Dengue bulletin, vol. 34, 2010

Dengue Bulletin – Volume 34, 2010 iii Acknowledgements The Editor, Dengue Bulletin, WHO/SEARO, gratefully thanks Donald S. Shepard, Ph.D., Professor, Heller School, Brandeis University Waltham, MA, USA for peer reviewing the manuscripts submitted for publication with assistance from Dr S.S. Vasan, Adjunct Professor, University of Malaya, Malaysia, and Head of Public Health, Oxitec Limited, UK. The editor also thanks Dr Chusak Prasittisuk, former Dengue Bulletin editor, and Dr Jai P. Narain, Director, Department of Communicable Diseases, World Health Organization, Regional Office for South-East Asia, for their support. In-house review: The manuscripts have also been reviewed in-house by Mr Nand L. Kalra regarding format, content, conclusions drawn, including condensation of tabular and illustrative materials for clear, concise and focused presentation, and bibliographic references. He was also involved in the final stages of printing of the Bulletin. The quality and scientific stature of the Dengue Bulletin is largely due to the conscientious efforts of the experts and editors and also due to the positive response of contributors to comments and suggestions.

Dengue Bulletin – Volume 34, 2010 v Contents Overview article Cost and burden of dengue and chikungunya from the Americas to Asia .........................................................................................1 Donald S. Shepard 1. Dengue in Brazil during 1999–2009: A review .................................................6 Luiz Tadeu Moraes Figueiredo 2. Economic cost of dengue public prevention activities in Puerto Rico ...............13 Carmen L. Pérez-Guerra, Yara A. Halasa, Reinaldo Rivera, Marisol Peña, Viani Ramírez, Martha Patricia Cano and Donald S. Shepard 3. Increase in dengue fever imported from Côte d’Ivoire and West Africa to France ......................................................................................24 Guy La Ruche, Charlotte Renaudat, Arnaud Tarantola, Valérie Caro, Martine Ledrans, Dominique Dejour-Salamanca, Laure Diancourt, Hugues Tolou, Marc Grandadam and Marc Gastellu-Etchegorry 4. Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India .................................................................................32 Tiina M. Murtola, S.S. Vasan, Tapasvi I. Puwar, Dipti Govil, Robert W. Field, Hong-Fei Gong, Ami Bhavsar-Vyas, Jose A. Suaya, Marion Howard, Donald S. Shepard, Vijay Kumar Kohli, P.B. Prajapati, Amarjit Singh and Dileep V. Mavalankar 5. Prevalence of chikungunya in the city of Ahmedabad, India, during the 2006 outbreak: A community-based study .....................................40 Tapasvi Puwar, Jay K. Sheth, Vijay Kohli and Rajpal Yadav 6. Prevalence of various symptoms and cost of treatment during the chikungunya epidemic in Ahmedabad, Gujarat, India, in 2006 .................46 Dileep Mavalankar, Dipti Govil, Neha Trivedi and Vinubhai Patel vi Dengue Bulletin – Volume 34, 2010 Contents 7. A private hospital-based study assessing knowledge, attitudes, practices and costs associated with dengue illness in Surat, India ....................54 Ami T. Bhavsar, Donald S. Shepard, Jose A. Suaya, Moses Mafowosofo, Clare L. Hurley and Marion W. Howard 8. Immediate cost of dengue to Malaysia and Thailand: An estimate ...................65 Lee Han Lim, S.S. Vasan, Luise Birgelen, Tiina M. Murtola, Hong-Fei Gong, Robert W. Field, Dileep V. Mavalankar, Nazni Wasi Ahmad, Lokman S. Hakim, Shahnaz Murad, Ng Chiu Wan, Lucy Lum Chai See, Jose A. Suaya and Donald S. Shepard 9. Cost of dengue in Thailand .............................................................................77 Sukhontha Kongsin, Sukhum Jiamton, Jose A. Suaya, Sirijitt Vasanawathana, Petcherut Sirisuvan and Donald S. Shepard 10. Clinical characterization, diagnosis and socioeconomic impact of hospitalized dengue in Cambodia ...................................................................89 Jose A. Suaya, Ngan Chantha, Rekol Huy, Binod K. Sah, Moh-Seng Chang, Duong Socheat, Philippe Buchy, Te Vantha, Ong Sivuth, Elizabeth Haileselassie and Donald S. Shepard Instructions for contributors .........................................................................104 Dengue Bulletin – Volume 34, 2010 1 Overview article Cost and burden of dengue and chikungunya from the Americas to Asia Donald S. Shepard# Brandeis University, Waltham, MA., USA Abstract The ten studies in this special issue document the substantial and growing burden of dengue in the Americas, Africa and Asia, and the burden of a chikungunya outbreak in India. Luiz Tadeu Moraes Figuedo’s paper on dengue in Brazil confirms the country’s worsening trend from 1999–2009, where cases rose at 6.2% per year and dengue deaths at 12.0% per year. Carmen Perez and co-workers, reporting on dengue vector control in Puerto Rico, found that 83% of the costs (US$ 1.97 per person per year) were funded by the lowest and often the least financed level of government: municipalities. Examining dengue cases imported into France, Guy LaRuche documented the alarming increase in cases originating from Cote d’Ivoire from only one case in 2006–07 to six cases in 2008. Using modeling and Monte Carlo simulations, Tiina Murtola and co-authors estimated the “immediate” cost of chikungunya and dengue in India at US$ 1.48 billion (range US$ 0.64 billion to US$ 3.60 billion). Tapasvi Puwar and co-workers, reporting on a 2006 household survey in Ahmedabad, found that only 23% of chikungunya cases sought care in public facilities, so that under-reporting must be considerable. Extending the analysis of this chikungunyua outbreak, Dileep Mavalankar and co-authors placed its economic cost at US$ 8.6-US$ 17.3 million. Ami T. Bhavsar and co-authors, studying dengue cases hospitalized at a private hospital in Surat, India, found that the economic cost of a case averaged US$ 585.57 (US$ 439.44 for direct medical costs and US$ 146.13 for indirect costs). Lee Han Lim and coworkers, estimated the “immediate” cost of dengue to Malaysia and Thailand at US$ 133 to $135 million, respectively. Sukhontha Kongsin and co-authors found that on a per capita basis, costs of dengue in Thailand in 2005 averaged US$ 3.55, of which 28% was due to vector control and 72% due to dengue illness. Examining the burden of dengue on households in Cambodia, Jose A Suaya and co-authors found that and 53% needed to sell household property to fund dengue treatment. Effective methods to prevent the disease would, therefore, result in important economic benefits in many tropical countries. Keywords: Dengue; chikungunya; economic cost; private sector; survey; simulation. # E-mail: shepard@brandeis.edu; Tel: +1-781-736-3975; Fax: +1-888-429-2672 This special issue (2010) addresses the cost and burden of dengue and chikungunya from the Americas to Asia. The World Health Organization (WHO), sponsor of the Dengue Bulletin, deserves commendation for its decision to publish 2 Dengue Bulletin – Volume 34, 2010 Cost and burden of dengue and chikungunya from the Americas to Asia this special issue (2010) in recognition of the importance and growing burden of dengue and chikungunya. While the combined efforts of international and national public health systems have been successful in controlling many infectious diseases, a few, unfortunately, remain stubbornly present. Dengue is among them. Important factors behind the increase in dengue incidence are increasing urbanization, crowding, and spiraling international travel.[1] Nevertheless new preventive strategies are now showing promise[1]. A dengue vaccine is entering Phase 3 clinical testing after successfully completing Phase 2 clinical testing. [2-3] A controlled release to test genetically modified mosquitoes in Asia is expected to begin in 2011.[4] Each of these control measures, however, requires resources to develop and implement. Quantification of the disease burden in both monetary and human terms is a key tool for health policy-makers. That tool allows them to assess trends over time, to compare dengue and chikungunya against other diseases, and to compare one geographical area with another. It can also allow health practitioners to compare one subgroup of patients with another or select one preventive strategy to guide prevention and treatment most appropriately. This special issue examines both dengue and chikungunya because the two related viral diseases have similar symptoms of acute fever and joint pain, and are transmitted by the same vectors, the mosquitoes Aedes aegypti and Aedes albopictus (also called the Asian Tiger Mosquito). The ten original papers in this special issue present a series of approaches and findings to contribute to measuring the cost and burden of dengue and chikungunya from the Americas to Asia. The papers are arranged geographically from west to east to reflect the longitudinal scope of these mosquito-borne diseases in the tropical regions of the world. The first paper, Dengue in Brazil during 1999–2009: A review by Luiz Tadeu Moraes Figueiredo, is based on the paper with the highest western longitude (60o west) in this special issue. Figueiredo’s paper focuses on the epidemiology of dengue over an 11-year period. The data show the remarkable year- to-year variation in incidence of the disease. The number of reported cases in the year of highest incidence (718 000 cases) is seven times the number reported for in the lowest year (113 000). In addition to the fluctuation, the author’s trend lines confirm disconcerting general upward trends corresponding to an annual rise of 6.2% in dengue cases and 12.0% in dengue deaths in the country. The second paper, Economic cost of dengue public prevention activities in Puerto Rico by Carmen Perez and colleagues, does not examine the disease itself, but the prevention and control activities, primarily through vector control. Perez and co-authors report that surveillance and vector control are implemented and funded through a combination of two levels of local government: the affected municipalities and the state (corresponding to provinces in some other countries). Across the study years, overall annual spending in the island totals US$ 1.97 per capita, of which 83% comes from the 12 municipalities with their own programmes and 17% from the state. Clean-up campaigns had the highest share of average expenditure, followed by fumigation, surveillance and inspection. Puerto Rico’s experience highlights the importance of the role of multiple levels of government in dengue prevention. Dengue Bulletin – Volume 34, 2010 3 Cost and burden of dengue and chikungunya from the Americas to Asia The third paper, Increase in dengue fever imported from Côte d’Ivoire and West Africa to France, by Guy La Ruche, provides a creative window on a disturbing trend – the increase in dengue in Africa. Reports to WHO document dengue transmission in Africa in recent history since 1948, with recent major outbreaks in Cape Verde (peaking in November 2009) and the Red Sea state of Sudan (peaking in March 2010). Yet the limited dengue diagnostic and surveillance systems provide few statistics. Using surveillance from international travellers for the years immediately preceding these outbreaks, La Ruche confirms a significant increase in dengue in Côte d’Ivoire from only one case in the 18-month study period in 2006–07 to six cases in 2008. This trend and the 148 imported cases to metropolitan France from 2006 to 2008 highlight the value of international cooperation in studying and controlling the disease. The fourth paper, A preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India, by Tiina Murtola and co-authors, is one of two papers in this special issue to use Monte Carlo simulations with existing data to extrapolate the annual burden of dengue or chikungunya to a state or national level. To address the fact that existing surveillance systems capture only a fraction of the actual cases, this paper develops the “RUHA” matrix by estimating shares of reported (R) and unreported (U) hospitalized (H) and ambulatory (A) dengue cases. The paper calculates that the immediate cost to households of chikungunya and dengue in the state of Gujarat was estimated to be 3.8 (range 1.6–9.1) billion Indian rupees (INR) per annum. The fifth paper, Prevalence of chikungunya in the city of Ahmedabad, India, during the 2006 outbreak: A community-based study, by Tapasvi Puwar and co-authors, describes an extensive household survey of 1301 households across 43 clusters. It ascertained the magnitude and characteristics of the disease in the city of 3.5 million persons. The authors found that 32.9% of the persons surveyed experienced the disease, of which the major symptoms were fever, chills, headache, joint swelling and itching. The prevalence was higher in slum neighbourhoods (where more than 40% of the population lived) than in ones with bungalows and apartments. The municipal authority officially reported 60 777 cases within the city limits. However, the survey found that only 23% of cases sought treatment within public facilities compared with 68% from private facilities and 9% that sought no treatment. Thus, they conclude that the official reports may substantially understate the total burden of illness. The sixth paper, Prevalence of various symptoms and cost of treatment during chikungunya epidemic in Ahmedabad, Gujarat, India, in 2006, by Dileep Mavalankar and co-authors, estimates the economic cost of the 2006 outbreak of the viral disease with further analyses of the survey data provided in the previous paper of Puwar et al. The sixth paper notes that the disease affected primarily working-age adults, with the highest number of cases occurring in the age decade of 30–39. For this neglected disease in this one city, the authors estimated that the immediate cost of the outbreak due to lost wages and treatment costs was approximately US$ 1.7 million based only on officially reported cases. Assuming that the actual number of cases may be 5 to 10 times the reported number, the actual economic cost of the chikungunya outbreak in Ahmedabad may have been US$ 8.6 to US$ 17.3 million, respectively. India is also the setting of the seventh paper, A private hospital-based study assessing knowledge, attitudes, practices and costs 4 Dengue Bulletin – Volume 34, 2010 Cost and burden of dengue and chikungunya from the Americas to Asia associated with dengue illness in Surat, India, by Ami T. Bhavsar and colleagues. As many private facilities and their patients are often reluctant to share their data with researchers, most studies of disease cost and burden are set in public facilities. This paper, set in a medium- sized private hospital in the city of Surat in Gujarat, India, is a welcome exception. The study found that for an average inpatient dengue episode in this facility, direct medical costs averaged US$ 439.44. Indirect costs added US$ 146.13, bringing the total cost per case to US$ 585.57. The study’s survey found considerable scope for enhanced prevention. Only 25% of respondents correctly answered that the dengue vector breeds in clean, stagnant water. Fully 93% of households stored water for daily use, a practice that facilitates mosquito breeding. Finally, the study demonstrates dramatically that dengue affects all economic strata. Ninety per cent of hospital patients came from the higher socioeconomic strata compared with only 39% of the population in urban Gujarat. The second paper modelling the economic impact of data from multiple sources is the eighth paper in this special issue, Immediate cost of dengue to Malaysia and Thailand: An estimate, by Lee Han Lim and co-workers. The study is notable for deriving comprehensive estimates of dengue cost in both countries and exploring the uncertainties in existing data. The authors found the “immediate” annual cost of dengue to be in the range of US$ 88 to US$ 215 million (mean of US$ 133 million) for Malaysia and US$ 56 to US$ 264 million (mean of US$ 135 million) for Thailand. In Malaysia, the most important parameters creating uncertainty in the immediate cost are the reporting rate, the hospitalization rate, and cost per ambulatory case. In Thailand, the corresponding parameters are cost per ambulatory case, cost per hospitalized case, and reporting rate. To improve estimates of dengue costs, future studies should also refine the estimates of the hospitalization rate in Malaysia and the cost per hospitalized case in Thailand. Similar to the study from Surat, the ninth study, Cost of dengue in Thailand by Sukhontha Kongsin and co-authors, is primarily based on a facility-level observational study of dengue patients. The authors found that the economic cost per non-fatal case in Khon Kaen Provincial Hospital in 2005 averaged US$ 573. Some previous economic studies have examined only household out-of-pocket payments for treatment. This study measured total resource use from all sources, including government, households and employers. On average, inpatient care cost US$ 418 per case, almost all is paid for by the government. Ambulatory care and direct non-medical care (mostly transportation) averaged US$ 49 and US$ 60 per case, respectively. Further, indirect costs (the value of time lost) averaged US$ 45 per case, mostly incurred by households. Thus, all payers incur substantial costs for a dengue case. Finally, the paper also estimated the cost of vector control, an important dimension that is empirically derived in very few studies. Overall, per capita costs of dengue in Thailand in 2005 were US$ 3.55, of which 28% was due to vector control and 72% due to dengue illness. The tenth study, Clinical characterization, diagnosis and socioeconomic impact of hospitalized dengue in Cambodia, by Jose A. Suaya and co-authors, uses the same data collection instruments and methods as the ninth paper (from Thailand). When measured in US dollars, the resource cost of a hospitalized dengue case in Cambodia (US$ 116) was substantially less than that in Thailand. Yet the economic hardship associated with a dengue hospitalization in Cambodia was extraordinarily high, with the Dengue Bulletin – Volume 34, 2010 5 Cost and burden of dengue and chikungunya from the Americas to Asia majority (88%) of households having these cases reporting a substantial adverse economic impact. For example, to pay for treatment, 39% of the households needed to borrow money beyond family or friends, and 53% needed to sell household property. These adverse effects are the combined results of considerably lower per capita income and the requirement in Cambodia that patients pay out-of-pocket for the majority of hospital costs, even in a government hospital. Thus, hospitalized dengue has major clinical and socioeconomic consequences in Cambodia. Overall, these ten studies document the substantial and growing burden of dengue in the Americas, Africa and Asia. Effective methods to prevent the disease would, therefore, result in important economic benefits in many tropical countries. References [1] Gubler, D.J. (1997) Dengue and dengue hemorrhagic fever; its history and resurgence as a global public health problem. In Dengue and Dengue Hemorrhagic Fever (Gubler, D.J. and Kuno, G., eds), pp. 1–22, CAB International Press. [2] Yoksan S. WHO’s efforts for development of a dengue vaccine. Dengue Bulletin 2008, 32: 1-16. [3] Sanofi Pasteur’s dengue vaccine in final stage of clinical development. Lyon, France: Sanofi Pasteur, 2010. Web: en.sanofi-aventis.com/ binaries/20101104_Dengue_en_tcm28- 29527.pdf [4] Lee HL, Joko H, Nazni WA, Vasan SS. Comparative life parameters of transgenic and wild strains of Aedes aegypti in the laboratory. Dengue Bulletin 2009, 33: 103-314. 6 Dengue Bulletin – Volume 34, 2010 Dengue in Brazil during 1999–2009: A review Luiz Tadeu Moraes Figueiredo# Vírus Research Unit of the School of Medicine, University of São Paulo, Ribeirão Preto, Av. Bandeirantes, 3900, 14049-900 Ribeirão Preto, SP, Brazil Abstract Over the period 1999 through 2009, the incidence of dengue in Brazil has shown an upward trend with more than 700 000 cases per year reported during 2002 and 2008. Despite efforts to control Aedes aegypti, a hyper-endemic situation has remained with large outbreaks involving all the four dengue serotypes (DENV-1 to 4). The average age of DHF/DSS cases has decreased, making it a disease affecting children. The number of fatalities has also increased. While other countries in the Americas have also witnessed an upward trend in case numbers, Brazil has been subject to more extreme peaks. Other arboviruses also cause sporadic human cases which are commonly confused with dengue and remain undiagnosed. Keywords: Dengue outbreaks; DHF/DSS; burden of dengue; Brazil; arbovirus. #E-mail: ltmfigue@fmrp.usp.br Introduction Brazil, the largest country in South America (area of 8 512 000 sq. km.), has a population of 192.9 million,[1] most of them living in urban areas of large cities infested by the mosquito Aedes aegypti and having frequent dengue outbreaks. The first outbreak occurred in the north-west Amazon region in the state of Roraima caused by DENV-1 in 1982.[2] Since then Brazil has become hyper-endemic for dengue and all the four serotypes (DENV-1 to 4) are circulating in the country. DENV are 60nm spherical enveloped viruses with surface glycoprotein projections, possessing a single- stranded RNA(+) genome of approximately 11 000 nucleotides. Viruses of the genus Flavivirus contain a single open reading frame RNA that encodes the 10 viral proteins, C-preM-E-NS1-ns2a-ns2b-NS3-ns4a-ns4b- NS5.[3] Epidemiological data Dengue viruses and the vector Aedes aegypti have spread throughout the country leading to outbreaks in all regions, including the most populated areas of Brazil, as shown in Figure 1. A sequence of outbreaks has followed the introduction into the country of DENV-1 genotype II in 1986, DENV-2 genotype I in 1990, and DENV-3 genotype III in 2000.[4] Recently, DENV-4 has been reported as causing acute febrile illness in Manaus city in the north of the country. Dengue Bulletin – Volume 34, 2010 7 Dengue in Brazil during 1999-2009 Figure 1: Map of Brazil showing the circulation of dengue virus serotypes by macro-region in 2008 [Source: Brazilian state public health laboratories (LACEN) and www.santiagosiqueira.pro.br] Distinct dengue virus serotypes and genotypes have been observed in molecular epidemiology studies, spreading from Rio de Janeiro toward the Northeast and Midwest in large outbreaks. Dengue viruses have also spread from Brazil to neighbouring countries such as Paraguay and Argentina. Two introductions of DENV-3 into the country were also observed, in Rio de Janeiro and in the North, both probably from the Caribbean.[5] In 2008 DENV-2 genotype I resurfaced in Rio de Janeiro leading to a huge outbreak with hundreds of DHF/DSS cases.[6] About five million dengue cases were reported in Brazil from 1985 to 2008.