB was 277 (95% CI: 208–347) per 100 000 population (among those aged ≥15 years), and for smear-positive TB it was 108 (95%  CI: 73–143) per 100  000 population. When extrapolated to all forms of TB and to all ages, prevalence was 240 (95% CI: 182–298) per 100 000 population. There was no significant difference between the three geographical strata (urban, rural and pastoralist). Other key results were: • the male to female ratio for TB prevalence was 1.5 for smear-positive TB and 1.2 for bacteriologically confirmed TB; • prevalence per 100 000 population increased with age and was highest in the age groups 45– 54 and 55–64 years; there were also relatively large proportions of bacteriologically confirmed cases in younger age groups; • among bacteriologically confirmed TB cases, 48% were symptom-screen negative, and among the smear-positive TB cases, 43% were symptom-screen negative; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 1.2 overall, but varied from 0.6 in those aged 25–34 years to 2.9 in the 55–64 year age group. The ratio was slightly higher for men than for women (1.3 versus 1.1); • among bacteriologically confirmed TB cases, 88% had no previous history of anti-TB treatment, and only 2.7% were on anti-TB treatment at the time of the survey; and • of the 54 bacteriologically confirmed and 25 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 23 (43%) and 13 (52%), respectively, had previously sought care in a public or private health facility for their symptoms. National TB prevalence surveys 2007–2016110 Implications of results The burden of smear-positive pulmonary TB was found to be lower than previously estimated. The observed smear-positive TB prevalence of 108 (95%  CI: 73– 143) per 100 000 population was about half of the level hypothesized to calculate the sample size of the survey, and about two thirds of the level estimated by WHO in 2009 (168 per 100 000 population). There were several reasons for the previous overestimation, including: • a lack of accurate population-based baseline data prior to the survey; • HIV prevalence in Ethiopia was previously assumed to be similar to the regional HIV prevalence for countries in sub-Saharan Africa, at 6%; results from sentinel HIV surveillance among women attending antenatal care suggested a level of 2.3% in 2009; and • expansion of the DOTS strategy and the presence of a high-quality nationwide treatment programme may have contributed to the low observed prevalence of smear-positive TB. On the basis of survey results, estimates of TB disease burden published by WHO were revised downwards, and the case detection rate, based on an updated estimate of TB incidence, was revised upwards. Although the survey revealed a lower TB prevalence than previously estimated, almost all TB cases had not been previously notified to the National TB Programme (NTP). In addition, 54% (58/107) of the previously undetected cases in the community were among the younger age groups (15–34 years), suggesting high levels of transmission (including to the children of young parents with TB). Other implications included: • a need to strengthen community screening of TB, to ensure earlier detection and treatment of cases; • a need to review the important role of chest X-ray screening in early detection of cases, given that half of bacteriologically confirmed cases did not report chronic cough and were only identified through such screening; • a need to expand the range of laboratory tests being used to diagnose TB, to include culture or Xpert® MTB/RIF, or both; at the time of the survey, the only laboratory test in widespread use for TB diagnosis was sputum smear microscopy, but more than 50% of survey cases were sputum smear-negative; • a need to understand that a smear-positive test result does not always indicate TB disease, especially in a community (as opposed to clinical) setting; of the 61 smear-positive participants, 27 were culture-negative and one had nontuberculous mycobacteria (NTM). In active TB case-finding activities, TB cannot be diagnosed based on smear examination alone; and • a need for more funding to implement better screening, wider use of chest X-ray and improvements to diagnostics. Photo credit: Marina Tadolini ETHIOPIA 111 Major successes, challenges and lessons learned This was the first-ever national TB prevalence survey in Ethiopia, and the first national survey in Africa in more than 50 years to be successfully implemented according to screening and diagnostic methods recommended in the 2011 edition of WHO’s handbook on national TB prevalence surveys (6). It only took just over 1 year between the decision to undertake a survey and the start of field operations. The population coverage (97%), participation rate (90%) and sputum collection rate (97%) were all very high. Major challenges in the early stages of the survey included mismanagement of sputum specimens, backlogs in culture inoculation and a high culture contamination rate (up to 15% for cultures in the first week of the survey). With strong leadership from the principal investigator and the survey’s technical working group, major efforts were made to address these challenges. Problems with management of sputum specimens were resolved, and the overall contamination rate for the survey was 6% (360/5868). Other challenges included delays in the procurement of chest X-ray equipment; difficulties in retaining radiologists during field operations; and the use of data management software that was not suited to the flow of data collection in a prevalence survey, which caused delays in data capture. Due to security and logistical challenges, 3% of the total population was excluded from the sampling frame (e.g. parts of Somaliland and areas bordering Eritrea). Only one specimen per participant was taken for culture; therefore, the prevalence of culture-positive TB may have been underestimated. Nonetheless, the relatively high culture contamination rate may have contributed to higher culture yields than those found in other African surveys that used culture with Löwenstein-Jensen media. An important lesson for future surveys was that the high level of commitment from different stakeholders was key to prompt survey preparation and implementation (the shortest preparation period of any survey in Africa in the period 2009–2015). This commitment had many benefits. For example, it ensured the early appointment of a full-time survey coordinator, close collaboration with the WHO Country Office and WHO headquarters, and excellent collaboration with the NTP. Other important benefits included Asia-Africa collaboration, combined with technical assistance from WHO and an independent consultant. Members of the survey team from Cambodia provided technical assistance to the Ethiopian survey team; the staff person leading WHO’s global work on national TB prevalence surveys made more than 10 visits during the course of the survey; and an independent consultant (funded by the Italian Cooperation) provided regular assistance throughout the survey, from protocol development to reporting of results. Photo credit: Zeleke Alebachew Photo credit: Marina Tadolini National TB prevalence surveys 2007–2016112 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. Kebede AH, Alebachew Z, Tsegaye F, Lemma E, Abebe A, Agonafir M et al. The first population-based national tuberculosis prevalence survey in Ethiopia, 2010–2011. Int J Tuberc Lung Dis. 2014;18(6):635–639 (https://www.ncbi.nlm.nih.gov/ pubmed/24903931, accessed April 2017). 5. First Ethiopian national population based tuberculosis prevalence survey. Addis Ababa: Ministry of Health, Federal Democratic Republic of Ethiopia; Ethiopian Health and Nutrition Research Institute; 2011 (http://www.ephi.gov.et/images/downloads/ Tuberculosis%20Prevalence%20Survey.pdf, accessed May 2017). 6. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 113 Name Role Organization Ifedayo Adetifa Principal investigator Medical Research Council (MRC) Unit-The Gambia Ma Ansu Kinteh Survey coordinator MRC Unit-The Gambia Martin Antonio Unit microbiologist and head of MRC TB Reference Laboratory MRC Unit-The Gambia Ramatoulie Manne Radiology coordinator MRC Unit-The Gambia Beatrice dei Alorse Radiology coordinator MRC Unit-The Gambia Simon Donkor Data manager MRC Unit-The Gambia Adedapo Bashorun Field team leader MRC Unit-The Gambia Christopher Linda Field team leader MRC Unit-The Gambia Semeeh Omoleke Field team leader MRC Unit-The Gambia Lindsay Kendall Biostatistician MRC Unit-The Gambia David Jeffries Biostatistician MRC Unit-The Gambia Edward Demba Scientific officer-mycobacteriology MRC Unit-The Gambia Catherine Bi Okoi Scientific officer-mycobacteriology MRC Unit-The Gambia Kodjovi Mlaga Scientific officer-mycobacteriology MRC Unit-The Gambia William dei Alorse Scientific officer-mycobacteriology MRC Unit-The Gambia Umberto D’Alessandro Epidemiologist/head of Disease Control and Elimination Theme MRC Unit-The Gambia Elina Cole Senior project administrator MRC Unit-The Gambia Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Charalampos Sismanidis Technical assistance (design and analysis) WHO headquarters Sian Floyd Technical assistance (analysis) London School of Hygiene & Tropical Medicine, UK Etienne Leroy Terquiem Technical assistance (radiology advisor) Consultant, France Jan van den Hombergh Technical assistance (radiology advisor) PharmAccess, Tanzania John Mayanda Technical assistance (radiology advisor) PharmAccess, Tanzania Bimbo Fasan Technical assistance (radiology advisor) Lagos state university teaching hospital, Nigeria Key people Surveyed clusters (N=80)a Data sources ■ The Gambian Survey of Tuberculosis Prevalence, Ministry of Health and Social Welfare. The Gambia, Medical Research Council Unit, April 2014. ■ Adetifa IM, Kendall L, Bashorun A, Linda C, Omoleke S, Jeffries D et al., A tuberculosis nationwide prevalence survey in Gambia, 2012, Bull World Health Organ 2016;94:433– 441. ■ Survey dataset. Summary statistics Participation rate 77% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 212 3.1 Prevalence:notification ratio (smear-positive TB, ≥15 years) 0.6 Finance Amount (US$) The Global Fund 1 844 198 Medical Research Council United Kingdom 16 979 Total budget 1 861 177 Survey organization and financing Implementing agency: The Medical Research Council Unit-The Gambia GAMBIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Summary statistics 2011–2013 National TB prevalence surveys 2007–2016114 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata No stratification was used, but final analysis accounted for urban/rural Sampling unit Region/enumeration area Sample size assumptions • Smear-positive prevalence 292 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.5 • k 0.5 • Response rate 85% • Sample size (estimated) 55 281 Number of clusters 80a Cluster size 700 Eligibility criteria • Age ≥15 years • Residency Residents who spent at least one night in the household in the 4 weeks before the census day; visitors who arrived in the household 4 weeks or more before the census day Survey design and methodology a Three clusters were replaced with back-up clusters due to a large uninhabited area in the urban part around the capital (one cluster) and to the military installations and area around the president’s residence (two clusters). Laboratory methodology Smear Two samples (spot/spot or spot/morning): direct preparation, FM (LED, auramine stain) Culture Two samples (spot/spot or spot/morning): concentrated preparation, MGIT media, sub-cultured onto a LJ slope for speciation purposes Identification of MTB MGIT™ TBc Identification Test TB drug susceptibility test Xpert MTB/RIF for all survey TB cases, not as part of the survey Xpert® MTB/RIF Done for all survey TB cases, not as part of the survey HIV test Not done Screening criteria Interviewa Cough ≥2 weeks Cough <2 weeks with ≥2 other TB symptomsb No cough with ≥3 other TB symptomsb Chest X-rayc Any lung or mediastinum abnormality Other Chest X-ray exempted a An in-depth interview about health-care seeking behaviour was done for TB symptomatic participants and for those with previous (within 5 years) or current history of TB. b Chest pain, night sweats, shortness of breath, loss of appetite, weight loss, fever, haemoptysis. c Mobile direct digital radiography. Field data collection Paper Database SQL Method of analysis MI+IPW Results first published in a report/paper April 2014 Official dissemination event May 2014 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 3 462 – Location of care sought • Consulted medical facility 1 706 49 Public facility 1 398 82 Private facility 220 13 Other (NGOs, MRC facility) 88 5.2 • Pharmacy 17 0.5 • Traditional centre 14 0.4 • Other 24 0.7 Self-treated N/A N/A No action taken 1 424 41 Unknown 277 8.0 Key survey results a Cough ≥2 weeks, or cough <2 weeks with ≥2 other TB symptoms, or no cough with ≥3 other TB symptoms. a TB treatment is available only in the public sector. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). Survey participants currently on TB treatment Number % Total participants currently on TB treatmenta 38 – • Treated in the public sector 38 100 • Treated in the private sector N/A N/A • Treated in other sector N/A N/A Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 4 5.0 Design effect k Smear-positive TB 1.8 1.3 Bacteriologically confirmed TB 1.6 0.7 Other sputum results Number % Total smear-positive participants 36 – Smear-positive participants without MTB confirmationa 8 22 Isolates with DR-TB (rifampicin resistance) detected 3 3.9 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 90 53–127 212 152–272 Male 148 88–208 333 233–433 Female 41 0–83 109 54–164 15–34 years 56 24–88 133 76–190 35–54 years 144 65–223 355 219–490 ≥55 years 159 0–367 329 99–558 Urban 96 43–148 266 164–368 Rural 86 32–140 109 54–164 a Age ≥15 years unless otherwise specified. GAMBIA 115 Survey flow: census to final outcomes Field operations: December 2011 to January 2013 Individuals enumerated in census 100 678 Eligible for sputum examination 5 948 (14%) Total bacteriologically confirmed cases 77 Eligible study population 55 832 (56%) Total participants 43 100 (77%) Interview and chest X-ray 42 942 (99.6%) Interview only 158 (0.4%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 1 026 (17%) Symptom positive, chest X-ray negative or N/A 2 436 (41%) Symptom negative, chest X-ray positive 2 384 (40%) Otherb 102 (1.7%) Symptom positive, chest X-ray positive 32 (41%) Symptom positive, chest X-ray negative or N/A 12 (16%) Symptom negative, chest X-ray positive 33 (43%) Other 0 (0%) Submitted specimens At least one specimen 5 436 (91%) Both specimens 5 309 (89%) Smear-positive casesd 34 (44%) Definite 28 Probable 6 Smear-negative casese 43 (56%) Definite 43 Probable N/A Laboratory result At least one culture result availablec 5 209 (88%) Symptom screening Cough ≥2 weeksa 962 (2.2%) Haemoptysis 210 (0.5%) Fever 6 637 (15%) Chest pain 6 551 (15%) Night sweats 1 595 (3.7%) Cough <2 weeks with ≥2 other TB symptomsa 1 372 (3.2%) No cough with ≥3 other TB symptomsa 1 128 (2.6%) Total symptom-screen positivea 3 462 (8.0%) Chest X-ray screening Normal 38 384 (89%) Abnormala 3 410 (7.9%) Other abnormality 1 148 (2.7%) Result not available N/A Total chest X-rays taken 42 942 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but two AFB positive or one AFB positive with chest X-ray suggestive of TB. e Definite: MTB confirmed by culture. Probable: no definition. Ineligible individuals 44 846 (45%) Children <15 years 41 561 (41%) Did not meet residency criteria 3 285 (3.2%) National TB prevalence surveys 2007–2016116 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–34 35–54 ≥55 50 100 150 200 250 300 350 400 0 Prevalence per 10 0 0 0 0 population 900 800 700 600 500 400 300 200 100 1 000 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 60 50 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB 600 300 400 500 200 100 0 15–34 35–54 ≥55 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0 15–34 35–54 Male Female≥55 Pr ev al en ce : no ti fi ca ti on ra ti o 0.1 0.2 0.3 0.4 0.5 0.7 0.6 0.8 0.9 1 .0 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 400 600 800 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 20 5 15 10 30 25 35 0 0 1 2 3 4 5 6 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data did not suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 0.96, variance 1.53, p=0.06). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Total (≥15) GAMBIA 117 Background Gambia, a small country in West Africa, had a population of 1.8 million in 2012, and an average gross national income (GNI) per person of US$ 520 per year, making it a low-income country (1). The prevalence of HIV in 2012 in the general population aged 15–49 years was 2.0% (95% confidence interval [CI]:1.6–2.4%) (2), and it was estimated that 16% (95% CI: 15–18%) of tuberculosis (TB) patients were coinfected with HIV (3). Gambia established a National Leprosy and TB Programme (NLTP) in 1984. The WHO-recommended DOTS strategy was adopted in the mid-1990s, and was eventually expanded to reach national coverage (4,5). The case notification rate for new TB cases peaked at 124 per 100 000 population in 2008 and then started to fall. The WHO estimate of TB prevalence in 2011 was 455 (95% CI: 225–764) per 100 000 population, with the case detection rate (notifications of new cases divided by incidence) estimated at 45% (95% CI: 38–55%). However, there was considerable uncertainty about the burden of TB disease, given that no national TB prevalence survey had previously been done, no direct measurements of TB mortality were available from vital registration, and the gap between notifications and incidence (due to underreporting or underdiagnosis of cases) had not been quantified and was difficult to estimate. National authorities in the Gambia considered that WHO estimates of TB incidence were too high. To better understand the burden of TB disease in the country, a decision to implement a national TB prevalence survey was taken in 2008–2009. The survey started in December 2011 and was completed in January 2013 (6,7).  Gambia was not one of the 22 global focus countries for national TB prevalence surveys selected by the WHO Global Task Force on TB Impact Measurement in December 2007. Nevertheless, the country was on the Task Force’s longer list of 53 countries considered to meet survey eligibility criteria. Key methods and results There were 80 survey clusters, with a target cluster size of 700 individuals. No stratification was used at the time of survey design, but urban and rural strata were examined during the analysis. A total of 100  678 individuals from 13 847 households were enumerated in the survey census, of whom 55 832 (56%) were eligible and invited to participate. Of these, 43 100 (77%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (8). A total of 5948 participants (14%) were eligible for sputum examination, of whom 5436 (91%) submitted at least one sputum specimen and 5309 (89%) submitted two sputum specimens. A total of 77 bacteriologically confirmed pulmonary TB cases was identified, including 34 cases of smear-positive TB. The prevalence of smear-positive TB was 90 (95% CI: 53–127) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 212 (95% CI: 152–272) per 100  000 population. When extrapolated to all forms of TB and to all ages, prevalence was 128 (95% CI: 94–162) per 100 000 population. The prevalence of smear-positive and bacteriologically confirmed TB was higher in urban areas than in rural areas. Photo credit: Ifedayo Adetifa National TB prevalence surveys 2007–2016118 Other key results were: • the male to female ratio for TB prevalence was 3.6 for smear-positive TB and 3.1 for bacteriologically confirmed TB; • prevalence per 100 000 population was highest in the older age groups; however, the absolute number of bacteriologically confirmed TB cases was relatively high in the younger age groups (15–34 years and 35–54 years); • among bacteriologically confirmed TB cases, 57% were symptom-screen positive, and of the smear-positive cases, 56% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 0.6 overall, but varied from 0.4 in those aged 15–34 years to 0.9 in the 35–54 year age group; it was slightly higher for men than for women (0.7 versus 0.5); • among bacteriologically confirmed TB cases, 16% had a previous history of anti-TB treatment and only 5.0% were on anti-TB treatment at the time of the survey; • of the 40 bacteriologically confirmed and 15 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 24 (60%) and 12 (80%), respectively, had previously sought care in a public or private health facility for their symptoms; and • three cases had rifampicin resistance based on Xpert® MTB/RIF testing; none of these rifampicin-resistant cases had a history of current or previous anti-TB treatment.   Implications of results The estimated TB prevalence for all ages based on survey results was 128 per 100 000 population (95% CI: 94–162). This was only one quarter of the pre-survey estimate published by WHO. Estimated incidence was revised downwards to a best estimate of 175 per 100 000 population (95% CI: 135–215), and the case detection rate was revised upwards, to 71% (95% CI: 70–73%). At the same time, stable TB case notification rates indicated that efforts in TB control were still not sufficient. This survey was the only one where the P:N ratio was less than one for all categories (all age groups, male and female). Possible explanations for this included an NTP that was able to efficiently detect and treat cases in the community as in high-resource settings, or over- diagnosis of smear-positive TB cases in routine health care services. Other implications included: • a need for interventions targeted or more tailored to men, especially those aged 35–54 years; • a need for expanded use of chest X-ray combined with better capacity to interpret results, and greater use of culture and Xpert MTB/RIF (or both), to improve detection of people with smear-negative culture-positive disease; Photo credit: Ifedayo Adetifa Photo credit: Ifedayo Adetifa GAMBIA 119 • a need to consider targeted interventions among older people, including paying particular attention to this group during contact investigations and active case finding (given the cultural acceptance of chronic cough among the elderly); and • a need for increased funding for implementation of targeted interventions, wider use and better interpretation of X-rays, and improvements to diagnostics. Major successes, challenges and lessons learned This was the first survey in Africa to have been outsourced by the NTP and conducted by a reputable research institute. In addition, it was the first survey for which results led to a statistically significant downward estimate of TB burden, thus confirming the value of undertaking a survey and validating the notion that WHO estimates were previously too high. One major challenge was ensuring participation, particularly in urban areas. Overall, the target of an 85% participation rate was not achieved. The survey was also prolonged to 13 months, and the start of the survey was delayed due to logistical problems. A particular difficulty was the procurement of mobile X-rays, due to a combination of the high unit cost and the need to adhere to European Union procurement rules and procedures given that the implementing agency for the survey (Medical Research Council Unit, The Gambia (MRCG)) was an affiliate of the Medical Research Council UK. However, following the survey, the radiology equipment was handed over to the government to help improve TB diagnosis. The MRCG laboratory had excellent capacity, made considerable efforts to ensure high-quality sputum samples and used best practices in the decontamination process. They pioneered the use of MGIT for primary diagnosis and identification of MTB. Nevertheless, culture contamination rates were relatively high (11%), in part because of the use of liquid culture. The relatively high contamination rate might have contributed to higher yields by culture (i.e. there were more smear- negative, culture-positive specimens than smear-positive, culture-positive ones). The contamination rate within the laboratory for routine samples was within the prescribed ranges for both solid and liquid media cultures. The survey was fully implemented by MRCG staff. Although the NLTP was represented by a designated liaison person (deputy programme manager) for implementation and on the survey steering committee (NLTP manager), their involvement was relatively limited. More active engagement would have helped to build greater ownership of survey results and strengthened use of the results in national strategic planning. Photo credit: Ifedayo Adetifa National TB prevalence surveys 2007–2016120 Financial policies of the funder – the Global Fund to Fight AIDS, Tuberculosis and Malaria – meant that key survey staff reached the end of their contracts before the survey report was ready. This contributed to delays in publication of the official survey report.   References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. The Gambian survey of tuberculosis prevalence (GAMSTEP). Banjul, The Gambia: Medical Research Council Unit, The Gambia; 2014. 7. Adetifa IM, Kendall L, Bashorun A, Linda C, Omoleke S, Jeffries D et al. A tuberculosis nationwide prevalence survey in Gambia, 2012. Bull World Health Organ. 2016;94(6):433–441 (https:// www.ncbi.nlm.nih.gov/pubmed/27274595, accessed August 2017). 8. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 121 Name Role Organization Frank Bonsu Principal investigator Ghana Health Service, National TB Control Programme (NTP) Kennedy Kwasi Addo Co-principal investigator Noguchi Memorial Institute, University of Ghana John Gyapong Co-investigator University of Ghana Ellis Owusu Dabo Co-investigator Kwame Nkrumah University of Science and Technology Kwadwo Koram Co-investigator Noguchi Memorial Institute, University of Ghana Augustina Badu Peprah Co-investigator Komfo anokye teaching hospital, Kumasi Raymond Yaw Gockah Survey coordinator Ghana Health Service, NTP Francisca Dzata NTP laboratory focal point Ghana Health Service, NTP Michael Omari Head of chest clinic laboratory Korle bu teaching hospital, Laboratory Robertson Adiei Cartographer Ghana Statistical Service, NTP Nii Nortey Hanson Nortey Deputy NTP manager Ghana Health Service, NTP Jane Amponsah Data manager Ghana Health Service, NTP Sauda Ahmed Assistant data manager Ghana Health Service, NTP Herve Awako ICT manager TABS Consult (data/IT management) Prince Boni Data planning TABS Consult (data/IT management) Zeleke Alebachew Technical assistance (report writing) Consultant, Ethiopia Irwin Law Technical assistance (survey advisor) WHO headquarters Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Charalampos Sismanidis Technical assistance (analysis) WHO headquarters Key people Surveyed clusters (N= 98)a Data sources ■ Bonsu F, Addo KK, Alebachew Z, Gyapong J, Badu-Peprah A, Gockah R et al. National population-based tuberculosis prevalence survey in Ghana, 2013. Int J Tuberc Lung Dis. 2020 Mar 1;24(3):321-328. ■ Survey dataset. Summary statistics Participation rate 91% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 356 1.4 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.5 Finance Amount (US$) The Global Fund 2 100 000 WHO 100 000 Total budget 2 200 000 Survey organization and financing Implementing agency: National TB Control Programme GHANA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2013 National TB prevalence surveys 2007–2016122 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Region/district/enumeration area Sample size assumptions • Smear-positive prevalence 270 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.4 • k 0.5 • Response rate 80% • Sample size (estimated) 63 905 Number of clusters 98 Cluster size 650 Eligibility criteria • Age ≥15 years • Residency Permanent residents who lived in the household at least 1 day in the past 14 days, or visitors who lived in the house- hold at least 7 days in the past 14 days Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): concentrated preparation, ZN Culture Two samples (spot, morning): concentrated preparation, LJ media and MGIT media (but only MGIT was used for study case definition) Identification of MTB PNB, capilia TB drug susceptibility test Xpert MTB/RIF Xpert® MTB/RIF Done for smear-positive specimens, and if cultures were contaminated (where specimens were available). HIV test Not done Screening criteria Interviewa Cough ≥2 weeks Chest X-rayb Any lung abnormality Other Chest X-ray exempted a An in-depth interview about health-care seeking behaviour and exposure to risk factors was done only for those with cough of two weeks or more, sputum production, current TB treatment or a history of TB. b Mobile digital radiography. Field data collection Paper/electronic Database Microsoft® Access Method of analysis MI+IPW Results first published in a paper March 2020 Official dissemination event March 2015 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 1 969 – Location of care sought • Consulted medical facility 793 40 Public facility 695 88 Private facility 61 7.7 Otherb 37 4.7 • Pharmacy 324 17 • Traditional center 20 1.0 Self-treated 567 29 No action taken 264 13 Unknown 1 0.1 Key survey results a Cough ≥2 weeks. b Faith based health facility. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b 1134 participants were tested with Xpert MTB/RIF and 11 were resistant to rifampicin. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 48 – • Treated in the public sector 42 88 • Treated in the private sector 1 2.1 • Treated in other sector (faith based health facility) 5 10 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 9 4.5 Other sputum results Number % Total smear-positive participants 198 – Smear-positive participants without MTB confirmationa 138 70 Isolates with DR-TB (rifampicin) detectedb 11 1.0 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 111 76–145 356 288–425 Male 198 133–264 431 327–536 Female 49 21–76 303 223–382 15–24 years 49 14–84 185 104–265 25–34 years 35 1.6–69 228 130–326 35–44 years 101 38–164 295 174–416 45–54 years 223 129–317 470 294–645 55–64 years 245 63–426 607 362–854 ≥65 years 212 77–347 908 597–1 219 Urban 142 89–195 293 216–372 Rural 75 39–111 429 315–542 a Age ≥15 years unless otherwise specified. Design effect k Smear-positive TB 1.5 0.9 Bacteriologically confirmed TB 2.0 0.7 GHANA 123 Survey flow: census to final outcomes Field operations: March to December 2013 Individuals enumerated in census 101 772 Eligible for sputum examination 8 298 (13%) Total bacteriologically confirmed cases 202 Eligible study population 67 757 (67%) Total participants 61 726 (91%) Interview and chest X-ray 59 718 (97%) Interview only 2 008 (3.3%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 771 (9.3%) Symptom positive, chest X-ray negative or N/A 1 198 (14%) Symptom negative, chest X-ray positive 4 387 (53%) Otherc 1 942 (23%) Symptom positive, chest X-ray positive 67 (33%) Symptom positive, chest X-ray negative or N/A 15 (7.4%) Symptom negative, chest X-ray positive 85 (42%) Otherc 35 (17%) Submitted specimens At least one specimen 8 126 (98%) Both specimens 7 706 (93%) Smear-positive casese 64 (32%) Definite 62 Probable 2 Smear-negative casesf 138 (68%) Definite 106 Probable 32 Laboratory result At least one culture result availabled 7 531 (91%) Symptom screening Cough ≥2 weeksa 1 969 (3.2%) Haemoptysis 273 (0.4%) Sputum production 2 274 (3.7%) Fever 1 405 (50%)b Chest pain 1 898 (67%)b Weight loss 1 242 (44%)b Night sweats 1 172 (42%)b Total symptom-screen positivea 1 969 (3.2%) Chest X-ray screening Normal 44 976 (75%) Abnormala 5 158 (8.6%) Other abnormality 9 584 (16%) Result not available N/A Total chest X-rays taken 59 718 a Eligible for sputum collection. b The denominator included only participants who had the in-depth interview (N=2 821). c Chest X-ray exempted and symptom-screen negative. d Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. e Definite: MTB confirmed by culture in at least one sample, or Xpert-positive in at least one sample with chest X-ray suggestive of TB. Probable: Xpert-positive in at least one sample without chest X-ray suggestive of TB. f Definite: MTB confirmed by culture and/or Xpert in two samples, or culture and/or Xpert in one sample with chest X-ray suggestive of TB. Probable: MTB confirmed by culture and/or Xpert in one sample without chest X-ray suggestive of TB. Ineligible individuals 34 015 (33%) Children <15 years 32 724 (32%) Did not meet residency criteria 1 291 (1.3%) National TB prevalence surveys 2007–2016124 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 200 300 400 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 800 900 1000 0 Prevalence per 10 0 0 0 0 population 2 000 6 000 8 000 10 000 12 000 4 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1 .5 1 .0 2.0 2.5 3.0 3.5 4.0 4.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB 1200 600 800 1000 400 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 Pr ev al en ce : no ti fi ca ti on ra ti o a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.06, variance 4.08, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimate of using smear-positive pulmonary TB notifications (2013) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 20 15 25 0 0 1 2 3 4 5 6 7 8 9 10 Total (≥15) GHANA 125 Background Ghana, in West Africa, had a population of 26 million in 2013, and a gross national income (GNI) per person of US$ 1740, making it a lower-middle-income country (1). In 2013, the prevalence of HIV in the general population aged 15–49 years was 1.5% (95% confidence interval [CI]: 1.2–2.0%) (2), and it was estimated that 24% (95% CI: 20– 27%) of tuberculosis (TB) patients were coinfected with HIV (3). The Ghana Tuberculosis Service was formally established with the appointment of its first director in 1959. It was restructured and renamed the National TB Control Programme (NTP) in 1994, the year in which implementation of the WHO-recommended DOTS strategy began (4,5). Three national strategic plans for TB control were implemented during the period 1994–2013. In its 2012 global TB report, WHO estimated that there were 20 000 (95% CI: 17 000–22 000) new cases of TB per year. Nonetheless, there was considerable uncertainty about estimates of the burden of TB disease, given that no national TB prevalence survey had been done since 1957; there were no direct measurements of TB mortality available from vital registration; and the gap between notifications and incidence (due to underreporting or underdiagnosis of cases) was not quantified and was difficult to estimate. A 2013 evaluation of TB surveillance using the WHO checklist of TB surveillance standards and benchmarks found that only four of the 13 standards expected from a high-performance surveillance system capable of providing direct and reliable measurements of the number of TB cases and deaths were fully met. In December 2007, Ghana was one of 22 global focus countries for a national TB prevalence survey selected by the WHO Global Task Force on TB Impact Measurement. In response, the Ministry of Health decided to implement a national TB prevalence survey in 2008, and secured funding from the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund). Field operations started in March 2013 and were completed in December 2013. Key methods and results There were 98 survey clusters in two strata (urban and rural), with a target cluster size of 650 individuals. A total of 101 772 individuals from 23 991 households were enumerated in the survey census, of whom 67 757 (67%) were eligible and invited to participate. Of these, 61 726 (91%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (6). A total of 8298 participants (13%) were eligible for sputum examination, of whom 8126 (98%) submitted at least one sputum specimen and 7706 (93%) submitted two sputum specimens. A total of 202 bacteriologically confirmed pulmonary TB cases was identified, including 64 cases of smear- positive TB. The prevalence of smear-positive TB was 111 (95% CI: 76–145) per 100  000 population (among those aged ≥15 years) and for bacteriologically confirmed TB it was 356 (95% CI: 288–425) per 100 000 population. Bacteriologically confirmed TB prevalence was generally higher in rural than urban areas. Other key results were: • the male to female ratio was 4.0 for smear- positive TB and 1.4 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age; however, the absolute number of TB cases was consistently high in all age groups; • among bacteriologically confirmed TB cases, 41% were symptom-screen positive, and of the smear-positive cases, 64% were symptom- screen positive; Photo credit: Irwin Law National TB prevalence surveys 2007–2016126 • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.5 overall, but varied from 0.9 in those aged 25–34 years to 3.8 in those aged 55–64 years, and it was higher for men than for women (3.1 versus 1.8); • among bacteriologically confirmed TB cases, 95% had no previous history of anti-TB treatment and only 4.5% were on anti-treatment at the time of the survey; and • of the 73 bacteriologically confirmed and 37 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 33 (45%) and 17 (46%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results The estimated TB prevalence (all forms, all ages) based on the survey (290 per 100 000 population; 95% CI: 196–384) was approximately three times higher than the pre-survey estimate (92 per 100 000 population; 95% CI: 44–158). Furthermore, the survey clearly revealed undiagnosed TB cases in the community, with many missed opportunities for diagnosis, including a high proportion of patients with chronic cough who visited both public and private health facilities but were not offered sputum examination. This was further compounded by the large proportion of people who self-treated, with a high usage of pharmacies as a first point of health care. Based on survey findings, the national TB control implementation strategy (TB strategic plan 2015–2020) was updated to include: • a revised screening and diagnostic algorithm that included chest X-ray and culture and/or Xpert® MTB/RIF in addition to smear microscopy and symptoms; • introduction of a policy to use chest X-ray as part of active TB case finding in vulnerable populations and in health-care settings; • wider use of Xpert MTB/RIF throughout the programme to detect bacteriologically confir- med cases and to exclude nontuberculous mycobacteria (NTM); and • targeting of TB screening activities to specific subpopulations, such as men and the elderly. In addition, the evidence of TB-related stigma and poor knowledge about TB in the general population prompted the development of a national communications strategy with stakeholders. The survey also highlighted gaps in the surveillance system that needed to be addressed; in particular, underreporting of smear-negative culture- positive TB cases to the NTP. Photo credit: Irwin Law GHANA 127 Culture and Xpert MTB/RIF testing showed that a smear- positive test result did not always indicate TB disease, especially in a community setting as opposed to a clinical setting. In active TB case finding, TB cannot be diagnosed based on smear examination alone. Major successes, challenges and lessons learned The major overarching success was that the first national TB prevalence survey in Ghana in more than 50 years was successfully implemented, with a high participation rate. A key part of the success story was that the survey was led and coordinated by the NTP, with stakeholders from research institutes, the national statistical service, universities and the Ministry of Health. The survey team also benefited from substantial technical assistance, coordinated by the WHO Global Task Force on TB Impact Measurement. Other successes included: • the survey enhanced national capacity to conduct culture examinations, drug susceptibility testing and action-oriented operational research; • collaboration with the private sector in data planning, management and storage made the survey one of the most technologically advanced (among those conducted in 2009–2016) in terms of data management; beyond the survey, this subsequently improved data management capacity within the NTP; and • the active community screening, specimen collection and transportation required in the survey improved working relationships between the NTP and research institutes. The survey faced several challenges. It took four years from the start of survey preparations in 2008 to reach the point at which field operations could be launched. The major reason for this delay was the substantial time taken to acquire digital X-ray units. When the survey was designed, the timely delivery of such units was expected from a large Netherlands-Ghana project to equip the district hospital network with digital equipment, based on a concessional loan and national counterpart funding. In practice, the project was not approved by the Dutch national parliament for several years and the NTP had to mobilize other funds to procure the X-ray units needed for the survey. During field operations there were logistical challenges. Transportation across harsh terrain caused some breakdowns in container X-ray units, which needed to be replaced with portable units that had shockproof boxes. In one of the two laboratories used in the survey, there was a Photo credit: Irwin Law National TB prevalence surveys 2007–2016128 breakdown of the biosafety cabinet (due to a blocked high efficiency particulate air [HEPA] filter), which may have caused specimen cross-contamination. The breakdown necessitated temporary suspension of laboratory work for maintenance, and thus delayed inoculation of the collected specimens. Furthermore, culture confirmation occurred in less than 85% of smear-positive survey cases, and the exclusion of solid culture and FM smear results (done in parallel with MGIT culture and Ziehl-Neelsen [ZN] smear) highlighted the challenges encountered by at least one of the two laboratories. Other challenges faced during the survey included a backlog of 20 000 chest X-rays that had to be read after field operations were completed; and delays in report writing and dissemination of results because survey staff had competing demands on their time. Future surveys would benefit from a dedicated budget and associated staff for report writing. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 129 Name Role Organization Dina Bisara Lolong Principal investigator National Institute of Health Research and Development (NIHRD) Francisca Srioetami Regional coordinator NIHRD Lamria Pangaribuan Regional coordinator NIHRD Ainur Rofiq Regional coordinator NIHRD Retno Kusuma Dewi Laboratory coordinator National TB Programme (NTP) Irfan Ediyanto Vice laboratory coordinator NTP Aziza G. Icksan Radiology coordinator Persahabatan hospital Narendro Arifia Data manager NIHRD Safrizal Field team leader The National TB Prevalence Survey team (NPS team) of NIHRD Darmawati Field team leader NPS team of NIHRD Risnawati Field team leader NPS team of NIHRD Ade Yoska Tilla Serihati Field team leader NPS team of NIHRD Elisabeth Bernadeth Field team leader NPS team of NIHRD Laura Valeria Field team leader NPS team of NIHRD M.N. Farid Technical assistance (statistics, data management) TB Operational Research Group (TORG) Pandu Riono Technical assistance (statistics) TORG Jubaedi Technical assistance (data management) WHO Indonesia M. Bintari Dwihardiani Technical assistance (survey advisor) WHO Indonesia Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Irwin Law Technical assistance (survey advisor) WHO headquarters Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Charalampos Sismanidis Technical assistance (analysis) WHO headquarters Philippe Glaziou Technical assistance (analysis) WHO headquarters Key people Surveyed clusters (N=156)a Data sources ■ Indonesia Tuberculosis Prevalence Survey 2013–2014. Ministry of Health, Republic of Indonesia; National Institute of Health Research and Development; in collaboration with Directorate General of Disease Control and Environment Health; 2015. Summary statistics Participation rate 89% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 759 2.3 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.3 Finance Amount (US$) The Global Fund 4 241 005 TB Care 1 379 576 Total budget 4 620 581 Survey organization and financing Implementing agency: National Institute of Health Research and Development INDONESIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2013–2014 National TB prevalence surveys 2007–2016130 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Three geographical regions (Sumatra, Java-Bali and others) Sampling unit Geographical region/village/census block Sample size assumptions • Smear-positive prevalence 156 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.5 • k 0.8 • Response rate 85% • Sample size (estimated) 78 000 Number of clusters 156 Cluster size 500 Eligibility criteria • Age ≥15 years • Residency Individuals who lived in the household for at least one month prior to the census Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): direct preparation, ZN Culture Two samples (spot, morning) for 52 clusters, one sample (morning) for 104 clusters: concentrated preparation, LJ media Identification of MTB MPT64 rapid test, niacin test TB drug susceptibility test Not done Xpert® MTB/RIF Done for smear-positive and non-conclusive culture samples HIV test Not done Screening criteria Interview Cough ≥2 weeks and/or haemoptysis Chest X-raya Any lung or pleura abnormality Other Chest X-ray exempted with any TB symptomsb a Direct digital radiography (portable). b Cough, haemoptysis, fever, chest pain, night sweats, loss of appetite, shortness of breath. Field data collection Paper/electronic Database Microsoft® Access Method of analysis MI+IPW Results first published in a report/paper September 2015 Official dissemination event October 2014 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 8 552 – Location of care sought • Consulted medical facility 2 231 26 Public facility 1 178 53 Private facility 672 30 Otherb 381 17 • Pharmacy, shop 2 636 31 • Traditional centre N/A N/A Self-treated N/A N/A No action taken 3 685 43 Unknown N/A N/A Key survey results a Cough ≥2 weeks and/or haemoptysis. b Nurse or midwife consultation. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 125 – • Treated in the public sector 68 54 • Treated in the private sector 52 42 • Treated in other sector 5 4.0 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 18 4.2 Design effect k Smear-positive TB 1.6 0.7 Bacteriologically confirmed TB 1.8 0.5 Other sputum results Number % Total smear-positive participants 291 – Smear-positive participants without MTB confirmationa 126 43 Isolates with MDR-TB detected N/A N/A Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 257 210–303 759 590–961 Male 393 315–471 1 083 873–1 337 Female 131 88–174 461 354–591 15–24 years 138 77–198 361 254–495 25–34 years 240 156–324 753 562–995 35–44 years 265 171–359 714 527–941 45–54 years 272 166–377 836 609–1 108 55–64 years 319 174–463 1 030 734–1 399 ≥65 years 528 292–763 1 582 1 123–2 154 Urban 282 220–345 846 678–1 048 Rural 231 163–300 674 512–874 a Age ≥15 years unless otherwise specified. INDONESIA 131 Survey flow: census to final outcomes Field operations: April 2013 to June 2014 Individuals enumerated in census 112 350 Eligible for sputum examination 15 446 (23%) Total bacteriologically confirmed cases 426 Eligible study population 76 576 (68%) Total participants 67 944 (89%) Interview and chest X-ray 64 338 (95%) Interview only 3 606 (5.3%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 4 459 (29%) Symptom positive, chest X-ray negative or N/A 4 093 (26%) Symptom negative, chest X-ray positive 6 743 (44%) Otherb 151 (1.0%) Symptom positive, chest X-ray positive 217 (51%) Symptom positive, chest X-ray negative or N/A 25 (5.8%) Symptom negative, chest X-ray positive 184 (43%) Other 0 (0%) Submitted specimens At least one specimen 15 141 (98%) Both specimens 14 604 (95%) Smear-positive casesd 165 (39%) Definite 165 Probable 0 Smear-negative casese 261 (61%) Definite 254 Probable 7 Laboratory result At least one culture result availablec 14 773 (96%) a Eligible for sputum collection. b 151 pregnant women reported at least one of following TB symptoms: cough, haemoptysis, fever, chest pain, night sweats, loss of appetite, shortness of breath. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture and/or Xpert. Probable: MTB not confirmed by culture and/or Xpert but chest X-ray suggestive of TB. e Definite: MTB confirmed by culture and/or Xpert. Probable: For six out of seven, cultures were identified by niacin but not MPT64, with chest X-ray suggestive of TB. One case was a pregnant participant who was Xpert-positive, but whose culture specimen was contaminated. Ineligible individuals 35 774 (32%) Children <15 years 33 206 (30%) Did not meet residency criteria 2 568 (2%) Symptom screening Cough ≥2 weeksa 8 377 (12%) Haemoptysisa 897 (1.3%) Sputum production 17 001 (25%) Chest pain 13 614 (20%) Fever 12 510 (18%) Total symptom-screen positivea 8 552 (13%) Chest X-ray screening Normal 46 712 (73%) Abnormala 11 202 (17%) Other abnormality 6 424 (10%) Result not available 0 (0%) Total chest X-rays taken 64 338 National TB prevalence surveys 2007–2016132 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 1.0 2.0 3.0 4.0 5.0 6.