[7] Figure 2 shows that the number of cases has fluctuated seven-fold from 113 000 to 781 000 cases per year over the period 1999 through 2009. In the years 2002 and 2008, the annual number of reported dengue cases exceeded Figure 2: Dengue (including DHF) cases in Brazil and elsewhere in the Americas 204 231 413 781 342 113 204 347 560 734 407y = 233.97 e 0.0622x y = 102.42 e 0.0741x 0 100 200 300 400 500 600 700 800 900 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009* Year Brazil Americas excluding Brazil Expon. (Brazil) Expon. (Americas excluding Brazil) C as es p er ye ar (i n th o u sa n d s) Source: Pan American Health Organization/World Health Organization 8 Dengue Bulletin – Volume 34, 2010 Dengue in Brazil during 1999-2009 700 000,[8] and the reported numbers have shown a general upward trend: an exponential curve fitted to these data shows a compound rate of growth of 6.2% per year. Similar data for nine countries in the Americas (Argentina, Columbia, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Peru and Venezuela) with complete data over the same years show an even more rapid upward trend (7.4% per year), but the peak values in Brazil have been much higher than those in the peak years for the other countries in the region. In 2009, however, the cases did not peak in Brazil as they did for other countries in the region, such as Argentina, which saw more cases in one year (25 000) than in the previous 10 years combined. The number of deaths from dengue in Brazil shows a similar upward trend (Figure 3). Exponential curves show average annual increases of 12% for Brazil and 38.8% for the other nine countries in the Americas. The figure for the highest year in Brazil (317 in 2007) is four times that for the lowest year (71 in 2004). Dengue diagnosis Diagnosis of dengue in Brazil has been made mostly by enzyme immunoassays detecting IgM-specific antibodies. Considering the increased number of dengue secondary infections it is possible that many patients Figure 3: Dengue deaths in Brazil and elsewhere in the Americas 82 92 140 255 164 71 159 193 317 306 y = 81.184e 0.1204x y = 3.4231e 0.3883x 0 50 100 150 200 250 300 350 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009* Year Brazil Americas excluding Brazil Expon. (Brazil) Expon. (Americas excluding Brazil) D ea th s p er ye ar Source: Pan American Health Organization/World Health Organization Dengue Bulletin – Volume 34, 2010 9 Dengue in Brazil during 1999-2009 presenting anamnestic responses, with low anti-dengue IgM serum levels, have been undiagnosed. In addition, virus isolation has been performed in only a small number of samples and diagnosis by reverse transcription followed by real-time polymerase chain reaction (RT-PCR) has been sporadic. To improve dengue diagnosis, Brazilian authorities could consider more widespread adoption of at least two routine diagnostic methods. Ideally, these methods should be simple, sensitive and able to diagnose dengue at the onset of symptoms. Examples are dengue virus NS1 protein detection by enzyme immunoassay, RT-PCR allowing virus types and genotypes to be determined, or RT-PCR allowing determination of the dengue virus load in the blood.[9] Other useful methods are serological tests for detection of anti- dengue IgG antibodies allowing differentiation between primary and secondary dengue infections. Dengue secondary infections and DHF/DSS The hyper-endemic dengue situation in Brazil has led to many secondary infections. In addition, the presence of DENV-3 genotype III and DENV-2 genotype I, both virulent, resulted in an increasing number of dengue haemorrhagic fever/dengue shock syndrome (DHF/DSS) cases. During 1999–2000 DHF/ DSS outbreaks were caused by DENV-2, Figure 4: DHF/DSS cases and deaths reported in Brazil, 1999–2008 72 888 679 2714 727 81 1395 642 1541 647 0 500 1,000 1,500 2,000 2,500 3,000 1999 [2] 2000 [2] 2001 [3] 2002 [3] 2003 [3] 2004 [3] 2005 [3] 2006 [3] 2007 [3] 2008 [2] Year and predominant dengue virus serotype (2 denotes dengue virus 2; 3 denotes dengue virus 3) Note: Data for 2008 are for a partial year. 0 20 40 60 80 100 120 140 160 180 DHF Reported Cases DHF Reported Deaths N u m b er o f D H F ca se s N u m b er o f D H F d ea th s Source: FVS, Brazilian Ministry of Health 10 Dengue Bulletin – Volume 34, 2010 Dengue in Brazil during 1999-2009 from 2001 to 2007 by DENV-3, and in 2008, again, by DENV-2. About 10 000 DHF/DSS cases with 1000 deaths have been reported in these last 10 years.[6] However, it is possible that a large number of DHF/DSS cases remain unreported. Figure 4 presents information on the number of dengue, DHF/DSS and fatal dengue cases, as well as the viral types predominant in each year from 1999 to 2008. Years with a high number of dengue cases also tend to experience a larger number of dengue deaths. In the last seven years, the average age of DHF/DSS patients has been decreasing. In large cities of the North such as Manaus, the disease is already predominant in children, and this trend has also been seen in the Northeast and in Rio de Janeiro.[6] Therefore, DHF/DSS in Brazil is gradually becoming a children’s disease, as it is in Asia,[10] although with a smaller number of patients. Dengue management and control The management of dengue patients, and especially DHF/DSS cases, has been a challenge in Brazil. The surveillance data show that case-fatality rates from DHF have reached as high as 20 per cent. The programme for dengue control aims at lowering Aedes aegypti infestation to under 5% of houses,[11] but the periodic outbreaks show the challenges in realizing this goal. Other arboviruses A final challenge in managing dengue cases in Brazil is that many patients with acute febrile illness diagnosed as having dengue fever are, probably, infected by other arboviruses. The nature of these infections remains unknown in most cases since the public health authorities focus on dengue diagnosis. Flaviviruses such as Saint Louis encephalitis were found to be causing an outbreak which included meningoencephalitis cases in São José do Rio Preto County, in rural Southeastern Brazil, in 2006 and 2007.[12] Despite the availability of an efficient vaccine, sylvatic yellow fever remains a worrisome problem in Brazil. Epizootics of sylvatic yellow fever were reported in the South, Southeast and Midwest regions in 2008–2009, and first observed to be causing a high mortality in sylvatic monkeys followed by the appearance of human disease: more than 100 individuals have been assaulted by yellow fever in the last two years.[13] Likewise, cacipacore, another flavivirus, has been reported to be causing acute febrile illness in the West[14] (Batista W., personal communication 2009). Oropouche orthobunyavirus has caused outbreaks in Manaus, a large city and capital of Amazon state (Mourão M.P., personal communication 2009) as well as in Pará[15] and Acre[16] states, during 2006, 2007 and 2008. Mayaro alphavirus is also producing dozens of acute febrile illness cases in the North of Brazil (Mourão M.P., personal communication 2009). Conclusions In conclusion, the burden of dengue in Brazil has increased during the decade 1999–2009. On average, the number of dengue cases and numbers of deaths have grown by 6.2% and 12% per year, respectively. These facts suggest that dengue is becoming more frequent and more severe. These trends parallel increases throughout the Americas. In both Brazil and other countries, the number of both dengue cases and deaths vary enormously on a year- to-year basis. Dengue Bulletin – Volume 34, 2010 11 Dengue in Brazil during 1999-2009 References [1] Instituto Brasileiro de Giographia e Estatistica (IBGE). Popclock Population. http://www.ibge. gov.br/english/}. Accessed: 14 May 2010. [2] Osanai CH. A epidemia de dengue em Boa Vista, Território Federal de Roraima, 1981-1982. Dissertação de mestrado, Escola Nacional de Saúde Pública, Rio de Janeiro, RJ, 1984. [3] Lindenbach B.D., Thiel H-J, Rice C.M. Flaviviridae: the viruses and their replication. In: Knipe D.M., Howley P.M. (eds), Fields Virology, 5th ed. Philadelphia: Lippincot Williams & Wilkins, p. 1101-1151, 2007. [4] Figueiredo LTM. Dengue in Brazil: Past, present and future perspective. Dengue Bulletin 2003; 27: 25-33. [5] Aquino V.H., Anatriello E., Silva E.V., Vasconcelos PFC, Vieira D.S., Batista W.C., Bobadilla M.L., Vasquez C., Moran M., LTM. F. Molecular epidemiology of dengue type 3 virus in Brazil and Paraguay. American Journal of Tropical Medicine and Hygiene 2006; 75: 910-5. [6] SVS. Casos de dengue no Brasil: Brasília: Ministério da Saúde, 2009. The efforts to control Aedes aegypti face many challenges, and the four dengue serotypes continue to cause large and successive outbreaks, characterizing a hyper- endemic situation in Brazil and more widely in the Americas. With a disproportionate increase in the number of children affected, dengue is becoming a children’s disease in Brazil. Furthermore, other arboviruses such as yellow fever, Saint Louis encephalitis, cacipacore, oropouche and mayaro are also causing outbreaks and sporadic human cases in the country. These infections are commonly confused with dengue and remain undiagnosed. This should be of concern to public health authorities, and should stimulate the use of arbovirus diagnostic methods in national and state public health laboratories. Acknowledgements The author thanks Clare Hurley, MM, and Eduardo Undurraga, MA, of Brandeis University for preparation of figures on the dengue burden. [7] Figueiredo RMP, Naveca FG, Bastos MS, Melo MN, Viana SS, Mourão MPG, Costa CA, IP F. Dengue Virus Type 4, Manaus, Brazil. Emerging Infectious Diseases 2008; 14: 667-9. [8] Pan American Health Organization/World Health Organization. 2007: Number of reported cases of dengue and dengue hemorrhagic fever (DHF), region of the Americas (by country and subregion). http:// www.paho.org/english/ad/dpc/cd/dengue- cases-2007.htm. Accessed: 21 May 2010. [9] Santos HWG, Poloni TRRS, Souza K.P., Muller VDM, Tremischin F, Nali LC, Fantinatti L.R., Amarilla A.A., Castro HLA, Nunes M.R., Vasconcelos P.F., Figueiredo LTM, VH. A. A simple one-step real-time RT-PCR for diagnosis of dengue virus infection. Journal of Medical Virology 2008; 80: 1426-33. [10] World Health Organziation (WHO). Dengue and dengue haemorrhagic fever, Fact Sheet No. 117. http://www.who.int/mediacentre/ factsheets/fs117/en/. Accessed: 21 May 2010. [11] FUNASA. Dengue. Instruções para pessoal de combate ao vetor. Manual de normas técnicas: Brasília: Ministério da Saúde, 2001. 12 Dengue Bulletin – Volume 34, 2010 Dengue in Brazil during 1999-2009 [12] Mondini A., Lázaro E., Cardeal ILS, Nunes S.H., Moreira C.C., Rahal P., Figueiredo LTM, Bronzoni RVM, Neto F.C., Nogueira M.L. Saint Louis encephalitis virus, Brazil. Emerging Infectious Diseases 2007;13:176-8. [13] SVS. Situação da Febre Amarela Silvestre no Brasil.: Brasília: Ministério da Saúde, 2007, 2008 e 2009. [14] Figueiredo LTM. The Brazilian Flaviviruses. Microbes and Infection 2000; 2: 1643-9. [15] Vasconcelos H.B., Azevedo R.S., Casseb S.M., Nunes-Neto J.P., Chiang J.O., Cantuária P., Segura M., Martins L.C., Montiero H.A., Rodregues S.G., Nunes M.R., Vasconcelos P.F. Oropouche fever epidemic in Northern Brazil: Epidemiology and molecular characterization of isolates. Journal of Clinical Virology 2009; 44: 129-33. [16] Terzian ACB, Bronzoni RVM, Drumond B.P., Silva-Nunes M., Silva N.S., Ferreira U.M., Sperança MA, ML N. Sporadic Oropouche virus infection: Acre, Brazil. Emerging Infectious Diseases 2009; 15: 338-50. Dengue Bulletin – Volume 34, 2010 13 Economic cost of dengue public prevention activities in Puerto Rico* Carmen L. Pérez-Guerraa#, Yara A. Halasab, Reinaldo Riveraa, Marisol Peñaa, Viani Ramíreza, Martha Patricia Canoc and Donald S. Shepardb aCenters for Disease Control and Prevention, Dengue Branch, San Juan, Puerto Rico 00920 USA bSchneider Institutes for Health Policy, Heller School, MS 035, Brandeis University, Waltham, MA 02454-9110 USA cPuerto Rico Department of Health, Rio Piedras, Puerto Rico, USA Abstract Dengue fever has become a major global public health problem in Puerto Rico. Approximately 5000 suspected cases were reported annually between 2002 and 2007. Vector control is currently the only approach to control the disease and includes prevention education, fumigation, inspections and clean-up campaigns. The annual cost to the public sector of dengue prevention, which includes surveillance and vector control activities, was estimated as part of a study of the economic burden of dengue in Puerto Rico. A telephonic survey was implemented to identify municipalities with vector control programmes and public agencies with dengue surveillance systems. Onsite interviews were conducted using a structured questionnaire. The economic cost of dengue was summarized by line item, function and year from 2002 through 2007. The Puerto Rico Department of Health (PRDH) and 12 municipalities out of 78 conducted vector control activities in different magnitudes during the study years. The cumulative cost of dengue vector control in the public sector was US$ 46.22 million for the years 2002–2007. PRDH spent an average of US$ 1.29 million ($0.33 per capita) per year, while the municipalities spent an average of US$ 6.41 million (US$ 1.64 per capita) per year. Clean-up campaigns had the highest share of average expenditure, followed by fumigation, surveillance and inspection. Puerto Rico’s per capita expenditure on dengue prevention activities is similar to that of other countries in the region. On average, Puerto Rico’s per capita spending on dengue illness is US$ 5.48 compared with US$ 1.97 spent on vector control. Keywords: Dengue; economic cost; control activities; public exchequer; Puerto Rico. *Supported by Sanofi Pasteur (Lyon, France) and the authors’ institutions #E-mail: cnp8@cdc.gov Introduction Dengue fever has become a major global public health problem.[1–6] It has spread to more than 100 countries within its usual tropical boundaries with the occurrence of frequent and cyclical epidemics, and also beyond to new and hitherto uncharted 14 Dengue Bulletin – Volume 34, 2010 Economic cost of dengue public prevention activities in Puerto Rico territories such as Africa and West Asia.[7–9] Studies estimated an annual 50 to 100 million cases of human infection of dengue fever, of which 500 000 are of dengue haemorrhagic fever, and 22 000 deaths occur per year, mainly in paediatric patients.[7–10] Dengue is caused by four related, but antigenically distinct, viruses (DENV-1, DENV-2, DENV-3 and DENV-4) belonging to the genus Flavivirus, family Flaviviridae[7,11-14]. Dengue viruses are transmitted in an endemic/ epidemic, human-mosquito-human, rural or urban cycle by the bite of infected mosquito vectors from the Aedes family; mainly, Aedes aegypti and Aedes albopictus.[14-15] Aedes aegypti is a very effective vector of dengue viruses and adaptive to exploiting the domestic environment. Female Aedes aegypti mosquitoes are very nervous feeders, they can often discontinue the feeding process and restart on the same or another individual soon after. Therefore, a single infected mosquito can transmit the dengue virus to multiple people over a short period of time.[14] Additionally, the female Aedes aegypti almost exclusively bites humans, most commonly feeds and rests indoors and is predisposed to lay eggs in man-made containers commonly found in and around homes such as flower vases, buckets, water storage containers, tyres and any trash items that collect water.[11,15] Infection with one dengue virus serotype results in lifelong homotypic immunity.[15] Cross-protection heterotypic immunity may last up to a year.[14-16] However, exposed individuals can theoretically be infected with all four serotypes with an increased risk of developing dengue haemorrhagic fever with subsequent infections.[15,16] From the epidemiological perspective, dengue includes four diseases. However, from the clinical perspective it is only one with five different presentations: non-specific febrile illness, classic dengue fever, dengue hemorrhagic fever, dengue shock syndrome and other severe dengue syndromes.[14] Potential approaches to reduce dengue infection include reduction of mosquito abundance, prevention of contact between the vector and humans, genetically manipulated vector mosquito and vaccine.[11,15] Currently, the only available control strategies are reducing mosquito abundance, reducing adult mosquito lifespan and preventing mosquito- human contact. Despite tremendous efforts by public health organizations in dengue endemic countries to achieve effective and sustainable control of Aedes aegypti, and thereby the outbreak of dengue cases, these efforts were very difficult to sustain due to uncontrolled urbanization, global spread of dengue viruses, lack of adequate resources for vector and dengue control programmes, the use of ineffective vector control methods, passive surveillance systems, the failure of governments to utilize data from active surveillance systems when available, and the need for substantial regional efforts to eliminate the vector in some regions.[2,3,4] Several approaches to eliminate the dengue vector have been implemented with various degrees of success. Findings on the effectiveness of these approaches are inconsistent. Many studies reported an effective reduction in the vector population using a community-based integrated approach tailored to local eco-epidemiological and sociocultural settings. This approach includes an educational component to increase knowledge and understanding of best practices. [17-19] A recent evaluation of larviciding from Cambodia found the programme effective Dengue Bulletin – Volume 34, 2010 15 Economic cost of dengue public prevention activities in Puerto Rico and cheaper in reducing dengue cases.[20] However, a recent systematic review found weak evidence that community-based dengue control programme alone or in combination with other control activities can enhance the effectiveness of the dengue control programme.[21] Puerto Rico gives an excellent example to study vector control costs. It has a passive and active surveillance system with strong laboratory capabilities in both serology and virology, although it lacks the ability to predict an epidemic. Dengue is a major public health concern in Puerto Rico. According to the Dengue Branch of the Centers for Disease Control and Prevention (CDC) in Puerto Rico, 30 482 suspected dengue cases were reported between 2002 and 2007. Of these, 9738 were confirmed dengue cases, with nearly 2.3% developing dengue haemorrhagic fever (DHF). In 2007, due to an island-wide dengue outbreak, 10 508 suspected dengue cases were reported, of which 3294 were dengue confirmed cases and 2.6% of the cases were DHF. The general infection rate was 8.6 laboratory-positive cases per 10 000 population in 2007, and the severity of dengue has been increasing in Puerto Rico.[22] In Puerto Rico, the vector control activities are carried out primarily by the Puerto Rico Department of Health (PRDH), which is responsible for island-wide outdoor fumigation to control Aedes aegypti adult population and premise inspection, and larviciding to control aquatic phases and larval habitats for the vector mosquito. Additionally, educational campaigns and epidemiological surveillance are conducted by the PRDH with support from the CDC’s Dengue Branch. In addition to PRDH and CDC’s Dengue Branch, several island municipalities have vector control programmes. Combined PRDH and these municipalities create the public vector control sector in Puerto Rico. The PRDH is the official organization responsible for disease prevention; its total per capita expenditure in controlling infectious diseases (of which dengue is one) for the year 2009 was US$ 217.51. In comparison, the overall PRDH per capita expenditure on tuberculosis for 2009 was US$ 0.21 and US$ 0.06 for diabetes, (Unpublished data, PRDH, 2009). Providing information on the cost of these activities is informative for policy-makers to allocate and/or reallocate scarce resources to projects which have the greatest benefit to society. Such data are also needed to balance existing and new potential strategies (e.g., vaccine and novel vector control technologies) against dengue, and to compare dengue with other diseases. Although studies from other localities, such as Panama and Thailand have emerged,[22,23] no previous report of the costs of vector control in Puerto Rico had been prepared. Therefore, it is important to understand the current structure of dengue prevention activities, including vector control activities, and the services provided covering the cost of different vector control activities to mitigate the disease or estimate the associated costs of the same in Puerto Rico. This paper presents a comprehensive analysis of dengue prevention activities, including vector control in Puerto Rico. It defines the main players involved in these activities and estimates the cost of these activities over a six-year period. This study is part of the economic burden of dengue in Puerto Rico, and seeks to inform future policy. 16 Dengue Bulletin – Volume 34, 2010 Economic cost of dengue public prevention activities in Puerto Rico Materials and methods This retrospective descriptive study was conducted between April 2009 and June 2010 to estimate the annual cost of dengue prevention activities, including surveillance and vector control initiatives conducted in Puerto Rico during the years 2002 through 2007. Each of the 78 municipalities was contacted by telephone to determine which ones had carried out an active vector control programme during the study period. Based on pilot studies in 2006 and 2008, a structured questionnaire was developed for collecting cost data from the PRDH and these jurisdictions. The questionnaire was divided into four parts. The first dealt with personnel cost, type of activities preformed by each worker and the time and effort allocated to dengue prevention activities. The second dealt with recurrent costs, including cost of chemicals used; personal protection equipment; maintenance of the buildings, vehicles and fumigation equipment; insurance of the buildings, vehicles and equipment; utilities, education materials, the aggregate cost of clean-up campaigns allocated to dengue, office supplies, advertisements and fuel. The third part focused on the amortized capital cost of buildings, spraying and fogging equipment, vehicles, laboratory equipment, and office equipment including computers and software. The fourth and final part requested data on the type and number of activities undertaken during the study years. Inputs from other institutions, volunteer efforts and donations to activities designed to control dengue were also tallied. The questionnaire was validated by review of officials from Carolina, the municipality with the largest programme, and sent to each jurisdiction by e-mail, mail and fax prior to an onsite personal interview. The onsite interviews were conducted between May 2009 and May 2010 with vector control personnel from the PRDH and the active municipalities. For each jurisdiction, an average of four visits were needed to interview executive directors and other key personnel to determine the level of effort given to dengue prevention and collect the cost data. Breakdowns of all resources were also estimated according to the following functions: inspections, fumigation, education, clean-up campaigns, surveillance and general management. Inspections denote visits to premises where a dengue case or a mosquito-larval habitat had been identified through a resident complaint or a vector control inspector. Whenever possible, the larval habitat (e.g. location of standing water) was removed or destroyed. When destruction is not possible, a larvicide is applied. Fumigation refers to outdoor spraying of insecticides (malathion, permethrin, aqualeur, etc.) with ultra-low volume (ULV) foggers or thermal fogging machine mounted on pick- up trucks. Fumigations take place according to the daily schedule in some municipalities or during a local or national outbreak when authorized by a municipal mayor or the PRDH, or in response to residents’ complaints about increased mosquito populations. Educational efforts include the preparation of leaflets, brochures, posters, videos, documentaries, newspaper articles, public service announcements on television and radio spots with the purpose of informing and educating the population about dengue, disease outbreaks and its prevention. Clean-up campaigns are conducted by the Waste Management Program of the municipalities to collect potential larval habitats for Aedes aegypti, in addition to its main Dengue Bulletin – Volume 34, 2010 17 Economic cost of dengue public prevention activities in Puerto Rico purpose of maintaining a clean environment in the cities. During a dengue outbreak, clean- up campaigns are organized by the municipal health officials and vector control programme personnel within the collaboration with other municipal employees. Surveillance denotes collecting, analysing and reporting the results from blood sample tests received from private and public clinics and hospitals island-wide, usually run by CDC/ PRDH Passive Dengue Surveillance System. General management refers to the facilitation of specific dengue prevention activities such as receiving complaints, scheduling inspection visits and larviciding of premises, educational activities, and fumigation. Purchase of vector control equipment and supplies as well as printing of educational materials, management and transportation of blood samples and reporting are also part of general management activities. Results All 78 municipalities responded to the telephonic interviews. The results of these interviews highlighted that in addition to PRDH, 12 municipalities have vector control programmes, and all have at least three activities to control and prevent dengue, namely outdoor fumigation, educational activities and clean-up campaigns. The history and magnitude of dengue vector control activities varies across jurisdiction; the programmes were well established in Guaynabo and Carolina while Ponce and Bayamon initiated their programmes in 2003. While all jurisdictions were active in clean-up campaigns, the Bayamon programme started in 2004, Lajas in 2005, Toa Baja in 2006 and Juana Diaz in 2007. Manati started their fumigation programme in 2006. Table 1 lists municipalities with dengue vector control programmes and the PRDH, their vector control activities, and the average population they served between 2002 and 2007. The funding and implementation of vector control is decentralized in Puerto Rico. The state pays, through the PRDH, for 16.8% of the total public expenditure of dengue prevention activities, while the 12 municipalities pay the overwhelming share of 83.2%. The PRDH spent an average US$ 1.29 million per year (US$ 0.33 per capita) while the 12 municipalities spent on average US$ 6.41 million per year (US$ 1.64 per capita). The results illustrate that only Carolina and Guaynabo have comprehensive vector control programmes, including fumigation, clean- up campaigns, inspection and surveillance, while the other ten municipalities have more limited programmes in terms of their scope and continuity of services, and the degree of focus on fumigation, education and clean-up campaigns. As illustrated in Table 2, the cumulative six-year cost of dengue vector control in the public sector was US$ 46.22 million for the years 2002 to 2007. The annual expense increased gradually from US$ 6.49 million (US$ 1.68 per capita) in 2002 to US$ 9.41 million (US$ 2.39 per capita) in 2007 due to the introduction of new vector control programmes in some municipalities and the expansion of existing vector control activities in others in response to an island-wide epidemic in 2007. The average annual cost of dengue vector control for 2002–2007 was US$ 7.70 million per year (US$ 1.97 per capita). 18 Dengue Bulletin – Volume 34, 2010 Economic cost of dengue public prevention activities in Puerto Rico Table 1: Dengue prevention activities by jurisdiction and population served Jurisdiction Fumigation Inspections Surveillance Education Clean-up campaigns Population average (2002– 2007)* Population (%) Puerto Rico Department of Health X X X X 3 901 234 100 Municipality .Bayamón X X X 222 350 5.70 .Cabo Rojo X X X 50 894 1.30 .Carolina X X X X X 187 314 4.80 .Dorado X X X 35 575 0.91 .Guayanilla X X X 23 569 0.60 .Guaynabo X X X X X 102 088 2.62 .Juana Díaz X X X 52 224 1.34 .Lajas X X X 27 316 0.70 .Manatí X X X 48 114 1.23 .Ponce X X X 182 659 4.68 .Salinas X X X 31 784 0.81 .Toa Baja X X X 94 814 2.43 *Source: Population Division, U.S. Census Bureau. Release Date: 19 March 2009. As shown in Figure 1, among the municipalities, Carolina had the highest average expenditure (US$ 4.197 million per year or US$ 22.41 per capita), while Juana Díaz had the lowest average expenditure (US$ 21 266 per year, or US$ 0.41 per capita). The range in per capita spending is enormous; with figures for the highest municipality (Carolina) being fifty times that of the lowest (Juana Diaz). As clean-up campaigns are often a particularly effective strategy, Carolina is noteworthy for spending 92% of its annual expenditure on this function. In contrast, the other 11 municipalities had an average annual budget of US$ 104 553 of which 52% was spent on clean-up campaigns. Supplementary tables illustrating the breakdown of vector control activities by the municipalities classified by line item, function and year are available on request. Analysing costs according to the line item reveals that recurrent costs were the highest, with an average of US$ 5.57 million per year Dengue Bulletin – Volume 34, 2010 19 Economic cost of dengue public prevention activities in Puerto Rico Table 2: Expenditure on dengue control by year and jurisdiction, 2002–2007 (in US$) Jurisdiction 2002 2003 2004 2005 2006 2007 Total Average PRDH* 1 329 191 1 404 541 1 412 076 977 661 1 034 167 1 621 606 7 779 243 1 296 540 Municipality Bayamon 0 124 765 190 834 235 214 253 089 265 466 1 069 369 178 228 Cabo Rojo 128 796 129 293 129 769 136 589 142 763 148 659 815 870 135 978 Carolina 3 599 397 3 787 785 4 006 174 4 389 204 4 544 182 4 859 969 25 186 711 4 197 785 Dorado 121 970 157 966 142 980 149 840 163 335 188 123 924 215 154 036 Guayanilla 56 038 62 906 62 248 66 475 66 733 68 172 382 571 63 762 Guaynabo 328 357 310 252 315 294 318 474 313 105 370 786 1 956 269 326 045 Juana Diaz 18 715 19 573 19 532 22 433 22 618 24 724 127 594 21 266 Lajas 0 0 0 162 749 164 980 174 172 501 900 83 650 Manati 683 106 622 507 603 926 671 557 720 659 856 871 4 158 626 693 104 Ponce 0 97 018 286 999 295 704 309 234 394 216 1 383 171 230 528 Salinas 151 820 169 911 199 841 212 709 234 993 236 981 1 206 255 201 043 Toa Baja 73 227 86 935 81 380 91 082 191 965 203 852 728 440 121 407 Total 6 490 618 6 973 452 7 451 053 7 729 690 8 161 824 9 413 598 46 220 234 7 703 372 *PRDH denotes Puerto Rico Department of Health. Figure 1: Annual per capita spending on dengue prevention activities by jurisdiction, 2002–2007 (in US$) $2.67 $2.70 $3.04 $3.19 $4.32 $6.32 $14.39 $22.41 $1.26 $1.28$0.80$0.41$0.33 $0 $5 $10 $15 $20 $25 D e p a rt m e n t o f H e al th Ju a n a D ia z B ay a m o n P o n ce T o a B aj a C ab o R o jo G u ay a n il la La ja s G u a yn a b o D o ra d o S al in a s M a n a ti C ar o lin a Jurisdiction Average US$ 1.97 per capita 20 Dengue Bulletin – Volume 34, 2010 Economic cost of dengue public prevention activities in Puerto Rico (72%); followed by personnel, with an average of US$ 1.91 million per year (25%); and amortized capital costs averaging US$ 0.22 million per year (3%). The breakdown of cost across line items is illustrated in Figure 2. Figure 2: Average annual spending on dengue prevention activities by line item, 2002–2007 Personnel cost, $1 913 679, 25% Recurrent cost, $5 565 396, 72% Capital cost, $224 297, 3% Figure 3 shows the breakdown of annual expenditure on dengue control by function. Clean-up campaigns had the highest annual average expenditure of US$ 4.99 million per year (65%), followed by fumigation with an average of US$ 1.01 million per year (13%), surveillance with an average of US$ 0.73 million per year (9%), and inspection with an average of US$ 0.43 million per year (6%), while education had the lowest expenditure with US$ 0.18 million per year (2%). Table 3 shows the average annual expenditure jointly by function and jurisdiction. It indicates that clean-up campaigns are the most variable component of costs. Discussion The overall results suggest that the public expenditure on dengue prevention activities in Puerto Rico, which is US$ 7.73 million (US$ 1.97 per capita), is broadly comparable to that of other countries in the region. For example, Panama spent US$ 5 million (US$ 1.56 per Figure 3: Average annual spending on dengue prevention activities by function, 2002–2007 Clean up campaigns (65%), 4 998 007US$ General management (5%), US$ 366 642 Education (2%), 176 780US$ Surveillance (9%), US$ 726 576 Inspection (6%), US$ 426 464 Fumigation (13%), US$ 1 008 904 Dengue Bulletin – Volume 34, 2010 21 Economic cost of dengue public prevention activities in Puerto Rico capita) during the epidemic year 2005, Brazil budgeted US$ 0.6 billion (US$ 3.61 per capita) for vector control in 1997, and the Brazilian state of Sao Paolo spent US$ 12.5 million (US$ 1.14 per capita) in 2005.[19,22,24] The additional cost of dengue prevention was studied in a few other countries including Cambodia where the annual vector control campaigns, focusing on larvicide and communication in Phnom Penh and Kandal (PP&K), cost US$ 0.6 million (US$ 0.2 per person living in PP&K).[19] To help sustain and improve dengue prevention activities, policy-makers need to understand the economic cost of dengue illness as well as the cost structure of the current vector control and surveillance programmes in the country. Recent literature provides a few examples of costs of dengue in endemic countries in the Americas. The total impact of dengue illness in the Americas is estimated at US$ 2.1 billion per year.[25] The economic impact of dengue illness for officially reported dengue cases in five countries in the Americas was estimated to range from US$ 135.2 million in Brazil to US$ 0.9 million in Panama. In three other countries, the costs were US$ 10.2 million in Venezuela, US$ 1.7 million in El Salvador, and US$ 1.2 million in Guatemala.[26] The variability among municipalities in per capita costs deserves further investigation. The per capita cost of vector control in Carolina is 11 times the average cost of US$ 1.97, followed by Manati which spent seven times the national average. DHPR spent only one sixth of the average national expenditure on dengue vector control followed by Juana Diaz which spent one fifth of the national average on vector control. The high level of spending in some municipalities of Puerto Rico on dengue vector control currently indicates the economic importance of the disease. The results underscore the value estimated through studying alternative technologies, including vaccines, personal protection methods against mosquito bites, and new vector control approaches, which collectively can help reduce the transmission and severity of this disease. This variability between municipalities illustrates their perspective on the importance of dengue compared with other public health problems and constraints that they face in prioritizing available resources. This natural variability provides an opportunity for further research to inform policy-makers about the effectiveness of various vector control strategies implemented in these jurisdictions. One limitation in this analysis is that the municipalities reported costs of clean-up campaigns only as lump-sum amounts. The reporting did not distinguish components within these aggregates, such as personnel, materials, transportation, etc. If more detailed data were available, the recurrent costs would have been smaller and the personnel costs larger. This limitation affects the breakdown by line item, but not the overall cost of dengue control. As part of the national effort to control dengue, CDC Dengue Branch helps support dengue surveillance. This cost was not included in this analysis. Therefore, the total cost of dengue prevention may be underestimated. Acknowledgements The authors are grateful to Drs Kay Tomashek and D. Fermin Arguello of the Centers for Disease Control for their advice and support, to Dr Jose Suaya of Brandeis University for assistance in study design, and to officials of the municipalities and the PRDH for their contribution towards supplying the requested data. 22 Dengue Bulletin – Volume 34, 2010 Economic cost of dengue public prevention activities in Puerto Rico References [1] Gubler DJ. Dengue and dengue haemorrhagic fever in Americas. Puerto Rico Health Science Journal 1987;6(2):107-11. [2] Gubler DJ. Aedes aegypti and Aedes aegypti- borne disease control in the 1990s: Top-down or bottom-up. American Journal of Tropical Medicine and Hygiene 1989;40(6):571-578. [3] Gubler DJ. Dengue and dengue haemorrhagic fever. Clinical Microbiology Reviews 1998;11(3):480-496. [4] Gubler DJ. Emergence of epidemic dengue/ dengue haemorrhagic fever as public health, social and economic problem in the 21st century. In: Scheld WM, Armstrong D, Hughes JM editors. Emerging infections 1. Washington: ASM Press; 1997.p.1-14. [5] Pinheiro FP, Corber SJ. Global situation of dengue and dengue hemorrhagic fever and its emergence in the Americas. World Health Statistics Quarterly 1997;50:161-169. [6] WHO. Dengue Guidelines for Diagnosis, Treatment, Prevention and Control. 2009. Available at: http://whqlibdoc.who.int/ publications/2009/9789241547871_eng. pdf. Accessed 17 Mar 2011. [7] Guzman A, Isturiz RE. Update on the global spread of dengue. International Journal of Antimicrobial Agents 2010;36S:S40-S42. [8] Periago MR, Guzman MG. Dengue and haemorrhagic dengue in the Americas. Revista Panamericana de Salud Pública 2007;21:187- 191. [9] Guzman MG, Kouri G. Dengue and haemorrhagic dengue research priorities. Revista Panamericana de Salud Pública. 2006;19:204-215. [10] WHO, Geneva 2006. Report of the Scientific Working Group meeting on Dengue. Geneva. www.who.int/tdr; TDR/SWG/08 2006. [11] Eisen L, Beaty BJ, Morrison AC, Scott TW. Proactive Vector control strategies and improved monitoring and evaluation practices for dengue prevention. Journal of Medical Entomological 2009;46(6):1245-1255. [12] Holmes EC, Twiddy SS. The origin, emergence and evolutionary genetics of dengue virus. Infection, Genetics and Evolution 2003;3(1):19- 28. [13] Westaway EG, Blok J. Taxonomy and evolutionary relationships of flaviviruses. In: Gubler DJ, Kuno G: Dengue and Dengue Haemorrhagic Fever, London, CAB International, 1997, p147. [14] Isturiz RE, Gubler DJ, Castillo JB. Dengue and dengue haemorrhagic fever in Latin America and the Caribbean. Infectious Disease Clinics of North America 2002;14(1):121-140. [15] Gubler DJ. Dengue. In Monath RP: Epidemiology of Arthropod-Borne Viral Disease, Vol 2. Boca Raton, FL, CRC Press, 1988, pp 223-260. [16] Winter PE, Nantapanich S, Nisalak A. Recurrence of epidemic dengue haemorrhagic fever in an insular setting. American Journal of Tropical Medicine and Hygiene 1969;18(4):573- 579. [17] Erlanger TE, Keiser J, Utzinger J. Effect of dengue vector control interventions on entomological parameters in developing countries: a systematic review and meta- analysis. Medical and Veterinary Entomology Journal 2008;22(3):203-21. [18] Tun-Lin W, Lenhart A, Nam VF, Rebollar-Tellez E, Morrison AC, Barbazan P, Cote M, Midega J, Sanchez F, Manrique-Saide P, Kroeger A, Nathan MB, Meheus F, Petzold M. Reducing costs and operational constraints of dengue vector control by targeting productive breeding places: a multi-country non-inferiority cluster randomized trail. Tropical Medicine and International Health 2009;14(9):1143-1153. Dengue Bulletin – Volume 34, 2010 23 Economic cost of dengue public prevention activities in Puerto Rico [19] Suaya JA, Shepard DS, Chang M-S, Caram M, Hoyer S, Socheat D, Chantha N, Nathan MB. Cost-effectiveness of annual targeted larviciding campaigns in Cambodia against the dengue vector Aedes aegypti. Tropical Medicine and International Health 2007;12(9):1026–1036. [20] Heintze C, Velasco Garrido M, Kroeger A. What do community-based dengue control programme achieve? A systematic review of published evaluations. Transactions of the Royal Society of Tropical Medicine and Hygiene 2007;101(4):317-25 [21] Tomashek KM, Rivera A, Munoz-Jordan JL, Hunsperger E, Santiago L, Padro O, Garcia E, Sun W. Description of a large island-wide outbreak of dengue in Puerto Rico, 2007. American Journal of Tropical Medicine and Hygiene 2009;81:467-74. [22] Armien B, Suaya JA, Quiroz E, Sah BK, Bayard V, Marchena L, Campos C, Shepard DS. Clinical characteristics and national economic cost of the 2005 dengue epidemic in Panama. American Journal of Tropical Medicine and Hygiene. 2008;79(3):364-71. [23] Kongsin S, Jaimton S, Suaya J, Vasanawathana S, Sirisuvan P, Shepard DS. Cost of dengue in Thailand. Dengue Bulletin 2010;34: 77-88. [24] Taliberti H, Zucchi P. Direct costs of the dengue fever control and prevention program in 2005 in the City of São Paulo. Revista Panamericana de Salud Pública 2010;27(3):175-80. [25] Shepard DS, Coudeville L, Halasa Y. Cost of dengue in the Americas. American Journal of Tropical Medicine and Hygiene 2011;84(2):200-207. [26] Suaya JA, Shepard DS, Siqueira JB, Martelli CT, Lum LCS, Tan LH, Kongsin S, Jiamton S, Garrido F, Montaya R, Armien B, Huy R, Castillo L, Caram M, Sah BK, Sughayyar R, Tyo KR, Halstead SB. Cost of dengue cases in eight countries in Americas and Asia, a prospective study. American Journal of Tropical Medicine and Hygiene 2009;80(5):845-55. 24 Dengue Bulletin – Volume 34, 2010 Increase in dengue fever imported from Côte d’Ivoire and West Africa to France Guy La Ruchea#, Charlotte Renaudatb, Arnaud Tarantolaa, Valérie Caroc, Martine Ledransa, Dominique Dejour-Salamancaa, Laure Diancourtc, Hugues Toloud, Marc Grandadamb,d and Marc Gastellu-Etchegorrya aInternational and Tropical Department, Institut de Veille Sanitaire, 12 rue du Val d’Osne, 94415 Saint-Maurice Cedex, France bInstitut Pasteur, National Reference Centre for Arboviruses, Paris, France cInstitut Pasteur, Genotyping of Pathogens and Public Health Platform, Paris, France dInstitut de Médecine Tropicale du Service de Santé des Armées, National Reference Centre for Arboviruses Associated Laboratory, Marseille, France Abstract Dengue is usually not considered a significant health problem in Africa because severe forms of dengue illness are rarely reported. In the absence of local surveillance data, the investigation of dengue cases imported to France contributed to document the circulation of dengue virus in this area. From 1 July 2006 to 31 December 2008, a total of 148 dengue cases imported to metropolitan France were reported through the mandatory notification system. Arthralgia and signs of severity (haemorrhage, thrombocytopenia) were less frequent in patients returning from West African countries. DENV-3 was isolated in two patients from Côte d’Ivoire in 2008. The number and proportion of patients returning from Côte d’Ivoire to France increased significantly in 2008 compared with the previous 18-month period. In parallel, the marginal increase in air travel does not explain the high increase observed in imported dengue cases to France. Our data illustrate increased dengue circulation and the emergence of DENV-3 in this area, with public health implications for epidemiological surveillance and case management locally. Keywords: Dengue; West Africa; Côte d’Ivoire; France; mandatory reporting; DENV-3. #E-mail: g.laruche@invs.sante.fr Introduction The genus Flavivirus includes pathogens with a huge impact on human health through diseases such as yellow fever, West Nile fever and dengue fever. Dengue can be caused by one of the four distinct serotypes of dengue virus (DENV). The transmission of dengue serotype 2 (DENV-2) in an enzootic cycle involving monkeys and several species of Dengue Bulletin – Volume 34, 2010 25 Dengue fever imported from West Africa to France mosquitoes has been documented in West African forests.[1,2] In West African countries (WAC), dengue can also occasionally be transmitted in an urban cycle involving humans and anthropophilic mosquitoes belonging to Aedes spp.[1,3] Little information, however, is available on the magnitude of transmission in urban environments in WAC. In the past 20 years, serosurveys evidenced circulation of mainly DENV-2 in WAC, especially in Senegal.[4,5] DENV-4 has also been isolated in Dakar residents.[4] Ten years ago, DENV-1 was isolated from a French soldier living in Abidjan, Côte d’Ivoire.[6] In the context of a yellow fever epidemic lasting from April to July 2008 in Abidjan, DENV-3 was typed by polymerase chain reaction (PCR) in one patient returning from Abidjan to France[7,8] and one returning to Japan[9]. To our knowledge, DENV-3 has not been previously isolated in patients from West African countries. Dengue is usually not considered a significant health problem in Africa because severe forms of dengue illness are rarely reported[10]. In a context of high malaria incidence and in the absence of dengue- specific preventive or curative treatments, most West African countries have not implemented specific dengue surveillance systems and often lack laboratory capacity to diagnose dengue infection. The circulation of dengue viruses in the area is likely, therefore, to be underestimated. Furthermore, the introduction of a new serotype, previously observed in the Americas, Asia, the South Pacific and East Africa[11], but never in West African countries, could have an important epidemic potential and be associated with a higher proportion of secondary infections and severe forms.