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 400 600 800 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 200 600 400 800 1000 1200 1400 1600 1800 0 Prevalence per 10 0 0 0 0 population 50 000 150 000 200 000 250 000 300 000 350 000 100 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1000 1500 2000 2500 500 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.73, variance 5.09, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using smear-positive pulmonary TB notifications (2013) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 20 5 15 10 30 25 35 0 0 1 2 3 4 5 6 7 8 9 10 11 Total (≥15) INDONESIA 133 Background Indonesia’s population was 251 million in 2013, making it the third most populous country in the world after China and India. In 2013, the average gross national income (GNI) per person was US$  3740, making Indonesia an upper-middle-income country (1). During the Millennium Development Goal (MDG) era (2000–2015), it consistently ranked among the top five countries in the world in terms of the estimated number of tuberculosis (TB) cases per year. It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2013, the prevalence of HIV in the general population aged 15–49 years was 0.4% (95% confidence interval [CI]: 0.4–0.5%) (2), and it was estimated that 4.7% (95% CI: 2.7–7.2%) of TB patients were coinfected with HIV (3). National TB control efforts started around 1970, with TB diagnosis and treatment in primary health-care facilities providing the backbone of the national TB programme (NTP). Indonesia adopted the WHO-recommended DOTS strategy in 1995 (4,5). The estimated burden of TB disease published by WHO in 2013 included a TB incidence rate of 183 (95% CI: 164–207) per 100 000 population (equivalent to about 0.5  million cases per year), a TB prevalence of 272 (95% CI: 138–450) per 100 000 population (equivalent to a best estimate of 680 000 cases) and a TB mortality rate of 25 (95% CI: 14–37) per 100 000 population. These estimates drew on notification data and a national TB prevalence survey conducted in 2004. However, the 2004 prevalence survey used a screening algorithm based only on symptoms (with no chest X-ray) and confirmation of TB using smear alone (without culture). In addition, the gap between notifications and incidence (which reflects underdiagnosis and underreporting of detected cases) was hard to quantify. It was recognized that many health facilities were detecting TB cases but not notifying them to national authorities (6,7). Given these limitations, and the size of the estimated TB burden as a proportion of the global total, Indonesia was one of the 22 global focus countries for a national TB prevalence survey selected by the WHO Global Task Force on TB Impact Measurement in December 2007. In Indonesia, it was also recognized that a national TB prevalence survey – adopting the 2011 WHO recommendations for screening and diagnostic methods (8) – would improve estimates of the burden of TB disease in the country. Survey planning started in January 2011, and the survey was implemented from April 2013 to June 2014. Key methods and results There were 156 clusters in three geographical strata (Sumatra, Java-Bali and others) and two population strata (urban and rural), with a target cluster size of 500 individuals. A total of 112 350 individuals from 34 947 households were enumerated in the survey census, of whom 76 576 (68%) were eligible and invited to participate. Of these, 67 944 (89%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (8). A total of 15 446 participants (23%) were eligible for sputum examination, of whom 15 141 (98%) submitted at least one sputum specimen and 14 604 (95%) submitted two sputum specimens. A total of 426 bacteriologically confirmed pulmonary TB cases was identified, including 165 (39%) cases of smear- positive TB. The prevalence of smear-positive TB was 257 (95% CI: 210–303) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 759 (95% CI: 590–961) per 100 000 population. The prevalence of bacteriologically confirmed TB was higher in urban areas (846 per 100 000 population; 95% CI: 678– 1048) than in rural areas (674 per 100 000 population; 95% CI: 512–874), and higher in Sumatra (913 per 100 000 population; 95% CI: 697–1177) and other regions (842 per 100 000 population; 95% CI: 635–1092) than in Java-Bali (593 per 100 000 population; 95% CI: 447–771). Photo credit: Irwin Law National TB prevalence surveys 2007–2016134 Other key results were: • the male to female ratio was 3.0 for smear- positive TB and 2.3 for bacteriologically confirmed TB; • prevalence per 100 000 population increased with age; however, the absolute number of bacteriologically confirmed TB cases was relatively high in the young and middle-age groups (25–54 years); • among bacteriologically confirmed TB cases, 57% were symptom-screen positive, and among smear-positive cases, 70% were symptom-screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.3 overall, but varied from 1.7 in those 55–64 years to 4.9 in those aged 65 years or more, and was higher for men than for women (2.9 versus 1.4); • among bacteriologically confirmed TB cases, 86% had no previous history of anti-TB treatment, and only 4.2% were on anti-TB treatment at the time of the survey; • of the 225 bacteriologically confirmed and 101 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 147 (65%) and 38 (38%), respectively, had previously sought care in a public or private health facility for their symptoms; and • only 20% of participants reported to be on anti- TB treatment were in the national TB electronic register (SITT); and this was also confirmed by the high level of under-reporting (41%) documented in the 2017 national inventory study (9). Implications of results The estimated TB prevalence of 660 (95% CI: 523–813) per 100  000 population (all forms of TB and all ages) based on the survey was much higher than the previous WHO estimate of 272 (95% CI: 138–450) per 100 000 population. TB incidence was re-estimated at 399 (95% CI: 274–546) per 100 000 population, equivalent to one million new cases per year and double the pre-survey estimate. The TB mortality rate was estimated at 41 (95% CI: 26–59) per 100 000 population, equivalent to 100 000 deaths per year (10). The new estimates were used as the basis for the National TB Strategic Plan 2015–2019 and for a proposal to the Global Fund to Fight AIDS, Tuberculosis and Malaria. Other implications included: • a need for TB case notification to be legally mandated with enforcement to address the underreporting of detected TB cases. Regulations, tools, implementation guidelines, Photo credit: Irwin Law INDONESIA 135 supervision mechanisms, and monitoring and evaluation tools should be prepared for this purpose; • a need for intensified case finding for TB, which has since become one of the major strategies of the NTP; • a need for improved access to health facilities, including via provision of universal health insurance, so that symptomatic individuals would be more likely to seek immediate treatment; • a need for the general population to be made more aware that anti-TB treatment in standard health facilities is free of charge, to encourage people to seek care promptly; • a need to use chest X-rays more widely, to improve case detection; for example, as part of community outreach or among key populations, such as prisoners, people living with HIV, people with comorbidities and the elderly; • a need to increase the number of qualified laboratories to improve access to, and the speed of, diagnosis, especially in rural areas where geographical barriers hinder the rapid delivery of specimens to referral laboratories; • a need for the NTP to implement innovative strategies to supervise TB service quality in all health facilities, including those in the private sector; • a need to understand that a positive smear result should not be the basis for providing anti-TB treatment (especially in the context of active case finding in the community rather than in a clinic) given the low positive predictive value of smear microscopy without confirmatory testing, compared with culture; and • a need for increased funding to implement all of the policy and programmatic measures listed above, especially given the major finding of the survey that the burden of TB disease was double the level previously estimated. Major successes, challenges and lessons learned The overarching major success was that the survey was successfully implemented with high quality and a high participation rate, and that it was the first in the country for decades to include chest X-ray screening combined with diagnosis using culture as well as smear microscopy. Several major challenges included those listed below. • The procurement process for chest X-ray equipment was slow. It took 18 months and delayed the start of the survey. Subsequently, setting up and using the chest X-ray equipment in the field generated some problems with data collection. These were partly alleviated by the availability of in-country servicing of the equipment, which facilitated timely repairs and troubleshooting. • Collecting and processing sputum specimens was demanding, and some contamination of specimens occurred. Morning specimens had higher rates of contamination (431 [3%] of 14  569 specimens) than spot specimens (47 [1%] of 4433). • Culture failed to grow in some settings, perhaps because of geographical challenges (e.g. poor road conditions and the difficulty of maintaining a cold chain in the context of high temperatures and humidity), poor sample handling and the limited number of laboratories. • The quality of laboratories may have varied, even though laboratory experts evaluated and validated the performance quality of the laboratories used in the survey. • Limited culture capacity meant that it was only possible to culture two specimens for every participant who submitted sputum samples in one third of survey clusters; in the remaining survey clusters, only one specimen was cultured. To mitigate this problem, Xpert MTB/ RIF was used when culture failed (e.g. from contamination in all tubes). Photo credit: Irwin Law National TB prevalence surveys 2007–2016136 • It took time for the updated estimates of TB disease burden to be officially accepted at the higher levels of the Ministry of Health (MoH). Thus, although events to disseminate results were held in October 2014 (shortly after field operations were completed), acceptance of results and publication of the survey report were delayed. After further discussions and briefings, the updated disease burden estimates were agreed in July 2015, and the survey report was published in September 2015. There were also some more minor challenges: • field operations in a few clusters were delayed by forest fires and volcanic activity; and • the participation rate was low in urban clusters, especially in economically wealthy areas in large cities. Most of the residents in these areas already had good access to health services, including annual health screening with chest X-ray, so the X-ray screening offered as part of the survey provided no incentive to participate. Important lessons learned for future surveys included: • even if the NTP or MoH is not directly involved in survey implementation, it is still important to ensure their involvement and ownership throughout the process, from design to dissemination of results. This facilitates survey implementation and rapid uptake and use of results; • although prior prevalence surveys can be used to help assess trends in TB disease burden, this is challenging when previous surveys have used different (and less sensitive) screening and diagnostic methods; and • to maintain high-quality laboratory services throughout the survey, laboratories need to be standardized and monitored frequently. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. Indonesia Tuberculosis Prevalence Survey 2013–2014. Ministry of Health, Republic of Indonesia; National Institute of Health Research and Development; in collaboration with Directorate General of Disease Control and Environment Health; 2015. 7. Tuberculosis Prevalence Survey in Indonesia 2004. National Institute of Health Research and Development, Ministry of Health- Republic of Indonesia; Jakarta, Indonesia; 2005. 8. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 9. World Health Organization. Global tuberculosis report 2018. Geneva: WHO; 2018; (http://apps.who.int/iris/bitstream/hand le/10665/274453/9789241565646-eng.pdf, accessed December 2018). 10. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015; (http://apps.who.int/iris/ bitstream/10665/191102/1/9789241565059_eng.pdf, accessed January 2018). 137 Name Role Organization Joseph Sitienei Principal investigator (PI) National Tuberculosis, Leprosy and Lung Disease Program (NTLD-P) Hillary Kipruto Co-PI WHO Kenya Jane Ong’ang’o Co-PI & study coordinator Kenya Medical Research Institute Bernard Langat Co-investigator NTLD-P Enos Masini Co-investigator NTLD-P Margaret Ndisha Co-investigator NTLD-P Faith Ngari Co-investigator NTLD-P Obadiah Njuguna Co-investigator NTLD-P Janet Agaya Co-investigator Kenya Medical Research Institute Jeremiah Chakaya Co-investigator Kenya Medical Research Institute Joel Kangangi Co-investigator WHO Kenya Maurice Maina Co-investigator United States Agency for International Development (USAID) Brenda Mungai Co-investigator Centre for Health Solutions, Kenya Rose Mwirigi Co-investigator National Tuberculosis Reference Laboratory Anja Vant’Hoog Co-investigator Academic Medical Centre of the University of Amsterdam Josephine Wahogo Co-investigator & laboratory manager National Tuberculosis Reference Laboratory Veronica Manduku Co-investigator & lead radiologist Kenya Medical Research Institute Geoffrey Okallo Data management team leader NTLD-P Richard Kiplimo  Data manager NTLD-P Amos Ndombi Data manager NTLD-P Dickson Kirathe IT manager NTLD-P Martin Githiomi IT officer NTLD-P Drusilla Nyaboke Logistician NTLD-P Maureen Kamene Kimenye Member of report writing committee NTLD-P Janice Njoroge Communication specialist Centre for Health Solutions, Kenya James Ng’ang’a Statistician Kenya National Bureau of Statistics Emily Bloss Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Martien W. Borgdorff Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Kenya Kevin Cain Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Kenya Julia Ershova Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Irwin Law Technical assistance (survey advisor) WHO headquarters Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Peou Satha Technical assistance (survey advisor) Consultant, Cambodia Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Sayori Kobayashi Technical assistance (data management) WHO headquarters Hazim Timimi Technical assistance (data management) WHO headquarters Key people Surveyed clusters (N=99)a Summary statistics Participation rate 83% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 558 2.3 Prevalence:notification ratio (Bacteriologically confirmed TB, ≥15 years) 3.5 Finance Amount (US$) The Global Fund/USAID TB ARC 30 627 The Global Fund 4 530 712 USAID 491 892 WHO/USAID 121 612 Total budget 5 174 843 Survey organization and financing Implementing agency: National Tuberculosis, Leprosy and Lung Disease Program KENYA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Data sources ■ Kenya Tuberculosis Prevalence Survey 2016, Survey Report. National Tuberculosis, Leprosy and Lung Disease Program, Ministry of Health, Republic of Kenya; 2018 (https://www. nltp.co.ke/survey-reports-2/). ■ Survey dataset. 2015–2016 National TB prevalence surveys 2007–2016138 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Urban, rural/enumeration area Sample size assumptions • Smear-positive prevalence 269 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.7 • k 0.6 • Response rate 85% • Sample size (estimated) 72 000 Number of clusters 100a Cluster size 720 Eligibility criteria • Age ≥15 years • Residency Residents who lived in the selected cluster for at least 30 consecutive days prior to the census Survey design and methodology a One cluster in Mandera was cancelled due to a security issue. Laboratory methodology Smear Two samples (spot, morning): direct preparation, FM (LED, auramine stain) Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB MPT64 rapid test TB drug susceptibility test Done Xpert® MTB/RIF Done for all morning samples and spot samples lacking a matching morning sample. HIV test HIV status was verbally obtained from participants. For prevalent TB cases, it was also obtained from the TB electronic and reporting system. Screening criteria Interviewa Cough ≥2 weeks Chest X-rayb Any lung abnormality Other Chest X-ray exempted a An in-depth interview about health-care seeking behaviour was done for all participants who reported any TB symptoms (cough, sputum production, haemoptysis, chest pain, fever, fatigue, night sweats, weight loss, shortness of breath). b Direct digital radiography. a The team used paper for field data collection throughout field operations in one cluster, due to the breakdown of the electronic system. Field data collection Papera/electronic Database SQL Method of analysis MI+IPW Results first published in a report/paper March 2018 Official dissemination event March 2017 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 230 174–286 558 455–662 Male 346 260–431 809 656–962 Female 138 79–196 359 258–460 15–24 years 218 133–303 360 242–478 25–34 years 259 164–353 716 526–906 35–44 years 297 164–430 602 422–782 45–54 years 234 101–367 607 432–781 55–64 years 118 24–211 587 372–803 ≥65 years 125 24–226 576 368–783 Urban 335 213–456 760 539–981 Rural 175 126–224 453 357–549 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 4 137 – Location of care soughtb 1 257 30 • Consulted medical facility Public facility 1047 – Private facility 198 – Other 3 – • Pharmacy 56 – • Traditional healer 9 – No action taken 2 763 67 Unknown 117 2.8 a Cough ≥2 weeks. b The subtotals do not add up to 1257 because participants could select more than one category. a Data were available only for participants who were eligible for sputum submission. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b DST was conducted for 225 participants. Survey participants currently on TB treatment Number % Total participants currently on TB treatmenta 62 – • Treated in the public sector 23 37 • Treated in the private sector 0 0 • Treated in other sector 1 1.6 • Treated in unknown sector 38 61 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 15 4.9 Design effect k Smear-positive TB 1.8 0.7 Bacteriologically confirmed TB 2.5 0.7 Other sputum results Number % Total smear-positive participants 141 – Smear-positive participants without MTB confirmationa 18 13 Isolates with MDR-TB detectedb 6 2.7 a Age ≥15 years unless otherwise specified. KENYA 139 Survey flow: census to final outcomes Field operations: August 2015 to July 2016 Individuals enumerated in census 126 389 Eligible for sputum examination 9 715 (15%) Total bacteriologically confirmed cases 305 Eligible study population 76 291 (60%) Total participants 63 050 (83%) Interview and chest X-ray 62 484 (99%) Interview only 566a (0.9%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 1 241 (13%) Symptom positive, chest X-ray negative or N/A 2 896 (30%) Symptom negative, chest X-ray positive 5 184 (53%) Otherc 394 (4.1%) Symptom positive, chest X-ray positive 115 (38%) Symptom positive, chest X-ray negative or N/A 32 (10%) Symptom negative, chest X-ray positive 154 (50%) Otherc 4 (1.3%) Submitted specimens At least one specimen 9 120 (94%) Both specimens 7 763 (80%) Smear-positive casese 123 (40%) Definite 123 Probable N/A Smear-negative casese 182 (60%) Definite 182 Probable N/A Laboratory result At least one culture result availabled 8 761 (90%) At least one Xpert MTB/RIF result available 8 936 (92%) Symptom screening Cough ≥2 weeksb 4 137 (6.6%) Sputum production 3 256 (5.2%) Haemoptysis 393 (0.6%) Chest pain 12 290 (19%) Fever 4 937 (7.8%) Total symptom-screen positivea 4 137 (6.6%) Chest X-ray screening Normal 50 935 (82%) Abnormalb 6 425 (10%) Other abnormality 5 124 (8.2%) Total chest X-rays taken 62 484 a 429 participants declined a chest X-ray, and 137 participants did not have a chest X-ray due to malfunctioning X-ray machines. b Eligible for sputum collection. c Chest X-ray exempted and symptom-screen negative. d Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. e Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. Ineligible individuals 50 098 (40%) Children <15 years 47 428 (38%) Did not meet residency criteria 2 670 (2.1%) National TB prevalence surveys 2007–2016140 10 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 80 90 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of bacteriologically confirmed TB prevalence to notifications by age and by sexc Age group (years) and sex 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 15 20 0 0 1 2 3 4 5 6 7 8 9 11 12 13 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 300 400 500 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 800 0 Prevalence per 10 0 0 0 0 population 10 000 30 000 40 000 50 000 60 000 20 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1000 500 700 600 800 900 400 300 200 100 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 3.08, variance 7.63, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using bacteriologically confirmed pulmonary TB notifications (2015) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Total (≥15) KENYA 141 Background Kenya had a population of 47 million in 2015. The average gross national income (GNI) per person was US$  1310 per year, making it a low-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. Kenya’s TB notification rate (new and relapse cases) increased from 50 per 100  000 population in 1990 to 287 per 100  000 population in 2005, then slowly decreased to 194 per 100 000 population in 2014. WHO estimates of TB incidence and prevalence in 2014 were 246 (95% confidence interval [CI]: 240–252) per 100 000 population and 266 (95% CI: 142–427) per 100 000 population, respectively (2). Like many other sub-Saharan African countries, from the mid-1980s Kenya was severely affected by the HIV/AIDS epidemic. From the mid-2000s, large investments in TB/ HIV collaborative activities resulted in a high proportion (>90%) of TB patients knowing their HIV status, and high uptake of antiretroviral therapy among coinfected patients. An integrated TB/HIV data collection system was implemented in Kenya in 2005, enabling the collection of HIV-related information as a standard part of TB diagnosis and treatment. The prevalence of HIV among TB patients with an HIV test result was 57% in 2005, declining to 39% in 2012 (2). In 2015, the prevalence of HIV in the general population aged 15–49 years was 5.6% (95% CI: 4.9–6.3%) (3), and it was estimated that 33% (95% CI: 32–35%) of TB patients were coinfected with HIV (4). A national TB prevalence survey (excluding the northern province and Nairobi) was implemented in Kenya in 1958–1959. As part of this survey, a tuberculin skin test was done for the whole population except infants aged under 1 month, and chest X-ray and bacteriological examinations (smear and culture) were done for all participants aged 10 years or more. The survey found a prevalence of approximately 3100 per 100 000 population, equivalent to 110 000 cases in the population of 3.5 million aged 10 years or more (5, 6). In December 2007, Kenya was one of the 22 global focus countries for a national TB prevalence survey that was selected by the WHO Global Task Force on TB Impact Measurement. In 2009, the Ministry of Health and the National Tuberculosis, Leprosy and Lung Disease Program (NTLD-P) decided to implement a second national TB prevalence survey. The survey started in July 2015 and was completed in July 2016 (6). Key methods and results There were 99 survey clusters in two strata (urban and rural), with a target cluster size of 720 individuals. A total of 126 389 individuals from 31 955 households were enumerated in the survey census, of whom 76 291 (60%) were eligible and invited to participate. Of these, 63 050 (83%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using a chest X-ray and an interview about symptoms (7). A total of 9715 participants (15%) were eligible for sputum examination; of these, 9120 (94%) submitted at least one sputum specimen and 7763 (80%) submitted two sputum specimens. This was one of the first surveys to test all screen-positive participants with both culture and Xpert® MTB/RIF. Valid Xpert MTB/RIF results were available for 8936 participants. Of these, 237 (2.7%) were Xpert positive for Mycobacterium tuberculosis, and six (2.5% of 237) were also rifampicin (RIF) resistant. Of 305 bacteriologically confirmed TB cases, 147 (48%) were confirmed by both culture and Xpert MTB/RIF, 68 (22%) were positive only by culture, and 90 (30%) were positive only by Xpert MTB/RIF.Photo credit: Irwin Law National TB prevalence surveys 2007–2016142 A total of 305 bacteriologically confirmed pulmonary TB cases were identified, including 123 (40%) cases of smear- positive TB. The prevalence of smear-positive TB was 230 (95% CI: 174–286) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 558 (95% CI: 455–662) per 100  000 population. Prevalence rates for smear-positive and bacteriologically confirmed TB were higher in urban areas than in rural areas. Other key results were: • the male to female ratio was 2.5 for smear- positive TB and 2.3 for bacteriologically confirmed TB; • the prevalence per 100 000 population was high for people aged 25 years or more, with a peak in those aged 25–34 years; the absolute number of bacteriologically confirmed TB cases in those aged under 45 years was relatively high; • among bacteriologically confirmed TB cases, 48% were symptom-screen positive, and among smear-positive TB cases, 69% were symptom- screen positive; • for bacteriologically confirmed TB, the ratio of prevalence to notifications (P:N ratio) was 3.5 overall, but varied from 2.8 in those aged 35– 44 years to 6.4 in those aged 65 years or more, and was slightly higher for men than women (3.8 versus 3.5); these findings were consistent with the 2013 TB inventory study, which found a high level of under-reporting (21%) of smear- positive TB cases especially in those over 55 years of age (8); • among bacteriologically confirmed TB cases, 72% had no previous history of anti-TB treatment and only 4.9% were on anti-TB treatment at the time of the survey; and • of the 139 bacteriologically confirmed and 52 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 52 (37%) and 34 (43%), respectively, had previously sought care in a public or private health facility for their symptoms; this was similar to findings from a patient-pathway analysis (PPA) in 2013, in which of those who sought care, 58% and 41% respectively initially sought care in a public or private health facility (9); and • although HIV testing was not done during field operations, the HIV status of bacteriologically confirmed TB cases was obtained from the national HIV electronic and reporting system; and of 305 bacteriologically confirmed TB cases, 41 (13%) were HIV-positive, 204 (67%) were HIV-negative and for 60 (20%) the status was unknown. Implications of results The survey showed that TB prevalence for all forms of TB and all ages, at 426 per 100  000 population (95% CI: 347–504), was significantly higher than the pre- survey estimate of 266 (95% CI: 142–427) per 100  000 population (3). The burden of TB was much higher than that reported through routine surveillance, especially among men and older age groups. The survey had several major programmatic, policy and funding implications: • the high prevalence in younger age groups, especially among men, suggested considerable active transmission of TB in the general community; • among confirmed TB cases, most (65%) of those with symptoms who did not seek treatment were men, consistent with greater reluctance among men to seek care for HIV (10); together with the finding that men had a higher burden of TB disease, this showed a need for innovative approaches to reduce barriers to accessing care and associated delays in diagnosis and treatment for TB among men; • among bacteriologically confirmed TB cases, more than half would have been missed if screening using the single criterion of cough of more than two weeks was relied upon; this Photo credit: Irwin Law KENYA 143 suggested that the screening criteria used in routine clinical settings should be reviewed and that expanded use of chest X-ray as a screening tool should be considered; • since more than half of the bacteriologically confirmed cases were smear-negative and were diagnosed by culture or Xpert MTB/RIF (or both), use of diagnostic tools besides smear microscopy should be expanded; • about 70% of participants who reported a chronic cough did not seek care, even though the Kenya Demographic Health Survey of 2014 found that about 80% of those aged 15–49 years knew that TB is spread through the air by coughing (11); nonetheless, the general population may be unaware of the actual symptoms of TB, and consequently delay seeking care; this suggested that improving community awareness about TB symptoms as well as the availability of free TB services at public health facilities could help to improve health care seeking behaviour; • the relatively high proportion of symptomatic cases who had sought some care before the survey but were not diagnosed with TB suggested a need to improve access to diagnostics and treatment, as well as a need to review the screening algorithm and develop strategies to improve patient awareness and health-care provider knowledge of TB symptoms; and • the prevalence of HIV infection among bacteriologically confirmed TB cases with known HIV status (17%; 41/2451) was lower than that reported among notified TB cases (33%) (12); this suggested that while there has been a strong focus on the TB/HIV programme, a large TB burden exists among those who are HIV-negative, for which more programmatic action is required. Major successes, challenges and lessons learned The national TB prevalence survey in Kenya 2015– 2016 was successfully implemented. This was the first African survey to use Xpert MTB/RIF and culture for all participants who screened positive, and despite the resulting increase in workload for the national reference laboratory, the survey demonstrated that using both tests was feasible. Good communication throughout the survey contributed to these achievements. This included high levels of community engagement (especially during visits prior to survey field operations) that fostered survey participation, and regular meetings and close collaboration between the NTLD-P, various implementing partners and technical agencies that facilitated effective project management and ownership of the final survey results. Challenges faced during the survey, and associated lessons learned, included those listed below. • The procurement process for digital chest X-ray machines by the WHO Regional Office for Africa was lengthy, which delayed the start of the survey by more than a year. Photo credit: Marina Tadolini 1 41 were HIV-positive and 204 were HIV-negative. National TB prevalence surveys 2007–2016144 • There was overreliance on the internet-based data management system in the field. Although electronic data collection in the field was innovative and efficient, enumeration data from the field had to be uploaded to the central server in Nairobi before other questionnaire data could be entered. For clusters that had good internet connection, this worked well, but for clusters with poor coverage, survey teams had to switch to a paper-based data collection system and were then faced with the issue of merging data from different systems. From about mid-way through field operations, development of a local area network in the field circumvented the need to upload data to the central server and improved the efficiency of electronic data management. • There were problems in linking laboratory and field data because the laboratory health information system was different from the one used by the survey itself, and mismatching of barcodes (this typically happened when they were handwritten). It took five months from the end of field operations to complete data cleaning. • One cluster (close to the border with Somalia) was cancelled due to security issues. • Budgetary limitations constrained the number of central chest X-rays that could be read centrally. Initially, it was planned that all chest X-rays would be read by the central radiologists, but in practice this was limited to only specific categories of images (i.e. all those with abnormal results; 10% of normal images, as determined by the field teams; and all images with discordant findings between two field readers). References 1. The World Bank. (https://data.worldbank.org/country, accessed January 2018). 2. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015 (http://apps.who.int/iris/ bitstream/10665/191102/1/9789241565059_eng.pdf, accessed January 2018). 3. UNAIDS. (http://aidsinfo.unaids.org/, accessed January 2018). 4. World Health Organization. Global tuberculosis database. Geneva: WHO; 2017 (http://www.who.int/tb/data/en/, accessed January 2018). 5. Roelsgaard E, Nyboe J. A tuberculosis survey in Kenya. Bulletin of the World Health Organization. 1961;25(6):851–870 (http://www. ncbi.nlm.nih.gov/pmc/articles/PMC2555628/, accessed February 2018). 6. Kenya Tuberculosis Prevalence Survey 2016. Survey Report. National Tuberculosis, Leprosy and Lung Disease Program, Ministry of Health, Republic of Kenya; 2018 (https://www.nltp. co.ke/survey-reports-2/). 7. World Health Organization. Tuberculosis prevalence surveys: a handbook. Geneva: WHO; 2011 (https://apps.who.int/iris/ bitstream/handle/10665/44481/9789241548168_eng.pdf, accessed August 2016). 8. Tollefson D, Ngari F, Mwakala M, Gethi D, Kipruto H, Cain K, Bloss E. Under-reporting of sputum smear-positive tuberculosis cases in Kenya. INt J Tuberc Lung Dis. 2016 Oct; 20(10):1334- 1341. 9. Masini E, Hanson C, Ogoro J, Brown J, Ngari F, Mingkwan P, Makayova J, Osberg M. Using patient-patheway analysis to inform a differentiated program response to tuberculosis: the case of Kenya. JID 2017;216(S7):S714-23. 10. UNAIDS. Get on the Fast-Track, The life-cycle approach to HIV, 2016. (http://www.unaids.org/sites/default/files/media_asset/Get- on-the-Fast-Track_en.pdf, accessed March 2018). 11. Republic of Kenya. Demographic and Health Survey. 2014 (https://dhsprogram.com/pubs/pdf/fr308/fr308.pdf, accessed February 2018). 12. World Health Organization. Global tuberculosis report 2016. Geneva: WHO; 2016 (https://apps.who.int/iris/bitstream/han dle/10665/250441/9789241565394-eng.pdf, accessed January 2018). 145 Name Role Organization Phannasinh Sylavanh Director and principal investigator National TB Control Programme Saveang Saisongkham Deputy director National TB Control Programme Phouvang Vangvichit Deputy director National TB Control Programme Soth Bounmala Survey coordinator/field team leader National TB Control Programme Phonenaly Chittamany Chief of statistics/field team leader National TB Control Programme Manikhone Ouanephongchaleune   Monitoring and evaluation/field team leader National TB Control Programme Bounkong Fongosa                           Monitoring and evaluation/field team leader National TB Control Programme Thavone Phengsavatdy Technical officer National TB Control Programme Liene Phonekeo Finance officer National TB Control Programme Donekham Inthavong Laboratory manager National TB Control Programme Phasouk Senephansiri Laboratory co-manager National TB Control Programme Oroth Rajphol Radiologist Mahosot hospital, Lao People's Democratic Republic (Lao PDR) Vongvilay Paphatsalang Radiologist Mahosot hospital, Lao People's Democratic Republic (Lao PDR) Vatthana Nanthana Country director advisor/translator Damien Foundation, Lao People's Democratic Republic (Lao PDR) Jacques Sebert Medical officer WHO Lao People's Democratic Republic (Lao PDR) Irwin Law Data manager/epidemiologist National TB Control Programme Fulgence Nzabintwali Technical assistance/laboratory co-manager National TB Control Programme Phimpha Paboriboune Scientific director Centre d’Infectiologie Christophe Merieux du Laos Vibol Iem Scientist Fondation Merieux, Lao People's Democratic Republic (Lao PDR) Pierre L’Her Technical assistance (pulmonologist, radiologist) Soutien Pneumologique International, France Etienne Leroy-Terquem Technical assistance (pulmonologist) Soutien Pneumologique International, France Charalampos Sismanidis Technical assistance (statistician) WHO headquarters Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Sang Jae Kim Technical assistance (laboratory advisor) Korean Institute of Tuberculosis, Republic of Korea Peou Satha Technical assistance (radiology and survey advisor) National Centre for TB and Leprosy Control, Cambodia Key people Surveyed clusters (N=50)a Data sources ■ Report of the first national tuberculosis prevalence survey in Lao PDR (2010–2011). Vientiane, Lao PDR: National Tuberculosis Centre, Department of Communicable Diseases, Ministry of Health - Lao PDR; 2014. ■ Law I, Sylavanh P, Bounmala S, Nzabintwali F, Paboriboune P, Iem V et al. The first national tuberculosis prevalence survey of Lao PDR (2010–2011). Trop Med Int Health. 2015;20(9):1146– 1154. ■ Survey dataset. Summary statistics Participation rate 85% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 595 2.3 Prevalence:notification ratio (smear-positive TB, ≥15 years) 3.5 Finance Amount (US$) The Global Fund 1 275 070 USAID 16 000 Total budget 1 291 070 Survey organization and financing Implementing agency: National TB Control Programme 2010–2012 a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. LAO PEOPLE’S DEMOCRATIC REPUBLIC National TB prevalence surveys 2007–2016146 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata No stratification was used, but final analysis accounted for urban and rural. Sampling unit Province/district/village/enumeration area Sample size assumptions • Smear-positive prevalence 251 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.3 • k 0.4 • Response rate 80% • Sample size (estimated) 40 000 Number of clusters 50 Cluster size 800 Eligibility criteria • Age ≥15 years • Residency Slept in the household for 14 days prior to the census Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): direct preparation, ZN Culture Two samples (spot, morning): direct preparation, Ogawa media Identification of MTB PNB, GenoType MTBDRplus (LPA) TB drug susceptibility test Done Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interviewa Cough ≥2 weeks within the past month and/or haemoptysis within the past month Chest X-rayb Any lung abnormality Other N/A a An in-depth interview about health-care seeking behaviour was done only for participants who had symptoms suggestive of TB. b Conventional radiography. Field data collection Paper Database Filemaker Pro 10 Method of analysis MI+IPW Results first published in a report/paper January 2014 Official dissemination event January 2013 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 3 239 – Location of care sought • Consulted medical facility 1 148 35 Public facility 990 86 Private facility 106 9.2 Otherb 52 4.5 • Pharmacy 690 21 • Traditional healer 26 0.8 Self-treated N/A N/A No action taken 1 210 37 Unknown 165 5.1 Key survey results a Cough ≥2 weeks and/or haemoptysis. b Village health volunteer (32), another country (17) and other (3). a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). b DST was done for 226 culture MTB-positive cases. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 42 – • Treated in the public sector 21 50 • Treated in the private sector 0 0 • Treated in unknown sector 21 50 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 6 2.5 Design effect k Smear-positive TB 2.2 0.7 Bacteriologically confirmed TB 3.2 0.7 Other sputum results Number % Total smear-positive participants 186 – Smear-positive participants without MTB confirmationa 92 50 Isolates with MDR-TB detectedb 2 0.9 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 278 199–356 595 457–733 Male 420 299–541 855 646–1 064 Female 152 88–215 366 254–477 15–24 years 80 11–149 145 41–249 25–34 years 184 16–352 292 120–464 35–44 years 201 98–304 484 307–661 45–54 years 412 234–590 714 461–968 55–64 years 513 279–747 1 131 704–1 557 ≥65 years 857 503–1 229 2 410 1 665–3 156 Urban 264 130–398 436 307–565 Rural 283 186–380 663 477–848 a Age ≥15 years unless otherwise specified. LAO PEOPLE’S DEMOCRATIC REPUBLIC 147 Survey flow: census to final outcomes Field operations: July 2010 to January 2012 a Eligible for sputum collection. b Symptom-screening results were not available for eight people. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but chest X-ray suggestive of TB. e Definite: MTB confirmed by two culture specimens, or by one culture with chest X-ray suggestive of TB. Probable: MTB confirmed by one culture with five or more colonies without chest X-ray suggestive of TB, or by one culture with less than five colonies and chest X-ray suggestive of TB. Individuals enumerated in census 78 819 Eligible for sputum examination 6 346 (16%) Total bacteriologically confirmed cases 237 Eligible study population 46 079 (59%) Total participants 39 212 (85%) Interview and chest X-ray 39 074 (99.6%) Interview only 49 (0.1%) Chest X-ray only 89 (0.3%) Symptom positive, chest X-ray positive 1 312 (21%) Symptom positive, chest X-ray negative or N/A 1 927 (30%) Symptom negative or N/A, chest X-ray positiveb 3 107 (49%) Other N/A Symptom positive, chest X-ray positive 111 (47%) Symptom positive, chest X-ray negative or N/A 7 (3.0%) Symptom negative, chest X-ray positive 119 (50%) Other N/A Submitted specimens At least one specimen 6 290 (99%) Both specimens 6 253 (98%) Smear-positive casesd 107 (45%) Definite 94 Probable 13 Smear-negative casese 130 (55%) Definite 101 Probable 29 Laboratory result At least one culture result availablec 6 251 (99%) Symptom screening Cough ≥2 weeksa 3 211 (8.2%) Haemoptysisa 991 (2.5%) Sputum production N/A Chest pain N/A Fever N/A Total symptom-screen positivea 3 239 (8.3%) Chest X-ray screening Normal 33 890 (87%) Abnormala 4 419 (11%) Other abnormality 854 (2.2%) Result not available N/A Total chest X-rays taken 39 163 Ineligible individuals 32 740 (41%) Children <15 years 21 517 (27%) Did not meet residency criteria 11 223 (14%) National TB prevalence surveys 2007–2016148 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 3.0 2.0 3.5 2.5 4.0 4.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 2000 400 600 800 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 500 1000 1500 2000 2500 3000 0 Prevalence per 10 0 0 0 0 population 1 000 3 000 4 000 5 000 6 000 2 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 3500 1500 2000 2500 3000 1000 500 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 4.74, variance 15.5, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 4 2 8 6 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1615 Total (≥15) LAO PEOPLE’S DEMOCRATIC REPUBLIC 149 Background Lao People’s Democratic Republic (PDR) is a landlocked country. In 2010, it had a population of 6.3 million and was one of the poorest countries in South-East Asia, with an average gross national income (GNI) per person of US$ 1000 per year, making it a lower-middle income country (1). The prevalence of HIV in the general population aged 15–49 years was 0.2% (95% confidence interval [CI]: 0.2–0.3%) (2), and it was estimated that 4.5% (95% CI: 3.7–5.4%) of tuberculosis (TB) patients were coinfected with HIV (3). The National TB Control Programme (NTP) was established in 1995. By 2005, the WHO-recommended DOTS strategy (4,5) had reached full country coverage across all 17 provinces and in all of the 140 district hospitals. As DOTS coverage expanded, the case notification rate (new and relapse cases) increased rapidly, from 41 per 100  000 population in 2000 to 65 per 100  000 population in 2005. Subsequently, the case notification rate stagnated, and the best estimate of the case detection rate (notifications of new cases divided by incidence) was 31% in 2011. Nonetheless, there was considerable uncertainty about the burden of TB disease, and the gap between notifications and incidence (due to underreporting or underdiagnosis of cases) was unclear (6–8). To better understand the burden of TB disease in the country, a decision to implement a national TB prevalence survey was taken in mid-2007. After three years of preparations, the survey was implemented from July 2010 to January 2012. Lao PDR was not one of the 22 global focus countries for national TB prevalence surveys identified by the WHO Global Task Force on TB Impact Measurement in December 2007. Nevertheless, Lao PDR was a regional priority for the WHO Western Pacific Region and was on the Task Force’s longer list of 53 countries considered to meet survey eligibility criteria. Key methods and results There were 50 survey clusters (no stratification was used at the time of survey design, but both urban and rural strata were examined during the analysis), with a target cluster size of 800 individuals. A total of 78  819 individuals from 14 800 households were enumerated in the survey census, of whom 46 079 (59%) were eligible and invited to participate. Of these, 39  212 (85%) did so. All participants were screened in line with the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (9). A total of 6346 people (16% of participants) were eligible for sputum examination, of whom 6290 (99%) submitted at least one sputum specimen and 6253 (99%) submitted two sputum specimens. Photo credit: Jacques Sebert National TB prevalence surveys 2007–2016150 A total of 237 bacteriologically confirmed pulmonary TB cases was identified, including 107 cases of smear- positive TB. The prevalence of smear-positive TB was 278 (95% CI: 199–356) per 100 000 population (among those aged ≥15 years) and for bacteriologically confirmed TB it was 595 (95% CI: 457–733) per 100  000 population. Prevalence in rural clusters was higher than in urban clusters. Other key results were: • the male to female ratio for TB prevalence was 2.8 for smear-positive TB and 2.3 for bacteriologically confirmed TB; • prevalence per 100 000 population increased with age; the absolute number of bacteriologically confirmed TB cases was highest in the group aged 65 years or more, and consistently high in other age groups; • among bacteriologically confirmed TB cases, 50% were symptom-screen positive, and among the smear-positive cases, 66% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 3.5 overall, but varied from 2.4 in those aged 35–44 and 55–64 years to 4.2 in the age group 15–24 years; the ratio was higher for men than for women (4.3 versus 2.6); • among bacteriologically confirmed TB cases, 6% had no previous history of anti-TB treatment, and only 3% were on anti-TB treatment at the time of the survey; and • of the 113 bacteriologically confirmed and 67 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 42 (37%) and 27 (43%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results Based on survey results, WHO estimated that the prevalence of TB (all ages, all forms of TB) in 2011 was 540 (95% CI: 353–767) per 100 000 population; estimates for previous years were also revised. The 2011 estimate was almost double the pre-survey WHO estimate that was used in the initial sampling design for the survey (289 per 100  000 population in 2007). The updated estimate of prevalence in 2011 was 64% lower than the revised 1990 estimate of 1490 (95% CI: 746–2490) per 100  000 population, indicating that the country had met the Millennium Development Goal target related to TB (that incidence should be falling by 2015) and the Stop TB Partnership target of halving TB prevalence between 1990 and 2015. Although it was not possible to quantify the relative contribution of the various factors that led to this decline, those considered to have played an important role included the countrywide expansion of DOTS and the associated availability of free anti-TB medication, increases in GNI per capita (from US$ 190 in 1990 to US$ 1120 in 2011) and improvements in overall Photo credit: Jacques Sebert LAO PEOPLE’S DEMOCRATIC REPUBLIC 151 people suggested a reluctance to seek care, possibly linked to health services that were not meeting the needs or expectations of this population. • Diagnostic services should be improved, progressing from a reliance on sputum smear microscopy to greater use of chest X-ray and either culture or rapid tests (e.g. Xpert® MTB/ RIF). • A smear-positive test result does not always indicate TB disease, especially in a community (as opposed to a clinic) setting. In active TB case finding, TB cannot be reliably diagnosed based on smear examination alone. • The ability of health-care workers to clinically recognize TB disease should be improved, given that one-third of symptomatic survey cases had already sought care in a public or private health facility, before being detected by the survey. Survey findings were used to prepare a funding application to the Global Fund to fight AIDS, TB and malaria, and to develop a new national strategic plan for TB. Major successes, challenges and lessons learned Major successes included completion of the survey with a small budget (US$  1.3 million), a high participation rate and the fact that many NTP staff were able to see, first-hand and often for the first time, the challenges of TB surveillance and case management in the more remote areas of the country. The survey was successfully implemented with the use of entirely conventional or traditional survey methods (i.e. paper-based data collection instruments, conventional chest X-ray systems and the Kudoh culture method with Ogawa media). Major challenges included the time taken to create the laboratory capacity needed for the survey (it took two years to refit the central-level laboratory), interruptions to funding, a need to mobilize additional funding towards the latter stages of the survey, and difficulties in ensuring that results were clearly understood and accepted by key stakeholders. It also took time to prepare the survey report due to the lack of staff needed for this task. Important lessons learned for future surveys included: • good financial planning is essential to ensure the smooth progress of a survey; • good technical assistance throughout survey preparations and implementation can help to ensure survey quality, especially when a survey living conditions (the Human Development Index was 0.397 in 1990 and 0.554 in 2011) (6,7). In common with other countries in Asia, the survey showed a markedly ageing TB epidemic, with prevalence in those aged 65 years or more as much as 10 times the level in those under 25 years of age. This suggested that transmission of infection was in decline and that endogenous re-activation of TB in older age groups, as opposed to new infections in the younger population, was likely to make a growing contribution to the overall TB burden. The survey had several major programmatic, policy and funding implications, which included those listed below. • It was clear that further efforts were needed to close gaps in case detection. The gap between prevalence and official notifications of new cases (the P:N ratio) was among the largest found in any survey conducted between 2009 and 2016. The particularly high P:N ratio for men compared with women, and for people aged under 35 years and 65 years or more, also indicated a need for interventions targeted to specific subpopulations. • In addition to programmatic efforts, the high P:N ratio indicated a broader need to strengthen the health system, and the overall availability and acceptability of diagnostic and treatment services. The chronicity of symptoms in older Photo credit: Irwin Law National TB prevalence surveys 2007–2016152 has not previously been conducted in the country; in Lao PDR, three full-time international staff based in the country provided support throughout, including the training and pilot phases, during which revisions were made to the protocol and data-collection tools; additional support was provided by staff involved in the Cambodian surveys (2002 and 2010–2011), and country missions were undertaken by staff from WHO headquarters and other technical partners including the Korean Institute of Tuberculosis, Centre d’Infectiologie Christophe Mérieux du Laos and Soutien Pneumologique International (France); and • a transparent and open communication strategy among all stakeholders helps to facilitate the adoption of new prevalence estimates (and programmatic implications).    