[12,13] In the absence of local surveillance data, the investigation of imported dengue cases to France contributes towards documenting dengue virus circulation in this area. Material and methods Notification by French physicians of clinically confirmed dengue infections became mandatory in July 2006 for surveillance purposes and to guide intervention in some French départements where Aedes albopictus, a potential vector of dengue virus, is established. Notifications seek to document viraemic cases and apply only to recent infections, i.e. cases where the onset of symptoms is less than seven days old. Cases are confirmed either by PCR or by serology (detection of dengue-specific IgM). This information is centralized by the Institut de Veille Sanitaire, the French institute for public health surveillance. We compared the characteristics of the patients returning to metropolitan France (continental France and Corsica) from West African countries or from other countries. Results From 1 July 2006 to 31 December 2008, a total of 148 dengue cases imported to metropolitan France (mainland France plus Corsica) were reported through the mandatory notification system. Of these patients, 69 (47%) came from the French Overseas Territories (58 from French West Indies, eight from French Polynesia, two from French Guiana and one from La Reunion Island), 42 (28%) came from Asia, 21 (14%) from Africa (16 from West Africa, four from Central Africa and one from East Africa) and 16 (11%) from South America and the Caribbean. Imported cases from Asia and from Africa increased whereas those from the French Overseas Territories and from South America and the Caribbean decreased in 2008 as compared with 2006–2007 (Figure 1). Among the 16 imported cases from West Africa (Figure 2), seven came from 26 Dengue Bulletin – Volume 34, 2010 Dengue fever imported from West Africa to France Figure 1: Mandatory notification of imported dengue cases to metropolitan France from July 2006 to December 2008, classified by region and year of acquisition 50% 61% 32% 27% 18% 39% 10% 9% 21% 13% 13% 8% 0% 20% 40% 60% 80% 100% 2006 (n=30) 2007 (n=56) 2008 (n=62) South America & Caribbean Africa Asia French Overseas Territories P er ce n ta g e o f ca se s Figure 2: Imported dengue cases from West African countries to metropolitan France notified from July 2006 to December 2008 1 (2006) 1 (2007) 2 (2008) (2008) No. imported cases (2007) Year of reporting Côte d’Ivoire Mali Burkina Faso Senegal 1 (2007) 6 (2008) Togo (2008) 1 (2007) 1 (2008) Abidjan 2 3 7 3 1 4 Dengue Bulletin – Volume 34, 2010 27 Dengue fever imported from West Africa to France Côte d’Ivoire, four from Burkina Faso, three from Mali, two from Senegal and one from Togo (one of these patients returned from a trip to Mali and Côte d’Ivoire); one case was reported in 2006, three cases in 2007 and 12 cases in 2008, a 300% increase in 2008 compared with 2006–2007. The six patients returning from Côte d’Ivoire in 2008 were individually contacted to document the places visited during their stay in that country. During the 15 days before the onset of symptoms, corresponding to the maximal incubation period, four of the six patients stayed exclusively in Abidjan whereas the remaining two stayed in Abidjan and Grand- Bassam, located along the coast at about 40 km from Abidjan. The number and the proportion of patients returning from Côte d’Ivoire increased significantly in 2008 compared with the previous 18 months. The increase was not statistically significant considering the patients returning from West African countries other than Côte d’Ivoire (Table 1). Data from the French Civil Aviation Authority show a relative stability in the number of air travellers between France and the West African countries where dengue cases were imported (+4.3% in 2008 compared with 2006–2007, see Figure 3) and an increase of 20.0% in air travellers between France and Côte d’Ivoire in 2008 compared with 2006–2007. This slight increase in air travel does not explain the high increase observed in imported dengue cases to France (Table 1). Baseline characteristics of the patients in terms of sex and age were similar in patients from West African countries when compared Figure 3: Quarterly number of imported dengue cases from West African countries to France (bars) and quarterly number of air travellers between West African countries and France (curve), July 2006 – December 2008 0 1 2 3 4 2006 Q3 2006 Q4 2007 Q1 2007 Q2 2007 Q3 2007 Q4 2008 Q1 2008 Q2 2008 Q3 2008 Q4 Quarter 0 50 000 100 000 150 000 200 000 250 000 300 000 N o . o f d en g u e ca se s fr o m W es t A fr ic an co u n tr ie s N o . o f ai r tr av el le rs b et w ee n Fr an ce an d W es t A fr ic an co u n tr ie s West African countries with imported dengue cases include Burkina Faso, Côte d’Ivoire, Mali, Senegal and Togo. Air travellers are from or to these West African countries. (Source for air travellers: French Civil Aviation Authority, 2009) 28 Dengue Bulletin – Volume 34, 2010 Dengue fever imported from West Africa to France Table 1: Mandatory notification of imported dengue cases to metropolitan France from July 2006 to December 2008, according to the period of acquisition Period of acquisition No. cases in the countries of travela Total of imported cases (%) Number of air travellersb Côte d’Ivoire West Africa excluding Côte d’Ivoire Côte d’Ivoire West Africa excluding Côte d’Ivoire Jul. 2006 – Dec. 2007 1/86 (1.1) 3/86 (3.5) 1/248 854 3/1 348 514 2008 6/62 (9.7) 6/62 (9.7) 6/199 053 6/911 425 P-valuec 0.022 0.11 0.033 0.10 aWest African countries with imported dengue cases include Burkina Faso, Côte d’Ivoire, Mali, Senegal and Togo. bAir travellers from or to the West African countries listed above. cComparison of percentages between the two periods by Fisher’s exact test or Chi-square test. (Source for air travellers: French Civil Aviation Authority, 2009) with patients returning from other countries. Pain, particularly arthralgia, was significantly less frequent in patients returning from West African countries. Signs of severity (haemorrhage, thrombocytopenia) seemed less frequent in patients returning from WAC, but the difference between the two groups did not reach statistical significance. Finally, two third of the patients were hospitalized (Table 2). Although biological diagnosis was performed either by PCR or by serology, dengue serotype determination by RT-PCR was only performed in three patients from WAC. DENV-1 was isolated in a patient from Burkina Faso in 2007, and DENV-3 was isolated in two patients from Côte d’Ivoire in 2008. For one of the DENV-3-infected patients returning to France in June 2008, data from the gene sequence of the protein E (GenBank accession no. FM213456) showed that this isolate presented a very strong identity (99.6%) with the sequence of the Japanese patient isolate.[9] A phylogenetic analysis limited to this region of the genome has shown that these two isolates are very likely related (an average of 99.2% identity) to the 2004 Saudi Arabia DENV-3 strain. Discussion There is a possible limitation to our study: we have no information on the completeness of the mandatory notification, which may be low in France, particularly during the first months after its introduction in 2006. The high proportion of hospital admissions may reflect better participation of hospital physicians compared with general practitioners. The lack of completeness, however, is probably not linked to the country of acquisition; if there is incomplete notification then it is regarding the overall number of dengue cases imported to France, which may be underestimated. Several factors contr ibute to an underestimation of the disease burden due to dengue infection in West Africa. Firstly, Dengue Bulletin – Volume 34, 2010 29 Dengue fever imported from West Africa to France disease surveillance data and systems are often lacking. Furthermore, limited access to health care may hinder the detection of symptomatic cases in West African countries. Finally, dengue infection diagnosis is difficult as it is often asymptomatic, and misdiagnoses are frequent because symptoms are non-specific and severe forms infrequent among imported cases, as suggested by a low frequency of haemorrhagic forms reported in our study, and tests are not widely available. Surveillance of imported cases of infectious diseases among travellers to developed countries can be used for the early detection of epidemic activity in specific areas, as shown by our study and other published studies.[14,15] Despite probable incompleteness, the mandatory notification surveillance system in France allows monitoring the trends of imported dengue cases from countries worldwide, especially those which lack surveillance systems for dengue. Since the beginning of 2008, dengue cases imported from West Africa, particularly from Côte d’Ivoire, are increasing. We conclude that this is due to increased dengue circulation in this area. This information was shared with the World Health Organization and local authorities, resulting in an investigation in Abidjan by the Global Outbreak Alert and Response Network. This investigation confirmed the emerging of DENV-3 and led Table 2: Dengue patient’s characteristics according to the place of acquisition, July 2006–December 2008 Characteristic Place of travel P-valuebWest Africaa n=16 Elsewhere n=132 Male 56.3% 47.7% 0.52 Median age (range) 51 years (19-71) 41 years (7-67) 0.29 Pain 87.5% 99.2% 0.03 Myalgia 53.8% 62.7% 0.37 Headache 30.8% 49.1% 0.21 Arthralgia 7.7% 40.0% 0.02 Abdominal pain 15.4% 7.3% 0.29 Minor haemorrhagic symptoms 12.5% 28.7% 0.14 Major haemorrhagic symptoms 0.0% 2.6% 0.69 Platelets ≤100,000/mL 26.7% 49.2% 0.10 Hospitalization 75.0% 61.3% 0.29 aWest African countries with imported dengue cases include Burkina Faso, Côte d’Ivoire, Mali, Senegal and Togo. bComparison of percentages between the two periods by Fisher’s exact test or Chi-square test; bold face represents a statistically significant result (P<0.05). 30 Dengue Bulletin – Volume 34, 2010 Dengue fever imported from West Africa to France to the transfer of technology for laboratory diagnosis in Abidjan.[16] The introduction of a new serotype may cause important epidemics and be associated with a higher proportion of severe forms in case of secondary dengue, i.e. sequential infection with two serotypes. This may especially be true because DENV-3 has previously been associated with high-severity outbreaks in many countries[11] and because it was isolated in the few severe forms ever reported in Africa.[13] These data have public health implications for epidemiological surveillance and management of patients locally. Further study is needed on the epidemiology and seasonal patterns of the four dengue serotypes in West African countries.[14] Acknowledgements We are indebted to Philippe Barboza, Sandra Cohuet and Philippe Desprès for reviewing the manuscript, and to Nicolas Berthet for his technical assistance in conducting typing of the dengue strain. We thank the physicians for providing the mandatory notification. References [1] Vasilakis N. Durbin A.P., da Rosa A.P., Munoz- Jordan J.L., Tesh R.B., Weaver S.C. Antigenic relationships between sylvatic and endemic dengue viruses. American Journal of Tropical Medicine and Hygiene. 2008 Jul;79(1):128- 32. [2] Vasilakis N., Tesh R.B., Weaver S.C., Sylvatic dengue virus type-2 activity in humans, Nigeria, 1966. Emerging Infectious Diseases. 2008 Mar;14(3):502-4. [3] Diallo M., Ba Y., Faye O., Soumare M.L., Dia I., Sall A.A. Vector competence of Aedes aegypti populations from Senegal for sylvatic and epidemic dengue 2 virus isolated in West Africa. Transactions of the Royal Society of Tropical Medicine and Hygiene. 2008 May;102(5):493-8. [4] Saluzzo JF, Cornet M, Castagnet P, Rey C, Digoutte JP. Isolation of dengue 2 and dengue 4 viruses from patients in Senegal. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1986;80(1):5. [5] Monlun E, Zeller H, Le Guenno B, Traoré- Lamizana M, Hervy JP, Adam F, Ferrara L, Fontenille D, Sylla R, Mondo M, Digoutte JP. Surveillance de la circulation des arbovirus d’intérêt médical dans la région du Sénégal oriental (1988-1991). Bulletin de la Societe de Pathologie Exotique. 1993;86(1):21-8. [6] Durand JP, Vallee L, de Pina JJ, Tolou H. Isolation of a dengue type 1 virus from a soldier in West Africa (Côte d’Ivoire). Emerging Infectious Diseases. 2000 Jan-Feb;6(1):83-4. [7] Parola P. Dengue/DHF update 2008 (32): France ex Cote d’Ivoire. ProMed. 2008 Aug 8. Available from http://www.promedmail.org, archive no. 20080808.2446. [8] Ninove L, Parola P, Baronti C, De Lamballerie X, Gautret P, Doudier B, Charrel RN. Dengue virus type 3 infection in traveler returning from West Africa. Emerging Infectious Diseases. 2009 Nov;15(11):1871-2. Dengue Bulletin – Volume 34, 2010 31 Dengue fever imported from West Africa to France [9] Takasaki T. Dengue/DHF update 2008 (35): Japan ex Cote d’Ivoire. ProMed. 2008 Aug 18. Available from http://www.promedmail. org, archive no. 20080818.2573. [10] Sang RC. Dengue in Africa. Report of the Scientific Working Group Meeting on Dengue, Geneva, 1-5 October, 2006. TDR/SWG/08. Geneva: World Health Organization; 2007; pp. 50-2 [cited 2009 Apr 14]. Available from http://www.who.int/tdr/publications/tdr- research-publications/swg-report-dengue/pdf/ swg_dengue_2.pdf [11] Messer WB, Gubler DJ, Harris E, Sivananthan K, de Silva AM. Emergence and global spread of a dengue serotype 3, subtype III virus. Emerging Infectious Diseases. 2003 Jul;9(7):800-9. [12] Alvarez M, Rodriguez-Roche R, Bernardo L, Vázquez S, Morier L, Gonzalez D, Castro O, Kouri G, Halstead SB, Guzman MG. Dengue haemorrhagic fever caused by sequential dengue 1-3 virus infections over a long-time interval: Havana epidemic, 2001-2002. American Journal of Tropical Medicine and Hygiene. 2006 Dec;75(6):1113-7. [13] Gubler DJ, Sather GE, Kuno G, Cabral JR. Dengue 3 virus transmission in Africa. American Journal of Tropical Medicine and Hygiene. 1986 Nov;35(6):1280-4. [14] Schwartz E, Weld LH, Wilder-Smith A, von Sonnenburg F, Keystone JS, Kain KC, Torresi J, Freedman DO. Seasonality, annual trends, and characteristics of dengue among ill returned travelers, 1997-2006. Emerging Infectious Diseases. 2008 Jul;14(7):1081-8. [15] Grandadam M. Dengue/DHF update 2008 (53): Cambodia, France ex Mali. ProMed. 2008 Dec 8. Available from http://www.promedmail. org, archive no. 20081208.3860. [16] Dengue in Africa: Emergence of DENV-3, Côte d’Ivoire, 2008. Weekly Epidemiolical Record. 2009 Mar 13;84(11-12):85-8. 32 Dengue Bulletin – Volume 34, 2010 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India Tiina M. Murtolaa, S.S. Vasanb,c, Tapasvi I. Puward, Dipti Govild, Robert W. Fielde, Hong-Fei Gongb, Ami Bhavsar-Vyasf,g, Jose A. Suayag, Marion Howardg, Donald S. Shepardg, Vijay Kumar Kohlih, P.B. Prajapatii, Amarjit Singhi and Dileep V. Mavalankard# aAalto University, Department of Mathematics and Systems Analysis, FI-00076 Aalto, Finland bOxitec Limited, 71 Milton Park, Oxford OX14 4RX, UK cCentre for Research in Biotechnology for Agriculture (CEBAR), University of Malaya, Kuala Lumpur, 50603, Malaysia dCentre for Management of Health Services, Indian Institute of Management, Vastrapur, Ahmedabad 380015, India eUniversity of Oxford, Department of Engineering Science, Parks Road, Oxford OX1 3PJ, UK fCentre for Micro Finance, 8th Floor, West Wing, Fountain Plaza, Khaleel Shirazi Estate, 31/2A, Pantheon Road, Egmore, Chennai 600 008, India gHeller School, MS035, Brandeis University, Waltham, MA 02454-9110, USA hAhmedabad Municipal Corporation, Sardar Patel Bhavan, Danapith, Ahmedabad 380 001, India iGovernment of Gujarat, Department of Health & Family Welfare, Commissionerate of Health, Medical Services & Medical Education, Dr Jivraj Mehta Bhawan, Block No.5, Sector-10, Gandhinagar 382 010, India Abstract This study aims to provide a preliminary estimate of the immediate cost of chikungunya and dengue to household in the Indian state of Gujarat. Combining nine earlier studies and data from interviews, we analysed the costs of non-fatal illness and of intervention programmes; building a more comprehensive picture of the immediate cost of these Aedes aegypti mosquito-borne diseases to Gujarat. The “RUHA matrix” was used to estimate the cost of illness by combining the shares of reported (R) and unreported (U) hospitalized (H) and ambulatory (A) cases of chikungunya and dengue with ambulatory and hospitalization costs per case and the number of reported cases. Using Monte Carlo sensitivity analysis, the immediate cost to households incurred on account of chikungunya and dengue to Gujarat was estimated to be 3.8 (range 1.6–9.1) billion Indian rupees (INR) per annum (US$ 90 million, range US$ 38 and US$ 217 million). It is hoped that this preliminary estimate will trigger more refined studies on cost of illness as well as cost-effectiveness of vaccines and other interventions to combat these neglected tropical diseases. Keywords: Burden of illness; chikungunya; dengue; immediate cost; Monte Carlo analysis; RUHA matrix; Gujarat. #E-mail: dileep@iimahd.ernet.in Dengue Bulletin – Volume 34, 2010 33 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India Introduction The number of dengue cases in the state of Gujarat, India, has followed an increasing trend since 2004. Several studies have estimated the costs of illness associated with dengue or chikungunya in different states of India, but the cost factors included tend to vary from study to study. In this paper, we make a preliminary estimate of the immediate cost of chikungunya and dengue to the state of Gujarat by combining available studies to include all major cost factors. Furthermore, we also analyse control costs to form a more comprehensive picture of the cost of these Aedes mosquito-borne diseases. Materials and methods The key components of the immediate cost of chikungunya and dengue to a society are: (i) cost of non-fatal illness; and (ii) cost of intervention programmes, which include vector control on Aedes mosquito, a fraction of the household insecticide market, and cost of research and development. Data on each cost parameter was collected from published and unpublished studies and from interviews with local authorities. Where direct data was not available, trends from other Asian countries were used. All cost estimates were inflation- adjusted to 2008 Indian rupees (INR). Costs in different countries were compared at the rate of 2008 US dollars (US$) and an exchange rate of 42 INR/US$ was used. Cost of illness was estimated by combining reported cases, and costs per case with a RUHA matrix (defined below). Data on reported dengue cases for the years 2003–2008 was used to adjust for year-to-year variations caused by the cyclical nature of dengue.(1) Chikungunya cases for 2006–08 were used to estimate the burden of an outbreak, which is assumed to occur cyclically.(2,3) Costs per ambulatory and hospitalized case were obtained from published and unpublished studies, which were compared with and combined to ensure consistency in factors included in the costs. The shares of reported (R) and unreported (U) hospitalized (H) and ambulatory (A) cases were estimated based on published literature and local information, and used to construct a RUHA matrix. For this study, chikungunya and dengue were assumed to be identical from the point of view of disease control and management. Monte Carlo sensitivity analysis was carried out (@Risk software version 5.0.1, Palisade Corporation, USA) to find out how uncertainties in each cost parameter affect the total cost to households of chikungunya and dengue. Sixty-five simulations, each with 10 000 iterations, were used, and for each iteration all parameters were independently drawn from Beta-PERT distribution. Beta-PERT was chosen because it places less emphasis on the direction of any possible skew compared to triangular distribution, but it is defined using the same parameters (minimum, most likely and maximum), which are easily understood and uncomplicated to estimate.(4) Results Reported cases The number of dengue cases in Gujarat reported by India’s National Vector Borne Disease Control Programme (NVBDCP) has followed an increasing trend since 2004 (Figure 1). In the years 2003–2008 the number of reported dengue cases has varied from 117 to 1023 with an annual average of 493. There was a major outbreak of chikungunya in Gujarat in 2006 with 76 012 reported cases, which declined to 3223 and 246 cases in 2007 34 Dengue Bulletin – Volume 34, 2010 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India Figure 1: Dengue cases in Gujarat 0 200 400 600 800 1 000 1 200 2003 2004 2005 2006 2007 2008 Year R ep o rt ed ca se s Source: NVBDCP, Government of India (since this article was written, the number of reported dengue cases have gone up to 2461 in 2009 and 2568 in 2010). and 2008, respectively (Source: NVBDCP). These three years were taken to represent the burden of a chikungunya epidemic, and the annual cost of chikungunya was calculated by assuming that similar epidemic peaks followed by two-year tails occur every seven years (range 4–20 considered).(2,3) Some discrepancies were noted between local and national data on reported cases, and these were taken into account in the Monte Carlo analysis. Cost per case data Costs per hospitalized and ambulatory case were derived from published and unpublished studies. The studies were compared to identify differences in cost factors included, and then combined to make cost estimates that include all main factors. The resulting minimum, most likely and maximum values for direct (including medical and non-medical) cost and indirect cost are shown in Table 1. The most likely values for costs per case in Gujarat sum up to US$ 300 and US$ 64 for hospitalized and ambulatory cases, respectively. These are generally consistent with those worked out by Suaya et al. in Malaysia (US$ 1259 and US$ 422, hospitalized and ambulatory, respectively)(5) when taking into account Malaysia’s roughly five times higher GDP per capita. RUHA matrix The RUHA matrix in Table 2 shows the characteristics of chikungunya and dengue cases in Gujarat. It has been constructed from the following data: Dengue Bulletin – Volume 34, 2010 35 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India Table 1: Costs per hospitalized and ambulatory case(6-9) (studies referred at 6–9) Cost component Cost per case (INR)a Range (INR)a Referencesb Hospitalized Direct cost Indirect cost Total 9790 2820 12 610 3300 – 155 640 0 – 31 020 [6-9],[8],[6] [6-9],[7],[6] Ambulatory Direct cost Indirect cost Total 1070 1610 2680 40 – 9500 0 – 23 080 [7-9],[8],[7] [7,9],[7],[7,9] INR denotes Indian Rupees. aValues inflation adjusted to 2008 Indian Rupees. bFor each item, references cited were used respectively for most likely, minimum, and maximum values. Table 2: RUHA matrix for Gujarat Setting Reported Unreported Total Hospitalised 1% 14% 15% Ambulatory 3% 82% 85% Total 4% 96% 100% Source: Authors’ estimates. (1) A reporting rate of 4–10% (expansion factor 10-27) was recently used by Garg et al. to estimate the burden of dengue in India;(7) this has been assumed to be applicable in Gujarat. A comparable reporting rate (3%) was found by attributing 1% of general fever cases (reported by Integrated Disease Surveillance System) to chikungunya or dengue. This percentage was based on data showing that at least 10% of tested cases are confirmed as chikungunya or dengue.(10, 11) The “confirmation rate” for general fever cases was taken as one tenth of this (i.e., 1%) to allow for smaller number of relevant symptoms. (2) Garg et al(7) used a hospitalisation rate of 9–20% for dengue cases in India based on Thailand data.(12) This range of rates agrees fairly well with chikungunya hospitalization rates of 6% and 13% found in studies in Ahmedabad city.(8,13) (3) The fraction of reported cases that are hospitalized was assumed to be 0.29 based on public sector case data in Ahmedabad in 2007.(10) Vector control costs In 2007–2008, the NVBDCP spent INR 73 million on measures to prevent and control chikungunya and dengue in Gujarat.(11) Additional spending by municipal corporations during that period was INR 44 million and INR 27 million in Ahmedabad and Surat, 36 Dengue Bulletin – Volume 34, 2010 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India respectively (assuming one third of Surat’s budget for the Vector-Borne Disease Control Programme is assigned for dengue). These public control cost estimates are conservative because they tend to focus on insecticides (possibly underestimating personnel costs) and because costs in districts other than Ahmedabad and Surat have not been estimated. This effect is partly cancelled out by using data from Malaysia(14) to estimate expenditure in Gujarat to be 2% to 6% of government vector control spending. Expenditure on household insecticides to prevent these Aedes mosquito- borne diseases was estimated indirectly using three independent methods, which give fairly consistent results (see Table 3). The annual cost (taken as the average of the most likely values of the three methods) is INR 95 million (range INR 39–320 million). Table 3: Household insecticide market estimates Method used Most likelyb (INR million) Rangeb (INR million) Coils market taken as 30-50% of total marketa 90 39–321 Insecticide (liquidator) cost estimated per day 127 42–253 Household insecticide market in Malaysia 68 40–105 Combined 95 39–321 INR denotes Indian Rupees. aCoils market data from Jyothy Laboratories Ltd Annual Report 2006-07; coils market percentage based on data from Malaysian CropLife & Public Health Association Estimates of Household Insecticide Sales, 2002 to 2006, Chooi Lam Khong, personal communication to Vasan S.S. (2007). bMost likely value is the average and range is the range of the three different approaches. Discussion Cost of chikungunya and dengue to Gujarat Monte Carlo sensitivity analysis carried out on the cost incurred by households on chikungunya and dengue resulted in a mean annual cost of INR 3.8 billion (range INR 1.6–9.1 billion), equivalent to a mean of US$ 90 million and a range of US$ 38 to 217 million. About 88% of this cost was due to chikungunya illness, 5% was due to dengue illness, and the remaining 7% is the cost due to intervention activities. The total immediate cost translates to approximately INR 67 per capita (range INR 29–161), or US$ 1.60 (range US$ 0.70–3.80). Comparable estimates of the cost of dengue were US$ 5.3 per capita in Malaysia(14) and US$ 6.2 per capita in Panama,(15) while Brazil spends US$ 4.3 per capita on dengue prevention alone.(16,17) The differences in these costs can be partially explained the fact that GDP per capita is five times higher in Malaysia, Panama and Brazil than in Gujarat. The high risk of chikungunya epidemics also increases the relative cost in Gujarat. Most of the variation of the total cost to households is caused by uncertainties in direct cost of hospitalization, ambulatory costs, chikungunya cyclicity (frequency of chikungunya epidemics) and reporting rate (Figure 2). Further studies are in progress to refine the estimates of ambulatory costs and reporting rate, improving this preliminary cost estimate. These two parameters have been observed to have comparable effects outside Gujarat (7,14) suggesting that improved understanding of them will help make more accurate economic cost estimates around Asia. Dengue Bulletin – Volume 34, 2010 37 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India Figure 2: Variation of total cost due to uncertainties in each parameter This study considers only the immediate cost of these Aedes mosquito-borne diseases. Previous studies indicate that long-term illness and deaths are associated with these diseases,(18,19) but the emotional and economic burden due to these is outside the scope of this study. These diseases can also have a long-term impact on education and economic growth,(20,21) per capita income,(22,23) foreign direct investment,(24,25) tourism,(26) etc. but these effects have not been taken into account in the cost estimates presented in this study. Furthermore, as this study was done from the household perspective, it does not include government subsidies to health centres, hospitals and other facilities that often pay for much of the original construction plus a share of the personnel and operating expenses. Another study in this issue of Dengue Bulletin also reports on the cost per case of dengue.(6) The resulting mean cost there (US$ 586) is dramatically higher than the mean from the present study ($28). The difference is due to several factors. The higher figure represented hospitalized cases, while the lower figure was mostly ambulatory cases. The higher figure is in the private sector and measures the full economic cost, while the lower value is simply the cost to households. It is hoped that this preliminary estimate will trigger more refined studies on cost of illness as well as cost-effectiveness of vaccines and other interventions to combat these neglected tropical diseases. Acknowledgements Funded in part by a grant to the Regents of the University of California from the Foundation 38 Dengue Bulletin – Volume 34, 2010 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India for the National Institutes of Health through the Grand Challenges in Global Health initiative. The authors thank their colleagues, especially Ms Luise Birgelen and Ms Aikaterini References [1] Cummings DAT, Irizarry RA, Huang NE, Endy TP, Nisalak A, Ungchusak K, et al. Travelling waves in the occurrence of dengue haemorrhagic fever in Thailand. Nature, 2004,427(6972):344-7. [2] Powers AM, Logue CH. Changing patterns of chikungunya virus: re-emergence of a zoonotic arbovirus. Journal of General Virology, 2007,88:2363-77. [3] World Heal th Organizat ion (WHO). Communicable Diseases Geneva: World Health Organization; [cited 9 Oct 2008]; Available from: http://www.searo.who.int/ en/Section10/Section2246.htm. [updated 16 May 2008]. [4] Vose D. Risk Analysis - A Quantitative Guide. Chichester: John Wiley; 2000. [5] Suaya JA, Shepard DS, Siqueira JB, Martelli CT, Lum LCS, Tan LH, et al. Cost of dengue cases in eight countries in Americas and Asia, a prospective study. American Journal of Tropical Medicine and Hygiene, 2009,80(5):845-55. [6] Bhavsar A, Shepard DS, Suaya JA, Mafowosofo M, Hurley CL, Howard MW. A private hospital- based study assessing knowledge, attitudes, practices and costs associated with dengue illness in. Dengue Bulletin, 2010,34: 54-64. [7] Garg P, Nagpal J, Kairnar P, Sinveratne SL. Economic burden of dengue infections in India. Trans R Soc Trop Med Hyg, 2008,102(6):570- 77. [8] Mavalankar DV, Govil D, Trivedi N, Patel VB, editors. Prevalence of various symptoms and cost of treatment during chikungunya epidemic in Ahmedabad, India. Proceedings of the 1st Symposium on the Burden of Neglected Diseases; 10 Sep 2008; Ahmedabad, India. [9] Public Health Resource Network, Gopalan SS. Household economic impact of an outbreak of chikungunya in Orissa, India: Extent of out of pocket healthcare expenditure and loss of productivity. In: Personal communication to Mavalankar DV, editor. 2009. [10] Kohli VK, editor. Entomological monitoring, sero-surveillance and preventive measures against dengue and chikungunya in Ahmedabad. Proceedings of the 1st Symposium on the Burden of Neglected Diseases; 10 Sep 2008; Ahmedabad, India. [11] Prajapati PB, Singh A, editors. Current scenario of vector borne diseases in Gujarat, India Proceedings of the 1st Symposium on the Burden of Neglected Diseases; 10 Sep 2008; Ahmedabad, India. [12] Anderson KB, Chunsuttiwat S, Nisalak A, Mammen Jr MP, Libraty DH, Rothman AL, et al. Burden of symptomatic dengue infection in children at primary shcool in Thailand: a prospective study. Lancet, 2007,369:1452-9. [13] Puwar TI, Seth JK, Yadav RS, Kohli VK, editors. Community-based investigation of suspected Chikungunya cases and intervention measures. Proceedings of the 1st Symposium on the Burden of Neglected Diseases; 10 Sep 2008; Ahmedabad, India. Mandaltsi of Oxford University, Ms Clare Hurley of Brandeis University, and Rosemary Susan Lees of the University of Malaya, for their comments. Dengue Bulletin – Volume 34, 2010 39 Preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India [14] Lee HL, Vasan SS, Birgelen L, Murtola TM, Gong HF, Field RW, et al. Immediate cost of dengue to Malaysia and Thailand. Dengue Bulletin, 2010,34:65-76. [15] Armien B, Suaya JA, Quiroz E, Sah BK, Bayard V, Marchena L, et al. Clinical characteristics and national economic cost of the 2005 dengue epidemic in Panama. American Journal of Tropical Medicine and Hygiene, 2008,79(3):364-71. [16] Ministério da Saúde. Dengue. [cited 15 August 2008]; Available from: http://www.senado. gov.br/web/ comissoes/cas/ap/AP20080401_ MinSaude_FabianoPimenta.pdf. [17] Programa Nacional de Control da Dengue. Instituído em 24 julho de 2002.: Ministério da Saúde, Brasil; 2002 [cited 15 August 2008]; Available from: http://portal.saude.gov.br/ portal/arquivos/ pdf/pncd_2002.pdf. [18] Krishnamurthy V. Chikungunya arthritis, [Editorial]. Indian Journal of Rheumatology, 2008,3(3):91-2. [19] Mavalankar D, Shastri P, Bandyopadhyay T, Parmar J, Ramani K. Increased mortallity rate associated with chikungunya epidemic, Ahmedabad, India. Emerging Infectious Diseases [serial on the Internet], ; 2008 [cited 30 July 2009]; Available from: http://www.cdc. gov/EID/content/14/3/412.htm. [20] Bloom DE, Canning D. Epidemics and Economics. Program on the global Demography of Aging working Paper No. 9. Harvard Initiative for global Health; 2006 [cited 30 July 2009]; Available from: http://www.hsph. harvard.edu/pgda/Working%20Papers/2006/ BLOOM_CANNINGWP9.2006.pdf. . [21] Bloom DE, Canning D, Sevilla J. The effect of health on economic growth: A production function approach. World Development, 2004,32(1):1-13. [22] Barro R, Sala-I-Martin X. Economic growth. New York: McGraw-Hill; 1995. [23] Bhagava A, Jamison DT, Lau LJ, Murray CJL. Modeling the effects of health on economic growth. Journal of Health Economics , 2001,20(3):423-40. [24] Alsan M, Bloom DE, Canning D. The effect of population health on foreign direct investment inflows to low- and middle-income countries. World Development, 2006,34(4):613-30. [25] Jones T. The Panama Canal: A Brief History. 1990 [cited 30 July 2009]; Available from: http:/ /www.i lovelanguages.com/tyler/ nonfiction/pan2.html. [26] Mavalankar D, Puwar T, Murtola T, Vasan S. Quantifying the impact of chikungunya and dengue on tourism revenues. W.P.No. 2009- 02-03. Ahmedabad, India: Indian Institute of Management Working Paper Series. 2009 [cited 30 July 2009]; Available from: http:// www.iimahd.ernet.in/publications/data/2009- 02-03Mavalankar.pdf. 40 Dengue Bulletin – Volume 34, 2010 Prevalence of chikungunya in the city of Ahmedabad, India, during the 2006 outbreak: A community-based study Tapasvi Puwara#, Jay K. Shetha, Vijay Kohlib and Rajpal Yadavc aDepartment of Community Medicine, Smt. Nathiba Hargovandas Lakhmichand Municipal Medical College, Ahmedabad, Gujarat, India bDepartment of Health, Ahmedabad Municipal Corporation, Ahmedabad, Gujarat, India cNational Institute of Malaria Research (ICMR), Field Unit, Nadiad, Gujarat, India Abstract Prevalence of chikungunya in the city of Ahmedabad, India during the 2006 outbreak was investigated to estimate the prevalence of suspected chikungunya cases to find out demographic parameters and proportion of various symptoms among suspected chikungunya cases, and to evaluate the effectiveness of control measures implemented by the public health sector. A total of 6667 people from 1301 households were surveyed. The prevalence of suspected chikungunya cases was 32.9% (31.8% – 34.2%). Prevalence was higher in females (p<0. 006), and also highest for the age group of 40 to 80 years. The north zone of the city, a densely populated slum, had significantly higher (p<0.001) numbers of suspected cases than the other zones. Case numbers were significantly higher in slums than in more affluent settings such as apartments and bungalows (p<0.001). Chills, headaches, joint swelling and itching were the main symptoms reported by the majority of cases. The majority (67.6%) of suspected cases sought treatment from the private sector and another 8.5% received no treatment at all. The hospitalization rate was found to be 6% for the study. Among mosquito control measures, visits by health workers (73.4%), application of larvicides (67.5%) and fogging (42.4%) achieved good coverage compared with information education and communication (IEC) measures (28.3%) and other activities. Keywords: Chikungunya; prevalence; India; community-based study; disease burden. #E-mail: drtapasvi@gmail.com Introduction Chikungunya fever is an arboviral disease transmitted to humans by infected mosquitoes of the Aedes genus.[1] Since the first isolated incident in Tanzania in 1952,[2] several outbreaks of the disease, caused by this alphavirus have been documented throughout South-East Asia.[3-6] Dengue Bulletin – Volume 34, 2010 41 Chikungunya outbreak during 2006 in Ahmedabad, India During an acute attack of chikungunya, patients usually develop a sudden onset of fever, headache and arthralgia[7] after an incubation period of 2–4 days (range 3–12 days). Symptoms are generally self-limiting and last for 2–3 days. However, arthralgia may persist for months or years.[1] During the last outbreak in India and other South-East Asian countries, the epidemic spread rapidly and affected many communities with an attack rate as high as 40%–60%.[8] Chikungunya re-emerged in India in December 2005 after a gap between epidemics of 32 years.[9] This particular outbreak was caused by the central/east African genotype[10] and infected around 1 400 000 people in India during 2006. It was estimated that some areas had attack rates of 45%.[11] Official figures from the government of India indicated that 1.39 million suspected chikungunya cases were identified from 152 districts.[12] Materials and methods The present study was conducted in Ahmedabad city during September 2006. A house-to-house survey was done to search for suspected chikungunya fever cases and collect information about reported symptoms. A definition of a suspect case was adopted from the Communicable Disease Alert (CD Alert) issue, Vol. 10 (2) of February 2006 published by the National Institute of Communicable Diseases (NICD), Delhi.[13] It defined suspected chikungunya as “…an acute illness characterized by sudden onset of fever with several of the following symptoms: joint pain, headache, backache, photophobia, arthralgia and rash”. This study was conducted during the peak of the outbreak confirmed by the National Institute of Virology (NIV), Pune, during April 2006. Therefore, the definition of a suspected case was modified and any person having acute fever and joint pain any time during the past one month was considered a suspect chikungunya case for the purpose of the study. Symptoms were reported as perceived by the individual personally or by the reporting adult family member in case the member was not available for the interview. The study was carried out using a standard questionnaire, pre-tested two weeks before the study. The questionnaire also included information on treatment-seeking practices, knowledge of prevention and control, as well as activities carried out by the local health department. All investigators for the study were trained for one day in how to identify a suspected case of chikungunya as per the adopted definition and how to fill in the questionnaire as per the study format. Senior faculty from the Municipal Medical College of the city checked every form at the end of each day and verified the information with investigators. Pre-testing of the questionnaire was done in the field on 612 people. The sample size was calculated on the basis of a pilot study which estimated the case prevalence to be greater than 30%. The sample size in the current study exceeds the requirement several-fold. The city of Ahmedabad is divided into five administrative zones and 43 wards. In an effort to make the sample selection representative of the entire city population, 30 houses were selected randomly from each ward irrespective of presence of suspected case. As more than 40% of the population of the city resides in slums, half of the houses in each ward were selected from slums while remaining half were selected from other localities including bungalows and apartments. 42 Dengue Bulletin – Volume 34, 2010 Chikungunya outbreak during 2006 in Ahmedabad, India Limitations of the study The study’s case definition was based entirely on self-reported cases. It was not feasible to do confirmatory serology tests for each and every self-reported case. Self-reported cases were also not cross-examined by medical persons. However, the questionnaire was filled in by intern doctors for each suspected case. Results A total of 1301 houses were surveyed covering a population of 6667. The prevalence of fever over the past one month was found to be 51.3% while the prevalence of joint pain was 41.7%. The prevalence of suspected chikungunya cases was found to be 32.9% (95% CI, 31.8–34.2). Table 1 shows the number of suspected cases was significantly higher in females (Chi square test: 7.35, p=0.006). Table 1: Gender-wise distribution of suspected chikungunya cases Gender Suspected chikungunya Total population Percentage Male 1090 3467 31.4 (29.4–32.6) Female 1106 3200 34.6 (33.4–36.6) Total 2196 6667 32.9 (31.8–34.2) Figures in brackets suggest upper and lower limits at 95% confidence interval. The mean age among suspected cases was found to be 33.7 years. Average family size was 5.17 and the average number of suspected cases per family was 1.7. The age- wise distribution of cases reveals the highest number of cases to be in the 20–30 years age group. However, age-specific prevalence rates clearly indicate an increase in the prevalence Figure 1: Age-wise distribution of suspected chikungunya cases and age-specific prevalence rate 0 100 200 300 400 500 600 Age group 0 5 10 15 20 25 30 35 40 45 50 Suspected cases Age-specific prevalence rate (%) N u m b er o f su sp ec te d ch ik u n g u n ya ca se s A g e sp ec if ic p re va le n ce ra te >80 – 90>70 – 80>60 – 70>50 – 60>40 – 50>30 – 40>20 – 30>10 – 20< = 10 Dengue Bulletin – Volume 34, 2010 43 Chikungunya outbreak during 2006 in Ahmedabad, India rate with increase in age. The highest age- specific prevalence rate of 44.4% was in the age group of 60–70 years as shown in Figure 1. The distribution of suspected cases by zone indicates a significantly higher number of cases from the north zone, which is densely populated and a reported starting point of the outbreak in Ahmedabad City. A comparison of house types revealed that a higher number of cases were from slums than others (Chi square test: 15.75, p<0.001). A variety of signs and symptoms were also reported during the household survey, the result of which is shown in Table 2. Table 2: Prevalence of various reported signs and symptoms Signs and symptoms N = 2196 Per cent of suspected cases (%) Headache 1002 46 Chills 974 44 Itching 629 29 Joint swelling (oedema) 612 28 Tenderness (palms/ soles) 602 27 Rashes 416 19 Oral ulceration 256 12 Lymphadenopathy 18 1 Analysis of treatment-seeking behaviour shows that the majority (67.6%) of cases sought treatment from private facilities; around 23% sought treatment from public facilities, while 8.5% sought no treatment. Over-the- counter drugs were used by 1% of suspected cases for their symptoms while 6% of cases required hospitalization. Three per cent of the population was acutely suffering from suspected chikungunya during the survey. An acute case of suspected chikungunya was taken as having signs and symptoms of suspected chikungunya in the last four days prior to the survey. Table 3 shows the activities carried out by the local health department during the last seven days as perceived by the family respondents. Table 3: Activities carried out by the local health department during the 2006 outbreak Activities Per cent of houses covered (%) Visit by any worker in the last seven days 73.4 Larviciding of Abate* 67.5 Fogging 42.4 General information, education and communication activities 28.3 Pamphlet distribution 17.8 Knowledge about how chikungunya spreads 53.7 Houses where mosquito repellents had been used in the past one month 58.0 *Abate: a larvicide that controls aquatic stages of vectors. Prevention and control activities including the visit by a worker for awareness, release of larvicides (Abate), and fogging and pamphlet distribution, were carried out extensively as a matter of priority for the local health department. Health workers were able to reach the majority of houses for prevention and control activities. Table 3 shows that 53.7% of respondents had accurate knowledge 44 Dengue Bulletin – Volume 34, 2010 Chikungunya outbreak during 2006 in Ahmedabad, India or information regarding the spread of the chikungunya disease, while 58.0% of houses had used mosquito repellents in the past month. Discussion During the 2006 chikungunya epidemic in India there was a controversy over the reported numbers of people affected by the disease. Many newspaper articles and nongovernmental organizations (NGOs) expressed skepticism about the numbers of chikungunya cases reported by the government.[14] The study also identified that the majority of suspected cases sought treatment from the private sector which could possibly explain the lower prevalence reported by the government officials. This community- based study found that in 2006–2007, the prevalence of suspected chikungunya in the Ahmedabad city was 32.9% (95% CI, 31.8-34.2). The study also revealed that a greater number of females suffered than males, and that those affected were from an older age group, with the highest age-specific prevalence rate being 60–70 years. The north zone had a significantly higher number of cases, which can be explained by the density of the population and the living conditions prevailing in that zone. Similarly, a favourable environment and lack of protective measures against the vector mosquito might have lead to a higher number of cases in slums compared with other housing types. The hospitalization rate in the study indicates that only 6% had a severe attack of the disease during the 2006 epidemic. The local health department made significant efforts to cover large numbers of houses and families for prevention and control activities. However, more studies using laboratory confirmation are needed to determine the prevalence of confirmed cases of chikungunya. Acknowledgements We sincerely thank the Ahmedabad Municipal Corporation and the Government of Gujarat for all the support they provided to carry out the study. We also like to thank faculty and intern doctors of Municipal Medical College, Ahmedabad. We appreciate the support given by scientists and staff members of the National Institute of Malaria Research (ICMR), Field Station, Nadiad, Gujarat. References [1] Pialoux G, Gaüzère BA, Jauréguiberry S, Strobel M. Chikungunya, an epidemic arbovirosis. Lancet Infect Diseases, 2007, 7(5), 319–327. [2] Ross RW. The Newala epidemic. III. The virus: Isolation, pathogenic properties and relationship to the epidemic. The Journal of Hygiene, 1956, 54: 177–191. [3] Robinson MC. An epidemic of virus disease in Southern Province, Tanganyika Territory, in 1952–53. I. Clinical features. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1955, 49: 28–32. Dengue Bulletin – Volume 34, 2010 45 Chikungunya outbreak during 2006 in Ahmedabad, India [4] Sergon K, Njuguna C, Kalani R, Ofula V, Onyango C, Konongoi L, Bedno S, Burke H, Dumilla AM, Konde J, Njenga MK, Sang R, Breiman RF. Seroprevalence of Chikungunya virus (CHIKV) infection on Lamu Island, Kenya, October 2004. American Journal of Tropical Medicine and Hygiene, 2008, 78: 333–337. [5] AbuBakar S, Sam IC, Wong PF, MatRahim N, Hooi PS, Roslan N. Reemergence of endemic chikungunya, Malaysia. Emerging Infectious Diseases, 2007, 13: 147–149. [6] Halstead SB, Udomsakdi S, Singharaj P, Nisalak A. Dengue chikungunya virus infection in man in Thailand, 1962–1964. III3. Clinical, epidemiologic, and virologic observations on disease in nonindigenous white persons. American Journal of Tropical Medicine and Hygiene, 1969, 18: 984–996. [7] Borgherini G, Poubeau P, Staikowsky F, Lory M, Le Moullec N, Becquart JP, Wengling C, Michault A, Paganin F. Outbreak of chikungunya on Reunion Island : early clinical and laboratory features in 157 adult patients. Clinical Infectious Diseases, 2007, 44: 1401–1407. [8] World Health Organization. Chikungunya in India. 