References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. Law I, Sylavanh P, Bounmala S, Nzabintwali F, Paboriboune P, Iem V et al. The first national tuberculosis prevalence survey of Lao PDR (2010–2011). Trop Med Int Health. 2015;20(9):1146–1154 (https://www.ncbi.nlm.nih.gov/pubmed/25939366, accessed July 2017). 7. Report of the first national tuberculosis prevalence survey in Lao PDR (2010–2011). Vientiane, Lao PDR: National Tuberculosis Centre, Department of Communicable Diseases, Ministry of Health - Lao PDR; 2014. 8. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf, accessed January 2018). 9. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 153 Key people Surveyed clusters (N=74)a Data sources ■ Malawi Tuberculosis Prevalence survey, technical report: Ministry of Health, National TB Control Programme; 2013– 2014. ■ Survey dataset. Summary statistics Participation rate 81% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 452 1.5 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.5 Finance Amount (US$) Ministry of Health, Malawi 1 023 244 The Global Fund 1 211 836 Total budget 2 235 080 Survey organization and financing Implementing agency: National TB Control Programme/Centre for Social Research, University of Malawi MALAWI a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Name Role Organization James Mpunga Principal investigator National TB Control Programme (NTP) Rhoda Banda Survey coordinator NTP Alister Munthali Co-principal investigator Centre for Social Research, University of Malawi Damson Kathyola Co-investigator Ministry of Health (MOH) Isaiah Dambe Co-investigator NTP Ishmael Nyasulu Co-investigator WHO Malawi Suzgo Mzumara Co-investigator (radiologist) MOH George B. Samuti Chief of laboratory Central Reference Laboratory, MOH Daniel Nyangulu Radiology coordinator MOH Charles Mandambwe Data manager NTP Masy Chiocha Data manager Centre for Social Research, University of Malawi Andrew Dimba Field team leader NTP Henry Kanyerere Field team leader NTP Lameck Mlauzi Field team leader NTP Sidon Konyani Technical assistance (epidemiologist) Centre for Social Research, University of Malawi Julia Ershova Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Irwin Law Technical assistance (survey advisor) WHO headquarters Patrick Moonan Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Peou Satha Technical assistance (radiology) CENAT, Cambodia Sian Floyd Technical assistance (analysis) London School of Hygiene & Tropical Medicine 2013–2014 National TB prevalence surveys 2007–2016154 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/semi-urban/rural Sampling unit Three major strata (urban, semi-urban, rural)/ward or area (urban), boma or town (semi-urban), traditional area (rural)/ enumeration area Sample size assumptions • Smear-positive prevalence 278 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.4 • k 0.5 • Response rate 80% • Sample size (estimated) 37 200 Number of clusters 74 Cluster size 500 Eligibility criteria • Age ≥15 years • Residency Slept in the household for at least 14 days before the census Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): concentrated preparationa, FM (LED, auramine stain), FM positives were re-confirmed by Xpert MTB/RIF. Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB Capilia TB drug susceptibility test Xpert MTB/RIF Xpert® MTB/RIF Any smear-positive specimens, and any specimens that were culture contaminated HIV test Not doneb Screening criteria Interviewa Any symptomsb ≥1 week Chest X-rayc Any lung abnormality Other N/A a An in-depth interview about health-care seeking behaviour was done only for those who screened positive. b Cough, sputum production, haemoptysis, chest pain, weight loss, night sweats, fatigue, fever, shortness of breath. c Conventional radiography. a Protocol violation, originally direct preparation. b Participants were interviewed about their HIV status. Field data collection Paper/electronic Database Microsoft® Access Method of analysis MI+IPW Results first published in a report/paper May 2016 Official dissemination event Pending Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 220 142–297 452 312–593 Male 303 176–431 546 335–757 Female 149 85–213 374 246–501 15–24 years 46 5.6–86 120 36–205 25–34 years 219 81–356 315 156–474 35–44 years 423 199–647 902 468–1 336 45–54 years 146 21–271 309 131–487 55–64 years 369 45–693 800 310–1 290 ≥65 years 645 261–1 028 1 564 888–2 241 Urban 555 281–830 1 014 486–1 542 Rural 169 96–242 373 239–506 Semi-urban 278 0–694 393 0–910 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 715 – Location of care sought • Consulted medical facility 1 280 47 Public facility 901 70 Private facility (including CHAMb) 379 30 • Pharmacy 32 1.2 • Traditional centre 41 1.5 • Other 4 0.1 Self-treated 236 8.7 No action taken 1 096 40 Unknown 26 1.0 a Any symptoms (cough, sputum production, haemoptysis, chest pain, weight loss, night sweats, fatigue, fever, shortness of breath) ≥1 week. b Christian Health Association of Malawi. a Data were available only for participants who were eligible for sputum submission. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b 358 participants were tested with Xpert MTB/RIF, and 9 were resistant to rifampicin. Survey participants currently on TB treatment Number % Total participants currently on TB treatmenta 12 – • Treated in the public sector 10 83 • Treated in the private sector (CHAM) 2 17 • Treated in unknown sector 0 0 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 4 3.0 Other sputum results Number % Total smear-positive participants 163 – Smear-positive participants without MTB confirmationa 101 62 Isolates with DR-TB (rifampicin) detectedb 9 4.7 a Age ≥15 years unless otherwise specified. Design effect k Smear-positive TB 2.1 1.1 Bacteriologically confirmed TB 3.2 1.1 MALAWI 155 Survey flow: census to final outcomes Field operations: June 2013 to May 2014 Individuals enumerated in census 68 220 Eligible for sputum examination 3 432 (11%) Total bacteriologically confirmed cases 132 Eligible study population 39 026 (57%) Total participants 31 579 (81%) Interview and chest X-ray 31 561 (99.9%) Interview only 18 (0.1%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 381 (11%) Symptom positive, chest X-ray negative or N/A 2 334 (68%) Symptom negative, chest X-ray positive 717b (21%) Other N/A Symptom positive, chest X-ray positive 25 (19%) Symptom positive, chest X-ray negative or N/A 67 (51%) Symptom negative, chest X-ray positive 40f (30%) Other N/A Submitted specimens At least one specimen 3 368 (98%) Both specimens 3 200 (93%) Smear-positive casesd 62 (47%) Definite 62 Probable N/A Smear-negative casese 70 (53%) Definite 70 Probable N/A Laboratory result At least one culture result availablec 3 327 (97%) Symptom screening Cough ≥2 weeks 1 192 (3.8%) Cough ≥1 week 2 047 (6.5%) Haemoptysis ≥1 week 69 (0.2%) Sputum production ≥1 week 1 101 (3.5%) Chest pain ≥1 week 1 039 (3.3%) Weight loss ≥1 week 370 (1.2%) Night sweats ≥1 week 415 (1.3%) Fatigue ≥1 week 496 (1.6%) Fever ≥1 week 401 (1.3%) Shortness of breath ≥1 week 431 (1.4%) Any symptoms (above) ≥1 weeka 2 715 (8.6%) Chest X-ray screening Normal 30 231 (96%) Abnormala 1 016 (3.2%) Other abnormality 312 (1.0%) Result not available 2 (<0.01%) Total chest X-rays taken 31 561 a Eligible for sputum collection. b Out of 717, 82 participants were defined as “chest X-ray abnormal but not suggestive of TB”, but were nonetheless requested to submit sputum samples. Teams were not consistent in their approach to sputum submission for participants with an abnormal chest X-ray (suggestive of TB or other abnormality). c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Smear-positive was defined as a specimen with ≥4 AFBs. Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. e Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. f Four out of 40 were “abnormal but not suggestive of TB” on chest X-ray. Ineligible individuals 29 194 (43%) Children <15 years 27 610 (41%) Did not meet residency criteria 1 584 (2.3%) National TB prevalence surveys 2007–2016156 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 200 300 400 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 200 400 600 800 1000 1200 1400 1600 1800 0 2 000 6 000 8 000 10 000 12 000 14 000 4 000 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB 2500 1500 2000 1000 500 0 15–24 25–34 35–44 45–54 55–64 ≥65 1.0 2.0 3.0 4.0 5.0 7.0 6.0 8.0 9.0 10.0 Prevalence per 10 0 0 0 0 population Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Pr ev al en ce : no ti fi ca ti on ra ti o a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 1.78, variance 5.82, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using smear-positive pulmonary TB notifications (2013) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 20 15 25 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Total (≥15) MALAWI 157 Background Malawi, in southern Africa, had a population of 16 million in 2013. The average gross national income (GNI) per person was US$ 390 per year, making it a low-income country (1). In 2013, the prevalence of HIV in the general population aged 15–49 years was 9.9% (95% confidence interval [CI]: 9.1–11%) (2), and it was estimated that 55% (95% CI: 49–62%) of tuberculosis (TB) patients were coinfected with HIV (3). The National Tuberculosis Control Programme (NTP) in Malawi began implementing what later became known as the DOTS strategy in 1984; it was one of the first model TB programmes in Africa. By 1999, DOTS had been expanded to all public health facilities, and facilities in the quasi-private sectors. TB remained one of the major public health concerns in Malawi throughout this period and worsened considerably following the emergence of the HIV epidemic in the late 1980s and 1990s. TB control was part of the Essential Health Package of the Malawi Government’s Health Sector Strategic Plan for 2011–2016 (4). WHO estimated that, in 2010, there were 219 (95% CI: 203–236) new cases of TB per 100 000 population per year, equivalent to a best estimate of 33 000 (95% CI: 31 000–35 000) new cases per year. Nonetheless, estimates of the burden of TB disease were uncertain because no national TB prevalence survey had ever been done, there were no direct measurements of TB mortality available from vital registration, and the gap between notifications and incidence (due to underreporting or underdiagnosis of cases) had not been quantified and was hard to estimate. Malawi was one of the 22 global focus countries for a national TB prevalence survey selected by the WHO Global Task Force on TB Impact Measurement in December 2007. To better understand the burden of TB, and with the new opportunity of funding from the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) and the national budget, the Ministry of Health decided in 2010 to implement a national TB prevalence survey. The survey started in June 2013 and was completed in May 2014. Key methods and results There were 74 survey clusters in three strata (urban, semi-urban and rural), with a target cluster size of 500 individuals. A total of 68  220 individuals from 16  380 households were enumerated in the survey census, of whom 39  026 (57%) were eligible and invited to parti- cipate. Of these, 31  579 (81%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (5). A total of 3432 participants (11%) were eligible for sputum examination; of these, 3368 (98%) submitted at least one sputum specimen and 3200 (93%) submitted two sputum specimens. A total of 132 bacteriologically confirmed pulmonary TB cases were identified, including 62 (47%) cases of smear-positive TB. The prevalence of smear-positive TB was 220 (95% CI: 142–297) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 452 (95% CI: 312–593) per 100 000 population. When extrapolated to all forms of TB and for all ages, prevalence was estimated as 362 (95% CI: 257–468) per 100  000 population. The prevalence per 100  000 population of both smear-positive and bacteriologically confirmed TB was higher in urban than in rural and semi-urban areas. Photo credit: Julia Ershova National TB prevalence surveys 2007–2016158 Other key results were: • the male to female ratio was 2.0 for smear- positive TB and 1.5 for bacteriologically confirmed TB; • prevalence per 100  000 population had two peaks, in those aged 35–44 years and the 65 years or over group; the absolute number of bacteriologically-confirmed cases was relatively high in the younger age groups (25–34 years and 35–44 years) and the elderly group (≥65 years); • among bacteriologically confirmed TB cases, 70% were symptom-screen positive, and among smear-positive cases, 66% were symptom-screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.5 overall, but varied from 1.1 in those aged 45–54 years to 9.0 in those aged 65 years or more, and was higher for men than women (2.9 versus 2.2); • among the bacteriologically confirmed TB cases, 90% had no previous history of anti- TB treatment and only 3.0% were on anti-TB treatment at the time of the survey; • of the 89 bacteriologically confirmed and 39 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 46 (52%) and 21 (54%), respectively, had previously sought care in a public or private health facility for their symptoms; and • All participants were asked whether they had ever been tested for HIV and, if willing, were asked to disclose their status; of the 31 579 participants, 19 703 (62%) disclosed their HIV status, and of those, 1840 (9.3%) reported being HIV-positive; and among 132 bacteriologically confirmed TB cases, 22 (17%) were HIV- positive, 52 (39%) were HIV-negative, and the status of the remaining 44% was unknown (all data were based on the verbal interview). Implications of results The prevalence of TB in Malawi was significantly higher than the pre-survey estimate of 140 (95% CI: 72–229) per 100 000 population (6). The fact that TB prevalence per 100 000 population increased with age suggested that the TB epidemic in Malawi had a downward trend. The elderly may also have more limited access to proper diagnosis and management. The survey had several major programmatic, policy and funding implications: • most TB cases in the community were HIV- negative, probably reflecting the effectiveness of TB and HIV interventions as well as a relatively poor detection rate of TB among HIV-negative people; • most undiagnosed TB patients with symptoms had not visited a medical service, indicating that TB diagnostic capacity was inadequate and needed to be strengthened; • the burden of TB was not evenly spread across the country: urban populations had a higher risk of acquiring and developing TB disease than did semi-urban and rural populations; active case finding strategies should be considered for these higher-risk populations; • TB case finding strategies better customized to men should be developed and implemented; and • microscopy contributed to only 47% of final TB diagnoses, suggesting that case detection and patient management would be improved by expanding the use of more sensitive and specific diagnostic tests. Photo credit: Ikushi Onozaki MALAWI 159 Major successes, challenges and lessons learned The major overarching success was that the first national TB prevalence survey in Malawi was successfully implemented, with a good participation rate. This was done using conventional tools (e.g. film-based portable chest X-ray equipment and paper-based data collection tools) as dictated by the relatively small budget provided by the Global Fund and the national government. Other successes included excellent collaboration between the NTP and the University of Malawi’s Centre for Social Research, and between the survey team, NTP and technical partners, including the United States Centers for Disease Control and Prevention (US-CDC), the London School of Hygiene and Tropical Medicine, and WHO, which strongly facilitated survey implementation. Given the challenges faced in some other countries, data management was effective, with on-site data entry in the field, timely data cleaning and validation, and continuous support from the US-CDC. The final validated data set was available within a few months of the completion of field operations. Challenges faced during the survey included those listed below. • It took two years to secure government funding to support field activities and more than a year to procure conventional X-ray equipment. During the survey, interruptions to disbursement of funds caused some delays in field operations. • A change of the lead technical adviser during the final stages of survey preparations meant that the survey team did not benefit from technical assistance during the pilot survey and the early stages of field operations. This contributed to some initial issues with data management, but these were subsequently rectified. • The suboptimal environment in which chest X-rays were often taken. X-ray units, and the chemical liquids used to develop and fix films, tended to overheat in hot conditions. Field operations were sometimes delayed while the units were allowed to cool down. In addition, individual identifiers were written on the films by hand after the images had been developed. This caused problems with later archiving and retrieval of images for central reading, and potentially caused some images to be mislabelled (i.e. labelled with the wrong participant's name). • Advice about sputum examination, which was not appropriate in the context of a prevalence survey, was provided to the central reference laboratory by an expert not directly involved in the survey. Although the intention was to conduct direct smear microscopy (to allow comparison with cases routinely detected by health services), in practice, centrifuged sediment was used for light-emitting diode (LED) fluorescent microscopy (FM). This was a protocol violation and resulted in many scanty smear-positive results. In 62% of the smear- positive specimens, Mycobacterium tuberculosis could not be detected by either culture or Xpert® MTB/RIF. In consultation with leading laboratory experts working with the Global Laboratory Initiative and the Supranational Reference Laboratory for Malawi, the survey re-categorized scanty 1–3 acid-fast bacilli (AFB) smears by concentrated LED FM as insignificant, and did not classify them as smear-positive. • There were incidents of laboratory cross- contamination. Of the specimens from 192 participants who were positive by culture or Xpert MTB/RIF (or both), one third were found to be clustered in the laboratory logbook; that is, consecutive specimens were positive for M.  tuberculosis. Following an extensive panel review of laboratory documents, chest X-rays and other information (e.g. data on family contacts), some laboratory cross-contamination was suspected. The panel concluded that a total of 60 participants with positive laboratory results should not be counted as TB cases. Of these 60, 29 had a very strong suspicion of cross- contamination and the remaining 31 had a single weak positive result (i.e. culture of fewer than five colonies) without other supportive evidence of TB disease other than symptoms. The final survey results may have underestimated TB prevalence. Photo credit: Julia Ershova National TB prevalence surveys 2007–2016160 • A substantial number of chest X-rays had to be read after field operations were completed. • There was a considerable delay in the writing of the survey report because no one was available to undertake this task. Important lessons for future surveys were: • all survey procedures must be closely monitored to prevent protocol violations, or to ensure that any violations are promptly corrected; • cross-contamination in the laboratory is a potential problem, and great care is needed to avoid cross-contamination compromising survey results; and • it is important to ensure that someone is available to prepare the survey report, and to include adequate funding for this activity in the budget. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. Malawi health sector strategic plan 2011–2016: Moving towards equity and equality. Ministry of Health, Government of Malawi; 2011 (http://www.healthpromotion.gov.mw/index. php/2013-08-12-12-52-31/2013-08-12-12-52-32/policies- strategies?download=6:malawi-health-sector-strategic- plan-2011-2016, accessed July 2017). 5. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 6. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf, accessed January 2018). 161 Name Role Organization Tugsdelger Sovd Principal investigator Ministry of Health Puntsag Banzragch Central panel team National Center for Communicable Diseases Naranbat Nyamadawa Survey consultant Mongolian Anti-Tuberculosis Coalition Naranzul Dambaa Survey coordinator National Center for Communicable Diseases Tsolmon Boldoo Data manager National Center for Communicable Diseases Bayasgalan Purev Central radiologist National Center for Communicable Diseases Buyankhishig Burneebaatar Laboratory doctor National Tuberculosis Reference Laboratory Oyuntuya Tumenbayar Laboratory doctor National Tuberculosis Reference Laboratory Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Yasunori Ichimura Technical assistance (survey advisor) Chiba University, Japan Norio Yamada Technical assistance (survey advisor) Research Institute of Tuberculosis, Japan Anti-Tuberculosis Association (RIT/JATA) M. Bintari Dwihardiani Technical assistance (survey advisor) WHO Indonesia M.N. Farid Technical assistance (survey advisor) Central Bureau of Statistics, Jakarta Satoshi Mitarai Technical assistance (laboratory advisor) RIT/JATA Soe Nyunt-U Technical/financial support WHO Mongolia Narantuya Jadambaa Technical/financial support WHO Mongolia Key people Surveyed clusters (N=98)a Data sources ■ Report of the first national tuberculosis prevalence survey in Mongolia (2014–2015). Ulaanbaatar city, Mongolia: Ministry of Health; 2016. ■ Survey dataset. Summary statistics Participation rate 84% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 560 2.8 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.5 Finance Amount (US$) Government of Mongolia 442 000 The Global Fund 617 000 WHO 34 700 Total budget 1 093 700 Survey organization and financing Implementing agency: National TB Programme, National Center for Communicable Diseases MONGOLIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2014–2015 National TB prevalence surveys 2007–2016162 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata City (Ulaanbaatar, Darkhan and Erdenet cities)/provincial center (except Darkhan and Orkhon provinces)/rural (all soums except provincial center soums) Sampling unit City: khoroo (sub-district) in UB city, bagh in Darkhan and Erdenet cities Provincial center: bagh (sub-soum) Rural: soum (sub-province) Sample size assumptions • Smear-positive prevalence 180 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.2 • k 0.5 • Response rate 85% • Sample size (estimated) 49 000 Number of clusters 98 Cluster size 600 (51 clusters in city strata); 500 (47 clusters in other strata) Eligibility criteria • Age ≥15 years • Residency Slept in the household for 14 days prior to census Survey design and methodology Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 204 143–265 560 455–665 Male 349 235–464 840 646–1 033 Female 68 38–99 299 225–372 15–24 years 135 42–228 555 362–748 25–34 years 281 152–410 634 431–837 35–44 years 208 94–323 472 289–655 45–54 years 197 77–318 527 344–711 55–64 years 170 63–277 562 372–752 ≥65 years 194 64–323 639 377–900 City 191 126–257 586 447–724 Provincial center 195 34–356 513 216–810 Rural 233 85–381 529 336–723 Laboratory methodology Smear Two samples (spot, morning): direct preparation, FM (LED, auramine stain). ZN for those smears that were FM positve Culture Two samples (spot, morning): direct preparation, Ogawa media Identification of MTB PNB, niacin test TB drug susceptibility test MTBDRplus testa Xpert® MTB/RIF Done for smear-positive specimens (from the early phase of field operations)b HIV test Not done Screening criteria Interview Cough ≥2 weeks Chest X-raya Any lung abnormality Other Chest X-ray exempted a Direct digital radiography by chest X-ray car and mobile apparatus. a Financial support was provided by Science and Technology Foundation Mongolia. b Xpert MTB/RIF was done for 84 out of 92 smear-positive specimens. Field data collection Paper Database Microsoft® Access Method of analysis MI+IPW Results first published in a report/paper December 2016 Official dissemination event March 2017 Analysis and reporting Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 546 – Location of care sought • Consulted medical facility 950 37 Public facility 920 97 Private facility 30 3.1 • Pharmacy 222 8.7 • Traditional medicine hospital 2 0.1 • Others 59 2.3 • Unspecified 104 4.1 No action taken 1 179 46 Unknown 30 1.2 a Cough ≥2 weeks. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b 234 culture-positive samples were tested with Genotype MTBDRplus. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 129 – • Treated in the public sector 126 98 • Treated in the private sector 0 0 • Treated in other sector 3 2.3 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 11 4.4 Design effect k Smear-positive TB 2.0 1.0 Bacteriologically confirmed TB 2.1 0.6 Other sputum results Number % Total smear-positive participants 92 – Smear-positive participants without MTB confirmationa 5 5.4 Isolates with MDR-TB detectedb 22 9.4 a Age ≥15 years unless otherwise specified. MONGOLIA 163 Survey flow: census to final outcomes Field operations: April 2014 to November 2015 (April to November 2014 for phase 1 (urban), April to November 2015 for phase 2 (rural)) Individuals enumerated in census 85 860 Eligible for sputum examination 10 359 (21%) Total bacteriologically confirmed cases 248 Eligible study population 60 031 (70%) Total participants 50 309 (84%) Interview and chest X-ray 49 406 (98%) Interview only 788 (1.6%) Chest X-ray only 115 (0.2%) Symptom positive, chest X-ray positive 817 (7.9%) Symptom positive, chest X-ray negative or N/A 1 729 (17%) Symptom negative, chest X-ray positive 7 064 (68%) Otherb 749 (7.2%) Symptom positive, chest X-ray positive 44 (18%) Symptom positive, chest X-ray negative or N/A 7 (2.8%) Symptom negative, chest X-ray positive 194 (78%) Otherb 3 (1.2%) Submitted specimens At least one specimen 9 546 (92%) Both specimens 9 473 (91%) Smear-positive casesd 88 (35%) Definite 87 Probable 1 Smear-negative casese 160 (65%) Definite 158 Probable 2 Laboratory result At least one culture result availablec 9 527 (92%) Symptom screening Cough ≥2 weeksa 2 546 (5.1%) Haemoptysis 777 (1.5%) Sputum production 6 481 (13%) Chest pain 6 451 (13%) Fever 1 280 (2.6%) Total symptom-screen positivea 2 546 (5.1%) Chest X-ray screening Normal 20 441 (41%) Abnormala 7 881 (16%) Other abnormality 21 199 (43%) Result not available N/A Total chest X-rays taken 49 521 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture and/or Xpert. Probable: MTB not confirmed by culture and/or Xpert but chest X-ray suggestive of TB. e Definite: MTB confirmed by culture. Probable: one scanty culture-positive without chest X-ray suggestive of TB but with chronic cough, and confirmed as TB cases by referral facilities. Ineligible individuals 25 829 (30%) Children <15 years 19 400 (23%) Did not meet residency criteria 6 429 (7.5%) National TB prevalence surveys 2007–2016164 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 80 90 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.5 2.0 3.0 3.5 4.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 250 500 750 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 0 Prevalence per 10 0 0 0 0 population 1 000 2 000 3 000 3 500 2 500 4 000 1 500 500 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1000 500 800 700 600 900 400 300 200 100 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.53, variance 5.49, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimate of using smear-positive pulmonary TB notifications (2014) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimated prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 20 25 15 0 0 1 2 3 4 5 6 7 8 9 10 Total (≥15) MONGOLIA 165 Background Mongolia is a landlocked country in East Asia that had a population of 2.9 million in 2014. The average gross national income (GNI) per person was US$  4260 per year, making it an upper-middle-income country (1). According to the Population and Housing Census of 2010, about 40% of the nation’s population lived in the capital city of Ulaanbaatar (2). In 2014, the prevalence of HIV in the general population aged 15–49 years was <0.1% (95% confidence interval [CI]: <0.1–<0.1%) (3), and it was estimated that 0.18% (95% CI: 0.17–0.20%) of TB patients were coinfected with HIV (4). The National Tuberculosis (TB) Programme (NTP) introduced the WHO-recommended DOTS strategy in 1994, and the country subsequently improved TB detection and treatment outcomes (5,6). The case notification rate (all types of TB) increased from 116 per 100 000 population in 1995 to 185 per 100 000 population in 2006, after which it decreased slowly. Treatment success was around 80–85% throughout the period 1999–2014. WHO estimated that the prevalence of TB was 254 (95% CI: 119–438) per 100 000 population in 2013. Although TB mortality declined from 3.2 per 100 000 population in 2000 to 1.9 per 100 000 population in 2013, TB remained the leading cause of death from communicable diseases in Mongolia (7). Between 1959 and 1961, and with the assistance of the Russian Federation, Mongolia undertook a large active TB screening programme that covered 88% of the total population. The survey estimated that 33% of the population had a positive tuberculin skin test result (8). No study of a similar magnitude had previously been conducted in the country. In 2011, in line with the Millennium Development Goals (MDGs) and the Regional strategy to Stop TB in the Western Pacific (9), as endorsed by WHO’s regional committee, the Government of Mongolia approved a 5-year national plan to stop and prevent TB. This plan included a national TB prevalence survey to measure the prevalence of bacteriologically confirmed pulmonary TB among those aged 15 years and more. Key methods and results Due to the scattered and sparse population in remote provinces and the cold winters, survey field operations were split into two phases: Phase 1 was mostly conducted in the capital city of Ulaanbaatar (2014); Phase 2 continued in the remote provinces (2015). Phase 1 was designed as an independent survey with a sample size large enough to provide TB prevalence estimates for the capital and urban areas, where most TB cases were notified (7). There were 98 survey clusters across three strata (city, provincial centre and rural), with a target cluster size of 600 individuals in cities and 500 individuals in other strata. A total of 85 860 individuals from 24 127 households were enumerated in the survey census, of whom 60 031 (70%) were eligible and invited to participate. Of these, 50  309 (84%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, chest X-ray and an interview about symptoms (10). A total of 10 359 people (21% of participants) were eligible for sputum examination; of these, 9546 (92%) submitted at least one sputum specimen and 9473 (91%) submitted two sputum specimens. A total of 248 bacteriologically confirmed pulmonary TB cases were identified, including 88 smear-positive TB cases. The prevalence of smear-positive TB was 204 (95% CI: 143–265) per 100 000 population among those aged ≥15 years, and for bacteriologically confirmed TB it was 560 (95% CI: 455–665) per 100  000 population. When extrapolated to all forms of TB and for all ages, prevalence was estimated as 757 (95% CI: 620–894) per 100 000 population. There was no significant variation in the prevalence of bacteriologically confirmed TB between the three strata, with the results being city, 586 (95% CI: 447–724) per 100  000 population; provincial centres, 513 (95% CI: 216–810) per 100 000 population; and rural sub-provinces, 529 (95% CI: 336–723) per 100 000 population. Other key results were: • the male to female ratio was 5.1 for smear- positive TB and 2.8 for bacteriologically confirmed TB; • prevalence per 100 000 population was high in all age groups; however, the absolute number of bacteriologically confirmed TB cases was relatively high in the young age groups (15–24 years and 25–34 years); National TB prevalence surveys 2007–2016166 • among bacteriologically confirmed TB cases, 21% were symptom-screen positive, and among the smear-positive cases, 34% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.5 overall, but varied from 1.4 in those aged 15–24 years to 3.5 in the 25–34 year age group, and was much higher for men than for women (3.8 versus 0.9); • among bacteriologically confirmed TB cases, 82% had no previous history of anti-TB treatment and only 4.4% were on anti-TB treatment at the time of the survey; and • of the 48 bacteriologically confirmed and 27 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 17 (35%) and 14 (52%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results Based on the first national TB prevalence survey, Mongolia was confirmed as a high TB burden country in the WHO Western Pacific Region, with considerable ongoing transmission in the community. The estimated national prevalence per 100 000 population was high, including among the younger age groups. These results suggested that TB should be reconsidered as a significant public health problem in Mongolia. The prevalence of bacteriologically confirmed prevalence was uniformly high across all strata. High prevalence with high notification rates in congested urban areas suggested a higher overall burden of TB in these places, especially in the sprawling residential areas with little infrastructure (known as the ger districts). The seasonal pattern of TB also indicated higher rates of transmission in the winter months, a time of year with higher air pollution in the ger districts. Increasing urbanization in the five years prior to the survey led to the expansion of static ger areas around the centre of Ulaanbaatar and a more densely populated environment with increased air pollution; the former may have increased TB transmission, and the latter may have contributed to delays in diagnosis because of the increased frequency of common coughs and reduced likelihood of suspecting TB as the cause. The high prevalence in provincial centres and rural (sub- provinces) areas indicated challenges related to access to health facilities and diagnostic services. The gap between prevalence and notification showed the limitations of existing approaches to case-finding, which relied upon symptom screening and smear microscopy. In the survey, smear-positive cases accounted for only 36% of bacteriologically confirmed cases; three-quarters of cases were symptom-screen negative, and were tested due to screening by chest X-ray (most smear-negative, culture-positive cases had small and atypical shadows in chest X-ray images). These findings suggested that access to high-quality chest X-rays should be improved, that Photo credit: M. Bintari Dwihardiani MONGOLIA 167 new diagnostic tools beyond smear such as Xpert® MTB/ RIF should be introduced, and that diagnostic services should be decentralized across the country. Because underreporting of detected cases to national authorities probably also contributed to the gap between prevalence and notifications, another identified priority was to strengthen the electronic reporting system with appropriate supervision. While strengthening TB control efforts in general, the importance of giving particular attention to risk groups with a high TB prevalence and to remote areas with poorer access was also recognized, and reflected in Mongolia’s 5-year national TB strategic plan for 2016–2020. Major successes, challenges and lessons learned Major successes included carrying out the first nationwide TB prevalence survey in Mongolia, and the first TB- related survey in the country for more than 50 years; achieving high population coverage (100%), with a participation rate of 84% and a sputum collection rate of more than 90%; reaching clusters located in remote areas with limited infrastructure; and examining all specimens to a high standard in one national reference laboratory. Specifically, specimens from remote clusters were transported using a nationwide sputum transportation system established in 2008; the overall culture (Ogawa) contamination rate was low (1.9%; 696/37  322 tubes); and all laboratory results were available, with an overall recovery rate of 87% (80 culture Mycobacterium tuberculosis [MTB]-positive among 92 smear-positive). Survey successes were facilitated by excellent leadership from the NTP; good collaboration between the Ministry of Health, the survey team and local authorities and health centres during field operations; the appointment of a full- time survey coordinator and data management team early in the process; and close collaboration with external partners including the Global Fund to Fight AIDS, TB and Malaria, the WHO country office and WHO headquarters. Good technical assistance throughout survey preparations and implementation helped to ensure the high quality of the survey, especially given that Mongolia had no previous experience of undertaking a survey of this magnitude. Experts in prevalence surveys visited more than 10 times during the course of the survey, and provided regular assistance throughout, from protocol development to reporting of final results. Good financial planning (especially with financial contributions from the government) was also vital in ensuring the smooth progress of the survey, including the ongoing maintenance of chest X-ray machines during field operations. Major challenges included interruptions to field operations during the long winter season; a lower participation rate among the young, men and urban clusters, especially in the wealthier parts of large cities; and postponement of field operations following a breakdown of both X-ray machines, since no backup machines were available. Photo credit: M. Bintari Dwihardiani Photo credit: M. Bintari Dwihardiani National TB prevalence surveys 2007–2016168 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. 2010 population and housing census of Mongolia. Census monograph, Ulaanbaatar: National Statistical Office of Mongolia; 2011. 3. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 4. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 5. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 6. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 7. Report of the first National Tuberculosis Prevalence Survey in Mongolia (2014–2015). Ulaanbaatar city, Mongolia: Ministry of Health; 2016. 8. Results of screening for TB in 1959–1961 in Mongolia. Mongolian Journal of Infectious Diseases. 2012;4(47). 9. WHO Regional Office for the Western Pacific. The regional strategic plan to stop TB in the Western Pacific. Manila, Philippines: WHO; 2000 (https://iris.wpro.who.int/bitstream/ handle/10665.1/9849/Regional_Strategic_Plan_to_Stop_TB_ WP_eng.pdf, accessed July 2017). 10. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 169 Name Role Organization Win Maung Vice-chair, lead SC, TC and CPD Director of Disease Control Thandar Lwin Survey coordinator, lead WC National Tuberculosis Programme (NTP) Tin Mi Mi Khaing SC and TC member Regional TB officer, Yangon Bo Myint SC and TC member Regional TB officer, Mandalay Tin Tin Mar TC and CPD member National TB Reference Laboratory (NTRL) Ti Ti TC member and laboratory advisor FIND Wint Wint Nyunt Lead laboratory unit NTRL San San Shein TC member, lead radiology unit Regional TB Centre, Mandalay Moe Zaw TC member, data manager NTP Hnin Wai Lwin Myo TC member, data management, WC NTP Si Thu Aung TC member, field team leader NTP Htay Lwin Field team leader State TB officer, Shan East Htar Htar Oo Field team leader NTP Thandar Thwin TC member, field team leader Regional TB Centre, Yangon Myo Zaw SC and TC member, monitoring & supervision WHO Myanmar Ikushi Onozaki SC, TC, CPD and WC member WHO headquarters Norio Yamada Technical assistance (epidemiology, analysis and WC) Research Institute of Tuberculosis, Japan Anti-Tuberculosis Association (RIT/JATA) Kosuke Okada SC and TC member, technical assistance (management) Japan International Cooperation Agency (JICA) Eva Nathanson Coordination (supply and logistics), WC member WHO Myanmar Key people Surveyed clusters (N=70)a Data sources ■ Report on national TB prevalence survey, 2009–2010, Myanmar. Ministry of Health, Department of Health, Government of Myanmar. ■ Survey dataset. Summary statistics Participation rate 89% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 613 2.5 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.1 Finance Amount (US$) WHO 15 000 Three diseases fund 270 000 JICA 114 000 Population Services International (PSI) 358 000 USAID 120 000 Total budget 877 000 Survey organization and financing Implementing agency: National Tuberculosis Programme MYANMAR CPD: central panel for diagnosis, SC: steering committee, TC: technical committee, WC: writing committee. a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2009–2010 National TB prevalence surveys 2007–2016170 Sampling design Sampling frame 293 out of 325 townshipsa Sampling design Multistage cluster sampling using PPS Strata Region/state Sampling unit Region, state/township/ward/village Sample size assumptions • Smear-positive prevalence 278 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.3 • k 0.4 • Response rate 90% • Sample size (estimated) 49 690 Number of clusters 70b Cluster size 710 Eligibility criteria • Age ≥15 years • Residency Individuals who lived in the household for ≥2 weeks at the time of the census Survey design and methodology a 32 townships were excluded from the sampling frame, mostly due to security issues. b Five clusters (Bokepyin, Kyarinnseikkyi, Nattalin, Mindat, Kunlon) were replaced by others within the same township during the pre-visit, due to security/transportation problems and a population aged 15 years and above that was too small. Laboratory methodology Smear Two samples (spot, morning): direct preparation FM (auramine stain), ZN for those smears that were FM positive Culture Two samples (spot, morning): direct preparation, Ogawa media Identification of MTB PNB, niacin, capilia TB drug susceptibility test Not done Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interview Cough ≥3 weeks and/or haemoptysis Chest X-raya Any lung abnormality Other Chest X-ray exempted a Conventional radiography (chest X-ray van or portable chest X-ray machine). Field data collection Paper Database Epi Info Method of analysis Classic survey analysis, logit model Results first published in a report/paper November 2011 Official dissemination event December 2010 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 242 186–315 613 502–748 Male 398 301–525 931 743–1 166 Female 122 77–194 367 288–469 15–24 years 43 18–103 95 48–187 25–34 years 190 131–274 469 339–648 35–44 years 350 231–530 739 579–944 45–54 years 304 189–489 811 591–1 111 55–64 years 373 248–560 858 619–1 189 ≥65 years 395 225–691 1 438 1 135–1 819 Region 192 137–267 523 421–649 State 369 236–578 838 560–1 252 Urban 331 216–506 903 662–1 232 Rural 216 154–304 527 410–677 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 1 691 – Location of care sought • Consulted medical facility 363 22 Public facility 197 54 Private facility (general practitioner, specialist) 166 46 • Pharmacy 271 16 • Traditional healer 243 14 Self-treated 307 18 No action taken 440 26 Other 39 2.3 Unknown 28 1.7 a Cough ≥3 weeks and/or haemoptysis. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). Survey participants currently on TB treatment Number % Total participants currently on TB treatment 79 – • Treated in the public sector 63 80 • Treated in the private sector (incl. general practitioner) 14 18 • Treated in unknown sector 2 2.5 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 13 4.2 Other sputum results Number % Total smear-positive participants 132 – Smear-positive participants without MTB confirmationa 16 12 Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. Design effect k Smear-positive TB 2.2 0.8 Bacteriologically confirmed TB 3.2 0.7 MYANMAR 171 Survey flow: census to final outcomes Field operations: June 2009 to April 2010 Individuals enumerated in census 93 806 Eligible for sputum examination 12 235 (24%) Total bacteriologically confirmed cases 311 Eligible study population 57 607 (61%) Total participants 51 367 (89%) Interview and chest X-ray 50 241 (98%) Interview only 1 126 (2.2%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 1 258 (10%) Symptom positive, chest X-ray negative or N/A 433 (3.5%) Symptom negative, chest X-ray positive 9 364 (77%) Otherb 1 180 (10%) Symptom positive, chest X-ray positive 65 (21%) Symptom positive, chest X-ray negative or N/A 1 (0.3%) Symptom negative, chest X-ray positive 231 (74%) Otherf 14 (4.5%) Submitted specimens At least one specimen 12 144 (99%) Both specimens N/A Smear-positive casesd 123 (40%) Definite 116 Probable 7 Smear-negative casese 188 (60%) Definite 188 Probable N/A Laboratory result At least one culture result availablec 12 027 (99%) Symptom screening Cough ≥3 weeksa 1 433 (2.8%) Haemoptysisa 285 (0.6%) Sputum production 9 953 (19%) Chest pain 6 827 (13%) Fever 3 122 (6.1%) Total symptom-screen positivea 1 691 (3.3%) Chest X-ray screening Normal 39 604 (79%) Abnormala 10 622 (21%) Other abnormality N/A Result not availablea 15 (0.03%) Total chest X-rays taken 50 241 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative (1096), corrective action (rechecked results of the interview and chest X-ray) (70), chest X-ray uninterpretable and symptom-screen negative (14). c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but two smear-positive, or one smear-positive with chest X-ray consistent with TB. e Definite: MTB confirmed by culture with at least one of the following conditions met: culture-positive (≥1 colony) in both two specimens, culture-positive (1–4 colonies) in one specimen and chest X-ray consistent with TB, or culture-positive (≥5 colonies) in one specimen. Probable: no definition. f Chest X-ray exempted and symptom-screen negative (12), symptom-screen negative and field chest X-ray negative (i.e. central chest X-ray healed TB) (1), symptom-screen negative and chest X-ray negative (field and central) (1) (the last two cases were from the corrective action). Ineligible individuals 36 199 (39%) Children <15 years 27 399 (29%) Did not meet residency criteria 8 800 (9.4%) National TB prevalence surveys 2007–2016172 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 400 600 800 1000 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 200 400 600 800 1000 1200 1400 1600 0 Prevalence per 10 0 0 0 0 population 10 000 30 000 40 000 50 000 60 000 20 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Fig. 2: TB prevalence per 100 000 population by age Age group (years) 2000 1000 1400 1200 1800 1600 800 600 400 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Bacteriologically confirmed TB Smear-positive TB a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 4.44, variance 13.61, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 6 2 4 10 12 8 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Total (≥15) MYANMAR 173 Background Myanmar is a country in South-East Asia that had a population of 51 million in 2009. It had an average gross national income (GNI) of US$ 630 per person per year, making it a low-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2009, the prevalence of HIV in the general population aged 15–49 years was 0.9% (95% confidence interval [CI]: 0.7–1.1%) (2), and it was estimated that 11% (95% CI: 6.4–16%) of TB patients were coinfected with HIV (3). The National TB Programme (NTP) introduced the WHO-recommended DOTS strategy in 1997 (4,5). As DOTS expanded, the case notification rate increased, from 67 (new and relapse cases) per 100 000 population in 2000 to 223 per 100 000 population in 2005. For smear- positive pulmonary TB specifically, the case notification rate increased from 38 per 100 000 population in 1999 to 76 per 100 000 population in 2005 (6). Myanmar carried out two national TB prevalence surveys before the introduction of DOTS: one in 1972 and one in 1994. The 1972 survey used chest X-ray (miniature photofluorography) and symptoms (cough, chest pain and haemoptysis) for screening and smear for diagnosis; culture testing was used in a limited number of clusters. The estimated prevalence of smear-positive pulmonary TB was 145 per 100  000 population among those aged 15 years or more. A tuberculin survey conducted at the same time suggested an annual risk of TB infection of 1.2%. Screening in the 1994 survey was based solely on symptoms, and diagnostic confirmation was limited to smear microscopy; chest X-ray and culture were conducted for a limited population but were not officially part of the protocol. The estimated prevalence of smear- positive pulmonary TB was 104 (95% CI: 72–137) per 100 000 population in participants aged 10 years or more. In the context of continuing increases in case notifications throughout the late 1990s and 2000s, the NTP initiated plans in 2005 for a third national TB prevalence survey, this time with chest X-ray screening and diagnosis based on culture as well as smear, with technical assistance from the Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) and WHO. However, the sudden termination of a grant from the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) in 2006 meant that survey operations were completed only in Yangon division (a pilot survey was also completed in Mandalay). A high prevalence of smear-positive and bacteriologically confirmed TB was confirmed in both the urban and rural parts of Yangon. The prevalence of smear-positive pulmonary TB was 279 (95% CI: 204–381) per 100 000 population among those aged 10 years or more, and the prevalence of bacteriologically confirmed TB was 534 (95% CI: 431– 661) per 100 000 population in the same age group (7). Based on these results, the NTP advocated further for a national prevalence survey. Myanmar was also one of the 22 global focus countries for national TB prevalence surveys selected by the WHO Global Task Force on TB Impact Measurement in December 2007. Planning restarted in 2008, and the survey was launched in June 2009 in close collaboration with four major partners: Three Diseases Fund, Japanese International Cooperation Agency, United States Agency for International Development, Population Services International and WHO. The survey was completed in April 2010 (8). Key methods and results There were 70 survey clusters in two strata (region and state, the latter having populations dominated by ethnic minorities), with a target cluster size of 710 individuals. A total of 93 806 individuals from 18 596 households were enumerated in the survey census, of whom 57 607 (61%) were eligible and invited to participate. Of these, 51 367 (89%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, chest X-ray and an interview about symptoms (9). A total of 12  235 participants (24%) were eligible for sputum examination, of whom 12 144 (99%) submitted at least one sputum specimen. A total of 311 bacteriologically confirmed pulmonary TB cases were identified, including 123 cases of smear- positive TB. The prevalence of smear-positive TB was 242 (95% CI: 186–315) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 613 (95% CI: 502–748) per 100 000 population. When extrapolated to all forms of TB and to all ages, prevalence was 544 (95% CI: 420–685) per 100 000 population. The prevalence of bacteriologically confirmed TB was much higher in the states (838 per 100 000 population; 95% CI: 560–1252) than in the regions (523 per 100 000 National TB prevalence surveys 2007–2016174 population; 95% CI: 421–649). The prevalence of bacteriologically confirmed TB was higher in urban areas (903 per 100  000 population; 95% CI: 662–1232) than in rural areas (527 per 100 000 population; 95% CI: 410– 677). Other key results were: • the male to female ratio was 3.3 for smear- positive TB and 2.5 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age; however, the absolute number of bacteriologically confirmed TB cases was relatively high in the young to middle age groups (25–54 years); • of the bacteriologically confirmed TB survey cases, 21% were symptom-screen positive, and among smear-positive cases, 34% were symptom-screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.1 overall, but varied from 0.7 in those aged 15–24 years to 3.0 in those aged 65 years or more, and was higher for men than for women (2.5 versus 1.6); • among bacteriologically confirmed TB survey cases, 86% had no previous history of anti- TB treatment and only 4.2% were on anti-TB treatment at the time of the survey; and • of the 60 bacteriologically confirmed and 37 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 21 (35%) and 14 (38%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results The 2009–2010 national TB prevalence survey revealed a high prevalence of TB in Myanmar despite efforts since the late 1990s to expand DOTS throughout the country. Although the estimated prevalence of smear-positive TB among all age groups (171 per 100 000 population; 95% CI: 131–223) was higher than the prevalence indicated in the 1994 survey, this did not mean that the burden of TB increased between 1994 and 2009 because the results were not directly comparable. The 1994 survey relied only on symptom screening and smear microscopy whereas the 2009–2010 survey used both chest X-ray and symptoms as screening tools. Prevalence results compared using the same screening criteria showed a 35% reduction in the prevalence of smear-positive pulmonary TB from 1994 to 2009–2010, suggesting that TB control efforts had a major impact. In the 2009–2010 survey, the difference between the total number of smear-positive pulmonary cases and the number of those with classic TB symptoms (i.e. a chronic cough), and between the total number of bacteriologically confirmed cases and the number of smear-positive cases, suggested that the case detection strategies used at the time of the survey had serious limitations and that a comprehensive review of approaches to case finding was warranted. For example, TB could be considered as part of the differential diagnosis of anyone with undiagnosed chronic symptoms, regardless of the presence of cough or any respiratory illness. The expansion of diagnostic tests including chest X-ray and culture was included in the national strategic plan for TB control 2011–2015. The finding that the prevalence of TB was higher in the states than in the regions suggested that specific efforts were needed to improve access to basic diagnostic services in the states, especially in the most remote areas. Among the symptomatic TB cases, 24% (16/66) chose to initially seek care in pharmacies or from a traditional healer. This suggested that incorporating these providers into formal TB control and care networks could help to detect cases earlier. The survey showed that chest X-ray was a more sensitive tool for TB detection than symptom screening. Therefore, anyone with an undiagnosed chest X-ray abnormality should be considered as a presumptive TB case and eligible for sputum examination. The diagnostic challenge was Photo credit: Kosuke Okada MYANMAR 175 further illustrated by the large share of smear-negative cases among all detected TB cases. Expanded use of Xpert® MTB/RIF was one of the strategies identified to address this challenge (major expansion of culture was not considered feasible, given the complexity of culture methods and the requirement for strict infection control measures). Survey results showed that specific measures were needed in congested urban areas where prevalence rates were highest. Examples that were identified included intensified collaboration with the private sector, since this provided services at convenient hours for those living in urban areas. Major successes, challenges and lessons learned The survey was successfully implemented with a high participation rate and with a comparatively low survey- specific budget (US$ 1 million, excluding the costs of NTP staff who worked on the survey). Even after the withdrawal of the financing initially committed by the Global Fund, resources were mobilized from other sources following intensive efforts by the NTP and the WHO Country Office in particular. Provisional results were shared with key officials and partners within 4 months of completing field operations. Final results, including updated estimates of TB disease burden (incidence, prevalence and mortality), were fully disseminated to national and international partners in December 2010. These estimates informed the subsequent revision of the national strategic plan for TB, and contributed to the mobilization of additional funding for TB care and treatment in Myanmar. Challenges included the need to exclude 32 of 325 townships from the sampling frame due to security issues; the fact that residency criteria for survey eligibility meant that 9.4% of those otherwise eligible were not included in the survey (mostly the mobile population, including seasonal workers); and low participation rates in a few areas, notably a few urban clusters and remote areas. These challenges affected the coverage and representativeness of the survey. Delays in procuring chest X-ray equipment delayed the start of survey operations, and some equipment then failed during field operations. The sputum cups that were used were not optimal for the purposes of culture testing and may have caused some laboratory cross-contamination. Unfortunately, no staff were available to write a paper to summarize the key results and lessons learned from the survey in a peer- reviewed journal. Survey results were analysed before guidance on analytical methods in the WHO handbook was finalized (9). The results from analyses that were restricted to survey participants (not taking into account those eligible but not participating in the survey) were used as the official survey results. Although the survey was carried out rigorously and had a high participation rate with few missing data, later analysis (that included more extended imputation for missing data) estimated TB prevalence to be about 10% higher than results in the official survey report. Important lessons learned included the value of the 2006 survey in Yangon for providing experience and expertise that were invaluable to the later national survey. Strong technical assistance from RIT/JATA and close collaboration with the WHO Country Office were also considered major contributions to the success of the survey. Photo credit: Ikushi Onozaki National TB prevalence surveys 2007–2016176 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. World Health Organization. Global tuberculosis report 2009. Geneva: WHO; 2009 (http://apps.who.int/iris/ bitstream/10665/44035/1/9789241563802_eng.pdf ?ua=1, accessed January 2018). 7. Annual report 2006. National Tuberculosis Programme, Myanmar; 2007. 8. Report on national TB prevalence survey, 2009–2010, Myanmar. Ministry of Health, Department of Health, Government of Myanmar;(https://www.who.int/tb/advisory_bodies/impact_ measurement_taskforce/prevalencesurveymyanmar_2009- 10report.pdf, accessed July 2017). 9. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 177 Name Role Organization Joshua Obasanya Principal investigator National TB and Leprosy Control Programme (NTBLCP) Emmanuel Idigbe Chairman technical committee Nigeria Institute of Medical Research, Lagos Chukwueme Nkemdilim Deputy survey coordinator National TB and Leprosy Control Programme Osahon Ogbweiwe Survey coordinator Consultant, Nigeria Philip Patrobas In-country technical advisor WHO Nigeria Awe Ayodele TB advisor to NTBLCP WHO Nigeria Abiola Tubi Laboratory manager Nigeria Centre for Disease Control (CDC-GAP) Babalola Akin Radiology coordinator Gwagwalada specialist hospital Gideon Zaphania Central data manager National TB and Leprosy Control Programme Samuel Ogiri Field team leader WHO-National professional officer North-Central zone Haruna Adamu Field team leader WHO-National professional officer North-East zone Moses Onoh Field team leader Medical advisor, The Leprosy Mission Nigeria Osakwe Puis Chijioke Field team leader WHO-National professional officer South-East zone Daniel Olusoji James Field team leader WHO-National professional officer South-West zone Jose Michael Madu Field team leader WHO-National professional officer South-South zone Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Eugene McCray Technical assistance (survey advisor) US Centers for Disease Control and Prevention (CDC) Charalampos Sismanidis Technical assistance (analysis) WHO headquarters Julia Ershova Technical assistance (data management) US Centers for Disease Control and Prevention (CDC) Daniella Cirillo Technical assistance (laboratory advisor) Supranational Reference Laboratory Milan Narayan Pendse Technical assistance (radiology) WHO headquarters (STOP-TB) Key people Surveyed clusters (N=70)a Data sources ■ Report first national TB prevalence survey 2012, Nigeria, Department of Public Health, Federal Republic of Nigeria. ■ Survey dataset. Summary statistics Participation rate 57% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 524 2.1 Prevalence:notification ratio (smear-positive TB, ≥15 years) 5.8 Finance Amount (US$) Ministry of Health, Nigeria 1 226 871 The Global Fund 1 465 283 WHO 375 650 Total budget 3 067 804 Survey organization and financing Implementing agency: National TB and Leprosy Control Programme NIGERIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2012 National TB prevalence surveys 2007–2016178 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Six geographical zones (north central, north east, north west, south east, south south, south west). Final analysis accounted for urban and rural areas. Sampling unit Six geographical zones/local government area/enumeration area Sample size assumptions • Smear-positive prevalence 346 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.5 • k 0.5 • Response rate 85% • Sample size (estimated) 49 000 Number of clusters 70a Cluster size 700 Eligibility criteria • Age ≥15 years • Residency Slept in the household ≥14 days prior to the census Survey design and methodology a Three clusters in the states of Borno and Yobe were excluded during field operations, due to security challenges. They were replaced with clusters in the states of Gombe, Bauchi and Adamawa. Laboratory methodology Smear Two samples (spot, morning): direct preparation, ZN Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB MPT 64 rapid test TB drug susceptibility test Not done Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interview Cough ≥2 weeks Chest X-raya Any lung abnormality Other N/A a Portable mobile X-ray unit (Min X-ray), computed radiography. Field data collection Paper/electronic Database Microsoft® Access Method of analysis MI+IPW Results first published in a report/paper November 2014 Official dissemination event November 2014 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 318 225–412 524 378–670 Male 484 333–635 751 538–965 Female 198 108–289 359 213–505 15–24 years 193 84–302 274 130–419 25–34 years 291 165–418 496 312–680 35–44 years 367 141–593 613 316–911 45–54 years 494 265–722 750 420–1 079 55–64 years 331 122–540 599 262–936 ≥65 years 332 106–559 660 318–1 003 Urban 413 269–556 663 441–884 Rural 182 111–254 323 191–456 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 466 – Location of care sought • Consulted medical facility 800 32 Public facility 628 79 Private facility 172 21 • Pharmacy 319 13 • Traditional centre 11 0.4 • Other 9 0.4 • Unspecified 3 0.1 Self-treated 680 28 No action taken 604 24 Unknown 40 1.6 a Cough ≥2 weeks. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). Survey participants currently on TB treatment Number % Total participants currently on TB treatment 82 – • Treated in the public sector 56 68 • Treated in the private sector 14 17 • Treated in other sector 5 6.0 • Treated in unknown sector 7 9.0 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 12 8.0 Other sputum results Number % Total smear-positive participants 184 – Smear-positive participants without MTB confirmationa 109 59 Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. Design effect k Smear-positive TB 2.6 0.9 Bacteriologically confirmed TB 2.6 0.7 NIGERIA 179 Survey flow: census to final outcomes Field operations: February to November 2012 Individuals enumerated in census 113 247 Eligible for sputum examination 4 688 (11%) Total bacteriologically confirmed cases 144 Eligible study population 77 797 (69%) Total participants 44 186 (57%) Interview and chest X-ray 43 199 (98%) Interview only 987 (2.2%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 746 (16%) Symptom positive, chest X-ray negative or N/A 1 720 (37%) Symptom negative, chest X-ray positive 2 222 (47%) Other N/A Symptom positive, chest X-ray positive 76 (53%) Symptom positive, chest X-ray negative or N/A 16 (11%) Symptom negative, chest X-ray positive 52 (36%) Other N/A Submitted specimens At least one specimen 4 558 (97%) Both specimens 4 133 (88%) Smear-positive casesc 107 (74%) Definite 75 Probable 32 Smear-negative casesd 37 (26%) Definite 35 Probable 2 Laboratory result At least one culture result available b 2 662 (57%) Symptom screening Cough ≥2 weeksa 2 466 (5.6%) Haemoptysis 288 (0.7%) Sputum production 3 551 (8.0%) Chest pain 6 813 (15%) Fever 8 493 (19%) Total symptom-screen positivea 2 466 (5.6%) Chest X-ray screening Normal 40 229 (93%) Abnormala 2 968 (6.9%) Other abnormality N/A Result not available 2 (<0.01%) Total chest X-rays taken 43 199 a Eligible for sputum collection. b Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. c Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture, but chest X-ray suggestive of TB. d Definite: MTB confirmed by culture. Probable: culture-positive without identification, and chest X-ray suggestive of TB. Ineligible individuals 35 450 (31%) Children <15 years 34 947 (31%) Did not meet residency criteria 503 (0.4%) National TB prevalence surveys 2007–2016180 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 200 300 400 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 800 0 Prevalence per 10 0 0 0 0 population 20 000 60 000 80 000 100 000 120 000 140 000 40 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 60 50 40 90 80 70 30 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB 1200 600 800 1000 400 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 1.0 2.0 3.0 4.0 6.0 5.0 7.0 8.0 9.0 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.06, variance 5.42, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 15 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Total (≥15) NIGERIA 181 Background Nigeria’s population was 168 million in 2012, making it the most populous country in Africa (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2012, Nigeria was a lower-middle income country with an average gross national income (GNI) per person of US$  2470 per year (1). At that time the prevalence of HIV in the general population aged 15–49 years was 3.4% (95% confidence interval [CI]: 3.1–3.6%) (2), and it was estimated that 24% (95% CI: 20–28%) of TB patients were coinfected with HIV (3). The National TB and Leprosy Control Programme (NTBLCP) was established in 1991 under Nigeria’s Ministry of Health (MoH). By the end of 2009, the number of DOTS centres represented 56% of the targeted number of 6261, and 1025 facilities contained laboratories with microscopes, equivalent to one centre for every 149 000 people. By 2012, DOTS was being implemented in areas that accounted for 85% of the country’s population. During DOTS expansion, case notifications consistently increased; however, they then plateaued between 2008 and 2012, despite an intensification of efforts to engage with public and private health-care providers outside the NTBLCP network (in 2012, these providers contributed 24% of case notifications). Based on WHO estimates, in 2012 Nigeria ranked fourth in Africa and 11th globally in terms of estimated incident cases per year. Nonetheless, there was considerable uncertainty about estimates of the burden of TB disease, given that no national TB prevalence survey had previously been carried out, that there were no direct measurements of TB mortality available from vital registration, and that the gap between notifications and incidence (due to underreporting or underdiagnosis of cases) had not been quantified and was hard to estimate. In December 2007, Nigeria was selected by the WHO Global Task Force on TB Impact Measurement as one of 22 global focus countries for a national TB prevalence survey, with the aim of better understanding the burden of TB disease at national and global levels. In 2008, the MoH decided to implement a national TB prevalence survey. The survey started in February 2012 and was completed in November 2012 (4). Key methods and results There were 70 survey clusters, with a target cluster size of 700 individuals (there were no strata, but urban and rural zones as well as six geographical zones were accounted for at the time of data analysis). A total of 113 247 individuals from 20  708 households were enumerated in the survey census, of whom 77 797 (69%) were eligible and invited to participate. Of these, 44 186 (57%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray (computed radiography with the imaging plate) and an interview about symptoms (5). A total of 4688 participants (11%) were eligible for sputum examination, of whom 4558 (97%) submitted at least one sputum specimen and 4133 (88%) submitted two sputum specimens. A total of 144 bacteriologically confirmed pulmonary TB cases were identified, including 107 cases of smear- positive TB. The prevalence of smear-positive TB was 318 (95% CI: 225–412) per 100 000 population (among those aged ≥15 years) and for bacteriologically confirmed TB it was 524 (95% CI: 378–670) per 100 000 population. Prevalence was significantly higher in urban than in rural areas. Other key results were: • the male to female ratio was 2.4 for smear- positive TB and 2.1 for bacteriologically confirmed TB; Photo credit: Philip Patrobas National TB prevalence surveys 2007–2016182 • prevalence per 100  000 population increased with age (with a peak among those aged 45–54 years); however, the absolute number of bacteriologically confirmed TB cases was relatively high in young age groups; • among bacteriologically confirmed TB cases, 64% were symptom-screen positive, and among the smear-positive cases, 75% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio)  was 5.8 overall, but varied from 4.1 in those aged 25–34 years to 7.7 among those aged 45–54 years, and was much higher for men then for women (7.2 versus 4.6); • among bacteriologically confirmed TB cases, 85% had no previous history of anti- TB treatment and only 8% were on anti-TB treatment at the time of the survey; and • of the 80 bacteriologically confirmed and 68 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 43 (54%) and 36 (53%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results The survey clearly demonstrated a high burden of TB disease in Nigeria, with an estimated prevalence of 323 (95% CI: 239–406) per 100 000 population (all forms of TB, all ages). The findings highlighted TB as a major public health problem that was much worse than previously thought, with a prevalence of 171 (95% CI: 44–382) per 100 000 population (6). The age distribution of cases and the high proportion of symptomatic cases in the community also demonstrated considerable ongoing transmission. After adjustments to include children and extrapulmonary TB, estimates of TB disease burden published by WHO were revised substantially upwards: estimates of TB prevalence were doubled, those for TB incidence trebled and those for TB mortality were increased fivefold compared with the previously estimated levels. The best estimate of the case detection rate (notifications of new and relapse cases divided by estimated incidence) was revised downwards. The survey had major programmatic, policy and funding implications. These included: • a need to substantially improve access to basic DOTS services to diagnose and treat people with TB; this was particularly evident from the high ratio of prevalent to notified TB cases (among the highest found in any survey conducted since 1990), and the fact that 75% of smear-positive cases already had typical TB symptoms but had either not yet sought care, or had sought health care but not been diagnosed; • a need for specific efforts in the hotspots where TB prevalence was highest – there was considerable variation in TB prevalence among survey clusters; and • a need for increased domestic funding at the federal government level, and especially at the state and local government authority levels, complemented by more international funding. Part of the reason for the large gap between the number of prevalent TB cases and the number of cases captured by the routine surveillance system could have been underreporting of detected cases. Possible solutions identified included strengthening linkages with all care providers and making TB notification mandatory by law. Photo credit: Philip Patrobas NIGERIA 183 Major successes, challenges and lessons learned The major success of this survey was that it was the first of its kind in Nigeria and contributed to a much better understanding and robust measurement of the burden of TB disease in the country. It was also implemented and completed in the face of several challenges beyond the control of the NTBLCP and the survey team. The biggest challenge outside the control of the NTBLCP and survey teams was the security situation in the country, which deteriorated during both preparations for and implementation of the survey. In August 2011, just as survey preparations (including all procurement) were almost complete and the survey was about to start, a terrorist attack occurred in the capital of Abuja. A bomb hidden in a car exploded underneath the United Nations (UN) building, killing 21 people and wounding 60 others (including WHO staff). Following this attack, the UN raised its rating of the security level and there was considerable debate about whether the survey should be cancelled. Eventually, the MoH decided to launch the survey in February 2012. Only three of the original list of randomly selected clusters had to be replaced due to the security measures in place; nevertheless, the security situation had other serious repercussions: • limited hours of operation for data collection during cluster operations (it was done from 7am to 5pm); • negative attitudes, including advice (or instructions) from some community leaders not to participate in the survey; • extremely limited access to the northern regions of the country (including the National TB Reference Laboratory used for the prevalence survey) for international staff; • no international technical assistance to the National TB Reference Laboratory in Zaria, although local staff from the United States Centers for Disease Control and Prevention continued to provide support; and • delays to the start of the survey that led to the expiry of the contract and associated licenses for X-ray software, which had to be re-procured. Linked to these repercussions, other major challenges included a low participation rate, especially in urban areas; a large number of positively-screened participants with missing culture results; and the possible under- performance of culture laboratories (related to lack of technical assistance). The implications of these challenges had to be investigated and adjusted for during analysis of survey data. Even with these data limitations, analyses that included imputation of missing data and sensitivity analysis showed that the limitations did not affect the main policy and programmatic implications drawn from the survey. For example, even if it was assumed that all those who refused to participate in the survey were healthy (i.e. without active TB disease), the number of detected TB cases still far exceeded the number of TB cases being detected and notified. Other challenges faced during the survey included: oversampling during field operations, despite the low participation rate (although field teams correctly registered the population in enumeration areas, regardless of their willingness to participate); survey investigators did not have access to national census data; staff from the Bureau of Statistics, who joined cluster operations, changed for each cluster; a 1-month suspension of field operations due to extreme rainfall; slow data entry from the field and in the laboratory; and delays in finalizing the survey report due to the departure of the survey coordinator and the lack of a full-time officer in the NTBLCP to oversee the survey. Important lessons learned for future surveys included the importance of ensuring that someone is available to prepare the survey report and of budgeting adequately for this activity; and working with the Bureau of Statistics, to inform them of the need to select clusters based on agreed survey methodology and not on their routine census activities. Photo credit: Philip Patrobas National TB prevalence surveys 2007–2016184 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. Report first national TB prevalence survey 2012 Nigeria: Department of Public Health, Federal Republic of Nigeria; 2014 (http://www.who.int/tb/publications/NigeriaReport_WEB_NEW. pdf, accessed May 2017). 5. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 6. World Health Organization. Global tuberculosis report 2012. Geneva: WHO; 2012 (http://www.who.int/tb/publications/ global_report/gtbr12_main.pdf, accessed June 2017). 185 Name Role Organization Ejaz Qadeer Principal investigator National TB Control Programme (NTP) Sabira Tahseen Co-principal investigator National TB Reference Laboratory (NTRL) Razia Fatima Co-principal investigator NTP Mohammad Asif Survey coordinator NTP Alamdar Hussain Rizvi Senior microbiologist NTRL, NTP Sabir Rehman Radiology coordinator NTP Zia Samad Data coordinator NTP Aisha Mariam Field team leader NTP Abdul Mannan Soomro Field team leader NTP Arshad Shamsi Field team leader NTP Riaz Ahmed Field team leader NTP Ghulam Nabi Shaikh Field team leader NTP Zulfiqar Ul Hassan Field team leader NTP Edine Tiemersma Technical assistance (survey advisor) KNCV Tuberculosis Foundation Masja Straetemans Technical assistance (survey advisor) KNCV Tuberculosis Foundation Nico Kalisvaart Technical assistance (data management) KNCV Tuberculosis Foundation Amal Bassili Technical assistance (survey advisor) WHO Eastern Mediterranean Regional Office (EMRO) Key people Surveyed clusters (N=95)a Data sources ■ Prevalence of pulmonary tuberculosis among the adult population of Pakistan 2010–2011. Ministry of Health. ■ Qadeer E, Fatima R, Yaqoob A, Tahseen S, Ul Haq M, Ghafoor A et al. Population based national tuberculosis prevalence survey among adults (≥15 years) in Pakistan, 2010–2011. PLoS One. 2016; 11(2). Summary statistics Participation rate 81% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 398 1.5 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.9 Finance Amount (US$) TB CAP 3 131 770 TB CARE 1 240 787 Total budget 4 372 557 Survey organization and financing Implementing agency: National TB Control Programme PAKISTAN a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2010–2011 National TB prevalence surveys 2007–2016186 Sampling design Sampling frame Whole country excluding the Federally Administered Tribal Areas, district Dera Bugti in Balochistan and 17 tehsils of Khyber Pakhtunkhwa Sampling design Multistage cluster sampling using PPS Strata No stratification was used, but the final analysis accounted for four provinces and two regions (Punjab, Sindh, Balochistan, Khyber Pakhtunkhwa, Azad Jammu and Kashmir, Gilgit-Baltistan), and urban/rural Sampling unit Province/district/tehsil/union council Sample size assumptions • Smear-positive prevalence 213 per 100 000 (≥15 years) • Precision 0.2 • Design effect 2.5 • k 0.7 • Response rate 85% • Sample size (estimated) 133 000 Number of clusters 95a Cluster size 1 400 Eligibility criteria • Age ≥15 years • Residency Individuals who slept in the household the night before the census Survey design and methodology a Three clusters in Balochistan (Lehri, Quetta, Awaran) were replaced by other clusters (Sharda in Azad-Jammu and Kashimir, Khan Pur in Punjab and Hub in Balochistan) due to security issues. Laboratory methodology Smear Two samples (spot, morning): direct preparation, ZN (spot sample was examined in the field site, morning sample in the central laboratory) Culture One sample (morning): direct preparation, Modified Kudoh method Identification of MTB PNB, MPB64: all culture-positive samples LPA, Xpert MTB/RIF: smear-positive with culture-negative or contaminated or N/A TB drug susceptibility test Done Xpert® MTB/RIF Done only for smear-positive with culture- negative or contaminated or N/A HIV test Not done Screening criteria Interviewa Cough ≥2 weeks or cough of any duration with no chest X-ray result Chest X-rayb Any lung abnormality Other TB treatment at the time of the survey a The first short screening consisted of questions about current TB treatment, current cough and its duration, and smoking behaviour. An in-depth interview (other TB symptoms and health-care seeking behaviour) was done only for those who screened positive. b Digital radiography (portable). Field data collection Paper/electronic Database EpiData version 3.1 Method of analysis MI+IPW Results first published in a report/paper March 2014 Official dissemination event March 2014 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 270 217–323 398 333–463 Male 352 273–431 484 392–577 Female 197 145–249 320 253–388 15–24 years 180 120–239 242 168–315 25–34 years 163 100–226 228 149–307 35–44 years 293 196–391 398 275–521 45–54 years 392 254–530 517 362–671 55–64 years 386 231–540 587 377–797 ≥65 years 691 439–942 1 369 1 028–1 710 Urban 209 147–270 310 234–386 Rural 321 241–401 471 377–564 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 5 417 – Location of care sought • Consulted medical facility N/A N/A Public facility N/A N/A Private facility N/A N/A Other N/A N/A • Pharmacy N/A N/A • Traditional healer N/A N/A No action taken N/A N/A Unknown N/A N/A a Cough ≥2 weeks or cough of any duration with no chest X-ray result. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and NAAT-negative. b 182 strains were tested. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 146 – • Treated in the public sector N/A N/A • Treated in the private sector N/A N/A • Treated in unknown sector N/A N/A Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 26 7.6 Design effect k Smear-positive TB 2.0 0.6 Bacteriologically confirmed TB 2.4 0.6 Other sputum results Number % Total smear-positive participants 236 – Smear-positive participants without MTB confirmationa 29 12 Isolates with MDR-TB detectedb 5 2.7 a Age ≥15 years unless otherwise specified. PAKISTAN 187 Survey flow: census to final outcomes Field operations: December 2010 to December 2011 Individuals enumerated in census 131 377 Eligible for sputum examination 10 471 (9.9%) Total bacteriologically confirmed cases 341 Eligible study population 131 329 (99.9%) Total participants 105 913 (81%) Interview and chest X-ray 100 984 (95%) Interview only 3 647 (3.5%) Chest X-ray only 1 282 (1.2%) Symptom positive, chest X-ray positive 2 819 (27%) Symptom positive, chest X-ray negative or N/A 2 598 (25%) Symptom negative, chest X-ray positive 5 042 (48%) Otherc 12 (0.1%) Symptom positive, chest X-ray positive 157 (46%) Symptom positive, chest X-ray negative or N/A 41 (12%) Symptom negative, chest X-ray positive 142 (42%) Otherc 1 (0.3%) Submitted specimens At least one specimen 8 521 (81%) Both specimens 6 810 (65%) Smear-positive casese 233 (68%) Definite 207 Probable 26 Smear-negative casesf 108 (32%) Definite 108 Probable N/A Laboratory result At least one culture result availabled 7 221 (69%) a Eligible for sputum collection. b The result was not entered on the form (933) or the form was not available (81). c Current TB treatment with symptom-screen negative and chest X-ray normal. d Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. e Definite: MTB confirmed by culture and/or NAAT. Probable: MTB not confirmed by culture and/or NAAT, but two smear-positive or one smear-positive with chest X-ray suggestive of TB. f Definite: MTB confirmed by culture (more than 5 colonies, or less than 5 colonies with chest X-ray suggestive of TB). Probable: no definition. Ineligible individuals 48 (0.04%) Children <15 years 45 (0.03%) Did not meet residency criteria 3 (<0.01%) Symptom screening Cough ≥2 weeksa 5 063 (4.8%) Cough (any duration) with no chest X-ray resulta 354 (0.3%) Haemoptysis N/A Sputum production N/A Chest pain N/A Fever N/A Total symptom-screen positivea 5 417 (5.2%) Chest X-ray screening Normal 93 391 (91%) Abnormala 7 861 (7.7%) Other abnormality N/A Result not availableb 1 014 (1.0%) Total chest X-rays taken 102 266 National TB prevalence surveys 2007–2016188 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 80 70 90 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 1.0 2.0 3.0 4.0 5.0 6.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 300 400 500 600 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 200 400 600 800 1000 1200 1400 1600 0 Prevalence per 10 0 0 0 0 population 20 000 60 000 80 000 100 000 120 000 40 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1800 1000 1200 1600 1400 600 800 200 400 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 3.59, variance 8.33, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 15 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Total (≥15) PAKISTAN 189 Background Pakistan’s population was 173 million in 2011, and it had an average gross national income (GNI) per person of US$ 1150 per year, making it a lower-middle-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2011, the prevalence of HIV in the general population aged 15–49 years was <0.1% (95% confidence interval [CI]: <0.1–<0.1%) (2), and it was estimated that 0.8% (95% CI: 0.6–0.9%) of TB patients were coinfected with HIV (3). The National TB Programme of Pakistan was revived in 2000 with a well-defined central unit, four TB control units at provincial level and one TB coordinator for each of the 139 districts. Implementation of the WHO- recommended DOTS strategy (4,5) began in 2000, and by 2005 full geographical coverage had been achieved in the public sector. The case notification rate (all forms of TB) increased from 7.7 per 100 000 population in 2000 to 153 per 100 000 population in 2010. In 2010, WHO estimated the incidence and prevalence of all forms of TB at 231 (95% CI: 190–276) per 100 000 population and 389 (95% CI: 191–657) per 100 000 population, respectively (6–8). Before 2010, three national TB prevalence surveys had been implemented: in 1960–1962, 1974–1978 and 1987– 1989. The screening and diagnostic methods used in the last of these surveys were a chest X-ray and an interview about symptoms, followed by smear microscopy for those reporting TB symptoms or with an abnormal X-ray. The prevalence of smear-positive pulmonary TB was estimated to be 170 per 100 000 population (7). In December 2007, Pakistan was one of the 22 global focus countries selected as a priority for a national TB prevalence survey by the WHO Global Task Force on TB Impact Measurement. National authorities had also recognized the value of a fourth national TB prevalence survey. At the same time, the feasibility of a survey was extensively discussed due to security concerns. Eventually, it was decided to implement a survey that excluded the country’s Federally Administered Tribal Areas and one district from Balochistan (Dera Bugti). Based on the most recent census data from these areas (from 1998), it was estimated that these excluded areas accounted for 6.5% of the country’s population. The fourth national TB prevalence survey started in December 2010 and was completed in December 2011. Photo credit: National TB Programme, Pakistan National TB prevalence surveys 2007–2016190 Key methods and results There were 95 clusters with a target size of 1400 individuals per cluster. No stratification was used at the time of survey design; however, results were later analysed separately for four provinces and two regions (Punjab, Sindh, Balochistan, Khyber Pakhtunkhwa, Azad Jammu and Kashmir, Gilgit-Baltistan), and for urban and rural areas. A total of 131  377 individuals from 33  324 households were enumerated in the survey census, of whom 131 329 (99.9%) were eligible and invited to participate. Of these, 105  913 (81%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, chest X-ray and an interview about symptoms (9). A total of 10 471 people (10% of participants) were eligible for sputum examination; of these 8521 (81%) submitted at least one sputum specimen and 6810 (65%) submitted two sputum specimens. A total of 341 cases of bacteriologically confirmed pulmonary TB were identified, including 233 cases of smear-positive TB. The prevalence of smear-positive pulmonary TB was 270 (95% CI: 217–323) per 100 000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 398 (95% CI: 333–463) per 100  000 population. The prevalence of bacteriologically confirmed TB was higher in rural areas (471 per 100 000 population; 95% CI: 377–564) than in urban areas (310 per 100 000 population; 95% CI: 234– 386). Other key results were: • the male to female ratio was 1.8 for smear- positive TB and 1.5 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age; however, the absolute number of bacteriologically confirmed TB cases was consistently high in all age groups; • among bacteriologically confirmed TB cases, 58% were symptom-screen positive, and among the smear-positive cases, 62% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio)  was 2.9 overall, but varied from 2.1 in those aged 25–34 years to 5.3 in those aged 65 years or more, and was higher for men than for women (3.8 versus 2.2); • among bacteriologically confirmed TB cases, only 7.6% were on anti-TB treatment at the time of the survey; no data about previous history of anti-TB treatment were available. No data were available on whether participants with symptoms had sought health care. Photo credit: National TB Programme, Pakistan PAKISTAN 191 Implications of results Based on survey results, the overall prevalence of TB (all forms of TB, all ages) was estimated as 342 (95% CI: 284–406) per 100 000 population. This was similar to (but more precise than) the pre-survey estimate published by WHO (389 per 100  000 population; 95% CI: 191–657) (8). This result showed that TB remained a major public health problem in Pakistan, requiring continued high- level political commitment and sustained funding. The high number of previously undiagnosed cases detected in the community, and the fact that 68% of these were smear-positive, suggested that people may not recognize the symptoms of TB, and that when they do seek care they may not be diagnosed. These findings indicated a need to implement strategies to increase population awareness of TB symptoms, and to improve the availability and quality of services for TB diagnosis and treatment in the community. It was also recognized that further analysis of health-care seeking behaviour and of awareness of TB among health-care providers could help the programme to design specific interventions. The higher prevalence of TB in older age groups and in rural areas demonstrated a need for improved case finding in these groups and areas in particular. One proposed option was active engagement of trained community health workers to help identify and refer people with TB symptoms to health services. Given the P:N ratio of 2.9, and results from a subsequent TB inventory study of the level of underreporting of detected TB cases in 2012, it was clear that both underreporting and underdiagnosis of TB needed to be addressed (10). With a large private sector, and less than 1% of private providers reporting TB cases to national authorities at the time of the survey, factors identified as being of vital importance were much greater engagement of the private sector and mandatory notification of cases. The survey was the first time the specimen transport system was successfully used for the transport of specimens via cold chain from field sites to the National TB Reference Laboratory (NTRL). This experience was used to improve specimen transportation undertaken as part of routine programmatic activities. Major successes, challenges and lessons learned Given the security  and geographical challenges, the successful implementation of the survey was a major achievement. Security concerns persisted throughout survey implementation. A complete security assessment was done and a handbook on security and safety was prepared in consultation with international experts in Photo credit: National TB Programme, PakistanPhoto credit: Masja Straetemans National TB prevalence surveys 2007–2016192 2010. Nonetheless, there were gaps in the monitoring and uptake of recommendations because those responsible for providing international technical assistance were not able to visit field sites. During the survey, three tehsils (administrative units) from Balochistan (Lehri, Quetta and Awaran) were replaced by three clusters (Sharda in Azad Jammu and Kashmir, Khan Pur in Punjab and Hub in Balochistan). The survey was also affected by a natural disaster. A major flood, mainly in Sindh province, affected a large part of the country, including 12 survey clusters. The flood damaged local infrastructure and displaced people. The survey field team visited the affected areas last, by which time the situation had improved. There was also a heatwave during field operations, during which it was an enormous challenge to maintain a cold chain for sputum transportation from clusters in remote villages to the NTRL. The NTRL achieved a lower than expected level of culture recovery, with only 69% (161/233) of the cases of smear- positive TB being confirmed by culture. Therefore, survey case definitions were amended to allow for the use of molecular tests (Genotype MTBDRplus or Xpert® MTB/ RIF, or both), which were then used to confirm whether individuals with a smear-positive result had TB. Another 46 of the 233 smear-positive TB cases (20%) were bacteriologically confirmed in this way. The survey faced major challenges with data management, and important lessons were learned that were subsequently used to help surveys in other countries. In particular, data collection in the field was based on multiple forms for each participant, rather than a single form. Besides needing to manage multiple forms, manual transcription errors when entering personal identification numbers (PINs) on the forms made it difficult to later match records (forms) for the same individual. Initially, about 8% of PINs were not available in the census register. It took more than a year of data cleaning and verification to produce the survey results. Data management challenges also delayed the follow-up of people with positive laboratory results. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. Prevalence of pulmonary tuberculosis among the adult population of Pakistan 2010–2011. Ministry of Health; (http://www.ntp.gov. pk/uploads/Prevalence_Report.pdf, accessed July 2017). 7. Qadeer E, Fatima R, Yaqoob A, Tahseen S, Ul Haq M, Ghafoor A et al. Population based national tuberculosis prevalence survey among adults (>15 years) in Pakistan, 2010–2011. PLOS ONE. 2016;11(2); (http://journals.plos.org/plosone/article?id=10.1371/ journal.pone.0148293, accessed January 2018). 8. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf, accessed January 2018). 9. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 10. Fatima R. Investigation of presumptive tuberculosis cases by private health providers: lessons learnt from a survey in Pakistan. 2014;4(2):110–112; (https://www.ncbi.nlm.nih.gov/pmc/articles/ PMC4539039/pdf/i2220-8372-4-2-110.pdf, accessed January 2018). 193 Name Role Organization Thelma E. Tupasi Principal investigator Tropical Disease Foundation, Inc. Ma. Imelda Quelapio Co-investigator Tropical Disease Foundation, Inc. Jennifer Chua Program manager Tropical Disease Foundation, Inc. Leilani Naval Administrative coordinator Tropical Disease Foundation, Inc. Nellie Mangubat Data processing coordinator Tropical Disease Foundation, Inc. Grace Egos Laboratory coordinator Tropical Disease Foundation, Inc. Lena Ablis Radiology coordinator Makati Medical Center Gerardo Beltran Radiology consultant Makati Medical Center Joselito Legaspi Radiology consultant Makati Medical Center Sistla Radhakrishna Biostatic consultant Consultant, Philippines Jesus Sarol Biostatistician University of the Philippines-Manila Ruffy Guilatco Data management staff Tropical Disease Foundation, Inc. Maricar Galipot Data management staff Tropical Disease Foundation, Inc. Genesis Ramos Data management staff Tropical Disease Foundation, Inc. Vivian Lofranco Field monitor Department of Health Onofre Edwin Merilles Field monitor Tropical Disease Foundation, Inc. Ruth Orillaza-Chi Field monitor Tropical Disease Foundation, Inc. Nona Rachel Mira Field monitor Tropical Disease Foundation, Inc. Virgil Belen Field monitor Tropical Disease Foundation, Inc. Albert Angelo Concepcion Field monitor Tropical Disease Foundation, Inc. Key people Surveyed clusters (N=50)a Data sources ■ Nationwide Tuberculosis Prevalence Survey 2007, final report, Republic of the Philippines. Tropical Disease Foundation Inc.; 2008. ■ Tupasi TE, Radhakrishna S, Chua JA, Mangubat NV, Guilatco R, Galipot M et al. Significant decline in the tuberculosis burden in the Philippines ten years after initiating DOTS. Int J Tuberc Lung Dis. 2009;13(10):1224–1230. ■ Floyd S, Sismanidis C. The 2007 Philippines nationwide TB survey confirmatory report of main results. London School of Hygiene & Tropical Medicine; 2008. ■ Survey dataset. Summary statistics Participation rate (chest X-ray) 90% Bacteriologically confirmed TB (≥10 years) • Prevalence per 100 000 population • Male:female ratio 660 2.6 Prevalence:notification ratio (smear-positive TB, ≥15 years) 1.9 Survey organization and financing Implementing agency: National TB Control Programme, Department of Health/ Tropical Disease Foundation, Inc. PHILIPPINES a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Finance Amount (US$) The Global Fund N/A World Health Organization N/A Total budget N/A 2007 National TB prevalence surveys 2007–2016194 Sampling design Sampling frame Whole countrya Sampling design Multistage cluster sampling using PPS Strata Metro Manila/other urban/rural Sampling unit Province/municipality/barangay Sample size assumptions • Smear-positive prevalence 300 per 100 000 (≥10 years) • Precision N/A • Design effect 1.25 • k 0.4 • Response rate 85% (for radiographic examination) • Sample size (estimated) 30 000 Number of clusters 50 Cluster size 600 Eligibility criteria • Age ≥10 years (chest X-ray) • Residency N/A Survey design and methodology a Four barangays in the "other urban" strata and 14 barangays in rural strata were excluded due to security issues and inaccessibility. Laboratory methodology Smear Three samples (spot, morning and morninga): direct preparation, FM (auramine stain) Cultureb Three samples (spot, morning and morninga): direct preparation for Ogawa media, and concentrated preparation for LJ media for pooled samples Identification of MTB Niacin, catalase test, nitrate production test TB drug susceptibility test Done Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interviewa The interview was conducted for each household about demographic and socio- economic factors, and also for participants ≥20 years about TB symptoms and the TB history. However, the interview result was not used as the selection criteria for sputum submission. Chest X-rayb Any lung abnormality (conducted for participants ≥10 years) Other N/A a An in-depth interview about health-care seeking behaviour was done for participants ≥20 years who reported any TB symptoms (cough more than 2 weeks, haemoptysis, chest or back pain, fever, night sweats, weight loss). b Conventional radiography. a Two morning samples were taken on the same day. b As per protocol, initially concentrated LJ was done for the first 37 participants. Due to the heavy workload for the laboratory, the method was changed to direct Ogawa and pooled concentrated LJ for the remaining participants. Field data collection Paper Database Microsoft® Access Method of analysis Multiple imputation Results first published in a report/paper July 2008 Official dissemination event July 2008 Analysis and reporting Key survey results Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positive N/A N/A Location of care sought • Consulted medical facility N/A N/A Public facility N/A N/A Private facility N/A N/A • Pharmacy N/A N/A • Traditional healer N/A N/A No action taken N/A N/A Unknown N/A N/A a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). b DST was done for 131 participants. Survey participants currently on TB treatment Number % Total participants currently on TB treatment N/A N/A • Treated in the public sector N/A N/A • Treated in the private sector N/A N/A • Treated in other sector N/A N/A Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment N/A N/A Design effect k Smear-positive TB 2.3 1.1 Bacteriologically confirmed TB 2.1 0.6 Other sputum results Number % Total smear-positive participants 55 – Smear-positive participants without MTB confirmationa 5 9.1 Isolates with MDR-TB detectedb 5 3.8 Prevalencea,b Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 280 190–370 660 530–800 Male 350 240–480 970 780–1 180 Female 190 120–290 370 260–510 10–19 years 20 0–90 180 90–310 20–29 years 220 100–420 500 300–770 30–39 years 240 100–470 600 370–920 40–49 years 460 250–790 1 100 750–1 560 ≥50 years 590 380–870 1 320 990–1 720 Metro Manila 430 0–870 640 160–1 120 Other urban 250 150–360 680 500–860 Rural 260 150–360 650 460–850 a Age ≥10 years. b Results for total prevalence and the three geographic strata are from the multiple imputation model; other numbers are crude estimates. PHILIPPINES 2007 195 Survey flow: census to final outcomes Field operations: July to December 2007 Individuals enumerated in census 30 667 Eligible for sputum examination 5 378c (26%) Total bacteriologically confirmed cases 136 Total participants Interview (TB symptom, ≥20 years) 15 242 (94%) Chest X-ray (≥10 years) 20 643 (90%) Eligible study population 22 867 (75%) 10–19 years 6 728 (22%) ≥20 years 16 139 (53%) Symptom positive, chest X-ray positive N/A Symptom positive, chest X-ray negative or N/A N/A Symptom negative, chest X-ray positive N/A Other N/A Symptom positive, chest X-ray positive N/A Symptom positive, chest X-ray negative or N/A N/A Symptom negative, chest X-ray positive N/A Other N/A Submitted specimens At least one specimen 5 173 (96%) Both specimens N/A Smear-positive casese 55 (40%) Definite 50 Probable 5 Smear-negative casesf 81 (60%) Definite 81 Probable N/A Laboratory result At least one culture result availabled 5 143 (96%) Symptom interview Cough ≥2 weeks 1 927 (13%) Haemoptysis 285 (1.9%) Chest pain 1 668 (11%) Fever 684 (4.5%) Night sweats 958 (6.3%) Chest X-ray screening Normal 15 962 (77%) Abnormalb 4 560 (22%) Other abnormality 1 (<0.01%) Result not available 120 (0.6%) Total chest X-rays taken 20 643 a BCG scar verification was undertaken for those aged 2 months–9 years, and a tuberculin skin test was done for two age groups (5–9 and 40–59 years) as well as those who had a chest X-ray suggestive of TB. b Only chest X-ray, and not a symptom interview, was used as a screening tool to determine eligibility for sputum collection. c The number eligible for sputum submission was more than the number with a field chest X-ray abnormality (i.e. 4560), because some consultant radiologists were involved in field screening and they found more chest X-ray abnormalities than field readers. d Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. e Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but chest X-ray suggestive of TB. f Definite: MTB confirmed by culture. Probable: no definition. Ineligible individuals Children <10 yearsa 7 800 (25%) Did not meet residency criteria N/A National TB prevalence surveys 2007–2016196 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female ≥5040–4930–3920–2910–19 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.5 3.