2006. Available from: URL: http://www. who.int/csr/don/2006_10_17/en/index.html. [9] Saxena S, Singh M, Mishra N, Lakshmi V. Resurgence of chikungunya virus in India: an emerging threat. Euro Surveillance, 2006.11(8):E060810.2. [Online] 2006. [Cited 5 February 2007] Available from: http:// www.eurosurveillance.org/ew/2006/060810. asp#2. [10] Yergolkar PN, Tandale BV, Arankalle VA, Sathe PS, Sudeep AB, Gandhe SS, Gokhle MD, Jacob GP, Hundekar SL, Mishra AC. Chikungunya outbreaks caused by African genotype, India. Emerging Infectious Diseases, 2006,12:1580–3. [11] Outbreak and spread of chikungunya. Weekly Epidemiological Record, No. 47, 2007, 82: 410. Available from http://www.who.int/ wer/2007/wer8247.pdf. [12] National Vector-Borne Disease Control Programme, Government of India, 2006. Available from http://www.nvbdcp.gov.in/Doc/ Facts%20about%20Chikungunya17806.pdf. [13] Chikungunya Fever. CD Alert, Monthly Newsletter of the National Institute of Communicable Diseases, 2006, 10(2):6. Available from http://nicd.nic.in/cdalert/ February-06.pdf. [14] Mavalankar DV, Shashtri P, Raman P. Chikungunya epidemic in India: A major public- health disaster. The Lancet Infectious Diseases, 2007;7;306-7 Available from http:// www.thelancet.com/journals/laninf/article/ PIIS1473-3099(07)70091-9/fulltext#article_ upsell. 46 Dengue Bulletin – Volume 34, 2010 Prevalence of various symptoms and cost of treatment during the chikungunya epidemic in Ahmedabad, Gujarat, India, in 2006 Dileep Mavalankara#, Dipti Govilb, Neha Trivedia and Vinubhai Patelc aIndian Institute of Management, Vastrapur, Ahmedabad 380015, Gujarat, India bIndian Institute of Health Management Research, Jaipur, Rajasthan, India c3B Vasu Apartments, Naranpura, Ahmedabad 380013, Gujarat, India Abstract Ahmedabad city in the State of Gujarat was one of the cities most affected by an epidemic of chikungunya. This paper attempts to document the epidemiological profile of chikungunya cases in the city of Ahmedabad in 2006. It also tries to study the economic burden of the disease on patients and understand their treatment-seeking behaviour. There was an average of 2.4 suspected chikungunya patients in a family of five members. Mean duration of illness was 19 days. Thirteen per cent of those suffering from chikungunya were admitted to hospital. The economic burden of chikungunya disease on a patient was US$ 28.5, out of which approximately US$ 15.4 was out-of-pocket expenditure on medical treatment, US$ 4.0 was on transportation. The average wage loss per patient (workers and non-workers) was US$ 10.4; with the average wage loss per worker being US$ 22.5. The cost of inpatient care was 1.5 times higher than the cost of outpatient care. An outbreak of an emerging disease creates unanticipated catastrophic health-care expenditure. Based on the study, the economic burden of chikungunya in Ahmedabad in 2006 (for 60 777 cases) was approximately US$ 1.7 million. The priorities when tackling such unexpected disease outbreaks should include strengthening of preventive and curative measures and continued research on the socioeconomic impact to help in policy-making and resource allocation. Keywords: Chikungunya; cost of treatment; Ahmedabad; India. #E-mail: dileep@iimahd.ernet.in Introduction In India, chikungunya fever is as old as it is globally. The epidemic was first reported during the 1960s in Kolkata in West Bengal, Pondicherry and Chennai in Tamil Nadu, Rajahmundry, Visakhapatnam and Kakinada in Andhra Pradesh, Sagar in Madhya Pradesh, and Nagpur in Maharashtra. Thereafter, sporadic cases also continued to be recorded, especially in Maharashtra state during 1973 and 2000.[1] The Indian health system was again recently challenged by the re-emergence of chikungunya disease after a gap of 32 years Dengue Bulletin – Volume 34, 2010 47 Cost of treatment of chikungunya cases in Ahmedabad in December 2005.[2] The National Institute of Communicable Diseases (NICD) Delhi issued a warning notice against the disease in March 2006. The states of Andhra Pradesh, Karnataka, Maharashtra, Madhya Pradesh, Tamil Nadu, Kerala, Gujarat, Andaman and Nicobar Islands and Orissa were affected by the disease[1] on a major scale. As per data from the National Vector-Borne Disease Control Programme (NCBDCP), 1 390 322 suspected and 1985 confirmed chikungunya fever cases were reported from 152 districts across the 13 states in India in the year 2006.[1] The state of Gujarat in India was also affected by the chikungunya epidemic between July and October 2006 with more than 72 589 officially reported cases.[1] Jamnagar was the first city in the state to be affected by the disease in April 2006 after focal epidemics in Maharashtra and Karnataka. The epidemic also affected Ahmedabad city1 extensively during the months of August–November 2006. Municipal health centers and hospitals in Ahmedabad reported 60 777 cases of chikungunya during the year (Source: National Vector Borne Disease Control Programme, Department of Health, Government of Gujarat). In the present paper, we document the epidemiological profile of chikungunya cases in the city of Ahmedabad, Gujarat, India. We also study the economic burden of chikungunya disease on patients and try to understand their treatment-seeking behaviour. 1 Ahmedabad is the largest city in the state of Gujarat and one of the largest urban agglomerations in India. As per the 2001 Indian census, the area under Ahmedabad Municipal Corporation has a population of 3 520 085. The population of the Ahmedabad urban agglomeration [which includes the region governed by Ahmedabad Urban Development Authority (AUDA)] is 4 525 013. The urban agglomeration accounts for 22 per cent of Gujarat's urban population. There are 43 wards in the city. Ahmedabad has a literacy rate of 79.9 per cent, which is the highest in Gujarat, where 87.8 per cent males and 71.1 per cent females are literate. Materials and methods We conducted a community-based study in the 43 wards of Ahmedabad city, Gujarat between the months of September and December 2006. A purposive sampling approach was used to identify the study population in each ward of the city. A structured questionnaire was filled in by health workers and link workers of the Municipal Corporation of Ahmedabad. Each health worker was asked to complete 10–15 questionnaires related to their work area. The households which reported to have cases of fever, joint pain, vomiting, skin rashes, etc. suggesting chikungunya symptoms were visited by these health workers for the interview. The information on each member of the household was covered in one interview schedule (one schedule per household). The respondent was the head of the household. Interviewers continued to visit households until they had filled in 10–15 questionnaires or visited households in their work area. The information on socioeconomic and demographic characteristics, treatment- seeking behaviour, cost of treatment, and loss in wages was collected. The data was coded, entered and analysed using SPSS Version 13.0 for Windows (SPSS com, Chicago, IL, USA). The cost of treatment was calculated taking into consideration the out-of-pocket expenditure in treatment (which includes consultation, blood test, medicine, hospital stay) and transportation. Expenditures were recorded in Indian rupees (INR) and later converted to US dollars at the exchange rate of 45.5 rupees per US dollar (the average value between April and December 2006.[3]) The total cost of the treatment incurred to a patient was calculated after taking into consideration the loss of wages to a household. 48 Dengue Bulletin – Volume 34, 2010 Cost of treatment of chikungunya cases in Ahmedabad A total of 3362 households were surveyed from 43 wards of the city, covering a population of 17 455, averaging 5.2 members per household. Almost half of the households earned their livelihood through daily wage work, whereas, 13% owned a business and one fifth had listed “service” as their primary occupation. Half of the members in a household (mean 2.6) suffered from fever, joint pain or other problems (cold and cough, vomiting, typhoid, malaria, skin problems/itching). For further analysis, the study considered only the cases with two symptoms, i.e. fever with joint pain, and the remaining cases (only fever and others) were excluded from the analysis. Limitation The aim of the study was to understand the profile of the population affected by chikungunya and not to measure the prevalence of the disease in the population. The information collected during the home visits was purely self reported and subjective. The symptoms were not verified by medical personnel. The information gathered may also be affected by the subjective understanding of the questionnaire by the interviewers. Results Profile of the patients Approximately 46% of the surveyed population (7944) suffered from fever coupled with joint pain or only joint pain (2.4 members per household). Out of the total number of those affected, 80% had fever along with joint pain (6359) where 20% reported to have only joint pain (1589). Among the patients, 54% were females and 46% males. The most seriously affected age group was 20–49 with a mean of 35 years (Figure 1). A total of 8.5% of the patients (aged 14+) were working: 81% male and 22% female. Most of the working male and female patients were daily wage-earners (45% and 16%, respectively) (Figure 2). Non-working female sufferers included those who were involved in non-wage-earning household work. The city of Ahmedabad is divided into five zones. The Central zone of the city was comparatively most affected (51%) compared with other zones: North zone (46%), East zone (47%), South zone (42%) and West zone (41%). Figure 1: Age distribution of suspected chikungunya cases (per cent) 5.2 14.4 19.5 21.0 18.2 11.4 10.3 1-9 10-19 20-29 30-39 40-49 50-59 60 + Dengue Bulletin – Volume 34, 2010 49 Cost of treatment of chikungunya cases in Ahmedabad Figure 2: Occupation of the sufferers by sex (> 14 years) (per cent) Service, 23.2Others 0.3 Not working, 19.2 Business, 12.6 Worker, 44.7 Service, 4.5 Others, 0.0 Not working, 77.8 Business, 2.0 Worker, 15.6 Male Female Figure 3: Month-wise per cent of suspected chikungunya cases, 2006 0.0 0.0 0.1 0.2 0.7 3.0 31.5 50.0 8.1 0.1 0.0 6.2 Ja n M ar ch A p ri l M ay Ju n e Ju ly A u g S ep t O c t N o v D e c N o R e sp Chikungunya illness and treatment- seeking The number of suspected chikungunya cases was the highest in the month of September (50%) followed by August (32%) (Figure 3). On average, chikungunya symptoms lasted for 19 days. Twenty-nine per cent of the patients reported that their illness had lasted for less than one week. A similar proportion suffered for one month or more (Table 1). A patient rested for an average of 12.8 days after the recovery. The majority of the patients consulted a doctor whereas 13% were admitted to the hospital for treatment. Forty-five per cent of the patients reported getting their blood tested. One quarter of the patients also practised home remedies (Figure 4). Cost of treatment, transportation and wage loss (in US$) On average the cost of chikungunya fever treatment was US$ 28.5 (INR 1296); including US$ 15.4 (INR 699) on medical treatment 50 Dengue Bulletin – Volume 34, 2010 Cost of treatment of chikungunya cases in Ahmedabad and US$ 4 (INR 183) on transportation. The average wage loss for patients working outside their homes (44% of surveyed population) was US$ 22.5 (INR 1024) (Table 2). Moreover, the average wage loss (workers and non-workers) per patient was US$ 10.4 (INR 473). The average cost incurred during the treatment differed for inpatients and outpatients (Table 3). The cost of treatment for outpatients was US$ 24.3, whereas the cost increased twofold for a patient who was admitted to hospital. The cost of transportation was also higher for inpatients than outpatients. An outpatient, who was working outside the home, lost an Figure 4: Treatment seeking behavior among chikungunya cases, Ahmedabad 2006 (per cent) 98.7 94.7 44.8 12.9 26.0 Consulted Doctor Medicines Blood Test Admitted in the Hospital Home Remedies Table 2: Total cost incurred during chikungunya (in US$) Mean (US$) SD Range N Cost of treatment 15.4 24.8 0-659.3 7837 Cost in transportation to hospital 4.0 8.8 0-153.8 7784 Wage loss (working outside the home for wages) 22.5 28.1 0-329.7 3238 Wage loss (total sufferers 14+) 10.4 22.2 0-329.7 7009 Total cost incurred during chikungunya* 28.5 38.7 0-747.3 7883 * Total may differ because of missing cases N=Number Table 1: Duration of Illness (per cent) Duration Fever with joint pain Only joint pain Total Less than one week 30.3 22.9 28.8 Two weeks 23.1 19.8 22.5 Three weeks 19.5 21.7 20.0 One month 14.6 19.7 15.6 More than one month 12.4 16.0 13.1 Total* 6274 1535 7811 * Total may differ because of missing cases Dengue Bulletin – Volume 34, 2010 51 Cost of treatment of chikungunya cases in Ahmedabad average wage of US$ 20.9 during his/her illness whereas this loss was recorded as US$ 32.7 for the average inpatient (working outside home). For all in- and outpatients (working and non-working), the average wage loss was US$ 15.3 and US$ 9.6, respectively. The total out-of-pocket and indirect cost of treatment was US$ 57.3 per inpatient and US$ 24.3 per outpatient. With an increased duration of illness, there was also an increase in the total cost of treatment (Table 4). Discussion Chikungunya is a “neglected” tropical arboviral infection in the developing world,[4] as there are limited studies on the impact of the disease on socioeconomic and medical determinants. Even though the disease is generally considered self-limiting and rarely life-threatening, the widespread occurrence of it causes substantial morbidity and economic loss due to ill health.[5,6,7] Studies indicate that an outbreak of an emerging disease creates unforeseen catastrophic health-care expenditure and reinforces the poverty and ill-health nexus.[5] Table 3: Average cost of treatment and wage loss during type of treatment (in US$) Admitted OPD Cost of treatment Mean 35.6 12.4 SD 52.4 15.2 Range 0-659.3 0-219.8 Cost of transportation Mean 8.0 3.4 SD 16.4 6.8 Range 0-153.9 0-153.9 Wage loss (working outside the home for wages) Mean 32.7 20.9 SD 38.6 25.8 Range 0-329.7 0-263.7 Wage loss (total sufferers 14+) Mean 15.3 9.6 SD 31.0 20.4 Range 0-329.7 0-263.8 Total cost incurred during chikungunya Mean 57.3 24.3 SD 71.2 28.9 Range 0-747.3 0-340.7 OPD: outpatient department SD: standard deviation Table 4: Total cost of treatment as per duration of illness (in US$) Duration of illness Cost of treatment Cost of transportation Wages loss (only working) Wage loss (total sufferers 14+) Total cost incurred One week or less 9.5 2.6 14.2 6.2 17.3 Two weeks 13.1 3.4 19.6 9.5 24.8 Three weeks 17.0 4.7 24.8 12.3 32.7 One month 20.4 5.0 29.9 13.8 37.7 More than one month 23.5 5.9 31.0 13.7 42.0 52 Dengue Bulletin – Volume 34, 2010 Cost of treatment of chikungunya cases in Ahmedabad Our study addresses the economic loss consequent to the chikungunya disease. It considers only a few direct (medicine, transportation and hospitalisation) and indirect costs (loss of pay) for the treatment of chikungunya. It does not address several other indirect costs related to days lost (for patient and relatives), such as food and lodging. In the present study, the economic burden of reported chikungunya disease on a patient was US$ 28.5, which includes US$ 15.4 as out- of-pocket expenditure on medical treatment, US$ 4 on transportation and US$ 10.4 due to wage loss. During 2006, the real per capita income of Gujarat was US$ 583 per annum (US$ 48.6 per month).[8] This study estimates that approximately 59% of the average monthly income was spent on the treatment of suspected chikungunya symptoms, which amounts to US$ 28.7 per month. This is an underestimate because our study did not estimate the opportunity costs of the illness (e.g. males/females working at home, children missing school time, etc.). A study by Gopalan and Das[5] calculated the loss of economic productivity due to chikungunya disease among the breadwinners in a village of Orissa. The median out-of- pocket health-care expenditure on treatment was US$ 84, which is much higher than the estimated cost in our study. An article on an epidemic in South India also shows a similar result.[7] It says that apart from the expenditure on treatment, lingering debilitation following chikungunya infection takes away the victim’s productivity and negatively impacts family incomes for weeks or months. Another study on a Chikungunya outbreak during 2005–2006 in India reveals another toll of the disease.[9] It suggests that poor people are more commonly affected, as malnutrition enhances individual susceptibility to infectious disease by lowering immunity. The loss of income during illness led to worsening poverty. The scope of the present study does not consider the effect of the fever on poverty or vice versa. However, the reported occupations of patients in the present study indicate that they may have a low standard of living since more than half of the surveyed families were working for daily wages. Based on our study, the economic burden of chikungunya in Ahmedabad in 2006 (for 60 777 cases) was approximately US$ 1.7 million (INR 78.7 million). The total cost of the epidemic to society may be much higher as this estimate does not cover unreported cases and several other indirect costs. Research also suggests that according to informal reports from various parts of the country, the incidence of disease was probably higher than the reported figures with the government.[5] With an assumption that the actual cases of chikungunya were 5 or 10 times higher than those reported, the total cost is pegged at US$ 8.6 million (INR 393.8 million) in the former case or US$ 17.3 million (INR 787.6 million) in the latter. The priorities of tackling such an unexpected disease outbreak should include the strengthening of preventive and curative measures. There is also a need for a detailed and systematic study of chikungunya disease along with its socioeconomic impact. Such studies will help in policy-making and resource allocation for mosquito-borne disease by the state and central governments. Dengue Bulletin – Volume 34, 2010 53 Cost of treatment of chikungunya cases in Ahmedabad References [1] Ministry of Health and Family Welfare. Chikungunya Fever. National Vector-Borne Disease Control Programme, Director- General of Health Services, 2009. Available at http://nvbdcp.gov.in/Doc/Facts%20about%20 Chikungunya17806.pdf. Accessed on 7th September 2010. [2] Lahariya C, Pradhan SK. Emergence of chikungunya virus in Indian subcontinent after 32 years: a review. Journal of Vector Borne Diseases, 2006, 43:151-160. [3] Antweiler W. Exchange Rate Review. Published by University of British Columbia, Vancouver BC, Canada, 2009. [4] LaBeaud AD. Why arboviruses can be neglected tropical diseases? PloS Neglected Tropical Disease, 2008, 2(6): e247. doi:10.1371/ journal.pntd.0000247. [5] Gopalan SS, Das A. Household economic impact of an emerging disease in terms of catastrophic out-of-pocket health care expenditure and loss of productivity: Investigation of an outbreak of chikungunya in Orissa, India. Journal of Vector-Borne Diseases, 2009, 46, 57-64. [6] Kr i shna MR, Reddy MR, Reddy SR. Chikungunya outbreak in Andhra Pradesh, South India. Current Science, 2006, 91 (5): 570-571. [7] Varshney V, Kannada D. Chikungunya chase. Down to Earth, 2008, October 1-15, 38-40. Available at http://www.downtoearth.org.in/ node/5132. [8] Gupta S. States performance in per capita income growth. ASSOCHAM ECO PULSE, January 2008. Available at http://www. assocham.org/arb/aep/states-per-capita- income.doc, accessed on 29th October 2009. [9] Kumar CJ, Babii CA, Krishnan BU, Kumar A, Joy S, Jose T, Philip A, Sambasivaiah K, Hegde BM. The socioeconomic impact of the chikungunya viral epidemic in India. Open Medicine, 2007, 1 (3), 150-152. 54 Dengue Bulletin – Volume 34, 2010 A private hospital-based study assessing knowledge, attitudes, practices and costs associated with dengue illness in Surat, India Ami T. Bhavsara, Donald S. Shepardb,#, Jose A. Suayab, Moses Mafowosofoc, Clare L. Hurleyb and Marion W. Howardc aGraduate Research Assistant, Brandeis University (at time of study), USA bSchneider Institutes for Health Policy, Heller School, Brandeis University, Waltham, MA 02454-9110, USA cHeller School for Social Policy and Management, Brandeis University, Waltham, MA, 02454-9110, USA Abstract We conducted a descriptive cross-sectional study at a private children’s and general hospital in Surat, India, during the 2006 dengue season (June to December 2006). We examined knowledge, attitudes and practices of patients associated with dengue and estimated the cost of a dengue episode for the study sample. Of 62 patients with suspected dengue, 40 consented to participate in this study. We interviewed 27 retrospectively in their homes and 13 prospectively during their hospitalization. Among the respondents, 75% were male, 85% were under the age of 14, 63% reported familiarity with dengue, but only 25% correctly knew that clean stagnant water is a breeding place for the dengue vector, and 23% did not know the mode of transmission for dengue infection. While 78% said that they take preventive steps against dengue, only 48% believed dengue to be a preventable illness. There was no significant relationship between their knowledge and prevention practices. Economic costs were measured in 2007 US dollars. For an average dengue episode, the mean costs were US$ 439.44 for direct medical care and US$ 146.13 for indirect costs, with a total cost of US$ 585.57. The indirect costs reflect an average burden of 50 days on a household due to days lost from school, work, and other activities by the patient and their caretakers. While other studies have examined public institutions, this is one of the first studies using data from a private hospital. This study shows that the cost of a dengue episode imposes substantial challenges, even on middle-class households. Keywords: Dengue; transmission; preventive practices; treatment cost; indirect cost; economic burden; Surat; India. #Corresponding author: Donald S. Shepard, shepard@brandeis.edu Dengue Bulletin – Volume 34, 2010 55 A KAP study associated with dengue illness in Surat, India Introduction Dengue is one of the leading causes of death and hospitalization among children in India. In recent times, a major outbreak associated with haemorrhagic manifestations occurred in Calcutta in 2004[1]. According to the WHO, dengue infection has existed in India for more than a century[2]. Communicable diseases are a major cause of death around the world. Despite the success of vaccination programmes for polio and many childhood diseases, other infections like AIDS, tuberculosis, malaria and dengue are still widespread in many developing countries[3]. Some of the studies on dengue in India have examined the cause of the disease and its impact, but previous cost analyses of dengue illness and its relation to preventive practices are virtually absent. Such an analysis would allow policy-makers to make decisions that take into account the impact or burden of the disease on society. The main purpose of this article is to assess the costs, knowledge, attitudes and practices (KAP) associated with dengue illness. The analysis will also inform future cost-effectiveness studies as it reveals the cost of illness that could be averted if either a vaccine is developed or preventive programmes are implemented(4). The analysis and conclusions could also be used by government agencies to develop financing programmes; for example, a micro-insurance programme that would help reduce the out-of-pocket expenditure incurred for treating the illness. In addition, Non-Governmental Organizations (NGOs) and community organizations working to improve health conditions in India could use the data to develop awareness programmes for preventing dengue illness. Materials and methods Study setting In August 2006, the city of Surat in Gujarat state in northwest India was hit by devastating floods. According to the local Sandesh newspaper, the city was flooded with 113 inches of rainfall in 48 hours; and more than 50 000 cases of malaria, cholera, diarrhoea, dengue and many other illnesses were reported.[5] Expecting a severe dengue season after these floods, Surat city was selected for this study. One of the most reputable private hospitals, Nirmal Children’s and General Hospital (NCGH), renowned in the region for dengue treatment, was chosen. This hospital- based study targeted all suspected dengue patients admitted in NCGH from within the limits of Surat Municipal Corporation from June 2006 to December 2006, the period of the 2006 dengue season. Interview data collection and analysis A standardized research instrument was adapted from the Dengue Burden of Illness Study[6] and the socioeconomic indicators section was adapted from a national survey in India called National Family Household Survey.[7] The instrument was then translated into Gujarati and pilot-tested at Brandeis University on Gujarati-speaking students. The prospective interviews (within hospital interviews) were conducted on or near the date of discharge. Patients who were discharged before the study were interviewed retrospectively at home. The database was in English. To protect the identity of the individuals interviewed, the database did not include any personal identifiers such as name and address, although a protected linkage to the subject’s 56 Dengue Bulletin – Volume 34, 2010 A KAP study associated with dengue illness in Surat, India ID number could allow researchers to check the original data if necessary. The interview data were analysed using the current version of SPSS statistical software.[8] Costing methodology All costs were first calculated in Indian rupees and then converted to US Dollars based on the mid-2007 dollar exchange rate of Rupees as derived from Bombay Stock Exchange Limited.[9] The total cost of a dengue illness episode comprises the direct cost and indirect cost of treating the illness. The direct cost includes both the cost of medical care and non-medical costs. First, the medical cost data were categorized by type of provider (public or private) and type of visit (ambulatory or hospitalization) and then summarized using SPSS. The cost of one day at a public hospital was assumed to be equal to the cost of one day in the general ward of NCGH. For this purpose, the cost of one day at NCGH was calculated on the basis of the average cost of hospitalization in the general ward and then dividing by the mean number of days for hospitalization during an episode of dengue illness. The cost of drugs was considered separately for a public hospital, assuming that the drug store is neither owned by the hospital nor located in the hospital compound. But for NCGH, the medical supplies and drug prices were included in the cost of treatment, as the medical supply store was owned by NCGH. Second, direct non-medical costs incurred by the hospitalized patients while seeking care with various providers was calculated. This cost comprised the costs incurred either by patients or their family members for transportation to and from the hospital and for lodging and food during hospitalization of the dengue patient. Indirect costs are the value of the patients’ or family members’ days lost or adversely affected by the illness at work, school, or home. Based on a dengue disease burden study[6] we valued the economic loss of one school day lost as the per day cost incurred by the Government of India for one day of public primary education for one pupil. We considered costs of primary education as the majority of the patients were attending this level of schooling. The public expenditure on education, i.e. the current spending per student in India as a percentage of per capita GDP, is 9.3%.[10] The GDP per capita of 3700 rupees.[11] The cost per child per year was divided by a total of 200 school days (the annual number based on the Indian school calendar after deducting school vacations and holidays) to estimate the cost of one day incurred by the Government for primary schools for one day. The per day cost was multiplied by the days of school absence to derive the total cost of school absence. Other days affected and their economic value For estimating the other days lost by the patient due to illness, we first calculated the total days of illness for the patient by estimating the days between the onset of fever and the date of interview (as interviews were taken on the day of or day before discharge). We then deducted the days of absence from work as they are already accounted for in workdays lost. The days of recovery were then added to arrive at the total of other days affected due to illness. The other days affected were calculated as the total number of days spent by each family member caring for the patient in terms of hours per day. To avoid confounding dengue Dengue Bulletin – Volume 34, 2010 57 A KAP study associated with dengue illness in Surat, India with other illnesses, we set the maximum for each member to 30 days less the number of days of work absence. The number of hours per day was limited to a maximum of eight hours per day for consistency with the number of working hours lost. The economic cost for other days affected due to illness was computed by multiplying the number of other days affected by the current minimum daily wage estimated for Gujarat, assuming that the cost of these days should be equal to the daily minimum wage earned by the people of Surat city. Days of absence from work and their economic value For evaluating the economic value of days of absence from work, we used the reported income lost if it was at least the amount that would have been lost at the minimum wage. However, some family members lost work days, but their loss of income was below the daily minimum wage or zero. This could have been due to paid sick leave or substitution by other family members for the time lost. For these members the loss of income was the product of the reported days of work lost and the daily minimum wage estimated for Gujarat, assuming that economic value of the work days lost should be at least equal to the daily minimum wage. Results A total of 40 patients with dengue illness were analysed. Table 1 describes the characteristics of the dengue cases interviewed and timing of initial medical care after onset of symptoms. The table shows that nearly two thirds (65%) of study patients had a clinical diagnosis Table 1: Characteristics of study patients, n=40 Characteristics of patients with dengue Per cent Clinical diagnosis Dengue fever Dengue haemorrhagic fever (DHF) DHF plus dengue shock syndrome Dengue shock syndrome alone 65 24 8 3 Type of respondent providing data Father of patient Mother of patient Both mother and father Other family members over the age of 18 Patient 38 30 10 20 3 Time to initiate medical care after onset of symptoms Less than 24 hours Between 24 and 48 hours Greater than 48 hours 63 25 13 58 Dengue Bulletin – Volume 34, 2010 A KAP study associated with dengue illness in Surat, India of dengue fever; the remaining 35% had a more severe form of the disease. Laboratory testing was performed on all cases. Out of 13 cases interviewed prospectively, six cases were laboratory confirmed (based on IgM positive). Although laboratory testing was also performed on the 27 cases who were interviewed retrospectively, the results were lost due to a flood and were not available for this study. The survey results, shown in Figure 1, indicate that a high percentage (90%) of the patients and their families were able to pay for Figure 1: Questionnaire responses from patient households concerning barriers to treatment, source of funds for treatment, knowledge about dengue disease, and practices related to dengue disease 65% 23% 18% 13% 3% 35% 90% 43% 20% 5% 10% 65% 63% 55% 37% 35% 25% 15% 23% 93% 78% 35% 13% 2% 0% 20% 40% 60% 80% 100% Barriers to obtaining care: Any barrier Money barrier Transportation barrier Distance barrier Fees barrier No barrier Source of funds for treatment: Current income and savings Borrowed money from relatives and friends Took out loan on interest Sold or mortgaged assets Other sources Knowledge about dengue: Learned through other sources Familiar with name of disease Knew dengue transmitted by mosquito bite Not familiar prior to onset of disease Learned through newspaper and television Knew mosquitoes breed in clean, stagnant water Knew dengue transmitted by Aedes aegypti Unfamiliar with symptoms of dengue Practices related to risk of dengue: Store water for daily use Take steps to prevent dengue in their house Use coil or liquid vaporizer to repel/kill mosquitoes Clean house Prevent water stagnation Per cent of patients' households Dengue Bulletin – Volume 34, 2010 59 A KAP study associated with dengue illness in Surat, India medical treatment using current income and savings—an indicator of their middle or upper socioeconomic status (SES). Also of interest is that 93% of households needed to store water for daily use, but only 25% knew that clean, stagnant water is a potential breeding site for the disease vector. To classify the SES of patients’ households more precisely, indicators were adapted from the household characteristics questionnaire of a national survey of India, the Standard of Living Index (SLI), taken from the 1998–1999 National Family Health Survey (NFHS-2).[7] Figure 2 shows the percentage of patients from low-, medium-, and high SES from the cohort (NCGH), from urban Gujarat and from south Delhi. Due to NCGH’s status as a private hospital, its proportion of patients with high standard of living (90%) is much higher than that in urban Gugarat (39%). Households with a high standard of living are better able to pay for the cost of treatment at a private medical facility such as NCG Hospital. The SES of the cohort was substantially higher than that of the average inhabitant of Gujarat in 1998–1999 [NFHS-2].[7] While comparable variables for NFHS-3 are not available, other data show moderate improvements in key assets from NFHS-2 to NFHS-3, but would not reach the level of the study participants [NFHS-3].[12] Direct medical cost Of the total 85 visits made by 40 patients, 34 of them were ambulatory visits and 51 were hospitalizations. Out of the total visits, only one visit was to a government hospital and the remaining 84 were to a private hospital. The results show that out of the total 34 ambulatory visits made for a dengue illness episode, none of the ambulatory visits was made to a public hospital, a mean of 0.85 visits were made in total for ambulatory care and the total cost for the ambulatory visits averaged US$ 33.92 for a dengue illness episode. Table 2, which reports household expenditures for medical care utilization and cost per dengue episode, shows that the total cost of the dengue illness episode per household was estimated at US$ 585.57, including direct and indirect costs. The figure seems plausible for India, and takes into account hospitalization in a private hospital, and includes the direct cost of care and the economic value of days of absence. This seems high, but may be due to the average number of people per household in the study group being six. In spite of the high number of days affected, the indirect total cost seems low because other days affected has been valued at daily minimum wage. This wage was used in our calculations because caregivers were often women who did not undertake paid work. The table also shows that the cost of treatment for householders at a private hospital far exceeds the costs at a public hospital. We can infer that most of the study cohort was a privileged group who could afford to choose private hospitalization. Figure 2: Socioeconomic status of the patients’ households based on Standard of Living Index (SLI) score in India 0% 20% 40% 60% 80% 100% Urban Gujarat South Delhi NCG Hospital Location of patient population High standard of living Medium standard of living Low standard of living P er ce n t 60 Dengue Bulletin – Volume 34, 2010 A KAP study associated with dengue illness in Surat, India According to a study done at the All India Institute of Medical Sciences (AIIMS), New Delhi, India, the out-of-pocket cost per outpatient visit was US$ 0.44 at a primary care facility and US$ 0.31 in the outpatient department of a secondary hospital, while at the tertiary hospital outpatient department, it was US$ 0.73. The cost of inpatient services at the secondary level was estimated as US$ 9.10 at the tertiary level it was US$ 20.20 per visit.[10,13] These costs are mainly from the Government’s perspective, whereas the current study emphasizes the household’s perspective. Also, we have explicitly taken into account the cost of drugs and medical supplies while calculating the cost per visit for private hospitals, although these may be included implicitly through providers’ charges. Direct non-medical cost The mean cost of transportation came to US$ 35.40; no one reported cost for lodging, and the mean cost of food expenses was also low at US$ 2.15. The total direct non-medical cost Table 2: Medical care utilization and costs per dengue episode for all members of the household Component Utilization Cost (US$) Mean SD Mean SD Direct medical care 0.9 0.6 34 39 Ambulatory care visits Inpatient care Days in public hospital Days in private hospital Total hospital days (all types) Prescriptions, laboratory, etc. 0.3 4.7 4.7 n.a. 0.2 5.2 5.2 n.a. 0.33 368 368 0.40 2 511 512 3 Total direct medical care n.a. n.a. 402 515 Direct non-medical cost n.a. n.a. 38 118 Indirect costs School lost Work for pay lost Other days affected (ages 0–14) Other days affected (ages 15+ only) Other days affected (all ages) 11.4 9.6 3.4 25.5 28.9 17.1 16.0 6.6 28.6 28.6 20 75 0 52 52 29 122 0 59 59 Subtotal 49.9 40.6 146 142 Total costs n.a. n.a. 586 583 Notes: This table considers only household expenditures, and does not consider government subsidies. SD denotes standard deviation; n.a. denotes not applicable. Dengue Bulletin – Volume 34, 2010 61 A KAP study associated with dengue illness in Surat, India averaged US$ 37.55, which increased the total direct cost for treating a dengue illness episode to a mean of US$ 439.44. Indirect cost Table 2 provides data for days of medical utilization and income lost for patients and their care givers during an episode of dengue. The cost per day for primary education was estimated to be US$ 1.72 calculated by dividing the annual national expenditure[10] for primary education by the annual number of pupil days.[11,14] The mean days of total school absence in all household members (11.4 days) combines days lost for the patient (an average of 10.7 days per household) and that for family members (an average of 0.7 days). The reported school absence was very low for family members because there was often a grandparent, uncle or aunt to help maintain the family routine, including school attendance of siblings of the sick child. Total cost per dengue episode The average cost of a dengue episode was estimated to be US$ 585.57. This amount is relatively high as private hospitals are expensive. Discussion This is the first study of treatment and costs of a dengue illness episode at a private institution in India. Although 90% of the households of patients treated belonged to a high socioeconomic status, dengue still imposed major financial burdens. To finance treatment, 43% of households reported that they borrowed money from friends or relatives and 20% had taken a loan. Difficulty with the cost of transportation and sufficiency of funds were the major challenges affecting care-seeking behaviour. Accessibility of health facilities, a major factor affecting health-seeking behaviour, was relatively good. Published data from 1991 reported that Surat city had 18 government hospitals, 80 nongovernment hospitals, 1398 allopathic doctors and 131 indigenous doctors. South Gujarat, Vadodara, Surat and Valsad districts have relatively high numbers of allopathic doctors.[15]. With growing urbanization and an increase in urban diseases, the numbers of hospitals and doctors serving the Surat area are also multiplying. This affects the treatment- seeking behaviour of people, as hospitals or dispensaries are available in the vicinity and in urban areas, and transportation to them is also easily available. This could explain why most patients were able to seek and receive care during the first 24 hours of onset of the illness. Since these data were collected from a private hospital, most of the patients were relatively well off; this fact helps explain why only a small percentage of people reported paying the fees of the medical provider to be difficult. Despite reported difficulties with the distance to hospital and the cost of transportation, the majority of patients in this study sought care promptly after the onset of illness. This behaviour is possibly due to the high SES of the households and the fact that we drew our study sample from hospitalized patients. In comparison, a study by Acharya et al. in south Delhi that used the Kuppuswamy scale to measure the SES of dengue patients’ families revealed a lower spectrum than NCGH.[16] The reason for this difference could 62 Dengue Bulletin – Volume 34, 2010 A KAP study associated with dengue illness in Surat, India be that the study in Delhi was conducted in an urban resettlement area which might not be a preferred place to dwell for people of high SES. A recent study on dengue costs in Pune, India, also focused on public sector institutions.[17] A study of chikungunya (another Aedes mosquito-borne viral illness) in Ahmedabad in this same issue of the Dengue Bulletin[18] has found a dramatically lower average household cost per case (US$ 25). Major reasons for the difference are that the Ahmedabad study did not count government subsidy to facilities, its subjects were primarily from the lower economic strata, its cases were mostly treated as outpatients, and it mostly describes public sector care.[18] Diverse settings, as illustrated by these two papers, can inform modelling studies in order to estimate overall immediate costs due to dengue and/or chikungunya, such as illustrated in papers[19,20] in this Dengue Bulletin. Collectively, the present study and the comparative literature show that dengue affects all economic strata. Considering the knowledge about dengue illness among study groups, it can be inferred that many people have heard about dengue illness but have little knowledge about breeding sites for the dengue vector. Knowledge about symptoms of dengue illness is also low. Data reveal that television and newspapers can play a major role in spreading awareness about the disease and schools have been effective in bringing about awareness in the communities. Our survey revealed that although the majority of households thought that dengue could be prevented and that cleanliness is important, respondents had little knowledge about the relationship between various household practices and dengue. Ambulatory costs in this study might seem high, but this is because the cost of a visit in this case also included the cost of drugs and medical supplies. Also of note is that all the ambulatory visits were done in private hospitals. Since this study was limited to only one private hospital, we have limited data for comparison between private and public medical facilities. While dengue affects people from all socioeconomic levels, for those with a low or medium standard of living private hospitalization is a costly option. Further study on the cost of treating a dengue illness episode in a public hospital is recommended. This will enable better cost comparisons for dengue treatment between these two sectors and will assist the government in selecting the kind of facilities it seeks to promote. In addition, it could help patients choose a facility when seeking treatment for dengue. The estimated cost of transportation is higher than might be expected because this cost not only represents the cost of taking the patient to various hospitals but also the daily cost of travel by family members to the hospital while the patient was hospitalized. The number of days affected is higher than is sometimes reported as it includes the days affected or lost by patients plus family members, and the average size of the household for the study was six people per family. The indirect total cost is relatively low because other days affected (which is a high number) are valued at daily minimum wage. For the most part, it was generally a woman who spent the large number of hours/days caring for a patient. Because the majority of women do not undertake paid work, other days affected were valued at daily minimum wage. To compare our results with those from similar studies elsewhere, we converted the results into international dollars (I$) to adjust Dengue Bulletin – Volume 34, 2010 63 A KAP study associated with dengue illness in Surat, India for purchasing power parity.[21] Our cost per case in NCGH of US$ 585.57 was I$ 1764. A multicountry study revealed that the total cost of dengue episodes for a hospitalized cohort using the same methods averaged I$ 1394 across eight countries.[6] Country-specific analyses found this cost ranged from I$ 752 in Guatemala to I$ 1988 in Malaysia. While out-of-pocket payments were substantial for patients in this study and the overall economic cost per case is above average, they are within the range of international experience. Improved knowledge about water storage and other preventive practices might help prevent some cases to reduce future dengue burden. References [1] National Institute of Communicable Diseases. CD Alert. Dengue/Dengue Haemorrhagic Fever, Delhi, India, 2004, 8: 2-4. [2] World Health Organization (WHO) and National Institute on Communicable Diseases (NICD). A report of meeting on vector borne disease surveillance with special reference to DengueNet. Delhi, India: National Institute of Communicable Diseases, 2004. (unpublished) [3] Jamison DT, Breman J, Measham AR, Alleyne G, Claeson M, Evans DB, Jha P, Mills A, Musgrove P. Disease Control Priorities in Developing Countries (2nd Edition). New York: Oxford University Press, 2006. [4] Shepard DS, Suaya JA, Halstead SB, Nathan MB, Gubler DJ, Mahoney RT, Wang DNC, Meltzer MI. Cost-effectiveness of a pediatric dengue vaccine. Vaccine, 2004; 22(9-10): 1275-80. [5] Flood situation in Surat. Sandesh Newspaper 8 Sept 2006; Available from: www.sandesh. com. Accessed 9 Sep 2006. [6] Suaya JA, Shepard DS, Siqueira JB, Martelli CT, Lum LC, Tan LH, Kongsin S, Jiamton S, Garrido F, Montoya R, Armien B, Huy R, Castillo L, Caram M, Sah BK, Sughayyar R, Tyo KR, Halstead SB. Cost of dengue cases in eight countries in Americas and Asia, a prospective study. American Journal of Tropical Medicine and Hygiene, 2009; 80(5): 845-55. [7] International Institute for Population Sciences (IIPS) and ORC Macro. National Family Health Survey (NFHS-2), 1998–99, IIPS, Mumbai, India 2000 Available from: http://www. nfhsindia.org. Accessed 14 October 2009. [8] SPSS. Version 14 for Windows: SPSS Inc., Chicago, IL, 2006. [9] Bombay Stock Exchange Limited. Key statistics, Dollar exchange rate, 2007. Available from: http://www.bseindia.com/about/st_key/ dollr_exg.asp. Accessed 28 Feb. 2007. [10] Ayre G, Callway R. Governance for sustainable development, a foundation for future, London: Earthscan, 2005. [11] Central Intelligence Agency. The World Fact Book, Rank-Order-GDP-Per Capita (PPP), 2007. Available from: http://www.cia.gov/cia/ publications/factbook/rankorder/2004rank. html. Accessed 29 March 2007. [12] International Institute for Population Sciences (IIPS) and ORC Macro. National Family Health Survey (NFHS-3), 2005–06, IIPS. Available from: http://www.nfhsindia.org. Accessed 14 October 2009. [13] Anand S, Hanson K. Disability-adjusted life years, a critical review. Journal of Health Economics, 1997; 16: 685-702. 