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 4 2 8 12 6 10 0 0 1 2 3 4 5 6 7 8 9 10 11 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 400300 500 600 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 ≥5040–4930–3920–2910–19 200 400 600 800 1000 1200 1400 0 Prevalence per 10 0 0 0 0 population 60 000 40 000 20 000 100 000 120 000 140 000 160 000 80 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1800 800 1000 1200 1400 1600 600 200 400 0 ≥5040–4930–3920–2910–19 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data did not suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.72, variance 4.70, p=0.08). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). As notification data in the WHO database was disaggregated by six age groups (as opposed to the five age groups used in the Philippines survey), crude prevalence rates for six age groups were recalculated for this figure. d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Total (≥15) PHILIPPINES 2007 197 Background The population of the Philippines was 88 million in 2007, and the average gross national income (GNI) per person was US$ 1900 per year, making it a lower-middle- income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2007, the prevalence of HIV in the general population aged 15–49 years was <0.1% (95% confidence interval [CI]: <0.1–<0.1%) (2), and it was estimated that 0.2 % (95% CI: 0.1–0.3%) of TB patients were coinfected with HIV (3). The National TB Control Programme (NTP) launched the WHO-recommended DOTS strategy in 1996 (4,5) and national coverage was achieved by 2005 (5,6). In 2005, WHO estimated TB incidence and prevalence as 291 per 100 000 population and 450 per 100 000 popula- tion, respectively; the notification rate (new and relapse TB cases) was 165 per 100 000 population and had not changed significantly since 2000. The case detection rate (notifications of new and relapse cases divided by estimated incidence in the same time period) was 55% in 2005 (8). The Philippines had previously undertaken two national TB prevalence surveys, one in 1981–1983 and one in 1997. In the 1981–1983 survey, the prevalence of smear- positive TB was 660 per 100  000 population (among those ≥10 years) and the prevalence of culture-positive TB was 860 per 100  000 population. The prevalence of smear-positive TB in the 1997 survey (360 per 100 000 population; 95% CI: 280–450) was lower than in the 1981–1983 survey, but the prevalence of culture-positive TB in the 1997 survey (960 per 100  000 population; 95% CI: 750–1160) had not significantly changed. Drug susceptibility testing of 188 isolates from the 1997 survey showed that 4.3% of survey cases had multidrug-resistant TB (1.5% of people with no previous TB history and 15% of previously treated cases) (6,9). The Philippines NTP undertook a third national TB prevalence survey in 2007 to determine the burden of TB and the impact of the DOTS programme, which had been launched 10 years previously. Key methods and results There were 50 survey clusters across three strata (Metro Manila, other urban and rural), with a target cluster size of 600 individuals. A total of 30  667 individuals from 6259 households were enumerated in the survey census, of whom 22 867 (75%) were eligible for chest X-ray and were invited to participate. Of these, 20 643 (90%) were screened by chest X-ray. A total of 5378 people (26% of participants) were eligible for sputum examination based on their chest X-ray result; of these 5173 (96%) submitted at least one sputum specimen. An interview about symptoms was undertaken for 15 242 participants aged 20 years or more; however, this was not considered to be a screening tool for sputum submission (6,9,10). A total of 136 bacteriologically confirmed pulmonary TB cases were identified, including 55 cases of smear-positive TB. The prevalence of smear-positive TB was 280 (95% CI: 190–370) per 100 000 population (among those aged ≥10 years), and for bacteriologically confirmed TB it was 660 (95% CI: 530–800) per 100  000 population. When extrapolated to all forms of TB and to all ages, prevalence was 576 (95% CI: 515–640) per 100 000 population. There was no significant variation in prevalence among the three strata. Photo credit: Leilani Naval National TB prevalence surveys 2007–2016198 Other key results were: • the male to female ratio was 1.8 for smear- positive TB and 2.6 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age, as did the absolute number of bacteriologically confirmed TB cases; • among bacteriologically confirmed TB cases who were interviewed using a symptom questionnaire, 42% had a chronic cough, and among the smear-positive TB cases who were interviewed using a symptom questionnaire, 59% had a chronic cough; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio)  was 1.9 overall, but varied from 1.2 in those aged 15–24 years to 3.2 in those aged 65 years or more, and was higher for women than men (2.1 versus 1.6); and • among bacteriologically confirmed TB cases, 72% had no previous history of anti-TB treatment. Data for those on anti-treatment at the time of the survey were not available. It was estimated that up to one third of participants with symptoms suggestive of TB had consulted a health facility and one quarter had taken no action. Implications of results Based on the prevalence surveys in 1997 and 2007, which followed a standard protocol and similar methodology, the prevalence of bacteriologically confirmed TB declined from 960 (95% CI: 750–1160) per 100 000 population to 660 (95% CI: 530–800) per 100 000 population. Smear- positive prevalence also declined, from 360 (95% CI: 280– 450) per 100 000 population to 280 (95% CI: 190–370) per 100 000 population. Between 1996 and 2007, the Philippi- nes NTP aggressively implemented its strategic plan for TB control in collaboration with private sector partners, increased its budgetary support, and continued to enhance the quality of DOTS services through training and retooling. The 2007 survey suggested that these efforts had contributed to a reduction in the burden of TB disease in the country. Nonetheless, in the 2007 survey, the prevalence of bacteriologically confirmed TB was 2.6 times higher among males than among females, and it increased with age. This shifting of the burden into older age groups mirrored results from other surveys in Asia and indicated a maturing epidemic. Specific efforts were still required to reduce the burden in males and older age groups. Photo credit: Leilani Naval PHILIPPINES 2007 199 Among the participants who reported TB symptoms in the 2007 survey, only one third had previously consulted health facilities; nearly half of them had chosen to self- medicate and the rest had not taken any action. Although in comparison to the 1997 survey the proportion of symptomatic participants who consulted health facilities increased marginally and the proportion who took no action dropped, the proportion who had self-medicated almost doubled. Among those in the 2007 survey who took no action, 45% considered their symptoms to be harmless, 39% could not afford the cost of treatment and 4% found the distance to a health facility to be a barrier. These findings highlighted a need to improve access to health facilities, social support and advocacy to communities (6). Major successes, challenges and lessons learned Major successes of the 2007 survey included completing the survey on time, despite challenges faced during field operations, and the high coverage of the survey’s screening and diagnostic tests (e.g. 90% of the 22 867 participants aged 10 years or older were examined by chest X-ray). Major challenges faced during the survey included the exclusion of some barangays (i.e. the smallest administrative unit) from the sampling frame because of security issues and inaccessibility, so the survey was not truly representative of the national population; it was difficult to define the study population in some congested areas because households were not clearly demarcated; logistical challenges were experienced in some barangays (e.g. households spread over several kilometres, or located in geographically challenging locations such as small islands or mountainous areas); and the quality of sputum samples was questionable in some clusters because of the absence of courier services, difficulties in maintaining the cold chain in tropical conditions and delays in processing specimens (this resulted in high specimen contamination rates; 6.9% of 13  926 specimens on Ogawa media and 8.3% of pooled specimens on LJ slopes (6)). Photo credit: Leilani Naval Photo credit: Leilani Naval National TB prevalence surveys 2007–2016200 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS. (http://aidsinfo.unaids.org/, accessed May 2017). 3. World Health Organization. Global tuberculosis database. Geneva: WHO; 2017 (http://www.who.int/tb/data/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. Nationwide Tuberculosis Prevalence Survey 2007, final report, Republic of the Philippines. Tropical Disease Foundation Inc.; 2008. 7. World Health Organization. Global Plan to Stop TB – Phase 1: 2001 to 2005. Geneva: WHO; (http://www.stoptb.org/global/plan/ plan0105.asp, accessed July 2017). 8. World Health Organization. Global Tuberculosis Control. Geneva: WHO; 2007 9. Tupasi TE, Radhakrishna S, Chua JA, Mangubat NV, Guilatco R, Galipot M et al. Significant decline in the tuberculosis burden in the Philippines ten years after initiating DOTS. Int J Tuberc Lung Dis. 2009;13(10):1224–1230. 10. Floyd S, Sismanidis C. The 2007 Philippines nationwide TB survey confirmatory report of main results. London School of Hygiene & Tropical Medicine; 2008. 201 Key people Surveyed clusters (N=106)a Data sources ■ National Tuberculosis Prevalence Survey 2016, Philippines: Department of Health, Republic of the Philippines; Foundation for the Advancement of Clinical Epidemiology, Inc.; Philippine Council for Health Research and Development; 2018. ■ Survey dataset. Summary statistics Participation rate 76% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 1 159 2.7 Prevalence:notification ratio (smear-positive TB, ≥15 years) 3.1 Finance Amount (US$) Department of Health, Philippines 1 987 462 The Global Fund 367 900 Total budget 2 355 362 Survey organization and financing Implementing agency: National TB Control Programme, Department of Health/ Philippine Council for Health Research and Development (PCHRD)/Foundation for the Advancement of Clinical Epidemiology (FACE, Inc.) PHILIPPINES a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Name Role Organization Mary Ann Lansang Principal investigator (PI) Foundation for the Advancement of Clinical Epidemiology (FACE, Inc.)/University of the Philippines Manila (UP Manila) Anna Marie Celina Garfin Chair of the technical working group Department of Health (DOH) Marissa Alejandria Co-PI FACE, Inc./UP Manila Myrna Mendoza Co-PI Foundation for the Control of Infectious Diseases, Inc. (FCID, Inc.)/UP Manila Jacinto Blas Mantaring III Co-PI FACE, Inc./UP Manila Noel Juban Co-PI, field coordinator FACE, Inc./UP Manila Sonia Salamat Co-field coordinator FCID, Inc./UP Manila Concepcion Ang Laboratory manager FCID, Inc./UP Manila Joseph Adrian Buensalido Laboratory manager FCID, Inc./UP Manila Johanna Patricia Cañal Radiology coordinator Philippine College of Radiology/UP Manila Maria Lourdes Amarillo Data manager FACE, Inc./UP Manila Olivia Sison Data manager FACE, Inc./UP Manila Jose Rene Cruz Field team leader FACE, Inc. Nori Jane Galagar Field team leader FACE, Inc. Anjo Benedict Fabellon Field team leader FACE, Inc. Rodelia Pascua Field team leader FACE, Inc. Allison Noel Field team leader FACE, Inc. Luis Anos Field team leader FACE, Inc. Aser Sisona Field team leader FACE, Inc. Irwin Law Technical assistance (survey advisor) WHO headquarters Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Yasunori Ichimura Technical assistance (survey advisor) Chiba University, Japan Marina Tadolini Technical assistance (survey advisor) Consultant, Italy M. Bintari Dwihardiani Technical assistance (survey advisor) Consultant, Indonesia Julia Ershova Technical assistance (data management) US Centers for Disease Control and Prevention (CDC) Hiroko Matsumoto Technical assistance (laboratory) Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) Tetsuhiro Sugamoto Technical assistance (laboratory) RIT/JATA 2016 National TB prevalence surveys 2007–2016202 Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 434 350–518 1 159 1 016–1 301 Male 673 528–819 1 713 1 482–1 943 Female 205 141–270 627 516–739 15–24 years 330 197–463 799 586–1 011 25–34 years 326 195–458 900 677–1 123 35–44 years 470 298–641 1 126 821–1 430 45–54 years 665 438–891 1 714 1 364–2 064 55–64 years 488 285–691 1 504 1 104–1 903 ≥65 years 503 310–696 1 659 1 261–2 058 NCR, 3, 4-Ab 599 451–747 1 358 1 103–1 612 Rest of Luzon 258 138–378 1 038 787–1 288 Visayas 471 261–680 1 234 873–1 594 Mindanao 268 173–364 856 686–1 026 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Four strata (National Capital Region, regions 3 and 4-A/rest of Luzon/Visayas/ Mindanao). In the final analysis, urban and rural were also considered. Sampling unit Four strata/province or HUC (Highly Urbanized Cities)/barangay Sample size assumptions • Smear-positive prevalence 260 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.8 • k 0.8 • Response rate 85% • Sample size (estimated) 54 000a Number of clusters 108b Cluster size 500 Eligibility criteria • Age ≥15 years • Residency Individuals who lived for at least two weeks in the household prior to the census Survey design and methodology a Six clusters (3000 individuals) were added to the original sample size (51 000), to ensure this sample size, in case of cancellation in the Autonomous Region in Muslim Mindanao due to security issues. b One cluster in Basilan province was excluded before field operations started, due to security issues. During field operations, three clusters (Sipangkot, Madaya and Maco barangays) were replaced by others from the same provinces, due to problems of accessibility and security. Another cluster (Holy Spirit barangay) was dropped because the board of directors of the private subdivision in the selected area refused to allow the survey team to do house-to-house mobilization and interviews. Laboratory methodology Smeara One or two samples (the morning sample was mainly used. If the morning sample’s volume was inadequate, the second spot sample was also used): direct preparation, FM (LED, auramine stain) Culturea One or two samples (the morning sample was mainly used. If the morning sample’s volume was inadequate, the second spot sample was also used): direct preparation, Ogawa media Identification of MTB MPT 64 rapid test TB drug susceptibility test Done Xpert® MTB/RIFa Done for all first spot specimensb HIV test Not done Screening criteria Interviewa Cough ≥2 weeks and/or haemoptysis Chest X-rayb Any lung or mediastinum abnormality Other Chest X-ray exempted a An in-depth interview about health-care seeking behaviour was done for participants who reported cough ≥2 weeks and/or haemoptysis. b Mobile digital X-ray machine. a All participants who were eligible for sputum examination were asked to submit two sputum samples (spot and morning) for smear, culture and Xpert MTB/RIF. The additional spot sample was collected when the volume of previous sputum specimens (first spot and/or morning) was less than 3ml. b If the first sample had an inadequate volume, a morning or second spot specimen was used. If all three specimens had less than 1ml each, the available specimens were pooled. Field data collection Electronic Database Epi Info Method of analysis MI+IPW Results first published in a report/paper May 2018 Official dissemination event August 2017 Analysis and reporting Key survey results Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 815 – Location of care soughtb • Consulted medical facility 530 – Public facility 359 67 Private facility 162 31 Other 3 0.6 Unspecified 6 1.1 • Pharmacy 4 – • Traditional healer 10 – Self-treatedb 1 130 – No action taken 1 142 41 Unknown 18 0.6 a Cough ≥2 weeks and/or haemoptysis. b Participants could answer more than one category. a Some participants answered more than one facility. The reason why they had multiple treatment places is unavailable. b Private pharmacy (23), relatives (1). a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b DST was done for 232 culture MTB-positive specimens. Survey participants currently on TB treatment Number % Total participants currently on TB treatmenta 170 – • Treated in the public sector 134 – • Treated in the private sector 15 – • Treated in other sectorb 24 – • Treated in unknown sector 1 – Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 30 6.4 Design effect k Smear-positive TB 1.7 0.6 Bacteriologically confirmed TB 2.0 0.4 Other sputum results Number % Total smear-positive participants 183 – Smear-positive participants without MTB confirmationa 10 5.5 Isolates with MDR-TB detectedb 9 3.9 a Age ≥15 years unless otherwise specified. b National Capital Region, regions 3 and 4-A. PHILIPPINES 2016 203 Survey flow: census to final outcomes Field operations: March to December 2016 Individuals enumerated in census 89 663 Eligible for sputum examination 18 597 (40%) Total bacteriologically confirmed cases 466 Eligible study population 61 466 (69%) Total participants 46 689 (76%) Interview and chest X-ray 41 444 (89%) Interview only 5 245 (11%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 1 444 (7.8%) Symptom positive, chest X-ray negative or N/A 1 371 (7.4%) Symptom negative, chest X-ray positive 10 702 (58%) Otherb 5 080 (27%) Symptom positive, chest X-ray positive 132 (28%) Symptom positive, chest X-ray negative or N/A 18 (3.9%) Symptom negative, chest X-ray positive 298 (64%) Otherf 18 (3.9%) Submitted specimens At least one specimen 16 242 (87%) Both (spot and morning) specimens 15 547 (84%) Third specimen 32 (0.2%) Smear-positive casesd 173 (37%) Definite 173 Probable N/A Smear-negative casesd 289 (62%) Definite 289 Probable N/A Smear unknown casese 4 (0.9%) Definite 4 Probable N/A Laboratory result At least one culture result availablec 15 776 (85%) At least one Xpert MTB/RIF result available 16 200 (87%) Symptom screening Cough ≥2 weeksa 2 458 (5.3%) Haemoptysisa 565 (1.2%) Fever 5 313 (11%) Weight loss 8 188 (18%) Night sweats 3 959 (8.5%) Total symptom-screen positivea 2 815 (6.0%) Chest X-ray screening Normal 22 625 (55%) Abnormala 12 146 (29%) Other abnormality 6 672 (16%) Result not available 1 (<0.01%) Total chest X-rays taken 41 444 a Eligible for sputum submission. b Chest X-ray exempted and symptom-screen negative (5079), poor chest X-ray image and symptom-screen negative (1). c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. e Definite: smear and culture not done, but MTB confirmed by Xpert. Probable: no definition. f Chest X-ray exempted and symptom-screen negative. Ineligible individuals 28 197 (31%) Children <15 years 27 885 (31%) Did not meet residency criteria 247 (0.3%) Missing data 65 (0.07%) National TB prevalence surveys 2007–2016204 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 5 10 20 15 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 500 700 900 1100 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 200 600 400 800 1000 1200 1400 1600 1800 0 Prevalence per 10 0 0 0 0 population 40 000 20 000 100 000 140 000 120 000 180 000 160 000 200 000 80 000 60 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 2500 1500 2000 1000 500 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 4.40, variance 9.12, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using smear-positive pulmonary TB notifications (2016) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Total (≥15) PHILIPPINES 2016 205 Background The Philippines had a population of 101 million in 2015 and was a lower-middle-income country with an average gross national income (GNI) per person of US$  3520 per year (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2015, the prevalence of HIV in the general population aged 15–49 years was <0.1% (95% confidence interval [CI]: <0.1–<0.1%) (2), and it was estimated that 0.9% (95% CI: 0.5–1.4%) of TB patients were coinfected with HIV (3). Using findings from the 2007 national TB prevalence survey as well as other data, WHO estimated TB incidence at 285 (95% CI: 228–342) per 100 000 population in 2008; this remained static up to 2014 (288 per 100 000 popula- tion; 95% CI: 254–324). Prevalence was estimated to have decreased slightly, from 548 (95% CI: 499–597) per 100 000 population in 2008 to 417 (95% CI: 367–471) per 100 000 population in 2014 (4,5). In December 2007, the Philippines was one of the 22 global focus countries selected by the WHO Global Task Force on TB Impact Measurement as a priority for a national TB prevalence survey during the period 2008–2015. The fourth national TB prevalence survey in the Philippines was conducted from March to December 2016 (6), following surveys in 1981–1983, 1997 and 2007. It was led by the National TB Control Programme, Department of Health and the Philippine Council for Health Research and Development, and was implemented by the Foundation for the Advancement of Clinical Epidemiology, Inc. The primary objective of the survey was to estimate the prevalence of pulmonary TB (bacteriologically confirmed; i.e. culture-positive TB or Xpert® MTB/RIF, or both) among the general population aged 15 years or more. Key methods and results There were 106 survey clusters in four strata – National Capital Region, regions 3 and 4-A; rest of Luzon; Visayas; and Mindanao. The target cluster size was 500 individuals. A total of 89 663 individuals from 19 707 households were enumerated in the survey census, of whom 61 466 (69%) were eligible and invited to participate. Of these, 46 689 (76%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using a chest X-ray and an interview about symptoms (7). A total of 18 597 participants (40%) were eligible for sputum examination; of these, 16  242 (87%) submitted at least one sputum specimen and 15  547 (84%) submitted two sputum specimens. Sputum specimens from 16  200 participants were tested with Xpert MTB/ RIF. Of these, 397 (2.5%) were Xpert positive for Mycobacterium tuberculosis, and of these, 29 (7.3%) were also rifampicin (RIF) resistant, 358 (90%) were RIF susceptible and 10 (2.5%) were indeterminate. Of 466 bacteriologically confirmed TB cases, 159 (34%) were confirmed by both culture and Xpert MTB/RIF, 69 (15%) only by culture and 238 (51%) only by Xpert MTB/RIF. Of the 466 bacteriologically confirmed TB cases, 173 (37%) were smear-positive. The prevalence of smear- positive TB was 434 (95% CI: 350–518) per 100  000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 1159 (95% CI: 1016–1301) per 100 000 population. Although there was no statistically significant variation between the four geographical strata, the highly urbanized strata (National Capital Region, regions 3 and 4-A) had the highest prevalence of bacteriologically confirmed TB (1358 per 100  000 population; 95% CI: 1103–1612), followed by Visayas (1234 per 100  000 population; 95% CI: 873– 1594), rest of Luzon (1038 per 100 000 population; 95% CI: 787–1288) and the more rural Mindanao (856 per 100 000 population; 95% CI: 686–1026). Photo credit: Julia Ershova National TB prevalence surveys 2007–2016206 Other key results were: • the male to female ratio was 3.3 for smear- positive TB and 2.7 for bacteriologically confirmed TB; • prevalence per 100 000 population was high in all age groups, especially in those aged 35 years or more, with the peak being in those aged 45−54 years; the absolute number of TB cases was high in the young and middle age groups (15−54 years); • among bacteriologically confirmed TB cases, 32% were symptom-screen positive, and among the smear-positive cases, 51% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio)  was 3.1 overall, but varied from 2.1 in those aged 55−64 years to 4.2 in the 15−24 years age group, and was higher for men than women (3.3 versus 2.5); • among bacteriologically confirmed TB cases, 82% had no previous history of anti-TB treatment and only 6.4% were on anti-TB treatment at the time of the survey; and • of the 138 bacteriologically confirmed and 82 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 35 (25%) and 22 (27%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results The prevalence of bacteriologically confirmed pulmonary TB was the highest of all national surveys implemented globally since 2007. Based on the survey, the estimated prevalence for all forms of TB and all ages was 982 (95% CI: 862–1100) per 100 000 population – this was almost 2.5 times higher than the pre-survey estimate (i.e. 417 per 100 000 population in 2014; 95% CI: 367–471) (5). Together with surveys in Bangladesh and Kenya, the 2016 survey in the Philippines was one of the first surveys to use both Xpert MTB/RIF and culture for all participants who screened positive. Although it was not surprising that the use of Xpert MTB/RIF increased the overall diagnostic yield, the prevalence of culture-confirmed TB alone was very high (587 per 100 000 population; 95% CI: 488−687) and showed that the Philippines was facing one of the highest burdens of TB in the world. When prevalence was extrapolated to all forms of TB and all ages, it was estimated that there were about 1 million people in the Philippines with TB in 2016, equivalent to 1 in 15 of all prevalent cases globally (6). Notwithstanding the limitation of a 76% participation rate, survey results were of high quality and provided a robust measurement of the burden of TB disease. Results from the 2016 prevalence survey were used to update estimates of TB incidence and mortality. The estimate of TB incidence after the survey was 554 (95% CI: 311–866) per 100 000 population in 2016, compared with the pre- survey WHO estimate (which had assumed a decline in incidence since 2007) of 288 (95% CI: 254–324) per 100 000 population in 2014; estimates for previous years were similarly revised upwards. The estimated mortality rate based on the survey was 21 (95% CI: 21–22) per 100 000 population in 2016, compared with a pre-survey estimate of 10 (95% CI: 9.1–11) in 2014; estimates for previous years were similarly revised upwards (5, 8). Photo credit: Irwin Law PHILIPPINES 2016 207 The sample size in 2016 was not designed to detect a specified effect size (e.g. 20% decline) in comparison with the 2007 survey, but rather to obtain an estimate of prevalence in 2016 with a specified precision. The 2016 survey was therefore not powered to detect small differences between it and the 2007 survey. Nonetheless, this limitation did not prevent an assessment of the trend in TB disease burden since 2007. Adjustments were made to ensure that the two datasets and methods were as comparable as possible, resulting in an upward adjustment of the 2007 survey results, to account for the more sensitive screening and diagnostic methods used in the 2016 survey. Based on these adjustments, the prevalence of culture-positive TB was 463 (95% CI: 333−592) per 100  000 population in 2007 and 512 (95% CI: 420−603) per 100 000 population in 2016 (6). The probability that prevalence did not decline over the period 2007–2016 was estimated at 75%. The lack of decline in TB prevalence since 2007 could be explained by a combination of case-detection gaps, significant delays in diagnosis, health system weaknesses, and broader social and economic influences on the TB epidemic. These broader influences included the level of poverty, with 22% of people living below the national poverty line in 2015; the level of undernourishment, with a prevalence of 14% in the general population in 2015 and no improvement since 2008; and low coverage of health insurance and social protection (e.g. coverage of only 4% in the poorest quintile in 2013), leading to financial barriers to accessing health services and high levels of TB-affected households facing catastrophic costs (35% in 2016−2017) (1, 9). At a broader level, the poor and disadvantaged require adequate social protection strategies and increased PhilHealth TB benefit packages to reduce catastrophic costs associated with TB, especially multidrug-resistant TB (MDR-TB).1 Based on TB prevalence survey findings, the National TB Control Programme (NTP) initiated the development of new strategies with a national multisectoral approach. These included: • introducing systematic screening among high- risk and vulnerable groups (including men, older age groups and those living in urban areas); • improving the use of tools for screening and diagnosis, coupled with improved training of health-care providers and health-care delivery; • initiatives to reduce geographic and financial barriers affecting access to health care; • greater engagement of public-private mix partnerships, including effective implementation of existing legislation on mandatory notification of TB cases; and • strengthening collaboration between the NTP and other health programmes, such as those for HIV, diabetes and lung health. In discussions towards the end of 2016, it was anticipated that these strategic actions would be implemented with the full support of the Department of Health, and full mobilization of the health sector. Measures that were agreed to be needed included the deployment of sufficient human resources at national and subnational levels; increased domestic funding; a presidential executive order for drug regulation; establishment of a high-level steering group; and ensuring financial protection (and sustained poverty alleviation efforts) for more than 90% of the poor through increased coverage of PhilHealth and expanded social protection programmes. Major successes, challenges and lessons learned Major successes in the survey included: • high-level commitment and excellent coordination by the implementing agency; • reaching remote hamlets and villages that were included in the sampling frame, based on efficient logistical management of field teams and equipment and use of digital X-ray Photo credit: Julia Ershova 1 PhilHealth is the national health insurance programme. National TB prevalence surveys 2007–2016208 machines, as well as effective use of social media and instant messaging; • regular supervision of field teams and laboratories by central staff, which helped to ensure the quality of survey operations and standardization across the teams; • double reading of each chest X-ray; that is, X-rays were read by one medical officer in the team and by another person (an off-site radiologist) who read the chest X-ray remotely, with a quick turnaround; • use of Xpert MTB/RIF, which made up for challenges associated with MTB culture processes; • almost 90% of specimens for culture being processed in 5 days or less; and • the availability of a large team of highly skilled people to clean and analyse data. Challenges faced during the survey included: • a low participation rate (76% compared with a target of 85%) despite extended hours for field operations including evenings and weekends; lower participation was observed in men, younger age groups, those living in urban areas and higher income groups, as well as during the two months preceding national elections; • the high sputum eligibility rate (40% of total participants screened) which led to a larger than expected laboratory workload; • difficulties in standardizing techniques across six laboratories; the culture recovery rate1 varied between 75% and 92% and contamination rates varied between 1.4% and 6.2% (6); and • logistical issues arose in maintaining cold storage during transport from the field to the laboratory, which may have affected culture results. During the preparation phase, one major lesson learned was the need for the implementing agency to have complete control of the design and implementation of the data management system. Initially, a private company was contracted to develop the system; however, because of the slow response times to adapt to changes in the survey protocols and data collection tools, plus ongoing costs, the company was replaced by an in-house team. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS. (http://aidsinfo.unaids.org/, accessed May 2017). 3. World Health Organization. Global tuberculosis database. Geneva: WHO; 2017 (http://www.who.int/tb/data/en/, accessed April 2017). 4. World Health Organization. Global tuberculosis control 2010. Geneva: World Health Organization; 2010 (http://apps.who.int/ iris/bitstream/10665/44425/1/9789241564069_eng.pdf, accessed February 2018). 5. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015 (http://apps.who.int/iris/ bitstream/10665/191102/1/9789241565059_eng.pdf, accessed July 2017). 6. National Tuberculosis Prevalence Survey 2016. Philippines: Department of Health, Republic of the Philippines; Foundation for the Advancement of Clinical Epidemiology, Inc.; Philippine Council for Health Research and Development; 2018. 7. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 8. World Health Organization. Global tuberculosis report 2017. Geneva: WHO; 2017 (http://www.who.int/tb/data/en/, accessed February 2018). 9. Survey to estimate the proportion of households experiencing catastrophic costs due to TB (report of the NTP Zonal Dissemination Forum: 2016 National TB Prevalence Survey and 2016 Catastrophic Cost Study, on 14 September 2017). Department of Health, Republic of the Philippines; 2017. 1 Recovery rate of MTB: number of smear-positive MTB that are culture positive out of the number of smear-positive specimens. 209 Name Role Organization Michel Gasana Principal investigator Tuberculosis & Other Respiratory Communicable Diseases Division-Kigali, Rwanda Claude Bernard Uwizeye Principal investigator US Centers for Disease Control and Prevention CDC-Kigali, Rwanda Eveline Klinkenberg Principal investigator KNCV Tuberculosis Foundation Pauline Basinga Principal investigator School of public health, National University of Rwanda Patrick Migambi Co-investigator and survey coordinator Tuberculosis & Other Respiratory Communicable Diseases Division-Kigali, Rwanda Julie Mugabekazi Co-investigator WHO Rwanda Védaste Ndahindwa Survey statistician School of public health, National University of Rwanda Elaine Kamanzi Survey laboratory activities coordinator National Reference Laboratory-Kigali, Rwanda Jules Kamugunga Mulinzi Survey data manager Tuberculosis & Other Respiratory Communicable Diseases Division-Kigali, Rwanda Alaine Umubyeyi Nyaruhirira Laboratory advisor Management Sciences for Health Louise Kalisa Survey radiology coordinator Kigali University Teaching Hospital-Kigali, Rwanda Calvin Mugabo Field team leader Tuberculosis & Other Respiratory Communicable Diseases Division-Kigali, Rwanda Liliane Umutesi Field team leader Tuberculosis & Other Respiratory Communicable Diseases Division-Kigali, Rwanda Ndeziki Mashengesho Field team leader Tuberculosis & Other Respiratory Communicable Diseases Division-Kigali, Rwanda Nico Kalisvaart Technical assistance (data management) KNCV Tuberculosis Foundation Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Peou Satha Technical assistance (radiology) CENAT, Cambodia Key people Surveyed clusters (N=73)a Data sources ■ The First National Tuberculosis Prevalence Survey 2012 in Rwanda, Institute of HIV/AIDS, Disease Prevention & Control, Tuberculosis & Other Respiratory Communicable Diseases Division, Republic of Rwanda, Ministry of Health, 2015. ■ Migambi P, Gasana M, Uwizeye CB, Kamanzi E, Ndahindwa V, Kalisvaart N, Klinkenberg E. Prevalence of tuberculosis in Rwanda: Results of the first nationwide survey in 2012 yielded important lessons for TB control. PLoS One. 2020 Apr 23;15(4):e0231372. ■ Survey dataset. Summary statistics Participation rate 96% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 119 3.9 Prevalence:notification ratio (smear-positive TB, ≥15 years) 1.3 Finance Amount (US$) The Global Fund 1 840 893 US CDC 415 000 KNCV 36 741 WHO/OGAC (PEPFAR) 75 778 Total budget 2 368 412 Survey organization and financing Implementing agency: Tuberculosis and Other Respiratory Communicable Diseases Division, Rwanda Biomedical Center, the Ministry of Health RWANDA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2012 National TB prevalence surveys 2007–2016210 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata No stratification was used, but the final analysis accounted for province (Kigali city, North, East, South, West) Sampling unit Province/administrative sector/umudugudu (village) Sample size assumptions • Smear-positive prevalence 304 per 100 000 (≥15 years) • Precision 0.23 • Design effect 1.7 • k 0.6 • Response rate 95% • Sample size (estimated) 44 500 Number of clusters 73a Cluster size 610 Eligibility criteria • Age ≥15 years • Residency Individuals who lived in the household for at least 1 month prior to the interview Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): direct preparation in the facility close to the survey site (and the National Reference Laboratory also examined smear with the concentrated preparation), FM (LED, auramine stain) Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB MPT64 rapid test TB drug susceptibility test Donea Xpert® MTB/RIF Not done HIV test Offered to those who screened positive Screening criteria Interviewa Cough (any duration) Chest X-rayb Any lung abnormality Other Chest X-ray exempted a An in-depth interview about health-care seeking behaviour was done only for those who screened positive by interview and/or chest X-ray. b Mobile chest X-ray truck, digital radiography. a Although the required number of clusters was 70, an additional 3 clusters were selected in Kigali to obtain more precise estimates. a 38 TB cases were tested. Field data collection Paper Database EpiData 3.1 Method of analysis MI+IPW Results first published in a report/paper June 2015 Official dissemination event January 2016 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 74 48–99 119 79–160 Male 142 88–196 208 139–278 Female 24 4.7–43 53 20–86 15–34 years 57 27–86 86 46–125 35–54 years 66 21–110 114 35–193 ≥55 years 159 54–263 262 104–421 Urban N/A N/A N/A N/A Rural N/A N/A N/A N/A Health-care seeking behaviour among participants who were symptom positive Number % Participants who were symptom positivea 2 855 – Location of care soughtb 921 32 • Consulted medical facility Public facility 941 – Private facility 48 – • Pharmacy 101 – • Traditional center 54 – • Other 38 – Self-treated 0 0 No action taken 1 934 68 Unknown N/A N/A a The in-depth interview identified 2855 participants who had a cough. This interview was in addition to the screening interview, and the extra participants who acknowledged a cough (304) were not included in the final screening outcomes. b The subtotals do not add up to 921 because participants could select more than one health facility or groups within a facility (e.g. public facility includes health center, district hospital, referral hospital and community health worker). a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). b DST was done for 38 TB cases. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 21 – • Treated in the public sector N/A N/A • Treated in the private sector N/A N/A • Treated in other sector N/A N/A • Treated in unknown sector N/A N/A Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 2 5.0 Design effect k Smear-positive TB 0.91 N/Aa Bacteriologically confirmed TB 1.3 0.7 Other sputum results Number % Total smear-positive participants 29 – Smear-positive participants without MTB confirmationa 7 24 Isolates with MDR-TB detectedb 2 5.2 a Age ≥15 years unless otherwise specified. a K could not be computed for smear-positive TB because the design effect was less than one. RWANDA 211 Survey flow: census to final outcomes Field operations: March to December 2012 Individuals enumerated in census 84 140 Eligible for sputum examination 4 747 (11%) Total bacteriologically confirmed cases 40 Eligible study population 45 058 (54%) Total participants 43 128 (96%) Interview and chest X-ray 43 062 (99.8%) Interview only 59 (0.1%) Chest X-ray only 7 (0.02%) Symptom positive, chest X-ray positive 545 (12%) Symptom positive, chest X-ray negative or N/A 2 092 (44%) Symptom negative, chest X-ray positive 2 107 (44%) Otherb 3 (0.1%) Symptom positive, chest X-ray positive 15 (38%) Symptom positive, chest X-ray negative or N/A 4 (10%) Symptom negative, chest X-ray positive 21 (52%) Other 0 (0%) Submitted specimens At least one specimen 4 700 (99%) Both specimens 4 412 (93%) Smear-positive casesd 27 (67%) Definite 22 Probable 5 Smear-negative casese 13 (33%) Definite 13 Probable N/A Laboratory result At least one culture result availablec 4 589 (97%) Symptom screening Cough (any duration)a 2 637 (6.1%) Haemoptysis 105 (0.2%) Sputum production 2 117 (4.9%) Fever 1 317 (3.1%) Weight loss 4 001 (9.3%) Total symptom-screen positivea 2 637 (6.1%) Chest X-ray screening Normal 40 408 (94%) Abnormala 2 652 (6.2%) Other abnormality N/A Result not available 9 (0.02%) Total chest X-rays taken 43 069 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but two smear-positive specimens or one smear-positive with chest X-ray suggestive of TB. e Definite: MTB confirmed by two cultures, or one culture with chest X-ray suggestive of TB. Probable: no definition. Ineligible individuals 39 082 (46%) Children <15 years 34 857 (41%) Did not meet residency criteria 4 225 (5.0%) National TB prevalence surveys 2007–2016212 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 150 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea 450 250 200 300 400 350 150 100 50 0 15–34 35–54 ≥55 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0 15–34 35–54 Male Female≥55 Pr ev al en ce : no ti fi ca ti on ra ti o 1.0 0.5 1.5 2.0 2.5 3.0 Age group (years) 0 15–34 35–54 ≥55 50 100 150 200 250 300 0 Prevalence per 10 0 0 0 0 population 3 500 3 000 2 500 2 000 1 500 1 000 500 4 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data did not suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 0.55, variance 0.61, p=0.27). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimate of using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 20 5 15 10 40 35 30 25 0 0 1 2 3 Total (≥15) RWANDA 213 Background Rwanda, in East Africa, had a population of 10 million in 2012, of which 85% lived in rural areas. The average gross national income (GNI) per person was US$ 640 per year, making it a low-income country (1). The prevalence of HIV in the general population aged 15–49 years was estimated at 3.1% (95% confidence interval [CI]: 2.7– 3.4%) in 2012 (2), and it was estimated that 26% (95% CI: 25–27%) of TB patients were coinfected with HIV (3). In 1990, the Programme National de Lutte contre la Tuberculose – Rwanda’s National Tuberculosis (TB) Control Programme (NTP) – was established within the Ministry of Health. At the same time, TB control activities were decentralized to the health-facility level (public and faith-based). The WHO-recommended DOTS strategy was implemented from the mid-1990s (4,5). In 2005, a community DOTS strategy was launched to help make services more accessible; it included increasing the role of community health workers in the detection and management of TB patients. Nationwide coverage for community TB care was achieved in 2010. Collaborative TB/HIV activities were launched in 2005. By December 2012, 99% of notified TB cases (all forms) knew their HIV status, and of these cases, 26% were HIV-positive. Of the TB patients living with HIV in 2012, 99% were initiated on co-trimoxazole prophylaxis, and 75% were on antiretroviral treatment (6). The total number of reported TB cases (all forms of TB) increased after 1995 and peaked at 8283 in 2006. Subsequently, TB case notifications fell year on year, to 6207 in 2012. The TB notification rate followed a similar downward trend; after a peak in 2006, it fell to 59 per 100 000 population in 2012 (and 37 per 100 000 population for smear-positive pulmonary TB) (7). In the 2013 WHO global TB report, the estimated prevalence of TB in 1990 was 356 (95% CI: 173–603) per 100 000 population for all forms of TB and 114 (95% CI: 61–183) per 100  000 population in 2012 (6). Over the same period, TB incidence was estimated to have fallen from 290 (95% CI: 259–323) per 100 000 population to 86 (95% CI: 77–96) per 100 000 population. The estimated TB case detection rate (for new and relapse cases) was 62% in 2012. However, there was no direct measurement of TB disease burden in Rwanda, and it was considered possible that the burden was lower than indicated in published estimates given the expansion in TB services and collaborative TB/HIV activities. In December 2007, Rwanda was one of the 22 global focus countries selected by the WHO Global Task Force on TB Impact Measurement as a priority for a national TB prevalence survey during the period 2008–2015. Following a decision by the NTP to carry out its first national TB prevalence survey, a survey was implemented between March and December 2012 (7). Photo credit: Kamugunga Jules National TB prevalence surveys 2007–2016214 Key methods and results There were 73 clusters in the survey, with a target cluster size of 610 individuals. Stratification was not used at the time of survey design; however, five provinces (Kigali city, North, East, South and West) were examined separately during the final analysis. A total of 84  140 individuals from 19 474 households were enumerated in the survey census, of whom 45 058 (54%) were eligible and invited to participate. Of these, 43  128 (96%) participated in the survey and were screened according to the 2011 algorithm recommended by WHO; that is, chest X-ray and a symptom screening interview (8). A total of 4747 people (11% of participants) were eligible for sputum examination, of whom 4700 (99%) submitted at least one sputum specimen and 4412 (93%) submitted two sputum specimens. A total of 40 bacteriologically confirmed pulmonary TB cases were identified, including 27 cases of smear-positive TB. The prevalence of smear-positive TB was 74 (95% CI: 48–99) per 100 000 population, and for bacteriologically confirmed pulmonary TB it was 119 (95% CI: 79–160) per 100 000 population (≥15 years). When extrapolated to all forms of TB and all ages, prevalence was estimated as 95 (95% CI: 66–124) per 100 000 population. Other key results were: • the male to female ratio was 5.9 for smear- positive TB and 3.9 for bacteriologically confirmed TB; • prevalence per 100 000 population increased with age; however, the absolute number of TB cases was relatively high in the young age group (15–34 years); • among bacteriologically confirmed TB cases, 48% were symptom-screen positive, and of the 27 smear-positive cases, 52% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio)  was 1.3 overall, but varied from 0.9 in those aged 35–54 years to 2.4 in the 55 years or more age group, and was higher for men than for women (1.8 versus 0.7); • among bacteriologically confirmed TB cases, 93% had no previous history of anti- TB treatment and only 5% were on anti-TB treatment at the time of the survey; • of the 17 bacteriologically confirmed and 12 smear-positive TB survey cases that screened positive for symptoms and were not on anti- TB treatment at the time of the survey, 2 (12%) and 2 (17%), respectively, had previously sought care in a public or private health facility for their symptoms; and • of those eligible for sputum examination, 94% (4445/4747) were offered HIV counselling and testing, of whom 5.2% (248/4747) refused; overall, 218 (4.9%) of those tested were HIV- positive, and 181 of the 218 (83%) already knew their HIV status; of 40 bacteriologically confirmed TB cases, 36 were tested for HIV and only 1 (2.8%) was HIV-positive. Implications of results The estimated prevalence of TB identified in the survey was lower than WHO estimates. This was a welcome finding, but also presented a challenge in terms of how to ensure continued funding to sustain efforts in TB control and further reduce this burden. It was recognized that finding and treating the remaining cases could require more costly interventions (on a per patient basis) than those used in the past. The prevalence of HIV among TB cases detected in the survey was low. This probably reflected two factors: the short duration of illness for HIV-positive TB cases that are untreated in the community, and the effective TB screening programme among people living with HIV. People at higher risk for TB – including people living with HIV as well as prisoners, refugees and students in boarding schools – were already a priority for the NTP at the time of the survey. However, all diagnosed cases should be used as an entry point to find additional cases, including through strengthened contact tracing and continued active case finding. Photo credit: Kamugunga Jules RWANDA 215 The higher burden of TB among men and the elderly was consistent with routine surveillance data. However, men were five times more likely than women to have TB, whereas among notified TB cases there were only twice as many men as women. These findings suggested under- diagnosis among men and the elderly, and associated differences in health-care seeking behaviour. Rwanda introduced a community health-insurance system in 1999 to improve access to health care. In 2012, 91% of the population was covered by this health insurance and 83% of the population could access a health-care facility within 2 hours of their home. Despite improving access to health care, survey data showed that people with TB or with symptoms meeting screening criteria did not always seek care, especially if they were poor, men or young adults. Overall, 70% of those with a cough who had not sought care at the time of the survey indicated that it was not important to do so; only 6% indicated that lack of money for transport was a barrier to accessing care (7). It appeared that people in the general community were not identifying themselves as being at risk of developing TB disease, and that innovative approaches would be needed to raise awareness and enhance care-seeking among individuals with a cough. The survey also suggested that the existing advocacy, communication and social mobilization strategy should be reviewed to incorporate innovative strategies to aid TB control. Possibilities that were identified included the use of role models or ambassadors, especially those with whom men could identify; raising awareness among health-care staff, given that only half of those who sought care for a chronic cough were asked to submit a sputum specimen for testing (7); improving health-care staff awareness that men and the elderly are more likely to have TB than other groups and are underdiagnosed; and strategic case finding among the elderly, for example through routine outpatient screening for TB in this age group. After the survey, the NTP defined five high risk groups that required greater attention: children under 15 years, people over 55 years, prisoners, people living with HIV (PLHIV) and contacts of TB cases. In addition, the NTP developed plans to use chest X-ray as a screening tool among prisoners and PLHIV, and for the scale-up of Xpert®MTB/RIF as a diagnostic tool. Contrary to expectations, one-third (16/54) of the participants with positive culture growth had nontuberculosis mycobacteria (NTM). This showed a need for further investigations to characterize the NTM problem in Rwanda, by conducting genotyping of the current cases, characterizing the affected population and determining the extent of the problem, as well as developing guidelines on treatment of NTM (such guidelines did not exist at the time of the survey). Overall, the survey showed that current efforts in TB prevention, diagnosis and treatment needed to be maintained while also being supplemented by new strategies, to ensure early detection and treatment of all cases, with a specific focus on key populations. Major successes, challenges and lessons learned The Rwandan survey showed that the country’s TB and TB/HIV services were well organized. However, since the number of detected cases was so small compared with the estimated burden when the survey was designed, it was hard to analyse in detail the characteristics of the detected TB cases. Most survey equipment was procured by the Rwanda Biomedical Center. Delays occurred in procurement despite the process starting early. The original plan was to import portable digital X-ray units, but this was not possible because the national radiation authority did not approve the units. Digital units in a container system were procured instead. During field operations, one digital container was accidentally dropped and required a service. Photo credit: Kamugunga Jules National TB prevalence surveys 2007–2016216 A major success was that the overall participation rate was very high, at 96%. The area in which achieving high participation was a challenge was Kigali (the capital city); as in other surveys, this made it more difficult to estimate TB prevalence in highly urbanized areas. Rwanda was also one of the first countries to provide high-quality data on TB/HIV coinfection with a large proportion of survey participants requesting to be tested. The survey was the first to successfully use a “paper-based horizontal data collection” approach. Without having the same individual survey form for each participant used throughout the screening process, individual data were collected independently and blindly from other information. However, more than in any other survey, this process required large and intensive amounts of human resource effort and a strong data management team. An external review confirmed that the central laboratory carried out culture examination in accordance with their standard operating procedures; that is, concentrated Löwenstein–Jensen media recommended by WHO. Nonetheless, compared with surveys that used liquid media (i.e. mycobacteria growth indicator tube), or solid media without centrifuge (i.e. Ogawa method), the yields by culture were limited. Of the 54 culture positive participants, there were 38 participants with culture- confirmed TB, and 3 of the 38 were excluded from the final case list (in total there were 35 definite survey cases and 5 probable cases). They had only an indication of TB in one of the collected samples, which was not confirmed by an indication in another sample or the central chest X-ray reading. Therefore, it is likely that the prevalence of bacteriologically confirmed TB was underestimated. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf, accessed January 2018). 7. The First National Tuberculosis Prevalence Survey 2012 in Rwanda, Institute of HIV/AIDS, Disease Prevention & Control, Tuberculosis & Other Respiratory Communicable Diseases Division, Republic of Rwanda, Ministry of Health, 2015. 8. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 217 Name Role Organization Igbal Ahmed Elbasheer Principal investigator Public Health Institute (PHI) Mona Hassen Mustafa Survey coordinator PHI Sawsan Mustafa Abdalla Survey coordinator PHI Heba Kamal Hamed Elneel NTP manager (coordination between the survey team and TB states coordinators) National TB Programme (NTP) Asrar Mohammed Abdelsalam Head of laboratory National Tuberculosis Reference Laboratory Majda Elsayed Central radiologist Consultant, Sudan Nahid Abdelgader Data manager PHI Abdelaeem Field team leader PHI Ahmed Elhaj Ali Field team leader PHI Alfakie Field team leader PHI Fatih Alrahaman Ali Abdel-rahaman Field team leader PHI Hasham Alamin Salem Field team leader PHI Hashim Salah Hamza Field team leader PHI Hozifa Omer Eljak Field team leader PHI Mohammed Osman Field team leader PHI Mustafa Field team leader PHI Nazar Alnoor Ibrahim Field team leader PHI Sami Abdel Hameed Field team leader PHI Sumia Yousif Mohammed Field team leader PHI Ayyed Muneam El-Dulaimi Technical support WHO Sudan Mai Mohammed Eltigany Technical support WHO Sudan Amal Bassili Technical assistance (survey advisor) WHO Eastern Mediterranean Regional Office (EMRO) Sabira Tahseen Technical assistance (survey advisor) Consultant, Pakistan Fasil Tsegaye Technical assistance (survey advisor) Consultant, Ethiopia Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Norio Yamada Technical assistance (statistics) Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) Kiyohiko Izumi Technical assistance (statistics) RIT/JATA Key people Surveyed clusters (N=109)a Data sources ■ Survey dataset. Summary statistics Participation rate 86% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 183 1.6 Prevalence:notification ratio (smear-positive TB, ≥15 years) 3.5 Finance Amount (US$) Government of Sudan 487 000 The Global Fund 1 400 709 Total budget 1 887 709 Survey organization and financing Implementing agency: Public Health Institute (PHI)/National TB Programme SUDAN a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2013–2014 National TB prevalence surveys 2007–2016218 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit State/administrative unit/popular administrative unit (PAU) Sample size assumptions • Smear-positive prevalence 239 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.9 • k 0.7 • Response rate 85% • Sample size (estimated) 91 131 Number of clusters 114a Cluster size 800 Eligibility criteria • Age ≥15 years • Residency Household members resident in the selected household for the past 6 months, and visitors who spent ≥3 weeks in the household prior to the census Survey design and methodology a 109 out of 114 PAUs were visited: four clusters (one in South Kordofan, two in Darfur State and one in Gazira) were cancelled due to security concerns and one due to non-compliance with eligibility criteria. Laboratory methodology Smear Two samples (spot, morning): direct preparation, FM (LED, auramine stain) Culture Two samples (spot, morning): direct preparation, Ogawa media Identification of MTB Capilia LPA for all smear-positive and all culture- positive samples TB drug susceptibility test Not done Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interviewa Cough ≥2 weeks Chest X-rayb Any lung abnormality Other Any current TB treatment, chest X-ray exempted a An in-depth interview about other TB symptoms and health-care seeking behaviour was done only for those who screened positive. b Direct digital (portable). Field data collection Paper/electronic Database CSPro Method of analysis MI+IPW Results first published in a report/paper Pending Official dissemination event Pending Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 87 52–121 183 128–238 Male 123 79–193 233 163–333 Female 58 32–105 143 98–208 15–24 years 27 9.2–82 49 21–111 25–34 years 122 68–220 250 154–407 35–44 years 157 88–279 249 154–403 45–54 years 67 27–168 227 132–390 55–64 years 87 31–246 255 134–486 ≥65 years 114 51–253 282 163–489 Urban 150 93–243 275 178–425 Rural 55 29–108 137 89–210 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 663 – Location of care sought • Consulted medical facility 1 308 49 Public facility 1 077 82 Private facility 90 6.9 Other (NGO) 141 11 • Pharmacy 52 2.0 • Traditional centre 49 1.8 No action taken 575 22 Other (unspecified) 69 2.6 Unknown 610 23 a Cough ≥2 weeks. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and LPA-negative. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 104 – • Treated in the public sector 69 66 • Treated in the private sector 1 1.0 • Treated in other sector 4 4.0 • Treated in unknown sector 30 29 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 8 7.1 Design effect k Smear-positive TB 1.8 1.1 Bacteriologically confirmed TB 2.7 1.1 Other sputum results Number % Total smear-positive participants 61 – Smear-positive participants without MTB confirmationa 4 6.6 Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. SUDAN 219 Survey flow: census to final outcomes Field operations: February 2013 to March 2014 Individuals enumerated in census 150 490 Eligible for sputum examination 17 541 (21%) Total bacteriologically confirmed cases 112 Eligible study population 96 979 (64%) Total participants 83 202 (86%) Interview and chest X-ray 76 904 (92%) Interview only 5 080 (6.1%) Chest X-ray only 1 026 (1.2%) Othera 192 (0.2%) Symptom positive, chest X-ray positive 1 823 (10%) Symptom positive, chest X-ray negative or N/A 840 (4.8%) Symptom negative or N/A, chest X-ray positive 9 838 (56%) Othere 5 040 (29%) Symptom positive, chest X-ray positive 43 (38%) Symptom positive, chest X-ray negative or N/A 8 (7.1%) Symptom negative or N/A, chest X-ray positive 45 (40%) Otheri 16 (14%) Submitted specimens At least one specimen 14 330 (82%) Both specimens 11 313 (65%) Laboratory result At least one culture result availablef 14 017 (80%) Symptom screening Cough ≥2 weeksb 2 663 (3.3%) Haemoptysis 221 (1.3%)c Sputum production 2 169 (12%)c Chest pain 4 039 (23%)c Fever 3 657 (21%)c Total symptom-screen positivea 2 663 (3.3%) Chest X-ray screening Normal 66 001 (85%) Abnormalb 11 661 (15%) Other abnormality N/A Result not availabled 268 (0.3%) Total chest X-rays taken 77 930 a Of 192 individuals, 80 were exempted from chest X-ray and were not screened by interview. In addition, 112 individuals did not attend the survey site and were not screened by interview or chest X-ray. Although a protocol violation, specimens were collected from these 112 individuals, and included in the final analysis (they were subsequently classified as off-site participants). b Eligible for sputum collection. c The denominator is 17 423 (on-site participants who screened positive). d Poor quality of film (13) and result missing (255). 13 (poor quality of film) out of 268 were asked to submit sputum. e Symptom-screen negative or missing and chest X-ray exempted (4899), symptom-screen negative and chest X-ray result N/A (13), symptom-screen negative or missing and chest X-ray negative but currently on TB treatment (10), off-site participants (112), symptom-screen negative or missing and chest X-ray negative, not currently on TB treatment but submitted sputum in error (6). f Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. g Definite: MTB confirmed by culture and/or LPA. Probable: no definition. h Definite: MTB confirmed by culture. Probable: no definition. i Symptom-screen negative and chest X-ray exempted (9), symptom-screen negative or missing and chest X-ray negative, not currently on TB treatment but submitted sputum in error (6), off-site participant (1). Smear-positive casesg 57 (51%) Definite 57 Probable N/A Smear-negative casesh 52 (46%) Definite 52 Probable N/A Smear unknown cases 3 (2.7%) (culture MTB-positive) Ineligible individuals 53 511 (36%) Children <15 years 51 549 (34%) Did not meet residency criteria 1 962 (1.3%) National TB prevalence surveys 2007–2016220 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 150 200 250 300 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 50 100 150 200 250 300 0 Prevalence per 10 0 0 0 0 population 4 000 2 000 8 000 12 000 10 000 14 000 16 000 6 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 500 250 400 350 300 450 200 150 100 50 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 1.03, variance 2.90, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using smear-positive pulmonary TB notifications (2013) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 55 50 45 40 35 30 25 20 15 10 5 0 0 1 2 3 4 5 6 7 8 Total (≥15) SUDAN 221 Background Sudan had a population of 39 million people in 2013, of which 88% were settled (i.e. in a permanent residence), including 33% in urban areas and 67% in rural areas (of whom 8% were nomads). The average gross national income (GNI) per person was US$ 1170, making it a lower- middle income country (1). In 2013, the prevalence of HIV in the general population aged 15–49 years was 0.2% (95% confidence interval [CI]: 0.2–0.3%) (2), and it was estimated that 4.4% (95% CI: 3.6–5.3%) of tuberculosis (TB) patients were coinfected with HIV (3). Nationwide coverage of DOTS was achieved in 2002. In March 2013, the Federal Ministry of Health of Sudan integrated management of disease-specific programmes, including TB, into a newly established Communicable and Noncommunicable Disease Administration (C&NCD) under the Directorate of Primary Health Care. State and locality TB programme officers were responsible for the implementation of TB control activities, including supervision of the TB management unit(s) in their area of responsibility. In 2013, there were 325 TB management units, usually from institutions that were part of the primary health-care network. The TB laboratory network was organized at three levels, with microscopy laboratories in each TB management unit, quality assurance performed by each of the 15 states, and one National TB Reference Laboratory. The treatment success rate for new smear-positive pulmonary TB cases was 80–82% between 2002 and 2010. This dropped to 70% in 2011 (4) because some cases could not be evaluated due to the conflict in West Darfur, but returned to 80% in 2014 (5). The case detection rate (notifications of new cases divided by estimated incidence) was estimated at 44% (95% CI: 37–54%) in 2012. Sudan was not one of the 22 global focus countries for national TB prevalence surveys identified by the WHO Global Task Force on TB Impact Measurement in December 2007. However, it was on the Task Force’s longer list of 53 countries considered to meet survey eligibility criteria. Given considerable uncertainty about estimates of the burden of TB disease, the lack of a previous national TB prevalence survey, the fact that no direct measurements of TB mortality were available from vital registration, and the difficulty in estimating the gap between notifications and incidence (due to underreporting or under-diagnosis of cases), it was decided to conduct the country’s first national TB prevalence survey. The survey started in February 2013 and was completed in March 2014. Key methods and results There were 109 survey clusters in two strata (urban and rural), with a target cluster size of 800 individuals. A total of 150 490 individuals from 24 837 households were enumerated in the survey census, of whom 96 979 (64%) were eligible and invited to participate. Of these, 83 202 (86%) did so. Almost all participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (6). Out of 83 202 participants, 112 were not screened either by interview or chest X-ray; instead, they submitted sputum at home when survey teams visited. A total of 17  541 participants (21%) were eligible for sputum examination, of whom 14 330 (82%) submitted at least one sputum specimen and 11 313 (65%) submitted two sputum specimens. A total of 112 bacteriologically confirmed pulmonary TB cases were identified, including 57 cases of smear- positive TB. The prevalence of smear-positive TB was Photo credit: Fasil Tsegaye National TB prevalence surveys 2007–2016222 87 (95% CI: 52–121) per 100  000 population (among those aged ≥15 years), and for bacteriologically confirmed TB it was 183 (95% CI: 128–238) per 100 000 population. When extrapolated to all forms of TB and to all ages, prevalence was 172 (95% CI: 122–222) per 100 000 population. The prevalence of smear-positive and bacteriologically confirmed TB was higher in urban than in rural areas. Other key results were: • the male to female ratio was 2.1 for smear- positive TB and 1.6 for bacteriologically confirmed TB; • the prevalence per 100  000 population was consistently high for people aged 25 years and over, and people in the age groups 25–34 and 35–44 years accounted for a relatively large proportion of the absolute number of bacteriologically confirmed TB cases; • among bacteriologically confirmed TB cases, 46% were symptom-screen positive, and among smear-positive cases, 56% were symptom-screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio)  was 3.5 overall, but varied from 1.6 in those aged 15–24 years to 6.0 in the 35–44 years age group, and was slightly higher for men than for women (3.7 versus 3.4); • among bacteriologically confirmed TB cases, 76% had no previous history of anti-TB treatment and 7.1% were on anti-TB treatment at the time of the survey; and • of the 44 bacteriologically confirmed and 26 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 30 (68%) and 16 (62%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results The TB prevalence survey confirmed a burden of disease similar to pre-survey estimates of prevalence, i.e. 207 (95% CI: 104–345) per 100 000 population in 2012 (4), while also producing more precise estimates. The survey showed that a high proportion of cases in the community had not yet reached TB diagnostic and treatment services, and high prevalence rates in the younger population (even in the context of low HIV prevalence among TB patients) confirmed ongoing transmission. Plausible explanations for a higher prevalence per 100 000 population in urban areas included large-scale displacement of people from rural areas due to insecurity and associated deterioration in economic conditions, in contrast with remote rural areas with nomadic populations and fewer opportunities to spread TB. The survey had several major programmatic, policy and funding implications including those listed below. • NTP services should be reoriented towards the hospital sector. Most survey cases that reported symptoms had sought treatment at general hospitals, rather than primary health centres (PHC), but in 2014 more than half of the country’s hospitals lacked TB diagnostic services. There was a need to strengthen hospitals to include TB diagnostic services, supported by strengthening PHC centres, especially for treatment monitoring. • Diagnosis with culture or other diagnostics beyond culture (e.g. Xpert® MTB/RIF) and use of chest X-ray as part of the screening algorithm for TB should be widely expanded. • Case-finding activities should be intensified, and targeted particularly towards those aged 25–44 years and urban areas. • There was a need to address inadequate knowledge of TB symptoms and the variable quality of services among health-care providers. • Increased funding was required to implement the above-listed policy and programmatic measures. Photo credit: Fasil Tsegaye SUDAN 223 Major successes, challenges and lessons learned The major overarching success was that Sudan’s first- ever national TB prevalence survey was successfully implemented, with a high participation rate. Advocacy through the media, and the involvement of stakeholders and community leaders at the state level, strongly facilitated survey participation. The survey faced several major challenges that included those listed below. • The harsh terrain and remoteness of much of the country made survey operations very demanding. This contributed to frequent breakdowns of equipment, problems with internet connectivity for electronic data collection (including the transfer of digital images from the field to the central level) and difficulties with the transportation of sputum samples. • Several clusters that were initially selected in Darfur state and other bordering states were later excluded due to security concerns, which had knock-on effects for the survey schedule and logistics. • There was a high turnover of staff. • Relatively few morning sputum specimens were collected. While at least one sputum specimen was obtained for 82% of participants who were eligible for sputum examination, there were fewer morning specimens than spot specimens. Since morning samples typically yield more bacteriologically confirmed results, this may have led to underestimation of TB prevalence. The specimen-based imputation model used in the analysis of data helped to compensate for this problem. • Backlogs delayed culture inoculation when field operations were accelerated without sufficient consideration of laboratory capacity (for 3460 out of 9664 morning specimens, inoculation occurred more than 7 days after collection). This may have resulted in some false-negative culture results, and contributed to the low culture confirmation of smear-positive survey cases. • Data entry errors occurred on the tablet computers used in the field. Considerable efforts were needed to fix these errors in the absence of routine recording of data on paper as well as electronically (the paper form for symptom screening was introduced from the fourth cluster onward). Important lessons learned for future surveys included: • paper records are valuable to back up electronic records; • different diagnostic techniques should be considered given the environmental challenges of maintaining the cold chain for specimens; and • Africa-Africa and Asia-Africa collaborations are valuable, as is technical assistance from international agencies; the coordinator of the national TB prevalence survey in Ethiopia and a laboratory expert from Pakistan both provided assistance; the WHO Regional Office for the Eastern Mediterranean provided assistance during survey preparations, including protocol development, and the Research Institute of Tuberculosis (Japan) helped to clean and analyse the data. Photo credit: Fasil Tsegaye National TB prevalence surveys 2007–2016224 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf, accessed January 2018). 5. World Health Organization. Global tuberculosis report 2014. Geneva: WHO; 2014 (http://apps.who.int/iris/ bitstream/10665/137094/1/9789241564809_eng.pdf, accessed January 2018). 6. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 225 Name Role Organization Sriprapa Nateniyom Principal investigator Bureau of Tuberculosis Sirinapha Jittimanee Survey coordinator Bureau of Tuberculosis Saijai Smithtikarn Laboratory coordinator Bureau of Tuberculosis Wilawan Dangsaart Radiology coordinator Bureau of Tuberculosis Wiriya Madasin Data manager Bureau of Tuberculosis Autagorn Chunmathong Field team leader Bureau of Tuberculosis Runjuan Sukkavee Field team leader Bangkok Metropolitan Administration Nuntaporn Meksawasdichai Field team leader Institute for Urban Disease Control and Prevention, Bangkok Pattana Pokaew Field team leader ODPC 1, Chiangmai Sakchai Chaiamahapurk Field team leader ODPC 2, Pitsanulok Pavasuth Chutjuntaravong Field team leader ODPC 3, Nakhonsawan Supaporn Wattanatoan Field team leader ODPC 4, Saraburi Ratree Dokkabowt Field team leader ODPC 5, Ratchaburi Ornnipa Iamsamang Field team leader ODPC 6, Chonburi Narong Wongba Field team leader ODPC 7, Konkaen Phalin Kamolwat Field team leader ODPC 9, Nakhonratchasima Walaya Sitti Field team leader ODPC 10, Ubonratchathani Kamonwan Imduang Field team leader ODPC 11, Nakhonsrithamarat Auyporn Petborisuit Field team leader ODPC 12, Songkhla Norio Yamada Technical assistance (data analysis) Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) Hataichanok Pukcharern Technical assistance (sampling methodology) National Statistics Office, Thailand Ikushi Onozaki Technical assistance (survey methodology) WHO headquarters Key people Surveyed clusters (N=83 (non-Bangkok))a Data sources ■ Survey dataset. Summary statistics Participation rate 79% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 242 3.3 Prevalence:notification ratio (smear-positive TB, ≥15 years) 1.8 Finance Amount (US$) Government of Thailand 100 080 The Global Fund 1 790 293 Total budget 1 890 373 Survey organization and financing Implementing agency: National TB Programme, Bureau of Tuberculosis THAILAND ODPC: The Office of Disease Prevention and Control a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2012–2013 National TB prevalence surveys 2007–2016226 Sampling design Sampling frame Whole country (but results applied only to non-Bangkok clusters due to the low rate participation within the Bangkok region) Sampling design Multistage cluster sampling using PPS Strata Bangkok, non-Bangkok (urban) and non- Bangkok (rural) Sampling unit Bangkok: three zonesa/enumeration areas Non-Bangkok: 12 regions (ODPCb)/provinces/ enumeration areas Sample size assumptions • Smear-positive prevalence 118 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.3 • k 0.5 • Response rate 85% • Sample size (estimated) 90 000 (Bangkok: 15 300, non-Bangkok: 74 700)c Number of clusters 100 (Bangkok: 17, non-Bangkok: 83) Cluster size 900 Eligibility criteria • Age ≥15 years • Residency 1) Permanent residents based on house registration or 2) temporary residents or non- residents who had slept in the household for at least 2 weeks before the census Survey design and methodology a The three zones included: i) the inner-most geographic region, ii) the surrounding districts, iii) districts on the edge of the Bangkok metropolitan area. b The Office of Disease Prevention and Control. c An additional 17% of the required sample size for the survey within Bangkok was calculated due to concerns about a low participation rate. Laboratory methodology Smear Two samples (spot, morning): direct preparation, ZN Culture Two samples (spot, morning): direct preparation, Ogawa modified Kudoh Identification of MTB Immunochromatographic assay TB drug susceptibility test Not done Xpert® MTB/RIF Done after the study for quality assurance only using smear-positive culture-negative samples HIV test Not done Screening criteria Interviewa Cough ≥2 weeks (3 points) Haemoptysis over the past month (3 points) Cough <2 weeks (2 points) Weight loss in the past month (1 point) Fever ≥1 week in the past two weeks (1 point) Night sweats in the past month (1 point) Screened positive: total score ≥3 or score ≥1 with chest X-ray exempted Chest X-rayb Any lung abnormality Other N/A a An in-depth interview about health-care seeking behaviour was done for those who screened positive and/or those who were currently on TB treatment. b Direct digital radiography. Field data collection Paper Database iDataFax Method of analysis MI+IPW Results first published in a report/paper Pending Official dissemination event Pending Analysis and reporting Key survey results (non-Bangkok survey) Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 104 55–195 242 176–332 Male 159 82–306 376 264–535 Female 51 23–117 115 71–184 15–24 years 22 2.2–215 218 86–555 25–34 years 126 44–362 186 91–380 35–44 years 49 18–132 134 69–259 45–54 years 109 27–439 265 146–482 55–64 years 164 92–293 295 187–463 ≥65 years 204 108–384 465 290–743 Urban 147 48–445 286 158–518 Rural 82 55–122 220 170–284 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 283 – Location of care sought • Consulted medical facility N/A N/A Public facility N/A N/A Private facility N/A N/A Other N/A N/A • Pharmacy N/A N/A • Traditional centre N/A N/A Self-treated N/A N/A No action taken N/A N/A Unknown N/A N/A a Clinical score ≥3 or score ≥1 with chest X-ray exempted. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). Survey participants currently on TB treatment Number % Total participants currently on TB treatment 66 – • Treated in the public sector 53 80 • Treated in the private sector 3 4.5 • Treated in other sector 3 4.5 • Treated in unknown sector 7 11 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 6 4.2 Design effect k Smear-positive TB 3.3 1.7 Bacteriologically confirmed TB 2.7 1.0 Other sputum results Number % Total smear-positive participants 75 – Smear-positive participants without MTB confirmationa 29 39 Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. THAILAND 227 Survey flow: census to final outcomes (non-Bangkok survey) Field operations: February to September 2012 Individuals enumerated in census N/A Eligible study population 78 839 Total participants 62 536 (79%) Interview and chest X-ray 61 965 (99%) Interview only 571 (1.0%) Chest X-ray only 0 (0%) a Eligible for sputum collection. b The result was missing or could not be read due to the poor quality of the chest X-ray. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture, but at least one smear-positive with chest X-ray suggestive of TB, or two smear-positive, or one smear-positive and confirmed as TB cases by referral health facilities. e Definite: MTB confirmed by culture (one significant culture-positive, or two scanty culture-positive, or one scanty culture-positive with chest X-ray suggestive of TB), or confirmed as TB cases by referral health facilities. Probable: no definition. Ineligible individuals N/A Children <15 years N/A Did not meet residency criteria N/A Eligible for sputum examination 6 050 (9.7%) Total bacteriologically confirmed cases 142 Symptom positive, chest X-ray positive 526 (8.7%) Symptom positive, chest X-ray negative or N/A 1 757 (29%) Symptom negative, chest X-ray positive 3 767 (62%) Other N/A Symptom positive, chest X-ray positive 42 (30%) Symptom positive, chest X-ray negative or N/A 6 (4.2%) Symptom negative, chest X-ray positive 94 (66%) Other N/A Submitted specimens At least one specimen 5 988 (99%) Both specimens 5 720 (95%) Smear-positive casesd 58 (41%) Definite 46 Probable 12 Smear-negative casese 84 (59%) Definite 84 Probable N/A Laboratory result At least one culture result availablec 5 821 (96%) Symptom screening Cough ≥2 weeks 1 775 (2.8%) Cough <2 weeks 4 620 (7.4%) Haemoptysis over the past month 121 (0.2%) Weight loss in the past month 1 679 (2.7%) Fever ≥1 week in the past two weeks 752 (1.2%) Night sweats in the past month 1 057 (1.7%) Clinical score ≥3a 2 260 (3.6%) Clinical score=1 or 2 with chest X-ray exempteda 23 (0.04%) Total symptom-screen positivea 2 283 (3.7%) Chest X-ray screening Normal 54 968 (88%) Abnormala 4 293 (6.9%) Other abnormality 2 682 (4.3%) Result not availableb 22 (0.04%) Total chest X-rays taken 61 965 National TB prevalence surveys 2007–2016228 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 80 90 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 150 200 250 300 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 50 100 150 250 300 200 350 400 450 500 0 Prevalence per 10 0 0 0 0 population 10 000 5 000 20 000 15 000 25 000 30 000 35 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 800 400 600 500 700 200 300 100 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 1.71, variance 3.16, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 15 5 10 25 20 0 0 1 2 3 4 5 6 7 8 Total (≥15) THAILAND 229 Background Thailand’s population was 67 million in 2012, and the average gross national income (GNI) per person was US$  5590 per year, making it an upper-middle-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2012, the prevalence of HIV in the general population aged 15–49 years was 1.2% (95% confidence interval [CI]: 1.0–1.4%) (2), and it was estimated that 14% (95% CI: 13–16%) of TB patients were coinfected with HIV (3). In 1996, the Government of Thailand began to implement the WHO-recommended DOTS strategy (4,5). By 2001, all districts had at least one public health-care facility implementing DOTS. The case notification rate for all forms of TB (new and relapse cases) decreased from 82 per 100 000 population in 1990 to 55 per 100 000 population in 2000, then increased to 101 per 100 000 population in 2010. The case detection rate (notifications of new cases divided by estimated incidence) reached 80% (95% CI: 67–97%) in 2011 (6). Surveillance data from routine TB notification and vital registration systems were available to estimate the burden of TB disease in Thailand; however, underreporting from hospitals and the private sector limited their accuracy. Before 2012, Thailand had already carried out four national TB prevalence surveys: in 1962, 1977, 1991–1992 and 2006. The observed prevalence of bacteriologically confirmed TB declined from 500 per 100 000 population (among those aged ≥15 years) in 1962 to 310 per 100 000 population (among those aged ≥15 years) in 1977 and 240 per 100 000 population (among those aged ≥10 years) in 1991 (7). Although the 2006 survey used interviews and chest X-rays for screening, and culture for diagnostic confirmation, the survey could not be used to estimate prevalence due to a low participation rate (56%), untimely reading of chest X-rays (this was only done after each field cluster operation), and a low sputum submission rate from eligible participants (19%). The fifth national survey was implemented in 2012–2013. Key methods and results (non-Bangkok survey) All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (8). The survey was undertaken in two phases: Phase 1 covered the non- Bangkok areas of the country from February to September 2012, and Phase 2 covered metropolitan Bangkok from April to July 2013. Due to political instability at the time, and the low participation rate in urban settings in previous surveys, the survey design anticipated operational difficulties in metropolitan Bangkok. Therefore, Phase 1 (non-Bangkok areas) was designed as an entirely independent survey that would provide a large enough sample to estimate TB prevalence in non-Bangkok areas; the estimated sample size was 74  700 in 83 clusters. Phase 2 (metropolitan Bangkok) was allocated 17 clusters to complement Phase 1, with an estimated sample size of 15 300. Phase 1 was successfully completed with a participation rate of 79%. In Phase 2, most residents were not available, resulting in a participation rate of 26%. Therefore, the national TB programme (NTP) and the survey team decided to report only on the results of the Phase 1 survey. To estimate national prevalence, the prevalence per 100 000 population in the urban clusters of Phase 1 was assumed to be similar to the prevalence per 100 000 population in the Bangkok region. Phase 1 included two strata (urban and rural), with a target cluster size of 900 individuals. No data were available on the numbers of individuals enumerated in the household census. 78  839 people were eligible and invited to participate. Of these, 62  536 (79%) did so. A total of 6050 participants (9.7%) were eligible for sputum examination based on chest X-ray and symptom screening. Of these, 5988 (99%) submitted at least one sputum specimen and 5720 (95%) submitted two sputum specimens. A total of 142 bacteriologically confirmed pulmonary TB cases were identified, including 58 cases of smear-positive TB. The prevalence of bacteriologically confirmed TB was 242 (95% CI: 176–332) per 100 000 population (among those aged ≥15 years), and for smear-positive TB it was 104 (95% CI: 55–195) per 100 000 population. There was no significant difference between urban strata (286 per 100 000 population; 95%  CI: 158–518) and rural strata (220 per 100 000 population; 95% CI: 170–284). National TB prevalence surveys 2007–2016230 Other key results were: • the male to female ratio for TB prevalence was 3.1 for smear-positive TB and 3.3 for bacteriologically confirmed TB; • the highest prevalence per 100 000 population was in those aged 45 years or more, and the absolute number of bacteriologically confirmed cases was also relatively high in the older age groups; • among bacteriologically confirmed TB cases, 34% were symptom-screen positive, and among the smear-positive cases, 48% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 1.8 overall, but varied from 0.8 in those aged 15–24 years to 2.7 in the 25–34 years age group, and was slightly higher for men than for women (1.9 versus 1.6); and • among bacteriologically confirmed TB cases, 95% had no previous history of anti-TB treatment, and only 4.2% were on anti-TB treatment at the time of the survey. Data on health-care seeking behaviour among TB cases were not available. Implications of results The survey showed that there was still a high burden of TB, and that the disease remained a public health threat. The updated national estimate (for all ages and all forms) of TB prevalence (236 per 100 000 population; 95% CI: 161–326) was higher than the pre-survey 2011 WHO estimate (182 per 100 000 population; 95% CI: 80–300) (9, 10). However, this did not necessarily mean that the burden of TB had been increasing. Combining the results with data from previous surveys, as well as adjusting for the fact that the 2012 survey methods were more sensitive than those of previous surveys (owing to the use of direct chest X-ray with a digital system and the improved quality of culture testing in regional laboratories guided by the National TB Reference Laboratory), TB prevalence was still estimated to be declining, although to only a limited extent. Assuming there were very few cases in those aged 10–15 years in 2012, the prevalence of bacteriologically confirmed TB in 2012 (242 per 100 000 population; 95%  CI: 176–332, ≥15 years) was similar to the estimate from the 1991–1992 survey (240 per 100 000 population, ≥10 years) (7). The 2012 survey even suggested that the prevalence of smear-negative culture-positive TB had increased. This may in part have been due to the impact of the HIV epidemic on the number of TB cases in the Photo credit: Sirin Jittimanee THAILAND 231 late 1990s, but it may also reflect the higher sensitivity of methods used to detect smear-negative culture-positive TB in the 2012 survey. In addition, since programmatic efforts prioritized the detection and treatment of smear- positive TB cases, the impact of TB control efforts was more likely to be seen in the prevalence of smear-positive TB. The observed prevalence of smear-positive TB in 1991–1992 (170 per 100 000 population, ≥10 years) was higher than the level found in 2012 (104 per 100  000 population; 95% CI: 55–195, ≥15 years) (7). Nonetheless, smear-positive TB accounted for only 41% of the total number of prevalent bacteriologically confirmed TB cases in the 2012 survey. Age-specific estimates of TB prevalence in 2012 also suggested a long-term decline in the burden of TB. Those aged 45 years or more accounted for more than two thirds of TB cases, suggesting that reactivation of infection from the past was playing a greater role than recent infection. However, an ageing population also contributed to a relatively slow decline in the overall burden of TB. The survey team traced the treatment provided to bacteriologically confirmed TB cases detected by the survey and found that only 45% (64/142) of patients had started treatment at a designated health centre or a nearby public hospital. Of the remainder, six died, four refused treatment and six were diagnosed as non-TB by hospitals; information was lacking on eight people who had moved outside of the survey site and on 54 (38%) for whom health facilities did not provide data, thereby limiting the analysis of treatment provision. The geographical variation in TB was also of concern. Among 142 bacteriologically confirmed patients detected by the survey, 81 (57%) were from the economically less- developed north-eastern region. Although confidence intervals were wide, results suggested that the level of TB prevalence in the north-eastern region could be more than twice that of other regions in Thailand. The classical pathway to detect TB (i.e. chronic cough recognition to diagnosis by smear) would only have detected one-fifth of the bacteriologically confirmed TB survey cases (26/142). This showed the need for wider use of chest X-rays and more sensitive tools, such as molecular technologies, in the diagnostic pathway. It was also evident that the case notification system needed improvement; for example, by introducing and monitoring mandatory notification of designated infectious diseases, including TB. Based on the survey, as many as 20% of the cases on anti-TB treatment may have been missed in the TB surveillance system. Major successes, challenges and lessons learned The national TB prevalence survey (non-Bangkok areas) was successfully carried out, and it provided the NTP and partners with a large and rich set of data. Estimates of TB prevalence based on the survey were more accurate and precise than those previously available, and trends were updated to show a slight decline overall. The delay in starting Phase 2 made it hard to find staff to write and publish the official survey report, because many survey team members had moved to other positions. Nonetheless, the results were used in a timely manner to update the national TB strategy and plan. It may not be possible to implement another national prevalence survey in Thailand in the future, given the difficulties in recruitment within the Bangkok region and in urban clusters, where the participation rate was only 65% overall. This was despite the extended hours of field activities in urban areas, including operating in the evenings and on weekends. Rapid urbanization means that the proportion of the population living in urban areas will continue to increase. In 2010, 44% of the population lived in urban areas, and this was projected to increase to 56% by 2020 (11). Thus, to accurately monitor the burden of, and trends in, TB disease, either alternative survey methods would need to be developed or (preferably) the Photo credit: Sirin Jittimanee National TB prevalence surveys 2007–2016232 surveillance system should be strengthened to meet the necessary quality and coverage standards, in particular to address the problem of underreporting of detected TB cases. The complexity of a multistage cluster sampling design and probable underrepresentation of the urban population complicated the analysis (only 15 of the 83 clusters were in urban settings). In addition, there was also a large difference between the registered population and the actual number of people enumerated in the survey. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013; (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf, accessed January 2018). 7. Sriyabhaya N, Payanandana V, Bamrungtrakul T, Konjanart S. Status of tuberculosis control in Thailand. Southeast Asian J Trop Med Public Health. 1993;24(3):410–419; (http://www.tm.mahidol. ac.th/seameo/1993-24-3/1993-24-3-410.pdf, accessed January 2018). 8. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011. (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 9. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015; (http://apps.who.int/iris/ bitstream/10665/191102/1/9789241565059_eng.pdf, accessed January 2018). 10. World Health Organization. Global tuberculosis control: WHO report 2011. Geneva: WHO; 2011; (http://apps.who.int/iris/ bitstream/10665/44728/1/9789241564380_eng.pdf, accessed January 2018). 11. Department of Economic and Social Affairs. World urbanization prospect: the 2014 revision (CD Rom edition). United Nations; 2014; (https://esa.un.org/unpd/wup/CD-ROM/, accessed January 2018) 233 Name Role Organization Frank Mugabe Principal investigator - policy Ministry of Health Elizeus Rutebemberwa Principal investigator - technical School of Public Health, Makerere University Bruce Kirenga Co-principal investigator School of Public Health, Makerere University Samuel Kasozi Study coordinator School of Public Health, Makerere University Harriet Kisembo Study investigator, lead radiologist Mulago hospital Okot Martin Nwang Study investigator Senior consultant pulmonologist and head of medical panel, Mulago hospital & complex William Worodria Study investigator Department of Medicine Mulago hospital & complex Abel Nkolo Study investigator WHO Uganda Emily Bloss Study investigator US Centers for Disease Control and Prevention (CDC) Moses Joloba Survey laboratory consultant and director National TB Reference Laboratory Kenneth Musisi Laboratory manager National TB Reference Laboratory Rogers Sekibira Data manager School of Public Health, Makerere University Ronald Anguzu Field team leader School of Public Health, Makerere University Annet Nagudi Field team leader School of Public Health, Makerere University Racheal Tumwebaze Field team leader School of Public Health, Makerere University Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Peou Satha Technical assistance (radiology) Consultant, Cambodia Julia Ershova Technical assistance (data management) US Centers for Disease Control and Prevention (CDC) Charalampos Sismanidis Technical assistance (analysis) WHO headquarters Key people Surveyed clusters (N=70)a Data sources ■ Report on the population-based survey of prevalence of tuberculosis disease in Uganda 2014–15. Kampala, Uganda: Makerere University School of Public Health (pending official publication). ■ Survey dataset. Summary statistics Participation rate 91% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 401 4.1 Prevalence:notification ratio (Bacteriologically confirmed TB, ≥15 years) 2.8 Finance Amount (US$) The Global Fund 2 841 452 Total budget 2 841 452 Survey organization and financing Implementing agency: Makerere University School of Public Health UGANDA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2014–2015 National TB prevalence surveys 2007–2016234 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Region/district/village Sample size assumptions   • Smear-positive prevalence 269 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.5 • k 0.6 • Response rate 85% • Sample size (estimated) 40 180 Number of clusters 70 Cluster size 580 Eligibility criteria   • Age ≥15 years • Residency Permanent residents who stayed at least one night in the past two weeks; temporary visitors who arrived at least two weeks before census day Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): direct preparation, ZN Culture Two samples (spot, morning): concentrated preparation, LJ media Identification of MTB MPT64 rapid test TB drug susceptibility test Not done as per protocol (post-survey study) Xpert® MTB/RIF Done on smear-positive specimens and/or if both samples were culture contaminated HIV test Offered to all participants who screened positive Screening criteria Interview Cough ≥2 weeks Chest X-raya Any lung abnormality Other Chest X-ray exempted a Conventional radiography. Field data collection Paper Database Microsoft® Access Method of analysis MI+IPW Results first published in a report/paper August 2017 Official dissemination event August 2017 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 174 111–238 401 292–509 Male 314 216–413 734 554–914 Female 70 25–114 178 109–248 15–24 years 124 50–198 228 117–338 25–34 years 191 98–284 442 291–592 35–44 years 294 162–425 624 379–869 45–54 years 164 25–303 565 280–850 55–64 years 254 26–481 636 277–995 ≥65 years 85 2–205 570 261–879 Urban 191 113–270 504 355–652 Rural 169 91–248 370 237–504 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 2 714 – Location of care sought • Consulted medical facility 1 201 44 Public facility 1 038 86 Private facility 146 12 Others (NGO) 17 1.4 • Pharmacy 421 16 • Traditional centre 11 0.4 Self-treated 22 0.8 No action taken 1 059 39 Unknown 0 0 a Cough ≥2 weeks. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 61 – • Treated in the public sector 57 93 • Treated in the private sector 4 7 • Treated in unknown sector 0 0 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 16 10 Design effect k Smear-positive TB 1.8 0.9 Bacteriologically confirmed TB 2.5 0.8 Other sputum results Number % Total smear-positive participants 91 – Smear-positive participants without MTB confirmationa 25 27 Isolates with MDR-TB detected 0 0 a Age ≥15 years unless otherwise specified. UGANDA 235 Survey flow: census to final outcomes Field operations: October 2014 to July 2015 Individuals enumerated in census 86 108 Eligible for sputum examination 5 142 (13%) Total bacteriologically confirmed cases 160 Eligible study population 45 293 (53%) Total participants 41 154 (91%) Interview and chest X-ray 41 003 (99.6%) Interview only 151 (0.4%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 552 (11%) Symptom positive, chest X-ray negative or N/A 2 162 (42%) Symptom negative, chest X-ray positive 2 298 (45%) Otherb 130 (2.5%) Symptom positive, chest X-ray positive 63 (39%) Symptom positive, chest X-ray negative or N/A 16 (10%) Symptom negative, chest X-ray positive 81 (51%) Other 0 (0%) Submitted specimens At least one specimen 4 844 (94%) Both specimens 4 532 (88%) Smear-positive casesd 66 (41%) Definite 66 Probable 0 Smear-negative casese 94 (59%) Definite 94 Probable 0 Laboratory result At least one culture result availablec 4 758 (93%) Symptom screening Cough ≥2 weeksa 2 714 (6.6%) Haemoptysis 263 (0.6%) Sputum production 4 819 (12%) Chest pain 12 142 (30%) Fever 3 593 (8.7%) Total symptom-screen positivea 2 714 (6.6%) Chest X-ray screening Normal 37 775 (92%) Abnormala 2 850 (6.9%) Other abnormality 378 (0.9%) Result not available 0 (0%) Total chest X-rays taken 41 003 a Eligible for sputum collection. b Chest X-ray exempted and symptom-screen negative. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture and/or Xpert. Probable: MTB not confirmed by culture and/or Xpert, but chest X-ray consistent with TB. e Definite: MTB confirmed by culture and/or Xpert with chest X-ray consistent with TB. Probable: culture weak positive (<20 colonies) in one sample and Xpert pending or N/A without negative evidence on chest X-ray (i.e. chest X-ray not taken). Ineligible individuals 40 815 (47%) Children <15 years 36 801 (43%) Did not meet residency criteria 4 014 (4.7%) National TB prevalence surveys 2007–2016236 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 100 150 200 300250 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 0 Prevalence per 10 0 0 0 0 population 5 000 15 000 20 000 25 000 10 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of bacteriologically confirmed TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 4.0 3.5 4.5 5.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB 1000 300 400 500 600 700 800 900 200 100 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.29, variance 5.42, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using bacteriologically confirmed pulmonary TB notifications (2015) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 5 10 15 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Total (≥15) UGANDA 237 Background The population of Uganda, in East Africa, was 37 million in 2014, with 48% aged under 15 years and 82% living in rural areas. The average gross national income (GNI) per person in 2014 was US$ 690, making it a low-income country (1). Uganda was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015). The prevalence of HIV in the general population aged 15–49 years in 2014 was 7.1% (95% confidence interval [CI]: 6.7–7.7%) (2), with declines evident since the early 1990s. In 2014, it was estimated that 45% (95% CI: 42–48%) of TB patients were coinfected with HIV (3). The National Tuberculosis and Leprosy Program (NTLP) was established in 1990, within the National Disease Control Department of the Ministry of Health (MoH). In 2015, the programme was staffed with a national programme manager and zonal TB and leprosy supervisors. The NTLP was responsible for policy formulation, planning, training, resource mobilization and setting standards for TB control. At the district level, TB control activities were the responsibility of district TB and leprosy supervisors, with oversight provided by the district health officer. At the time of the survey design in 2008, WHO estimated that there were 311 (95% CI: 249–373) new TB cases per 100  000 population per year, equivalent to 98  356 new cases of TB per year (95% CI: 78 685–118 027). However, there was considerable uncertainty about estimates of the burden of TB disease, given that no national TB prevalence survey had ever been done, no direct measurements of TB mortality were available from vital registration, and the gap between notifications and incidence (due to underreporting or under-diagnosis of cases) had not been quantified and was difficult to estimate. For these reasons, as well as Uganda’s share of the global and regional TB burden, the country was one of the 22 global focus countries for a national TB prevalence survey selected by the WHO Global Task Force on TB Impact Measurement in December 2007. With the new opportunity of funding for a survey from the Global Fund to Fight AIDS, Tuberculosis and Malaria, in 2008 the MoH decided to implement a national TB prevalence survey. Following various challenges and delays, survey field operations started in October 2014 and were completed in July 2015 (4). Photo credit: Julia Ershova National TB prevalence surveys 2007–2016238 Key methods and results There were 70 survey clusters in two strata (urban and rural), with a target cluster size of 580 individuals. A total of 86 108 individuals from 17 535 households were enumerated in the survey census, of whom 45 293 (53%) were eligible and invited to participate. Of these, 41 154 (91%) did so. All participants were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and an interview about symptoms (5). A total of 5142 participants (13%) were eligible for sputum examination, of whom 4844 (94%) submitted at least one sputum specimen and 4532 (88%) submitted two sputum specimens. A total of 160 bacteriologically confirmed pulmonary TB cases were identified, including 66 cases of smear-positive TB. The prevalence of smear-positive TB was 174 (95% CI: 111–238) per 100 000 population and for bacteriologically confirmed TB it was 401 (95% CI: 292–509) per 100 000 population. The prevalence of bacteriologically confir- med TB was highest in those aged 35–44 years, at 624 (95% CI: 379–869) per 100 000 population. The prevalence of bacteriologically confirmed TB was higher in urban areas than in rural areas: 504 (95% CI: 355–652) per 100 000 population in urban areas and 370 (95% CI: 237–504) per 100 000 population in rural areas. Other key results were: • the male to female ratio was 4.5 for smear- positive TB and 4.1 for bacteriologically confirmed TB; • prevalence increased with age, up to the age group 35–44 years, and it was consistently high in older age groups; however, the absolute number of bacteriologically confirmed TB cases was relatively high in younger age groups; • of the TB survey cases, 49% were symptom- screen positive, and of the smear-positive cases, 55% were symptom-screen positive; • for bacteriologically confirmed TB, the ratio of prevalence to notifications (P:N ratio) was 2.8 overall, but varied from 2.4 in those aged 45–54 years to 4.6 in the 65 years and over age group, and was higher for men than for women (3.7 versus 2.0); • of the TB survey cases, 84% had no previous (or current) history of anti-TB treatment and 10% were on anti-TB treatment at the time of the survey; and • of the 66 bacteriologically confirmed and 20 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 27 (41%) and 14 (70%), respectively, had previously sought care in a public or private health facility for their symptoms. A total of 5142 individuals who screened positive were eligible for HIV testing, but 756 (15%) of these individuals were not tested during the survey. For those who were tested, 422 (9.6%) were found to be HIV-positive. Of the 160 bacteriologically confirmed cases, 15 (9%) did not have an HIV test and, of those tested, 39 (27%) were HIV- positive. Implications of results The estimated TB prevalence for all ages and all forms of TB based on the results from the survey (253 per 100 000 population; 95% CI: 191–315) was much higher than the pre-survey estimates (154 per 100 000 population; 95% CI: 85–243) (6). At the time of the survey, Uganda’s NTLP screened for TB disease using chronic cough (i.e. cough >2 weeks). However, in the prevalence survey, half of the bacteriologically confirmed cases were initially identified for diagnostic testing based only on chest X-ray screening. This suggested that the NTLP should seriously consider ways of improving access to chest X-ray services. Since it was recognized that it might take time to expand such screening, a need for more research about how to improve symptom screening was also identified. Photo credit: Marina Tadolini UGANDA 239 Urban areas had a higher prevalence per 100 000 popu- lation than rural areas, and there were three times more cases of TB among men than women. Thus, the NTLP needed to give more attention to ensuring access to screening and enrolment on treatment among men and for people living in urban areas. The TB/HIV data showed that integration of HIV services with anti-TB treatment should be continued. Of participants with smear-positive specimens, 28% (25/91) did not have Mycobacterium tuberculosis (MTB). Thus, a smear-positive result alone was not adequate for the detection of TB cases, especially in the context of intensified case finding or active case detection strategies. Before the survey, Uganda was in the list of 22 HBCs as defined by WHO. The survey identified a higher prevalence than expected, but the results were available only after a new list of 30 HBCs was defined by WHO for the period 2016–2020. The need for good communication between all levels of WHO, the NTLP and the MoH to determine the consequences of Uganda not being in the list of 30 TB HBCs (although it remained on the list of high TB/HIV burden countries) was recognized. Major successes, challenges and lessons learned It was a major success to implement the country’s first national TB prevalence survey. Survey preparations started in 2008, but there were long delays primarily due to the challenge of securing funding. Whereas prevalence surveys in other countries introduced technologies such as digital chest X-ray or electronic data collection in the field, the Ugandan survey used only conventional chest X-ray equipment and paper-based data collection because of limitations on funding and time. Data were entered into the database at central level upon completion of each cluster operation. Nonetheless, the quality of the survey was exceptionally high. There were no major delays in survey implementation resulting from major accidents or equipment failure. However, the X-ray machine often had to be restarted due to excessive humidity and heat, which affected the auto- film processor. Therefore, X-ray examinations were often interrupted, and participants were kept waiting. A high participation rate, both in rural and urban clusters, was achieved due to the dedication of the central and field teams, community involvement and careful preparation of survey operations (especially in big cities). A high sputum collection rate was also achieved. Uganda’s survey was also one of the first to provide high-quality data on Photo credit: Irwin Law Photo credit: Julia Ershova National TB prevalence surveys 2007–2016240 TB/HIV coinfection with a large proportion of survey participants requesting to be tested. The biggest advantage of the survey in Uganda was that the country had its own National TB Reference Laboratory in Kampala – one of only a few laboratories in Africa qualified as a supranational reference laboratory. This laboratory produced highly reliable results for both Xpert® MTB/RIF and culture testing. A large data management team at the central level cleaned and validated data in a timely and systematic way. This allowed the final validated dataset to be available within a few months of the end of the survey. Given the elements described above, the survey in Uganda was one of the highest quality prevalence surveys in Africa. Clear demarcation between the terms of reference of the NTLP manager and the head of research team that actually undertook the survey facilitated smooth field operations and post-survey work. The survey team lacked experience in conducting a TB prevalence survey, because those trained through a WHO workshop had left the survey team by the start of the survey. This lack of experience resulted in some census sampling errors and lack of community participation during the pilot. Nonetheless, major challenges were ultimately solved by the survey team, with the support of intensive technical assistance during the early phases of the survey. References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. Report on the population-based survey of prevalence of tuberculosis disease in Uganda 2014–15. Kampala, Uganda: Makerere University School of Public Health 2018. 5. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 6. World Health Organization. Global tuberculosis report 2014. Geneva: WHO; 2014 (http://apps.who.int/iris/ bitstream/10665/137094/1/9789241564809_eng.pdf ?ua=1, accessed January 2018). 241 Name Role Organization Saidi M. Egwaga Principal investigator (PI) National Tuberculosis and Leprosy Programme (NTLP) Godfrey S. Mfinanga Co-PI National Institute for Medical Research (NIMR), Muhimbili Medical Research Center Deusdedit V. Kamara Survey coordinator NTLP Senkoro Mbazi Assistant survey coordinator NIMR, Muhimbili Medical Research Center Ahmed Khatib Programme manager Zanzibar Tuberculosis and Leprosy Programme Basra Doulla Laboratory manager Central TB Reference Laboratory (CTRL), NTLP Lulu Fundikira Radiology coordinator Muhimbili University of Health and Allied Sciences (MUHAS) Raymond P. Shirima Data manager NTLP Blasdus F. Njako Field team leader NTLP Msaki John Field team leader NIMR Rahim Ishumi Field team leader NIMR Lugano Mtafya Field team leader NIMR Moses Ringo Field team leader NIMR Frank van Leth Technical assistance (survey advisor) KNCV Tuberculosis Foundation Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Charalampos Sismanidis Technical assistance (analysis) WHO headquarters Key people Surveyed clusters (N=62)a Data sources ■ The First National Tuberculosis Prevalence Survey in the United Republic of Tanzania, final report: Ministry of Health and social welfare; 2013. ■ M. Senkoro, S. Mfinanga, S. Egwaga, R. Mtandu, D.V. Kamara, D. Basra, et al. Prevalence of pulmonary tuberculosis in adult population of Tanzania: a national survey, 2012. Int J Tuberc Lung Dis 20(8):1014–1021. ■ M. Senkoro, S.G. Hinderaker, S.G. Mfinanga, N. Range, D.V. Kamara, S. Egwaga, et al. Health care-seeking behaviour among people with cough in Tanzania: findings from a tuberculosis prevalence survey. Int J Tuberc Lung Dis 19(6):640–646. ■ Survey dataset. ■ Summary statistics Participation rate 77% Smear-positive TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 275 2.3 Prevalence:notification ratio (smear-positive TB, ≥15 years) 3.0 Finance Amount (US$) PATH/USAIDS 29 673 The Global Fund 2 611 312 Other partners 521 184 MOH, Tanzania 200 000 Total budget 3 362 169 Survey organization and financing Implementing agency: The National Tuberculosis and Leprosy Programme (NTLP) UNITED REPUBLIC OF TANZANIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2011–2012 National TB prevalence surveys 2007–2016242 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/semi-urban/rural/Zanzibar Sampling unit Four strata/district/ward Sample size assumptions • Smear-positive prevalence 261 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.6 • k 0.6 • Response rate 80% • Sample size (estimated) 46 792 Number of clusters 62 Cluster size 750 Eligibility criteria • Age ≥15 years • Residency Slept for the past 2 weeks in the household prior to the census Survey design and methodology Laboratory methodology Smear Three samples (spot, morning and spot; both spot samples were examined in the field, and a morning sample was examined in the central laboratory): direct preparation, FM (LED, auramine stain) Culture One sample (morning): concentrated preparation, LJ media Identification of MTB PNB TB drug susceptibility test Done at the Antwerp SRL, not as part of the original protocol Xpert® MTB/RIF Done only for smear-positive slides to confirm the presence of MTB at the Antwerp SRL, not as part of the original protocol HIV test Done for participants who screened positive Screening criteria Interviewa Cough ≥2 weeks or haemoptysis or fever ≥2 weeks or weight loss or excessive night sweats Chest X-rayb Any lung (or mediastinum) abnormality Other N/A a An in-depth interview was done only for those who screened positive, to obtain information on demographics, risk factors for TB, knowledge about TB and health- care seeking behaviour. b Mobile X-ray unit, computed radiography. a Reported prevalence results are based on a re-analysis by WHO. Field data collection Paper Database EpiData Method of analysis Cluster-levela Results first published in a report/paper August 2013 Official dissemination event July 2013 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 275 232–326 N/A N/A Male 407 319–494 N/A N/A Female 179 130–228 N/A N/A 15–24 years 51 13–88 N/A N/A 25–34 years 280 178–381 N/A N/A 35–44 years 316 199–433 N/A N/A 45–54 years 241 123–359 N/A N/A 55–64 years 462 264–660 N/A N/A ≥65 years 662 436–888 N/A N/A Urban 328 184–471 N/A N/A Semi-urban 302 201–404 N/A N/A Rural 268 210–327 N/A N/A Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 3 388 – Location of care sought • Consulted medical facility 481 14 Public facility (incl. mission hospital) 445 93 Private facility 36 7.5 • Pharmacy 147 4.3 • Traditional centre 11 0.3 • Otherb 412 12 Self-treated N/A N/A No action taken 1 688 50 Unknown 649 19 a Data on health-care seeking behaviour were only available for participants who reported cough ≥2 weeks and/or haemoptysis. b This included 257 dispensaries and 155 unspecified locations. a This could not be calculated because not all 162 smear-positive participants were tested with culture and/or Xpert MTB/RIF. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 88 – • Treated in the public sector N/A N/A • Treated in the private sector N/A N/A • Treated in other sector N/A N/A Smear-positive TB cases detected by the survey who were currently on TB treatment 5 3.7 Design effect k Smear-positive TB 1.9 0.6 Bacteriologically confirmed TB N/A N/A Other sputum results Number % Total smear-positive participants 162 – Smear-positive participants without MTB confirmationa N/A N/A Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. No TB cases were identified in Zanzibar. UNITED REPUBLIC OF TANZANIA 243 Survey flow: census to final outcomes Field operations: December 2011 to November 2012 Individuals enumerated in census 137 547 Eligible for sputum examination 6 302 (13%) Eligible study population 65 664 (48%) Total participants 50 447 (77%) Interview and chest X-ray 47 455 (94%) Interview only 2 383 (4.7%) Chest X-ray only 609 (1.2%) Symptom positive, chest X-ray positive 804 (13%) Symptom positive, chest X-ray negative or N/A 3 459 (55%) Symptom negative, chest X-ray positive 2 039 (32%) Other N/A Symptom positive, chest X-ray positive 55 (41%) Symptom positive, chest X-ray negative or N/A 18 (13%) Symptom negative, chest X-ray positive 48 (36%) Otherd 13 (9.7%) Submitted specimens At least one specimen 5 768 (92%) Three specimens 4 705 (75%) Smear-positive casesc 134 Definite 120 Probable 14 Laboratory result At least one culture result availableb 4 723 (75%) Symptom screening Cough ≥2 weeksa 3 238 (6.5%) Haemoptysisa 536 (1.1%) Fever ≥2 weeka 932 (1.9%) Weight lossa 531 (1.1%) Night sweatsa 1 188 (2.4%) Total symptom-screen positivea 4 263 (8.6%) Chest X-ray screening Normal 45 123 (94%) Abnormala 2 843 (5.9%) Other abnormality N/A Result not available 98 (0.2%) Total chest X-rays taken 48 064 a Eligible for sputum collection. b Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. c Definite: MTB confirmed by culture (NTRL) and/or Xpert (Antwerp SRL). Probable: MTB not confirmed by culture or Xpert, but chest X-ray final reading “consistent with TB”. Please see the main text for further details. d 13 were screened negative, and these people were not part of the total number of people eligible for sputum examination. The reason for their sputum submission was unknown. Ineligible individuals 71 883 (52%) Children <15 years 57 081 (41%) Did not meet residency criteria 14 802 (11%) National TB prevalence surveys 2007–2016244 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 1.0 2.0 3.0 4.0 6.0 5.0 7.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 4: Cluster variation of the number of smear-positive TB casesb Number of smear-positive TB cases N um be r of c lu st er s 10 5 20 15 0 0 1 2 3 4 5 6 7 8 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 400 600 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Smear-positive TB Fig. 3: Estimated number of smear-positive TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 700 0 Prevalence per 10 0 0 0 0 population 10 000 15 000 20 000 25 000 5000 Es ti m at ed n um be r of s m ea r- po si ti ve T B ca se s 1000 500 400 700 600 900 800 300 100 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of smear-positive TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 2.16, variance 4.24, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Total (≥15) UNITED REPUBLIC OF TANZANIA 245 Background In 2012, the United Republic of Tanzania, in East Africa, had a population of 48 million. It was a low-income country with an average gross national income (GNI) per person of US$  780 per year (1). The United Republic of Tanzania was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2012, the prevalence of HIV in the general population aged 15–49 years was 5.2% (95% confidence interval [CI]: 4.6–5.9%) (2), and it was estimated that 38% (95% CI: 33–44%) of TB patients were coinfected with HIV (3).   The United Republic of Tanzania’s National Tuberculosis and Leprosy Programme (NTLP) was established in 1977. During the 1980s it became the first country in the world to use an approach to TB control that later became known as the DOTS strategy, and was considered a “model” DOTS programme (4). Before the national TB prevalence survey of 2012, WHO estimated the incidence of all forms of TB at 169 (95% CI: 159–180) per 100 000 population and the prevalence at 177 (95% CI: 93–286) per 100 000 population (5). These estimates were primarily based on data from case notifications, corrected for detection and reporting gaps of the surveillance system of TB cases, as best understood by experts. To move away from expert opinion and instead use a robust, nationally representative, direct measurement to estimate the burden of TB disease in the country, it was decided to conduct a national TB prevalence survey. A detailed survey protocol was first developed in 2006, but funds were only secured in 2010. In the intervening period, the United Republic of Tanzania became one of the 22 global focus countries for a national TB prevalence survey selected by the WHO Global Task Force on TB Impact Measurement. The survey started in December 2011 and was completed in November 2012 (6). Key methods and results In 2012, the United Republic of Tanzania became only the second country in Africa to complete a national TB prevalence survey that used the screening and diagnostic methods recommended in the latest guidance issued by WHO (7). There were 62 clusters sampled in four strata (urban, semi-urban, rural and Zanzibar) across the country, with a target cluster size of 750 individuals. A total of 137 547 individuals was enumerated in the survey census, of whom 65   664 (48%) were eligible (non-residents and children were ineligible) and invited to participate. Of these, 50 447 (77%) did so. They were screened according to the 2011 algorithm recommended by WHO; that is, using both a chest X-ray and an interview about symptoms (7). A total of 6302 participants (13%) screened positive and were eligible for sputum examination; of these, 5768 (92%) submitted at least one sputum specimen, and 4705 (75%) submitted three sputum specimens.   A total of 134 smear-positive pulmonary TB cases were identified in the survey. This translated into an estimate of smear-positive TB prevalence in the country, among those aged 15 years or more, of 275 (95% CI: 232–326) per 100 000 population in 2012. There were no significant Photo credit: Agatha Anthony Photo credit: Agatha Anthony National TB prevalence surveys 2007–2016246 2 To extrapolate the prevalence of bacteriologically confirmed TB among those aged ≥15 years to the prevalence of TB for all ages and all forms of TB, it was assumed that 45% of the general population were children, that extrapulmonary TB accounted for 23% (SD 9%) of all TB cases (based on 2008-2012 notification data) and that the ratio of childhood to adult TB was 0.07 (SD 0.03). differences among three geographical strata (urban, semi-urban and rural) in the level of smear-positive TB prevalence among those aged 15 years or more. No TB cases were identified in Zanzibar. Other key results were: • the male to female ratio for TB prevalence was 2.3 for smear-positive pulmonary TB; • prevalence increased with age, with a notably high level per 100 000 population (and estimated number of smear-positive cases) in those aged 25–44 years; • among smear-positive pulmonary TB cases, 44% were symptom-screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 3.0 overall, but varied from 1.3 in those aged 15– 24 years to 6.6 in those aged 65 years or more; the ratio was slightly higher for men than for women (3.3 versus 2.8); • among smear-positive pulmonary TB cases, 93% had no previous history of anti-TB treatment, and only 3.7% were receiving treatment at the time of the survey; and • among participants who screened-positive, 5.0% (318 of 6302) tested HIV-positive. No data on health-care seeking behaviour among smear- positive TB cases were collected. Unfortunately, the number of bacteriologically confirmed TB cases could not be validated. In the survey, field teams prioritized the detection of smear-positive individuals in the field and the early treatment of these people. Microscopy laboratories – using light-emitting diode fluorescence microscopy (LED FM) – were set up in every cluster site by senior laboratory staff from the Central TB Reference Laboratory (CTRL). However, this had an unintended negative effect on the quality of samples collected, and on testing by culture. For example, many samples (possibly as many as half) took more than a week to reach the CTRL; and a preliminary survey report stated there were 100 TB cases with a smear- positive result1 and 73 with a culture-positive result (6) - such a finding had never previously been observed in a national TB prevalence survey that followed the 2011 WHO guidelines (7). This in turn led to concerns that the number of culture-positive TB cases from the survey was underestimated. Following discussion and review of laboratory results with external partners including the country’s Supranational Reference Laboratory (SRL) in Antwerp, Belgium, as well as WHO, there was consensus that the culture results from the prevalence survey could not be used (8–10). In an attempt to confirm the validity of the smear- positive test results, and following discussions between the NTLP in the United Republic of Tanzania, the head of the Central TB Reference Laboratory (CTRL), the survey team, the KNCV Tuberculosis Foundation (the main technical partner) and WHO, it was agreed to send all specimen slides classified as smear-positive for testing using Xpert® MTB/RIF to the SRL Antwerp. A positive Xpert result would exclude false-positive microscopy as well as NTM (without the presence of MTB in the case of mixed infection). Results from the SRL (which became available in September 2014) concluded that “...an estimate of prevalence based on microscopy-positives could be justified” (11). The final case count, combining SRL Antwerp and survey CTRL results, was a total of 134 smear-positive TB cases, compared with 100 from the initial analysis based only on survey CTRL findings (8–10). This final count is the one used in this profile. To estimate the prevalence of bacteriologically confirmed TB among those aged 15 years or more, data from the neighbouring countries of Ethiopia, Malawi, Rwanda, Uganda and Zambia were used. From surveys in these five countries, the combined estimate of the ratio of bacteriologically confirmed to smear-positive TB was 2.16:1 (standard deviation [SD]: 0.46).2 This ratio was applied to the smear-positive prevalence estimate for the United Republic of Tanzania, resulting in an estimate of the prevalence of bacteriologically confirmed TB of 590 (95% CI: 330–860) per 100 000 population. A further step of extrapolation to all forms of TB and all ages resulted in an estimated TB prevalence of 443 (95% CI: 258–629) per 100 000 population. 1 The NTLP’s smear-positive case definition: two smear-positive specimens regardless of culture result, one smear-positive specimen with chest X-ray abnormality consistent with TB, or smear-positive specimen with a culture positive result. UNITED REPUBLIC OF TANZANIA 247 Implications of results The estimated prevalence of TB (443 per 100  000 population; 95% CI: 258–629, all forms, all ages) was higher than the pre-survey WHO estimate (2012) of 177 (95% CI: 93–286) per 100 000 population (5). However, the re-estimated time series of prevalence showed a continual decline since 2005. A striking finding of the survey was that 52% of the identified smear-positive TB cases were aged 45 years or more. This indicated that prevalent TB was largely driven by progression from a much earlier acquisition of a latent infection. In contrast, routine programmatic data from 2012 showed that only 28% of notified TB cases were aged 45 years or more, indicating important gaps in the detection of cases in the middle to older age groups. The large proportion of prevalent TB cases in older age groups points towards a historic positive effect of NTLP control strategies; however, differences with the estimated number of notified TB cases suggested a need for the NTLP to reassess its screening and diagnostic strategies (for example, to widen the range of symptoms considered when screening for TB in routine practice, and expand the use of chest X-ray), and to create better community awareness about the symptoms of TB. The strong emphasis of the NTLP on TB/HIV activities may have taken attention away from a large, unidentified population of older HIV-negative people with TB. The post-survey estimate of the case detection rate (notifications of new cases divided by estimated incidence) in 2015 was only 36% (95% CI: 21–77), compared with a pre-survey estimate of 79% (95% CI: 74–84%) (12,13). Major successes, challenges and lessons learned Due to serious limitations with culture examinations, it was difficult to accurately estimate the burden of bacteriologically confirmed TB. Nonetheless, collaborative post-survey activities with all partners made the survey results (especially estimates of the prevalence of smear-positive pulmonary TB) useful for the NTLP. One key message was the age distribution of prevalent cases, which suggested an epidemiological shift towards older people and potentially reactivation of previous infection, which has been a sign of effective population- wide TB control activities in other countries in the past (14) and is consistent with a shift observed more recently in Asian countries that have implemented prevalence surveys. The identification of a higher burden of TB disease among those who were HIV-negative compared with those who were HIV-positive was also helpful. The survey team, to their credit, published their results in a peer-reviewed journal (14–16). In addition to the major challenge with culture testing, other challenges and associated lessons learned included: • smear microscopy in the field is technically feasible, but in practice it can be fraught with potential contamination issues; • multiple paper forms were used for data collection (including handwritten individual identifiers), and administrative errors made it difficult or impossible to match the personal identifiers on these forms with laboratory specimens and other clinical information; digital data entry and barcoding is vital to ensure the quality of data management in future surveys; • there was erroneous oversampling of study participants in the initial clusters, to increase participation. Such protocol violations need to be avoided; • some recommendations from external monitoring missions were not implemented in a timely manner, thus potentially impacting on survey quality; • vehicles with computerized radiography equipment had to be checked and serviced during the survey and given the technology used, additional manual steps and human resources were required to develop images before reading Photo credit: Agatha Anthony National TB prevalence surveys 2007–2016248 compared with more recent digital radiography systems; • the agreement to reach consensus between key stakeholders on the actual number of TB cases took more than a year; and • there were delays in disbursement of funds to support survey operations. References 1. The World Bank. (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS. (http://aidsinfo.unaids.org/, accessed May 2017). 3. World Health Organization. Global tuberculosis database. Geneva: WHO; 2017 (http://www.who.int/tb/data/en/, accessed April 2017). 4. Enarson D A. Principles of IUATLD collaborative tuberculosis progammes. Bull Int Union Tuberc Lung Dis 1991; 66 (4): 195– 200. 5. World Health Organization. Global tuberculosis report 2012. Geneva: WHO; 2012 (http://www.who.int/tb/publications/ global_report/gtbr12_main.pdf, accessed May 2017) 6. United Republic of Tanzania – Ministry of Health and Social Welfare. The first national tuberculosis prevalence survey in the United Republic of Tanzania – interim report. Tanzania: 2013. 7. World Health Organization. Tuberculosis prevalence surveys: a handbook. Geneva: WHO; 2011 (https://apps.who.int/iris/ bitstream/handle/10665/44481/9789241548168_eng.pdf, accessed August 2016). 8. Supranational Reference Laboratory. Final report. Technical assistance and support. Antwerp, Belgium: SRL; 2013. 9. Supranational Reference Laboratory. Final report on Xpert MTB/ RIF testing of survey sputum smears positive for acid-fast bacilli. Antwerp, Belgium: SRL; 2013. 10. World Health Organization. Addendum to the report of the first national TB prevalence survey of the United Republic of Tanzania. Geneva: WHO; 2015. 11. Van Deun, A. Final report on Xpert MTB/RIF testing of survey sputum smears positive for acid-fast bacilli by the Supra-National TB Reference Laboratory (SRL) in Antwerp, Belgium. 3 September 2014. 12. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/bitstream/ handle/10665/91355/9789241564656_eng.pdf ?sequence=1, accessed May 2017) 13. World Health Organization. Global tuberculosis report 2015. Geneva: WHO; 2015 (http://apps.who.int/iris/bitstream/handle/ 10665/191102/9789241565059_eng.pdf;jsessionid=DA5FF20A 3580FAAD2F9FA0267F9D1724?sequence=1, accessed May 2017) 14. Senkoro M, Mfinanga S, Egwaga S, Mtandu R, Kamara DV, Basra D et al. Prevalence of pulmonary tuberculosis in adult population of Tanzania: a national survey, 2012. Int J Tuberc Lung Dis. 2016;20(8):1014–1021 (https://www.ncbi.nlm.nih.gov/ pubmed/27393533, accessed January 2018). 15. Senkoro M, Hinderaker SG, Mfinanga SG, Range N, Kamara DV, Egwaga S et al. Health care-seeking behaviour among people with cough in Tanzania: findings from a tuberculosis prevalence survey. Int J Tuberc Lung Dis. 2015;19(6):640–646 (https://www. ncbi.nlm.nih.gov/pubmed/25946352, accessed January 2018). 16. Senkoro M, Kumar AM, Chinnakali P, Mfinanga SG, Egwaga S, Kamara V et al. Population impact of factors associated with prevalent pulmonary tuberculosis in Tanzania. Int J Tuberc Lung Dis. 2016;20(10):1326–1333 (https://www.ncbi.nlm.nih.gov/ pubmed/27725043, accessed January 2018). 249 Name Role Organization Dinh Ngoc Sy Principal investigator National Tuberculosis Control Programme/National lung hospital Nguyen Viet Nhung Principal investigator, survey coordinator National Tuberculosis Control Programme/National lung hospital Nguyen Binh Hoa Survey coordinator, data manager, field team leader National Tuberculosis Control Programme/National lung hospital Nguyen Van Hung Laboratory manager National Tuberculosis Control Programme/National lung hospital Do Trong Nghia Radiology manager National Tuberculosis Control Programme/National lung hospital Nguyen Van Cu Field team leader National Tuberculosis Control Programme/National lung hospital Chu Manh Dung Field team leader National Tuberculosis Control Programme/National lung hospital Nguyen Cong Chi Field team leader National Tuberculosis Control Programme/National lung hospital Ha Thuc Van Field team leader National Tuberculosis Control Programme/Danang Hospital for TB and Lung Diseases Bao Thuyet Field team leader National Tuberculosis Control Programme/Danang Hospital for TB and Lung Diseases Vu Ngoc Tuan Field team leader National Tuberculosis Control Programme/Pham Ngoc Thach hospital Pham Vuong Khac Thai Field team leader National Tuberculosis Control Programme/Pham Ngoc Thach hospital Tran Ngoc Thach Field team leader National Tuberculosis Control Programme/Pham Ngoc Thach hospital Thai Anh Sam Field team leader National Tuberculosis Control Programme/Pham Ngoc Thach hospital Frank G.J. Cobelens Technical assistance (survey advisor) Academic Medical Center, University of Amsterdam Martien W. Borgdorff Technical assistance (survey advisor) Academic Medical Center, University of Amsterdam Edine W. Tiemersma Technical assistance (data analysis) KNCV Tuberculosis Foundation Nico Kalisvaart Technical assistance (data management) KNCV Tuberculosis Foundation Agnes Gebhard Technical assistance (analysis on social economic status) KNCV Tuberculosis Foundation Marleen Vree Technical assistance (analysis) Landsteiner Institute, Medical Center Haaglanden, The Hague, The Netherlands Key people Surveyed clusters (N=70) a Data sources ■ Viet Nam National Tuberculosis Programme, national TB prevalence survey in Vietnam, 2006–2007. Ministry of Health, Viet Nam; Hanoi November 2008. ■ Hoa NB, Sy DN, Nhung NV, Tiemersma EW, Borgdorff MW, Cobelens FG. National survey of tuberculosis prevalence in Viet Nam. Bull World Health Organ. 2010;88(4):273–280. ■ Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, Tiemersma EW. Yield of interview screening and chest X-ray abnormalities in a tuberculosis prevalence survey. Int J Tuberc Lung Dis. 2012;16(6):762–767. ■ Hoa NB, Tiemersma EW, Sy DN, Nhung NV, Vree M, Borgdorff MW et al. Health-seeking behaviour among adults with prolonged cough in Vietnam. Trop Med Int Health. 2011;16(10):1260–1267. Summary statistics Participation rate 91% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 307 4.5 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.3 Finance Amount (US$) Government of Viet Nam (Ministry of Health) 893 000 Government of the Netherlands 92 000 The Global Fund 10 000 WHO 57 000 Total budget 1 052 000 Survey organization and financing Implementing agency: National Tuberculosis Control Programme VIET NAM a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2006–2007 National TB prevalence surveys 2007–2016250 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban, rural and remote districts Sampling unit District/commune/sub-commune Sample size assumptions • Smear-positive prevalence 100 per 100 000 (≥15 years) • Precision 0.2 • Design effect 1.5 • k 0.6 • Response rate 80% • Sample size (estimated) 105 000 Number of clusters 70 Cluster size 1 500 Eligibility criteria • Age ≥15 years • Residency Lived in the household for at least three months prior to the census Survey design and methodology Laboratory methodology Smear Three samples (one spot immediately, one early morning and one spot on or before the transport day): direct preparation, ZN Culture One sample (morning): concentrated preparation, LJ media Identification of MTBa Niacin TB drug susceptibility test Doneb Xpert® MTB/RIF Not done HIV test Not done Screening criteria Interviewa Productive cough ≥2 weeks Chest X-rayb Any lung abnormality Other Current TB treatment or history of TB in preceding two years or chest X-ray exempted a An in-depth interview about TB related symptoms and health-care seeking behaviour was conducted among people who screened positive on any one of the screening criteria. b Mobile mass miniature radiography system based on photofluorography and mobile direct digital X-ray unit (slot scan system). a Species identification was done for positive cultures. b All M.tuberculosis isolates were tested for resistance to isoniazid, rifampicin, ethambutol and streptomycin but results were not officially reported, since measurement of levels of drug resistance was not a primary objective of the survey. Field data collection Paper Database EpiData version 3.1 Method of analysis Adjustment of standard errors for cluster design Results first published in a report/paper November 2008 Official dissemination event November 2008 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 197 150–244 307 249–366 Male 351 262–440 536 431–642 Female 69 39–99 118 75–161 15–24 years 42 12–73 55 22–88 25–34 years 84 24–143 136 63–210 35–44 years 247 157–337 321 222–420 45–54 years 234 145–322 344 228–460 55–64 years 329 187–470 599 288–910 ≥65 years 429 239–620 764 492–1 037 Urban 203 132–274 282 188–376 Rural 219 145–294 344 259–430 Remote 134 92–177 232 131–333 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 4 172 – Location of care sought • Consulted medical facility 1 228 29 Public facility 1 029 84 Private facility 199 16 • Pharmacy 671 16 • Traditional centre 25 0.6 Self-treated N/A N/A No action taken 2 248 54 Unknown N/A N/A a Productive cough ≥2 weeks. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A). Survey participants currently on TB treatment Number % Total participants currently on TB treatment 64 – • Treated in the public sector 46 72 • Treated in the private sector 2 3.1 • Treated in unknown sector 16 25 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 10 3.7 Other sputum results Number % Total smear-positive participants 186 – Smear-positive participants without MTB confirmationa 49 26 Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. Design effect k Smear-positive TB 2.7 0.8 Bacteriologically confirmed TB 2.6 0.6 VIET NAM 251 Survey flow: census to final outcomes Field operations: September 2006 to July 2007 Individuals enumerated in censusa 114 389 Eligible for sputum examination 8 005 (8.5%) Total bacteriologically confirmed cases 269 Eligible study population 103 924 (91%) Total participants 94 179 (91%) Interview and chest X-ray 87 314 (93%) Interview only 6 444 (6.8%) Chest X-ray only 421 (0.5%) Symptom positive, chest X-ray positive 518 (6.5%) Symptom positive, chest X-ray negative or N/A 3 522 (44%) Symptom negative, chest X-ray positive 2 972 (37%) Otherf 993 (12%) Symptom positive, chest X-ray positive 48 (18%) Symptom positive, chest X-ray negative or N/A 23 (8.6%) Symptom negative, chest X-ray positive 181 (67%) Otherj 17 (6.3%) Submitted specimens At least one specimen 7 648 (96%) At least two specimens 7 554 (94%) Three specimens 7 117 (89%) Smear-positive casesh 174 (65%) Definite 137 Probable 37 Smear-negative casesi 95 (35%) Definite 95 Probable N/A Laboratory result At least one culture result availableg 7 257 (91%) Symptom screening Productive cough ≥2 weeksc 4 172 (4.5%) Haemoptysis 134 (1.8%)d Chest pain 2 012 (27%)d Fever 820 (11%)d Night sweats 470 (6.2%)d Weight loss 885 (12%)d Total symptom-screen positivea 4 172 (4.5%) Chest X-ray screening Normal 83 223 (95%) Abnormalc 3 681 (4.1%) Other abnormality 738 (0.8%) Result not availablee 93 (0.1%) Total chest X-rays taken 87 735 a There were 137 549 individuals in the census, and only adults (≥15 years) were counted for the prevalence survey. b 23 160 children (6–14 years) from the same population as the prevalence survey participated in a concurrent tuberculin survey (see reference 11). c Eligible for sputum collection. d The denominator is the number of participants who had an in-depth interview (N=7580). e The results were not recorded. f Currently on TB treatment, including participants who screened positive (symptom and/or chest X-ray) (64), history of TB in the two years preceding the survey, including participants who screened positive (symptom and/or chest X-ray) (364), chest X-ray exempted and symptom-screen negative (58), and participants who were not eligible for sputum submission but submitted sputum based on the team leader’s decision (507). g Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. h Definite: MTB confirmed by culture. Probable: MTB not confirmed by culture but either two or more positive smears, one positive smear with chest X-ray consistent with TB. i Definite: MTB confirmed by culture. Probable: no definition. j One was chest X-ray exempted and 16 were included due to the team leader’s decision. Ineligible individuals 10 465 (9.1%) Children <15 yearsb N/A Did not meet residency criteria 10 465 (9.1%) National TB prevalence surveys 2007–2016252 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 300 400 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 800 700 900 0 Prevalence per 10 0 0 0 0 population 10 000 5 000 30 000 35 000 25 000 40 000 45 000 50 000 20 000 15 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Fig. 2: TB prevalence per 100 000 population by age Age group (years) Bacteriologically confirmed TB Smear-positive TB 1200 600 800 1000 400 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 500 600 Pr ev al en ce : no ti fi ca ti on ra ti o Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 3.84, variance 9.35, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using notifications obtained from the WHO global TB database, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 5 15 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Total (≥15) VIET NAM 253 Background Viet Nam is the easternmost country on the Indochinese Peninsula in South-East Asia, and in 2006 its population was 83 million. The average gross national income (GNI) per person was US$ 760 per year, making it a low-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. In 2005, the prevalence of HIV in the general population aged 15–49 years was 0.4% (95% confidence interval [CI]: 0.3–0.5%) (2), and it was estimated that 7.0% (95% CI: 6.2–7.8%) of TB patients were coinfected with HIV (3). In 1995, the National TB Control Programme (NTP) in Viet Nam began implementing the WHO-recommended DOTS strategy (4,5) and achieved nationwide DOTS coverage in 1999 (6). Based on mathematical models, it was predicted that TB prevalence and incidence would start to decline in Viet Nam when 70% of its new smear- positive TB cases were detected and 85% of cases were successfully treated (7). According to WHO estimates, Viet Nam reached and exceeded these targets in 1997 (8). However, its notification rate (new and relapse) increased from 73 per 100 000 population in 1990 to 111 per 100 000 population in 2000. A small decrease in TB notification rates among women and older adults was offset by an increase among young men, resulting in stabilization of notification rates during this period (8). Given that the epidemiology of TB in Viet Nam did not follow the predicted pattern, and that estimates of TB incidence (used as the denominator for estimates of the case detection rate) were based on tuberculin surveys1 conducted in the 1990s, the NTP decided to implement a national TB prevalence survey in 2006–2007. The objectives of the survey were to obtain a direct measure of the burden of TB disease, and to better understand the epidemiology of TB and the effectiveness of TB control efforts in Viet Nam. Key methods and results There were 70 clusters in three strata (urban, rural and remote), with a target cluster size of 1500 individuals. A total of 114  389 individuals from 34  271 households were enumerated in the survey census, of whom 103 924 (91%) were eligible and invited to participate. Of these, 94 179 (91%) were screened by chest X-ray and symptom screening interview, in line with the WHO 2011 algorithm (9). A total of 8005 people (8.5% of participants) were eligible for sputum examination. Of these, 7648 (96%) submitted at least one sputum specimen and 7117 (89%) submitted three sputum specimens (10,11). A total of 269 bacteriologically confirmed pulmonary TB cases were identified, including 174 cases of smear- positive TB. The prevalence of smear-positive TB was 197 (95% CI: 150–244) per 100 000 population (among those aged ≥15 years) and for bacteriologically confirmed TB it was 307 (95% CI: 249–366) per 100 000 population. When extrapolated to all forms of TB and to all ages, prevalence was estimated as 266 (95% CI: 117–477) per 100 000 population. There was no significant variation in prevalence between the three strata (urban, rural and remote). However, in the middle geographical zone of the country, where there are more remote and mountainous areas, prevalence was 209 (95% CI: 132–287) per 100 000 population, which was significantly lower than the level of 286 (95% CI: 218–355) per 100 000 population in the 1 Tuberculin surveys were used to estimate the annual risk of infection but did not provide a direct measure of the burden of TB disease. Photo credit: Frank Cobelens National TB prevalence surveys 2007–2016254 northern zone and of 367 (95% CI: 249–486) per 100 000 population in the southern zone (10,11). Other key results were: • the male to female ratio was 5.1 for smear- positive TB and 4.5 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age, and the absolute number of bacteriologically TB cases was also relatively large in older age groups (≥35 years); • among bacteriologically confirmed TB cases, 26% were symptom-screen positive, and among the smear-positive cases, 53% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.3, but varied from 1.3 in those aged 15–34 years to 2.9 in the 35–44 year age group, and was higher for men than for women (2.7 versus 1.6); • among bacteriologically confirmed TB cases, 79% had no previous history of anti-TB treatment and only 4% were on anti-treatment at the time of the survey; and • of the 196 bacteriologically confirmed and 134 smear-positive TB survey cases that screened positive for symptoms,2 and were not on anti-TB treatment at the time of the survey, 62 (32%) and 46 (34%), respectively, had previously sought care in a public or private health facility for their symptoms. Implications of results The survey found that the prevalence of smear-positive TB among those aged 15 years or more was 197 (95% CI: 150–244) per 100 000 population. Assuming that there were no smear-positive TB cases in those aged under 15 years, the national prevalence of smear-positive TB (all age groups) was 145 (95% CI: 110–180) per 100 000 population. Therefore, the prevalence of smear-positive TB was 1.6 times higher than the level of 90 per 100 000 population in 2006 that had been estimated prior to the survey (based on data from tuberculin surveys). The prevalence survey demonstrated that the previous estimates based on tuberculin survey data from the 1990s were too low. Nonetheless, the distribution of TB by age, sex and geography was similar to patterns observed in case notification data. Specifically, the burden of TB was much higher in men than in women, and the epidemic was a progressively ageing one, with the highest burden found in the oldest age groups. The survey also confirmed a relatively low burden in the remote, mountainous areas compared with urban and low-lying rural areas. To address the high TB burden in older people and men, active case finding efforts were expanded, with specific attention paid to those groups. Only about one half of the smear-positive TB cases found in the survey reported a productive cough of ≥2 weeks duration. Given that detection of TB cases in health facilities used a screening algorithm based on the presence of a persistent productive cough, a large proportion of TB cases would not have met the standard screening criteria. Furthermore, over a third of the bacteriologically confirmed cases were smear-negative, so that without culture (which was not routinely done), many cases could not be confirmed. These findings highlighted important limitations in the TB screening and diagnostic algorithms used for routine care (i.e. a presumptive TB case was identified only by symptoms – mainly a cough for ≥2 weeks). They also highlighted the need to widen the eligibility criteria for smear examination to other TB- related symptoms in addition to cough, and to expand the use of culture for TB diagnosis. Broader symptom screening criteria and greater use of chest X-ray were implemented in Viet Nam following the prevalence survey. The survey also showed that nearly 30% (1228/4172) of people with prolonged productive cough had visited a health-care provider, and of these, 84% (1029) had visited a public health-care facility. A common first point of contact was a pharmacy, which highlighted the important role this sector could play in TB case-finding activities, especially through the referral of a person with 2 Health-care seeking behaviour data were available for survey cases with any TB symptoms, not just for those with chronic cough. Photo credit: Nguyen Binh Hoa VIET NAM 255 presumptive TB to appropriate health-care providers. In addition, TB patients waited on average about one month before seeking care, which demonstrated the need to increase awareness in the general population about TB symptoms, and the need to improve the diagnostic practices of providers to ensure appropriate and timely diagnosis and management of TB. Major successes, challenges and lessons learned The survey provided the first ever direct measurement of TB disease burden at the national level in Viet Nam. It also provided a large amount of other information about the TB epidemic, much of which was published in a timely fashion (11–15). For example, data collected during, or in association with, the survey provided information about: • the relationship between TB and household expenditure (as a proxy for socio-economic status); • health-care seeking behaviour among people with presumptive TB; • the distribution and frequency of mycobacteria other than TB; • the yield of interview screening and chest X-ray abnormalities; • diagnosis and treatment of TB in the private sector; • comparisons between TB prevalence and the annual risk of tuberculous infection; and • the quality and coverage of the national TB surveillance system. The results were also used to evaluate and improve approaches to TB control within Viet Nam, and the experience gained during the survey helped to build global and regional capacity to conduct prevalence surveys. The inclusion criteria used in the survey posed some challenges. Specifically, adults who were not present in the sampled clusters for at least three months, or who were incarcerated or who lived in military barracks (i.e. the mobile population), were not included in the survey. As a result, it was not known how well the survey sample represented the mobile population, and therefore the total Vietnamese population. The proportion of young men in the mobile population was relatively high at the time of the survey; thus, this group was underrepresented in the survey sample. The prevalence of TB among men of this age group was higher than that among women of the same age, but lower than the prevalence among older men. Photo credit: Frank Cobelens National TB prevalence surveys 2007–2016256 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 5. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 6. Huong NT, Duong BD, Co NV, Quy HT, Tung LB, Bosman M et al. Establishment and development of the National Tuberculosis Control Programme in Vietnam. Int J Tuberc Lung Dis. 2005;9(2):151–156 (https://www.ncbi.nlm.nih.gov/ pubmed/15732733, accessed January 2018). 7. Dye C, Maher D, Weil D, Espinal M, Raviglione M. Targets for global tuberculosis control. Int J Tuberc Lung Dis. 2006;10(4):460– 462 (https://www.ncbi.nlm.nih.gov/pubmed/16602414, accessed January 2018). 8. Vree M, Bui DD, Dinh NS, Nguyen VC, Borgdorff MW, Cobelens FG. Tuberculosis trends, Vietnam. Emerg Infect Dis. 2007;13(5):796–797 (https://www.ncbi.nlm.nih.gov/pmc/articles/ PMC2738473/, accessed January 2018). 9. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 10. Viet Nam National Tuberculosis Programme, national TB prevalence survey in Vietnam. Ministry of Health, Viet Nam. 11. Hoa NB, Sy DN, Nhung NV, Tiemersma EW, Borgdorff MW, Cobelens FG. National survey of tuberculosis prevalence in Viet Nam. Bull World Health Organ. 2010;88(4):273–280 (http://www. who.int/bulletin/volumes/88/4/09-067801/en/, accessed January 2018). 12. Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, Tiemersma EW. Yield of interview screening and chest X-ray abnormalities in a tuberculosis prevalence survey. Int J Tuberc Lung Dis. 2012;16(6):762–767 (https://www.ncbi.nlm.nih.gov/ pubmed/22507287, accessed January 2018). 13. Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, Tiemersma EW. First national tuberculin survey in Viet Nam: characteristics and association with tuberculosis prevalence. Int J Tuberc Lung Dis. 2013;17(6):738–744 (https://www.ncbi.nlm.nih. gov/pubmed/23676155, accessed January 2018). 14. Hoa NB, Tiemersma EW, Sy DN, Nhung NV, Vree M, Borgdorff MW et al. Health-seeking behaviour among adults with prolonged cough in Vietnam. Trop Med Int Health. 2011;16(10):1260–1267 (https://www.ncbi.nlm.nih.gov/pubmed/21692960, accessed January 2018). 15. van Leth F, Guilatco RS, Hossain S, Van’t Hoog AH, Hoa NB, van der Werf MJ et al. Measuring socio-economic data in tuberculosis prevalence surveys. Int J Tuberc Lung Dis. 2011;15 Suppl 2:S58– 63 (https://www.ncbi.nlm.nih.gov/pubmed/21740660, accessed January 2018). 257 Name Role Organization Nathan Kapata Principal investigator National TB and Leprosy Control Programme Pascalina Chanda Kapata Survey coordinator and co-principal investigator Ministry of Health William Ngosa Assistant survey coordinator Ministry of Health Mine Metitiri Assistant survey coordinator Ministry of Health Lutinala Nalomba Mulenga Chest diseases laboratory team lead Ministry of Health Mathias Tembo Tropical Diseases Research Centre laboratory team lead Ministry of Health Patrick Katemangwe University teaching hospital laboratory team lead Ministry of Health Mazyanga Mazuba Liwewe HIV laboratory team lead Ministry of Health Veronica Sunkuntu Radiology team lead Ministry of Health Chris Silavwe Data manager Ministry of Health Chitani Mbewe Field team leader Ministry of Health Sam Msariri Field team leader Ministry of Health Mashina Chomba Field team leader Ministry of Health Jane Shawa Field team leader Ministry of Health Eveline Klinkenberg Technical assistance (survey advisor) KNCV Tuberculosis Foundation Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Charalampos Sismanidis Technical assistance (survey advisor) WHO headquarters Nico Kalisvaart Technical assistance (data management) KNCV Tuberculosis Foundation Julia Ershova Technical assistance (data management) US Centers for Disease Control and Prevention (CDC) Key people Surveyed clusters (N=66)a Data sources ■ National tuberculosis prevalence survey 2013–2014 technical report. Zambia: Ministry of Health, Government of the Republic of Zambia; 2015. ■ Kapata N, Chanda-Kapata P, Ngosa W, Metitiri M, Klinkenberg E, Kalisvaart N et al. The prevalence of tuberculosis in Zambia: Results from the first national TB prevalence survey, 2013–2014. PLoS One. 2016;11(1):e0146392. Summary statistics Participation rate 84% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 638 1.7 Prevalence:notification ratio (smear-positive TB, ≥15 years) 2.0 Finance Amount (US$) Government of Zambia 1 639 303 USAID 2 000 000 US CDC 1 737 264 Total budget 5 376 567 Survey organization and financing Implementing agency: National TB and Leprosy Control Programme ZAMBIA a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. 2013–2014 National TB prevalence surveys 2007–2016258 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Province/district/ward/census supervisory area Sample size assumptions • Smear-positive prevalence 199 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.5 • k 0.6 • Response rate 85% • Sample size (estimated) 54 400 Number of clusters 66 Cluster size 825 Eligibility criteria • Age ≥15 years • Residency Slept in the household 24 hours prior to the census Survey design and methodology Laboratory methodology Smear Two samples (spot, morning): concentrated preparation, FM (auramine stain) Culture Two samples (spot, morning): concentrated preparation, MGIT media Identification of MTB Capilia TB drug susceptibility test Not done Xpert® MTB/RIF Donea HIV test Done for participants who consented Screening criteria Interviewa Cough ≥2 weeks or fever ≥2 weeks or chest pain ≥2 weeks Chest X-rayb Any lung abnormality Other N/A a An in-depth interview about other TB symptoms and health-care seeking behaviour was undertaken only for those who screened positive. b Direct digital radiography (portable). a Xpert MTB/RIF was conducted for all smear-positive, some smear-negative with culture contaminated, or smear-negative culture indeterminate but chest X-ray suggestive of TB. Field data collection Electronic Database Microsoft® Access Method of analysis MI+IPW Results first published in a report/paper September 2015 Official dissemination event January 2016 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 319 232–406 638 502–774 Male 445 309–580 833 641–1 024 Female 221 139–303 487 353–621 15–24 years 154 71–236 285 157–412 25–34 years 422 245–599 664 337–891 35–44 years 496 315–676 947 660–1 237 45–54 years 323 139–507 926 611–1 240 55–64 years 333 149–517 708 401–1 013 ≥65 years 288 91–485 876 535–1 218 Urban 583 391–775 993 714–1 273 Rural 187 130–243 460 344–577 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 4 453 – Location of care sought • Consulted medical facility 1 829 41 Public facility 1 680 92 Private facility 75 4.1 Other facility 74 4.0 • Pharmacy 16 0.4 • Traditional centre, Faith based organization 1 0.02 Self-treated N/A N/A No action taken 2 534 57 Unknown 73 1.6 a Cough ≥2 weeks or fever ≥2 weeks or chest pain ≥2 weeks. a Data were available only for participants who were eligible for sputum submission. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. Survey participants currently on TB treatment Number % Total participants currently on TB treatmenta 114 – • Treated in the public sector 61 54 • Treated in the private sector 1 0.9 • Treated in unknown sector 52 45 Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 7 2.6 Design effect k Smear-positive TB 2.3 0.8 Bacteriologically confirmed TB 3.3 0.7 Other sputum results Number % Total smear-positive participants 356 – Smear-positive participants without MTB confirmationa 221 62 Isolates with MDR-TB detected N/A N/A a Age ≥15 years unless otherwise specified. ZAMBIA 259 Survey flow: census to final outcomes Field operations: August 2013 to July 2014 Individuals enumerated in census 98 458 Eligible for sputum examination 6 708 (15%) Total bacteriologically confirmed cases 265 Eligible study population 54 830 (56%) Total participants 46 099 (84%) Interview and chest X-ray 45 633 (99%) Interview only 466 (1.0%) Chest X-ray only 0 (0%) Symptom positive, chest X-ray positive 1 505 (22%) Symptom positive, chest X-ray negative or N/A 2 948 (43%) Symptom negative, chest X-ray positive 2 255 (34%) Other N/A Symptom positive, chest X-ray positive 115 (44%) Symptom positive, chest X-ray negative or N/A 46 (17%) Symptom negative, chest X-ray positive 104 (39%) Other N/A Submitted specimens At least one specimen 6 154 (92%) Both specimens 4 057 (61%) Smear-positive casesd 135 (51%) Definite 135 Probable N/A Smear-negative casesd 130 (49%) Definite 130 Probable N/A Laboratory result At least one culture result availablec 5 785 (86%) Symptom screening Cough ≥2 weeksa 2 405 (5.2%) Haemoptysis N/A Sputum production N/A Chest pain ≥2 weeksa 3 426 (7.4%) Fever ≥2 weeksa 1 030 (2.2%) Total symptom-screen positivea 4 453 (9.7%) Chest X-ray screening Normal 41 662 (91%) Abnormala 3 741 (8.2%) Other abnormality N/A Result not availableb 230 (0.5%) Total chest X-rays taken 45 633 a Eligible for sputum collection. b Results were not interpretable (19), missing (136), or not available for other non-specified reasons (75). Among 230 participants, 19 who had uninterpretable chest X-ray images were requested to submit sputum samples. c Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. d Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. Ineligible individuals 43 628 (44%) Children <15 years 43 424 (44%) Did not meet residency criteria 204 (0.2%) National TB prevalence surveys 2007–2016260 Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 200 300 400 500 Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 400 300 500 600 700 800 900 1000 0 Prevalence per 10 0 0 0 0 population 2 000 8 000 10 000 12 000 14 000 16 000 4 000 6 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 90 80 70 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of smear-positive TB prevalence to notifications by age and by sexc Age group (years) and sex 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB 1400 600 800 1200 1000 400 200 0 15–24 25–34 35–44 45–54 55–64 ≥65 Data source: smear-positive pulmonary TB notified cases from NTP data (2014) (including cases diagnosed by Xpert MTB/RIF) a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data suggested that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 4.02, variance 17.8, p<0.05). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using smear-positive pulmonary TB notifications (2014) obtained from the NTP (including TB cases diagnosed by Xpert MTB/RIF), and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 6 2 4 10 8 12 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Total (≥15) ZAMBIA 261 Background Zambia, a landlocked country in Southern Africa, had a population of 16 million in 2014. Its average gross national income (GNI) per person was US$ 1740, making it a lower-middle-income country (1). It was one of the top 30 high TB burden countries (HBCs) defined by WHO for the period 2016–2020. The prevalence of HIV in the general population aged 15–49 years was 13% (95% confidence interval [CI]: 13–14%) in 2014 (2), and it was estimated that 61% (95% CI: 55–66%) of TB patients were coinfected with HIV (3). Zambia’s National TB Control Programme (NTP) was established in 1964. It operated as a vertical programme in the health sector until 1993, when it was combined with the AIDS and sexually transmitted infections (STI) programme. After initiating implementation of the DOTS strategy in 1995, the NTP moved from hospitals to community-based programmes. However, during a period of health sector decentralization in the late 1990s, the national programme almost collapsed. There was a loss of structure, staff training and guidance at all levels, and interruptions to drug supplies were frequent. The NTP was reorganized in 2000 and subsequently strengthened (4). The HIV epidemic led to a dramatic increase in the TB notification rate throughout the 1990s, from 216 cases per 100  000 population in 1990 to 524 cases per 100 000 population in 2004. Subsequently, notification rates started to fall, to a level of 365 cases per 100  000 population in 2010. The highest notification rate was in Lusaka (the capital city), followed by areas in the Copperbelt and southern provinces (especially along the railway lines). WHO estimated that, in 2012, TB incidence was 427 (95% CI: 385–470) per 100  000 population, prevalence was 388 (95% CI: 197–642) per 100  000 population and the case detection rate (notifications of new cases divided by estimated incidence) was 68% (95% CI: 62–75%) (5). In 2012, there was no direct measurement of TB disease burden in Zambia, and routine notification data were the main source of information to assess progress towards TB targets. However, the gap between notifications and incidence due to underreporting and underdiagnosis of cases was difficult to estimate. It was also recognized that the HIV epidemic had increased the level of TB disease burden, and that this might have been exacerbated by growing levels of poverty. For these reasons, it was decided to carry out Zambia’s first national TB prevalence survey, to obtain a direct measure of TB disease burden in the community, inform policy-makers and provide baseline data for future evaluation of programmatic achievements. Survey preparations began in 2008, but funding was delayed when the Global Fund to Fight AIDS, Tuberculosis and Malaria suspended all funding to Zambia, with the exception of essential activities. Survey preparations resumed in 2012, in close collaboration with the KNCV Tuberculosis Foundation, following agreement that the US Government would provide financial support for the survey through the TB CARE project. The survey started in August 2013 and was completed in July 2014 (6–9). Key methods and results There were 66 survey clusters in two strata (urban and rural), with a target cluster size of 825 individuals. A total of 98 458 individuals in 17 485 households were enumerated in the survey census, of which 54 830 (56%) were eligible and invited to participate. All 46 099 participants (84% of the total eligible) were screened according to the 2011 algorithm recommended by WHO; that is, chest X-ray and a symptom screening interview (10). A total of 6708 people (15% of participants) were eligible for sputum examination, of whom 6154 (92%) submitted at least one sputum specimen and 4057 (61%) submitted two sputum specimens. A total of 265 bacteriologically confirmed pulmonary TB cases were identified, including 135 cases of smear-positive TB. The estimated prevalence of smear-positive TB was 319 (95% CI: 232–406) per 100  000 population among those aged 15 years or more, and for bacteriologically confirmed TB it was 638 (95% CI: 502–774) per 100 000 population. When extrapolated to all forms of TB and to all ages, prevalence was 455 (95% CI: 366–544) per Photo credit: National TB Programme, Zambia National TB prevalence surveys 2007–2016262 100  000 population. The prevalence of bacteriologically confirmed TB was higher in urban areas than in rural areas (993 versus 460 per 100 000 population). Other key results were: • the male to female ratio was 2.0  for smear- positive TB and 1.7 for bacteriologically confirmed TB; • prevalence per 100  000 population increased with age, up to the 35–44 years age group, and was consistently high above 45 years; however, the absolute number of TB cases was relatively high in the younger age groups; • of the bacteriologically confirmed TB cases, 61% were symptom-screen positive, and of the smear-positive cases, 67% were symptom- screen positive; • for smear-positive pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.0 overall, but varied from 1.6 in those aged 45–54 years to 3.2 in the 65 years and over age group, and was slightly higher for men than for women (2.1 versus 2.0); and • of the bacteriologically confirmed TB cases, 14% of cases had no previous history of anti- TB treatment and only 2.6% were on anti-TB treatment at the time of the survey. Data on health-care seeking behaviour among survey TB cases were not available. The risk of TB was also analysed in terms of socio- economic status using wealth tertiles. In rural areas, the risk of TB was higher among the highest wealth tertile than among the lowest and middle tertiles. The opposite was true in urban areas, where the lowest and middle wealth tertiles had a prevalence that was twice as high as that found in the highest wealth tertile. HIV pre-test counselling was conducted for 44  761 (97%) of the 46 099 survey participants. Of those who underwent pre-test counselling, 30 605 (68%) consented to be tested, of whom 30 584 (99.9%) were tested. Of those tested, 2063 (6.7%) were HIV-positive, 28 431 were HIV-negative and 90 had an indeterminate result. Of 265 bacteriologically confirmed TB cases, 134 were tested for HIV and 36 (27%) were HIV-positive. HIV prevalence was four times higher among individuals with bacteriologically confirmed TB (27% [95% CI: 17–36%]) than among those without (6.5% [95% CI: 5.4–7.5%]). The prevalence of both smear-positive and bacteriologically confirmed pulmonary TB among HIV- positive people was more than four times higher than among HIV-negative people. However, there were still more TB cases among HIV-negative people, highlighting the high burden of TB at community level irrespective of the HIV epidemic. Photo credit: Julia Ershova Table 1: Bacteriologically confirmed TB prevalence by wealth index Table 2: Pulmonary TB prevalence by HIV status Wealth indexa Bacteriologically confirmed TB Prevalence per 100 000 population 95% CI Rural Highest 610 423–797 Middle 364 224–505 Lowest 483 294–672 Urban Highest 603 386–820 Middle 1251 911–1592 Lowest 1208 750–1666 HIV status Prevalence per 100 000 population 95% CI Smear-positive TB HIV-positive 887 424–1350 HIV-negative 182 129–236 Bacteriologically confirmed TB HIV-positive 1726 1029–2423 HIV-negative 387 294–480 a Please refer to the official report for an explanation of how the index was derived. ZAMBIA 263 Implications of results The survey showed that the prevalence of bacteriologically confirmed TB was higher than that estimated before the survey. In particular, the burden of TB among HIV- negative people had been underestimated. The survey also showed that many TB cases were likely to be missed (or detected late) when services rely on passive case finding alone. Of the symptomatic cases found during the survey, 97% (258/265) were not yet on treatment. The fact that half of the symptomatic cases not on treatment had already sought care for their symptoms also demonstrated a need to strengthen health services; for example, by raising health worker awareness of TB symptoms and by making diagnostics more widely available. Just over half of survey cases (51%; 134/265) were in those aged 25–44 years, with prevalence peaking in those aged 35–44 years. The economic consequences of this disease burden warrant further investigation, especially in the context of the End TB Strategy milestone for 2020, that no TB-affected households face catastrophic costs as a result of TB disease (11). The finding that 49% (130/265) of survey cases were smear-negative and that 39% of survey cases did not meet symptom screening criteria (despite using a wider range of screening symptoms than most other surveys) indicated a need to improve capacity to diagnose cases of culture-positive but smear-negative TB and to carefully assess the use of chest X-ray during the diagnostic process (particularly in the context of active case finding). Just over one third of the survey cases (34%; 89/258) that had not been detected before the survey were in the Copperbelt province, highlighting that coverage of diagnostic and treatment services needed to be improved in this particular “hotspot”. The socio-demographic disparities evident in the survey results also showed a need for more targeted efforts for certain population groups: men, the urban poor and those living in densely populated farming areas. The large number of symptomatic participants with nontuberculous mycobacteria (NTM) showed that NTM should be better characterized in Zambia, to enable the appropriate management of clinically relevant cases of NTM. Of the 6123 culture results available, 923 (15%) were NTM. Just over half (478/923) of individuals with a positive result had an abnormal chest X-ray, and 71% (655/923) were symptomatic (i.e. had either cough, chest pain or fever). The fact that a large proportion (62%, 221/356) of participants with a smear-positive result were found not to have Mycobacterium tuberculosis (MTB) showed Photo credit: Julia Ershova National TB prevalence surveys 2007–2016264 that a smear-positive result alone may not be adequate for the detection of TB cases, especially in the context of intensified case finding or active case detection strategies. Major successes, challenges and lessons learned Field operations were implemented within the expected timeframe with minimal interruptions and with a high participation rate overall (84%). One of the reasons for this was that the hard-to-reach rural areas were covered in the early part of the survey, during the drier parts of the year. The participation rate was lower (~49%) in the early stages of the survey (a common finding in surveys); this lower rate was also linked to the long distances to be travelled in the more remote and sparsely populated parts of the country. The survey teams encountered some myths and misconceptions among community members about TB, which had some impact on the overall participation rate. Measures implemented to improve participation rates included use of in-cluster community sensitization, involvement of the local political and traditional leadership, and ongoing community education using mainstream television and community radio stations to disseminate the objectives and procedures of the survey. The 7-day period used for cluster operations was manageable, but sometimes required field staff to work long hours, depending on the flow of participants. For example, in rural areas, participants tended to report to the cluster site later in the day, meaning that the teams had to work late into the night to clear the queue of participants, which in turn meant that transport had to be provided to those coming from locations far from the main survey camp site. This was a valuable lesson for future surveys, because without providing such support the participation rate might have been lower. The survey in Zambia was one of the first surveys to use digitalized data management from the household census through to central X-ray reading and laboratory management. It became the first national TB prevalence survey that used hand-held size apparatus to collect data in field conditions. There were initial problems with data- capture devices (e.g. pairing up of barcode scanners, short battery life and fragile barcodes), but these were resolved during the early stages of the survey. The efficiency of the fully digitalized data management system, use of direct digital chest X-ray units and good overall organization meant that the time between completion of field operations and dissemination of results and final publication was relatively short (14 months). References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. The National Tuberculosis and Leprosy Programme, TB manual (3rd ed). Zambia: Ministry of Health;(http://www.who.int/hiv/ pub/guidelines/zambia_tb.pdf?ua=1, accessed July 2017). 5. World Health Organization. Global tuberculosis report 2013. Geneva: WHO; 2013 (http://apps.who.int/iris/ bitstream/10665/91355/1/9789241564656_eng.pdf ?ua=1, accessed January 2018). 6. Chanda-Kapata P, Kapata N, Klinkenberg E, Mulenga L, Tembo M, Katemangwe P et al. Non-tuberculous mycobacteria (NTM) in Zambia: prevalence, clinical, radiological and microbiological characteristics. BMC Infect Dis. 2015;15:500 (https://www.ncbi. nlm.nih.gov/pubmed/26545357, accessed May 2017). 7. Chanda-Kapata P, Kapata N, Klinkenberg E, William N, Mazyanga L, Musukwa K et al. The adult prevalence of HIV in Zambia: results from a population based mobile testing survey conducted in 2013–2014. AIDS Res Ther. 2016;13:4 (https://www.ncbi.nlm. nih.gov/pubmed/26793264, accessed May 2017). 8. Kapata N, Chanda-Kapata P, Ngosa W, Metitiri M, Klinkenberg E, Kalisvaart N et al. The prevalence of tuberculosis in Zambia: Results from the first national TB prevalence survey, 2013–2014. PLOS ONE. 2016;11(1):e0146392 (https://www.ncbi.nlm.nih.gov/ pubmed/26771588, accessed May 2017). 9. National tuberculosis prevalence survey 2013–2014 technical report. Zambia: Ministry of Health, Government of the Republic of Zambia; 2015. 10. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017). 11. World Health Organization. The End TB Strategy. Geneva, WHO; 2014 (http://www.who.int/tb/strategy/en/, accessed July 2017). 265 Key people Surveyed clusters (N=75)a Data sources ■ Republic of Zimbabwe – Ministry of Health and Child Care. Report of the First National Population-based Tuberculosis Prevalence Survey. Republic of Zimbabwe, Ministry of Health and Child Care, August 2015. ■ Survey dataset. Summary statistics Participation rate 78% Bacteriologically confirmed TB (≥15 years) • Prevalence per 100 000 population • Male:female ratio 344 1.4 Prevalence:notification ratio (Bacteriologically confirmed TB, ≥15 years) 2.5 Finance Amount (US$) The Global Fund 3 464 437 Total budget 3 464 437 Survey organization and financing Implementing agency: The National Tuberculosis Control Programme/Biomedical Research and Training Institute ZIMBABWE a The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Name Role Organization Charles Sandy Principal investigator National Tuberculosis Control Programme (NTP) Junior Mutsvangwa Co-principal investigator Biomedical Research and Training Institute Ronnie Matambo Survey coordinator Biomedical Research and Training Institute Dumisani Ndlovu Radiology coordinator Biomedical Research and Training Institute Ellen Munemo Laboratory manager National Microbiology Reference Laboratory Eve Marima Data manager The Zimbabwe National Statistics Agency (ZIMSTAT) Hebert Mutunzi Technical working group member, laboratory NTP Mkhokeli Ngwenya Technical working group member, survey design NTP Joconiah Chirenda Technical working group member, survey design University of Zimbabwe, College of Health Sciences Nicholas Siziba Technical working group member, M&E NTP Peter Shiri Technical working group member, M&E NTP Martin Mapfurira NTP officer NTP Patrick Hazangwe Technical assistance WHO Zimbabwe Wilfred Nkhoma Technical assistance (survey advisor) WHO Regional Office for Africa (AFRO) Ikushi Onozaki Technical assistance (survey advisor) WHO headquarters Marina Tadolini Technical assistance (survey advisor) Consultant, Italy Fasil Tsegaye Technical assistance (survey advisor) International Union Against Tuberculosis and Lung Disease Kunihiko Ito Technical assistance (radiology) Research Institute of Tuberculosis/Japan Anti-Tuberculosis Association (RIT/JATA) Mourad Gumusboga Technical assistance (laboratory advisor) Supranational Reference Laboratory, Antwerp Belgium Hazim Timimi Technical assistance (data management) WHO headquarters Norio Yamada Technical assistance (analysis) RIT/JATA Kosuke Okada Technical assistance (reporting) RIT/JATA 2014 National TB prevalence surveys 2007–2016266 Sampling design Sampling frame Whole country Sampling design Multistage cluster sampling using PPS Strata Urban/rural Sampling unit Ward/enumeration area Sample size assumptions • Smear-positive prevalence 190 per 100 000 (≥15 years) • Precision 0.25 • Design effect 1.2 • k 0.4 • Response rate 85% • Sample size (estimated) 44 951 Number of clusters 75a Cluster size 600 Eligibility criteria • Age ≥15 years • Residency Permanent residents who had slept at least one night out of the last 14 days on the day of census, or non-residents who had slept in the household for 14 days or more before the day of the census Survey design and methodology a Two clusters (Macheke and Chiredzi) were replaced with other communities within the same district (same strata), due to difficulties in reaching the cluster site following severe rainfall as well as community apathy after long delays due to a breakdown of the mobile X-ray unit. Laboratory methodology Smear Two samples (spot, morning): concentrated preparation, FM (LED, auramine stain) Culture Two samples (spot, morning): concentrated preparation, LJ media and MGIT media (for all) Identification of MTB MPT64 rapid test TB drug susceptibility test Done using Xpert MTB/RIF (all smear- positive and culture-positive specimens) Xpert® MTB/RIF All smear-positive specimens (and all culture-positive specimens for rifampicin resistance testing) HIV test Done at referral centre for all bacteriologically confirmed TB cases Screening criteria Interviewa Cough of any duration, current night sweats, haemoptysis at any time in the past 12 months prior to study Chest X-rayb Any lung abnormality Other Chest X-ray exempted a An in-depth interview about health-care seeking behaviour was done only for those who screened positive. b Chest X-ray truck, mobile digital radiography. Field data collection Paper Database CSPro Method of analysis MI+IPW Results first published in a report/paper August 2015 Official dissemination event March 2017 Analysis and reporting Key survey results Prevalencea Smear-positive TB Bacteriologically confirmed TB Number per 100 000 population 95% CI Number per 100 000 population 95% CI Total 82 47–118 344 268–420 Male 103 43–163 413 303–523 Female 65 27–104 288 189–386 15–24 years 52 21–131 129 68–245 25–34 years 138 70–274 373 255–546 35–44 years 85 34–215 546 371–804 45–54 years 75 23–248 310 168–570 55–64 years 35 4.5–277 490 276–869 ≥65 years 47 6.6–341 547 310–962 Urban 116 38–193 355 228–482 Rural 64 36–114 337 243–431 Health-care seeking behaviour among participants who were symptom-screen positive Number % Participants who were symptom-screen positivea 1 833 – Location of care soughtb 486 26 • Consulted medical facility Public facility 438 – Private facility 45 – • Pharmacy 17 – • Traditional centre, faith healer 13 – Self-treated N/A N/A No action taken 1 130 62 Unknown 217 12 a Cough (any duration), current night sweats, and/or haemoptysis at any time in the last 12 months prior to the survey. b Participants could answer more than one category. a This includes culture with no evidence of MTB (i.e. culture-negative, NTM, contaminated, N/A) and Xpert-negative. b Xpert MTB/RIF was done only to test for rifampicin resistance among 107 bacteriologically confirmed cases. Survey participants currently on TB treatment Number % Total participants currently on TB treatment 84 – • Treated in the public sector N/A N/A • Treated in the private sector N/A N/A • Treated in other sector N/A N/A Bacteriologically confirmed TB cases detected by the survey who were currently on TB treatment 2 1.9 Design effect k Smear-positive TB 0.97 N/Aa Bacteriologically confirmed TB 1.1 0.3 Other sputum results Number % Total smear-positive participants 206 – Smear-positive participants without MTB confirmationa 183 89 Isolates with DR-TB (rifampicin) detectedb 13 12 a Age ≥15 years unless otherwise specified. a k could not be calculated for smear-positive TB because the design effect was less than one. ZIMBABWE 267 Survey flow: census to final outcomes Field operations: January 2014 to December 2014 Individuals enumerated in census 85 636 Eligible for sputum examination 5 820 (17%) Total bacteriologically confirmed cases 107 Eligible study population 43 478 (51%) Total participants 33 736 (78%) Interview and chest X-ray 32 468 (96%) Interview only 1 226 (3.6%) Chest X-ray only 42 (0.1%) Symptom positive, chest X-ray positive 628 (11%) Symptom positive, chest X-ray negative or N/A 1 205 (21%) Symptom negative, chest X-ray positive 2 803 (48%) Otherc 1 184 (20%) Symptom positive, chest X-ray positive 29 (27%) Symptom positive, chest X-ray negative or N/A 10 (9.4%) Symptom negative, chest X-ray positive 64 (60%) Otherc 4 (3.7%) Submitted specimens At least one specimen 5 705 (98%) Both specimens 5 441 (94%) Smear-positive casese 23 (22%) Definite 23 Probable N/A Smear-negative casese 84 (78%) Definite 84 Probable N/A Laboratory result At least one culture result availabled 5 680 (98%) Symptom screening Cough of any durationa 1 415 (4.2%) Haemoptysisa 310 (0.9%) Sputum production N/A Chest pain N/A Fever 283 (0.8%) Night sweatsa 560 (1.7%) Total symptom-screen positivea 1 833 (5.4%) Chest X-ray screening Normal 27 214 (84%) Abnormala 3 431 (11%) Other abnormality 1 856 (5.7%) Result not availableb 9 (0.02%) Total chest X-rays taken 32 510 a Eligible for sputum collection. b Chest X-ray taken but results were missing. c Chest X-ray exempted and symptom-screen negative. d Cultures that were either contaminated and/or missing without a definitive result (i.e. MTB, NTM, negative) were excluded. e Definite: MTB confirmed by culture and/or Xpert. Probable: no definition. Ineligible individuals 42 158 (49%) Children <15 years 35 100 (41%) Did not meet residency criteria 7 058 (8.2%) National TB prevalence surveys 2007–2016268 Fig. 1: Participation rate by age and sex 100 Pa rt ic ip at io n ra te ( % ) Age group (years) 80 90 70 60 Male Female 15–24 25–34 35–44 45–54 55–64 ≥65 Fig. 5: Ratio of bacteriologically confirmed TB prevalence to notifications by age and by sexc Age group (years) and sex 1.0 2.0 3.0 4.0 5.0 6.0 0 15–24 25–34 35–44 45–54 55–64 Male Female≥65 Pr ev al en ce : no ti fi ca ti on ra ti o Fig. 6: Estimates of TB prevalence (all ages, all forms of TB) before (in blue) and after (in red) the national TB prevalence surveyd Prevalence per 100 000 population 300200 400 500 600 Fig. 2: TB prevalence per 100 000 population by age Pr ev al en ce p er 1 0 0 0 0 0 p op ul at io n Age group (years) Bacteriologically confirmed TB Smear-positive TB Fig. 3: Estimated number of bacteriologically confirmed TB cases and prevalence per 100 000 population, by agea Age group (years) 0 15–24 25–34 35–44 45–54 55–64 ≥65 100 200 300 400 500 600 0 Prevalence per 10 0 0 0 0 population 2 000 6 000 8 000 10 000 12 000 4 000 Es ti m at ed n um be r of b ac te ri ol og ic al ly co nf ir m ed T B ca se s 1000 500 600 700 900 800 400 300 200 100 0 15–24 25–34 35–44 45–54 55–64 ≥65 a The estimated number of bacteriologically confirmed TB cases is the product of age-specific prevalence and population estimates from the UN Population Division (2015 revision). b The data did not suggest that the distribution of cases by cluster (blue bars) was significantly different from the theoretical distribution (red line) (mean 1.43, variance 1.90, p=0.06). The theoretical distribution assumes cases are distributed at random i.e. no clustering effect. c Notification rates were estimated using bacteriologically confirmed pulmonary TB notifications (2014) obtained from the NTP, and population estimates from the UN Population Division (2015 revision). d The blue bar denotes the best estimate of prevalence and its range that was indirectly derived from the estimate of incidence previously published by WHO, adjusted to the year of the prevalence survey based on previously published trends in incidence. Fig. 4: Cluster variation of the number of bacteriologically confirmed TB casesb Number of bacteriologically confirmed TB cases N um be r of c lu st er s 10 5 20 15 25 0 0 1 2 3 4 5 6 Total (≥15) ZIMBABWE 269 Background Zimbabwe is a landlocked country in Southern Africa. In 2014, it had a population of 15 million, and a gross national income (GNI) per person of US$ 840, making it a low-income country (1). It was one of the 22 high tuberculosis (TB) burden countries (HBCs) defined by WHO as a top priority for global efforts in TB control in 1998 and throughout the Millennium Development Goal (MDG) era (2000–2015), and one of the top 30 HBCs defined by WHO for the period 2016–2020. The prevalence of HIV in the general population aged 15–49 years was 15% (95% confidence interval [CI]: 14–16%) in 2014 (2), and it was estimated that 69% (95% CI: 64–74%) of TB patients were coinfected with HIV (3). Zimbabwe’s National TB Programme (NTP) was established in the late 1960s (4). The WHO-recommended DOTS strategy (5,6) was adopted in 1996 and nationwide coverage was achieved in 1998 (7,8). In 2014, there were 220 functional TB diagnostic centres within the public health system, and TB diagnosis and treatment was provided free of charge within the public health sector. The case notification rate declined from 2004, reaching a low of 302 per 100  000 population in 2007, probably influenced by health-system challenges in the context of an economic recession. With improved TB financing starting from 2008, case notifications increased, likely reflecting a mixture of better access to services and improved disease surveillance. TB notifications started to decline again from 2011 (332 per 100  000 population). In 2013, before the national TB prevalence survey, the TB prevalence was estimated as 409 (95% CI: 235–630) per 100  000 population; TB incidence was estimated at 552 (95% CI: 474–643) per 100 000 population; and the case detection rate (notifications of new cases divided by estimated incidence) was estimated at 42% (95% CI: 36–49) (7). However, these estimates were not informed by any direct measurement of disease burden. The NTP initiated preparations for a national TB prevalence survey in 2012, with financing from the Global Fund to Fight AIDS, Tuberculosis and Malaria. The objective was to obtain a direct measurement of the burden of TB disease, and better quantification of the gap between this burden and case notifications. The survey started in January 2014 and was completed in December 2014 (8). Key methods and results There were 75 survey clusters in two strata (urban and rural), with a target cluster size of 600 individuals. A total of 85 636 individuals from 19 629 households were enumerated in the survey census, of whom 43 478 (51%) were eligible and invited to participate. Of these, 33 736 (78%) participated and were screened according to the 2011 algorithm recommended by WHO; that is, using chest X-ray and a symptom screening interview (9). A total of 5820 people (17% of participants) were eligible for sputum examination, of whom 5705 (98%) submitted at least one sputum specimen and 5441 (94%) submitted two sputum specimens. The Zimbabwean survey was one of only a few national surveys during the period 2009–2015 that used the mycobacteria growth indicator tube (MGIT) for culture, and in which smear-positive specimens were tested with Xpert® MTB/RIF. A total of 107 bacteriologically confirmed pulmonary TB cases were identified, including 23 cases of smear- positive TB. Among the survey population of people aged 15 years or more, the prevalence of smear-positive TB was 82 (95% CI: 47–118) per 100 000 population, and for bacteriologically confirmed TB it was 344 (95% CI: 268–420) per 100  000 population. When extrapolated to all forms of TB and to all ages, prevalence was 275 (95% CI: 217–334) per 100 000 population. There was no significant difference between the two strata; in urban areas the prevalence of bacteriologically confirmed was 355 (95% CI: 228–482) per 100 000 population, and in rural areas it was 337 (95% CI: 243–431) per 100  000 population. Photo credit: Charles Sandy National TB prevalence surveys 2007–2016270 Other key results were: • the male to female ratio was 1.6 for smear- positive TB and 1.4 for bacteriologically confirmed TB; • prevalence per 100  000 population had two peaks, in those aged 35–44 years and ≥65 years; however, the absolute number of TB cases was relatively high in younger age groups; • among bacteriologically confirmed TB cases, 36% were symptom-screen positive, and among the smear-positive cases, 61% were symptom- screen positive; • for bacteriologically confirmed pulmonary TB, the ratio of prevalence to notifications (P:N ratio) was 2.5 overall, but varied from 1.5 in those aged 45–54 years to 5.3 in the 65 years or over age group, and was slightly lower for men than for women (2.4 versus 2.7); • among bacteriologically confirmed TB cases, 81% had no previous history of anti-TB treatment and only 1.9% were on anti-TB treatment at the time of the survey; and • of the 38 bacteriologically confirmed and 13 smear-positive TB survey cases that screened positive for symptoms and were not on anti-TB treatment at the time of the survey, 17 (45%) and 7 (54%), respectively, had previously sought care in a public or private health facility for their symptoms. Although not directly part of the survey itself, the HIV status of the bacteriologically confirmed TB cases was obtained from referral centres. Of the 107 cases, 42 (39%) were HIV-positive, 41 (38%) were HIV-negative and the HIV status of the remaining 24 (22%) was unknown. The proportion of cases who were HIV-positive (39%) was higher than the proportion in the population aged 15–49 years (15%) (2), but lower than in the clinical setting, where it was 60% (19 290 / 32 018), based on NTP notification of TB cases (all age groups) by HIV status in 2014. Implications of results The smear-positive prevalence of 82 per 100 000 popula- tion was much lower than the estimated prevalence used during the sample size calculation (i.e. 190 per 100 000 in the adult population). The survey showed the challenges that the NTP faces in detecting cases. Two peaks in TB prevalence per 100 000 population were observed: one in those aged 35–44 years and the other in those aged 65 years or more. Apart from the impact of HIV, other factors contributing to the higher prevalence in the former group probably included higher rates of urbanization and mixing in congregated settings. It was recognized that the NTP should strengthen TB/ HIV activities in collaboration with the national HIV/ AIDS programme. The high prevalence among the elderly indicated that intensified efforts to detect cases in this subpopulation might be warranted. Although there were as many TB cases in urban areas as in rural areas, case notification rates were lower in rural areas. Possible explanations included poorer accessibility to medical services, and challenges with diagnosis and clinical management in rural areas; for example, TB diagnosis in rural areas was more dependent on smear microscopy since not all district hospitals were equipped with X-ray machines (or more advanced diagnostic tools) at the time of the survey. Proposed solutions included referral mechanisms from health centres to district hospitals, or outreach services to the community by mobile teams. Photo credit: Marina Tadolini ZIMBABWE 271 The number of smear-negative culture-positive TB cases (84 cases) was almost four times the number of smear- positive culture-positive cases (23 cases). The former group cannot be definitively diagnosed under normal programmatic conditions if routine diagnostic services rely on smear microscopy alone. Among the smear- positive participants, “smear-positive” but “culture/Xpert- negative/non-TB” accounted for 89% (183/206). This finding highlighted major concerns about the positive predictive value of smear examination in the context of routine health services. The survey thus demonstrated that the diagnostic services available at the time of the survey (which depended mostly on smear microscopy) needed to be thoroughly reviewed. For example, there was a need to assess the role of chest X-ray for individuals with severe or chronic respiratory symptoms (or both), and to expand referral services, so that presumptive TB cases with negative smears could access care at facilities equipped with chest X-ray, culture or Xpert MTB/RIF. Major successes, challenges and lessons learned The survey was successfully implemented within one calendar year, and preliminary results were available within six  months of completing field operations. Although the survey team in Zimbabwe did not participate in preparatory workshops organized by WHO for global focus countries, two visits to prevalence surveys in Malawi and Rwanda, and technical assistance from the Ethiopian deputy survey coordinator, greatly assisted the team’s understanding of how to organize and undertake a survey. This external technical support was vital in ensuring a good-quality survey. Despite a high contamination rate (1432 (13%) out of 11 138 samples – spot and morning – with MGIT), the performance of culture testing was high with the support of the Supranational Reference Laboratory in Antwerp, Belgium to ensure quality management of culture testing. Challenges faced during the survey included: • a participation rate that was lower-than-targeted (i.e. 78%), especially in men aged 15–54 years and women aged 15–24 years; factors that affected participation included damage to the digital chest X-ray system in a container due to poor road conditions; hot weather conditions which discouraged participation; and the presence of some religious groups who objected to any modern medical interventions; • retrieval of X-ray images was sometimes problematic because the archiving and communication system was controlled by the X-ray unit supplier in the Netherlands; there was also a backlog in central reading of X-rays due to the limited access to the internet; these challenges were resolved through in- country technical assistance provided by the Research Institute of Tuberculosis/Japan Anti- Tuberculosis Association; • delays in the communication of laboratory results and follow-up of confirmed TB cases resulted in delayed case management and loss to follow-up; as a result, not all confirmed TB cases were tested for HIV as planned; and • the lack of clarity of defined roles, responsibilities and deliverables during survey preparations among the four key partners – the survey team, the laboratory, the NTP and the Zimbabwe National Statistics Agency (ZIMSTAT); survey implementation and data management were done by the Biomedical Research Institute and ZIMSTAT, respectively; however, the delayed sharing of datasets and different data management processes between the two agencies made survey management a challenge; other data management issues related to excessive delays caused by double data entry, and the lack of a barcoding system during field data collection. Photo credit: Charles Sandy National TB prevalence surveys 2007–2016272 References 1. The World Bank (https://data.worldbank.org/country, accessed April 2017). 2. UNAIDS (http://aidsinfo.unaids.org/, accessed April 2017). 3. World Health Organization. Global Tuberculosis Database. 2017. (http://www.who.int/tb/country/en/, accessed April 2017). 4. Republic of Zimbabwe – Ministry of Health and Child Welfare. National Tuberculosis Control Program, External Review Report. Republic of Zimbabwe, Ministry of Health and Child Welfare; 2011. 5. WHO Tuberculosis Programme.  (1994).  WHO Tuberculosis Programme: framework for effective tuberculosis control.  World Health Organization.  (http://www.who.int/iris/ handle/10665/58717, accessed January 2018). 6. World Health Organization. Global tuberculosis programme. Global tuberculosis control report 1997. Geneva: WHO; (https:// apps.who.int/iris/bitstream/handle/10665/63354/WHO_ TB_97.225_(part1).pdf?sequence=1, accessed January 2018). 7. World Health Organization. Global tuberculosis report 2014. Geneva: WHO; 2014 (http://apps.who.int/iris/ bitstream/10665/137094/1/9789241564809_eng.pdf ?ua=1, accessed January 2018). 8. Republic of Zimbabwe – Ministry of Health and Child Care. Report of the First National Population-based Tuberculosis Prevalence Survey. Republic of Zimbabwe, Ministry of Health and Child Care, 2015. 9. World Health Organization. Tuberculosis prevalence surveys: a handbook (WHO/HTM/TB/2010.17). Geneva: WHO; 2011 (https://apps.who.int/iris/bitstream/handle/10665/44481/ 9789241548168_eng.pdf, accessed August 2017).