64 Dengue Bulletin – Volume 34, 2010 A KAP study associated with dengue illness in Surat, India [14] World Bank. Summary Education Profile: India. Available from: http://devdata.worldbank. org/edstats/SummaryEducationProfiles/ C o u n t r y D a t a / G e t S h o w D a t a . asp?sCtry=IND,India. Accessed 29 Mar 2007. [15] National Council of Applied Economic Research. West and Central India: Human Development Report. New Delhi, India: Oxford Press, 2002. [16] Acharya A, Goswami K, Srinath S, Goswami S. Awareness about dengue syndrome and related preventive practices amongst residents of an urban resettlement colony of South Delhi. Journal of Vector Borne Diseases, 2005; 42: 122-7. [17] Garg P, Nagpal J, Kairnar P, Sinveratne SL. Economic burden of dengue infections in India. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2008, 102(6): 570-77. [18] Mavalankar DV, Govil D, Trivedi N, Patel VB. Prevalence of various symptoms and cost of treatment during chikungunya epidemic in Ahmedabad, India. Dengue Bulletin. 2010;34: 46-53. [19] Murtola TM, Vasan SS, Puwar TI, Govil D, Field RW, Gong HF, Bhavsar AT, Suaya JA, Howard M, Shepard DS, Kohli VK, Prajapati PB, Singh A, Mavalankar DV. A preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India. Dengue Bulletin, 2010, 34: 32-39. [20] Lee HL, Vasan SS, Birgelen L, Murtola TM, Gong HF, Field RW, Mavalankar DV, Nazni WA, Hakim LS, Murad S, Ng CW, Lum LCS, Suaya JA, Shepard DS. Immediate cost of dengue to Malaysia and Thailand: An Estimate. Dengue Bulletin, 2010;34:65-76. [21] World Bank. World Development Indicators. Available from: http://ddp-ext.worldbank. org/ext/DDPQQ/member.do?method=getM embers&userid=1&queryId=135. Accessed 10 Nov 2009. Dengue Bulletin – Volume 34, 2010 65 Immediate cost of dengue to Malaysia and Thailand: An estimate Lee Han Lima#, S.S. Vasanb,c, Luise Birgelenc,d, Tiina M. Murtolae, Hong-Fei Gongc, Robert W. Fieldd, Dileep V. Mavalankarf, Nazni Wasi Ahmada, Lokman S. Hakima, Shahnaz Murada, Ng Chiu Wang, Lucy Lum Chai Seeh, Jose A. Suayai and Donald S. Shepardi aWHO Collaborating Centre for Ecology, Taxonomy and Control of Vectors of Malaria, Filariasis and Dengue, Institute for Medical Research, Medical Entomology Section, Ministry of Health Malaysia, Jalan Pahang, 50588 Kuala Lumpur, Malaysia bUniversity of Malaya, Centre for Research in Biotechnology for Agriculture (CEBAR), IPS Building (Level 5, Block B), Kuala Lumpur, 50603, Malaysia cOxitec Limited, 71 Milton Park, Oxford OX14 4RX, UK dUniversity of Oxford, Department of Engineering Science, Parks Road, Oxford OX1 3PJ, UK eAalto University, Department of Mathematics and Systems Analysis, FI-00076 Aalto, Finland fCentre for Management of Health Services, Indian Institute of Management, Vastrapur, Ahmedabad 380015, India gUniversity of Malaya, Department of Social and Preventive Medicine, Faculty of Medicine, Kuala Lumpur, 50603, Malaysia hUniversity of Malaya, Department of Paediatrics, Faculty of Medicine, Kuala Lumpur, 50603, Malaysia iBrandeis University, Schneider Institutes for Health Policy, Heller School, MS035, Waltham, MA 02454-9110, USA Abstract We have analysed the average annual cost of dengue in Malaysia during the period 2002–2007 and in Thailand between 2000 and 2005. The key cost components, estimated by combining existing data from both published and unpublished studies, consist of: (i) costs of non-fatal illness; (ii) vector (Aedes mosquitoes) control costs; and (iii) research and development (R&D) costs incurred by government institutions. We found the immediate cost of dengue to Malaysia to be in the range of US$ 88 million to US$ 215 million (mean US$ 133 million) per annum. For Thailand, the corresponding range is US$ 56 million to US$ 264 million (mean US$ 135 million) per annum. For the period analysed, Thailand has 3.6 times more total cases of dengue, but Malaysia has a 4.6 times higher cost per case. In Malaysia, the most important parameters creating uncertainty in the immediate cost are reporting rate, hospitalization rate, and cost per ambulatory case. The corresponding parameters in Thailand are cost per ambulatory case, cost per hospitalized case, and reporting rate. Better estimates of cost per ambulatory case and reporting rate are therefore needed for both countries. Future studies should also refine the estimates of hospitalization rate in Malaysia and the cost per hospitalized case in Thailand. #E-mail: leehl@imr.gov.my; Tel and fax: +603 2616-2688 66 Dengue Bulletin – Volume 34, 2010 Immediate cost of dengue to Malaysia and Thailand: An estimate Introduction In two previous publications we developed the RUHA framework (defined below) to incorporate disease underreporting in estimating the cost of illness from certain mosquito-borne diseases, and applied it to dengue and chikungunya in India.[1,2] This paper uses this approach to examine the immediate costs of dengue in two heavily affected countries in South-East Asia, viz., Malaysia and Thailand. Methods Estimating key cost parameters Previous studies have reported dengue costs for a specific year, but we have analysed the average cost over several years to reduce the effect of year-to-year fluctuations caused by the tendency of dengue to occur in cycles.[3] Based on data availability, we have chosen the period 2002–2007 for Malaysia and 2000– 2005 for Thailand. From a societal perspective, the key cost components consisted of: (i) costs of non-fatal illness (i.e. direct and indirect health-care costs due to hospitalized and ambulatory cases, both reported and unreported); (ii) vector (Aedes) control costs (i.e. inspections and enforcement, fumigation, household insecticides, private sector vector control and the government’s communication- for-behavioural-impact (COMBI) efforts); and (iii) research and development (R&D) costs incurred by government institutions. This is consistent with Haddix et al.[4] who have considered cost of illness and cost of intervention measures in their analysis. These components were estimated by combining data from various studies (both published and unpublished) where available. If unavailable, trends/data from comparable countries in the region were used to obtain a range of values, which were then subjected to a Monte Carlo sensitivity analysis. For comparison, the impact of a major outbreak on tourism revenues[2] has been included in the epilogue. All cost estimates have been inflation-adjusted to 2008 Malaysian Ringgit (MYR) and Thai Baht (THB), and final values have been reported in US dollars, at the exchange rates for 2008 (exchange rates 3.22 MYR/US$ and 33.6 THB/US$). Illness costs Illness costs are determined by three inputs: (i) a “RUHA matrix” (defined below); (ii) the reported number of cases, and (iii) direct costs (e.g. medical treatment, drugs, and transportation) and indirect costs (e.g. income and value of time lost by patients/ household members) of hospitalized as well as ambulatory cases. The RUHA matrix shows the shares of reported (R) and unreported (U) hospitalized (H) and ambulatory (A) cases. Malaysia’s immediate cost of dengue is substantial and is equivalent to 3%–7% of the government’s spending on health care. According to our estimates the illness costs due to dengue are 11 times (range 5 to 28 times) the amount of government spending on Aedes vector control in Malaysia, and 13 times (range 1 to 106 times) the government’s spending on Aedes vector control in Thailand. This relationship shows that increased investment on prevention could potentially generate large offsets in illness costs. In addition to the immediate costs reported here, dengue may also adversely impact tourism and create emotional and long-term burdens on families affected by illness and deaths. Keywords: Burden; dengue; immediate costs; Malaysia; Monte Carlo; RUHA matrix; Thailand. Dengue Bulletin – Volume 34, 2010 67 Immediate cost of dengue to Malaysia and Thailand: An estimate We have analysed these proportions using dimensionless analysis (or ratio analysis). We have used the ‘RUHA matrix’ approach, similar to Mavalankar et al.,[1] as a useful tool for combining cost studies, filling gaps in data and making comparisons across countries. Monte Carlo sensitivity analysis A Monte Carlo sensitivity analysis was carried out (@Risk software version 5.0.1, Palisade Corporation, USA) to determine how uncertainties in cost parameters affect the total cost of illness (similar to Rafael et al.[5] and Mavalankar et al.[1]). Fifty-five and 50 simulations were performed (five per input) for Malaysia and Thailand respectively, each with 10 000 iterations, and the mean of the total cost was used as a tracking variable in the simulations. For each iteration, values for the parameters were independently drawn from the Beta-PERT distribution.[6] The Beta-PERT distribution is defined by minimum, most likely and maximum values which are relatively uncomplicated to estimate; it is also considered preferable to the triangular distribution (for instance, Rafael et al.[5]) because the smooth shape of the curve places less emphasis in the direction of any possible skew. Results RUHA matrix for Malaysia and Thailand Table 1, panel 1a presents an initial RUHA matrix for Malaysia based on a preliminary study in a public hospital by Shepard et al.[7] Comparing patterns from neighbouring Thailand (Table 2), in which a careful population-based study was reported,[8,9] Table 1, panel 1a appears to have underestimated the UA cases because it did not account for cases treated solely in private clinics and private hospitals. The Ministry of Health, Malaysia[10] has estimated that the public sector bears 39.5% of the ambulatory care costs; so we have refined the RUHA matrix by adding 69 private sector ambulatory cases (all assumed unreported) to the 45 public sector ambulatory cases studied by Shepard et al.[7] This refined RUHA matrix for Malaysia (Table 1, panel 1b) is more similar to the one constructed for Thailand (Table 2). Reported cases The number of reported cases for Malaysia and Thailand for recent years are shown in Table 3. The number of reported cases in Malaysia during 2002–2007 varied from 31 545 to 50 341 (mean 37 793) per annum, and these were assumed to be mostly hospitalized cases in public sector facilities.[7] Unreported cases have been assumed to vary from year to year in proportion to reported cases. In Thailand, the number of reported cases during 2000–2005 varied substantially – from 18 617 to 139 327 (mean 68 028) per annum. Costs per case Both hospitalized and ambulatory cases have direct and indirect cost components, and their estimates from public sector hospitals in Malaysia[12] are shown in Table 4. As these authors have estimated the ambulatory cost from reported cases, the reporting rate of ambulatory cases has been modelled to have a medium to large correlation (following Cohen 1988[15]) with the average cost of an ambulatory case. This assumption would result in a conservative estimate of the national cost. For Thailand, the maximum, minimum and most likely values of hospitalized and ambulatory costs per case have been taken from previous reports.[9,12,16] 68 Dengue Bulletin – Volume 34, 2010 Immediate cost of dengue to Malaysia and Thailand: An estimate Table 1: RUHA matrices for Malaysia Reported Unreported Sum 1a: Preliminary (not accounting for UA cases in private facilities) Hospitalized 43% 21% 64% Ambulatory 5% 31% 36% Sum 48% 52% 100% 1b: Refined (accounting for UA cases in private facilities)a Hospitalized 28% 13% 41%b Ambulatory 3% 56% 59% Sum 31%c 69% 100% Notes to Table 1: In Table 1a, 100% = 124 cases; bold values from studies; values not in bold are calculated; Source: Shepard et al.[7] In Table 1b, 100% = 193 cases; bold values from studies; values not in bold are calculated. aWe assumed the public sector’s share for ambulatory care for subsequently hospitalized patients of 39%[7] would apply to all ambulatory cases. Recognizing the uncertainty in our assumption, our simulations varied this parameter over the range of 30% to 50%. As home-care patients (who do not visit any health-care facilities) were not included, this RUHA matrix is a conservative estimate. bBased on Anderson et al.,[8] the 25% hospitalization rate has been taken to be the most likely estimate for the share of dengue patients who are hospitalized in Thailand. Their study included home-care patients, therefore 25% was also taken to be the most likely estimate for Malaysia. 41% (obtained in the revised RUHA matrix) has been taken as the maximum estimate of the share of dengue patients hospitalized in Malaysia. The data from Malaysia reported by Lum et al.,[11] Suaya et al.,[12] and Shepard et al.[7] are consistent with this assumption. The corresponding hospitalization rate of 20% was used by Suaya et al.[13] for Cambodia, which was taken as the minimum for Malaysia and Thailand. cThe range of 7%–20% was considered by Clark et al.[9] in their study of Thailand, so this range has been used for that country, with 10% as the most likely value. The values of 10%–31% have been used as the range for Malaysia with 20% as the most likely estimate. Table 2: RUHA matrix for Thailanda Reported Unreported Sum Hospitalized 9%b 16% 25%cd Ambulatory 1% 74% 75% Sum 10%e 90% 100% Notes to Table 2: Bold values from studies; values not in bold are calculated; Sourcesa: Anderson et al.[8] and Clark et al.[9]. These two sources have been assumed to be consistent with each other. Garg et al.[14] have also combined percentages from these two studies. bRatio of hospitalized cases to reported cases has been taken as 0.9 based on preliminary data from Malaysia summarized by Shepard et al[7]: out of 60 reported cases 54 were hospitalized. cSource: Anderson et al.[8] dA comparable hospitalization rate of 20% was used by Suaya et al.[12] for Cambodia. eSource: Clark et al.[9] report a range 7%–20%. Dengue Bulletin – Volume 34, 2010 69 Immediate cost of dengue to Malaysia and Thailand: An estimate Aedes control costs Table 5 shows the cost parameters that contribute to vector (Aedes) control costs. A large correlation (following Cohen[15]) was used between government expenditure on Aedes control (parameters in section 1.1 in Table 5) and reported annual dengue cases. Market size data did not indicate any obvious dependence between household insecticide spending and the number of reported dengue cases; hence, no correlation was used to relate these. The total market size of household insecticides has been reported by the Malaysian CropLife & Public Health Association (MCPA), and we have considered 10% of these household insecticides to be used for dengue and chikungunya control. Table 4: Direct and indirect costs per case for hospitalized and ambulatory cases (US$) Cost component Malaysia a (most likely) Malaysia (range) Thailandb (most likely) Thailandc (range) Hospitalized Direct Indirect 1259 1101 158 1101–1259 Present in all cases 0% – 100%d 208 118–794 Ambulatory Direct Indirect 174 283 139 139–422 0%–100%e Present in all cases 60.5 22.4–341.2 Notes to Table 4: aThe University Malaya Medical Centre (UMMC), collaborating with two Ministry of Health hospitals (in Ampang and Klang), found that the number of investigations and management of patients were more or less similar in all locations studied (Lucy Lum, unpublished data, 2007). The length of hospital stay was also comparable (in fact slightly shorter in UMMC), therefore, the costs per case in Table 3 would result in a conservative estimate when applied to the entire public and private sector health-care network. Our study also assumes that not all of the direct and indirect costs would apply to all the cases, leading to a conservative estimate of the cost of dengue. bThe most likely estimates of hospitalized and ambulatory costs in Thailand have been taken from Okanurak et al.[17] cMinimum costs per case in Thailand have been calculated as the sum of private costs from Clark et al.[9] and public costs from Okanurak et al.[17] Maximum cost per hospitalised case in Thailand has been taken from Suaya et al.[12] Maximum cost per ambulatory case has been calculated based on the results of Suaya et al.[12] using the ratio of hospitalized cost to ambulatory cost. These authors have recently published their work.[18] dThis approach is similar to Anderson et al.[8] in assuming that indirect costs vary. 100% is taken as the most likely value in our simulation. eThe variation is assumed to be dependent on the reporting rate of ambulatory cases. For the most likely value used in our simulation, home-care patients are assumed to incur 0%, while others incur 100% of direct cost. Table 3: Reported dengue cases obtained from Malaysia and Thailand Year Malaysia Thailand 2000 NC 18 617 2001 NC 139 327 2002 32 767 114 800 2003 31 545 62 767 2004 33 895 38 367 2005 39 654 34 291 2006 38 556 NC 2007 50 341 NC NC: Not considered for the study as data were not available. 70 Dengue Bulletin – Volume 34, 2010 Immediate cost of dengue to Malaysia and Thailand: An estimate The Malaysian household insecticides market size has been adjusted for relative GDP and population in order to estimate the total size of the household insecticides market in Thailand. In order to be conservative, only 7.5% of this market has been considered to be relevant for dengue. Discussion Immediate cost of dengue to Malaysia and Thailand Using Monte Carlo analysis, we have estimated the mean economic cost of dengue to Malaysia as US$ 133 million (range US$ 88 million to US$ 215 million) per annum. Figure 1 shows the variation of total cost resulting from each parameter varying from minimum to maximum. The most important parameters accounting for the variation of the total cost are the reporting rate, the annual variation in reported cases, the hospitalization rate, and cost per ambulatory case. The bulk of the cost (78%) comes from illness; hence a key factor is the low reporting rate for dengue, especially for ambulatory cases. For Thailand, the estimated mean economic cost of dengue is US$ 135 million (range US$ 56 million to US$ 264 million) per annum (Figure 2). In this case, the most important parameters accounting for the variation of the total cost are cost per ambulatory case, cost per hospitalized case, and reporting rate. The first two factors have become significant due to large differences in the estimates reported by Okanurak et al.,[17] Table 5: Cost parameters determining the cost of Aedes control and R&D expenditurea All values in US$ million Malaysia Thailand Cost parameter Most likely Range Most likely Range 1. Aedes control 1.1 1.2 1.3 1.4 Inspections, enforcement and fumigation Household insecticides (aerosols, mats, coils, liquids, baits, etc.) Private sector vector control Government’s ComBI efforts 8.96a 7.20 0.43 8.85 6.27–11.15 6.49–16.16 0.29–0.57 5.59–12.42 8.50 6.72 0.41 11.01b 2.50–19.02 4.06–10.04 0.20–0.61 6.95–15.45b Sub-total 25.44 18.64–40.30 26.64 13.71–45.12 2. R&D 2.1 2.2 On Aedes control Other 0.52 0.12 0.31–0.72 0.03–0.21 Sub-total 0.64 0.34–0.93 0.61c 0.30–1.21c Notes to Table 5: aSuaya and Shepard.[12] ComBI denotes communication-for-behavioural-impact bEstimated by multiplying the Malaysian value by 2.31 (for relative size of the population) and by 0.54 (for relative size of the GDP per capita). cIn Malaysia, R&D costs were found to be 7% of item 1.1 (for the most likely value). We have assumed the same for Thailand, and taken its range to vary by 50% and 100% on either side. Dengue Bulletin – Volume 34, 2010 71 Immediate cost of dengue to Malaysia and Thailand: An estimate Clark et al.[9] and Suaya et al.[12] – indicating the need for further studies. Estimates of cost per ambulatory case and reporting rate need to be improved in both countries. In addition, we need better estimates of the hospitalization rate in Malaysia and the cost per hospitalized case in Thailand. Both Figure 1 and Figure 2 show the variation in the mean economic cost of dengue due to “annual variation in the number of reported cases”. This shows that policy-makers should take a longer term view of dengue when allocating funds as the number of cases fluctuate from year to year, especially in Thailand. Comparing the RUHA matrices for Malaysia (Tables 1) with Thailand (Table 2), we recommend that future studies address the uncertainty in the number of unreported cases, especially unreported ambulatory cases, and that studies similar to Shepard et al.,[7] Suaya et al.,[12] and Kongsin et al.[16] should be Figure 1: Sensitivity analysis of the annual cost of dengue to Malaysia (Bars show range from possible variation in each factor) 72 Dengue Bulletin – Volume 34, 2010 Immediate cost of dengue to Malaysia and Thailand: An estimate carried out to assess the variation in costs per case across the healthcare network. Another article in this issue of the Dengue Bulletin also develops an estimate of the national cost of dengue in Thailand.[16] That estimate, US$ 158 million with a standard deviation of US$ 33 million, is close to the central estimate for Thailand in this article. While both articles rely on some of the same primary data (the initial observations collected by Kongsin et al.[16]), the merger of these data with other information and assumptions was done independently in the two articles. Despite the uncertainties, this agreement suggests that the present findings are plausible. Comparison between countries The mean number of reported cases per annum in Thailand (68 028 cases for 2000– 2005) has been 1.8 times that of Malaysia (37 793 cases for 2002–2007). The reporting rates in Thailand (10%) and Malaysia (20%), given in the notes to Table 1, differ by a factor of 1 to 2. Thus, Thailand has 3.6 times as many total cases, The average cost per case in Malaysia (US$ 445.25) is higher than that of Thailand (US$ 97.38), as shown in Table 4. so Malaysia has a 4.6 times the cost per case. These two offsetting factors are the reason why the total immediate cost of dengue is comparable between these countries, with Figure 2: Sensitivity analysis of the annual cost of dengue to Thailand (Bars show range from possible variation in each factor) Dengue Bulletin – Volume 34, 2010 73 Immediate cost of dengue to Malaysia and Thailand: An estimate the bulk of immediate costs in both countries due to illness – 78% in Malaysia and 76% in Thailand. The mean annual cost of dengue per capita is US$ 5.3 for Malaysia and US$ 2.0 for Thailand. This is not unexpected because the GDP per capita of Malaysia is 1.9 times that of Thailand. The immediate cost of dengue (US$ 88 million to US$ 215 million per annum) to Malaysia is substantial and amounts to US$ 3.5 to US$ 8.5 per capita and is equivalent to 3%–7% of government spending on health- care.[19] The mean cost of US$ 5.3 per capita for Malaysia compares closely to the US$ 6.2 estimated by Armien et al.[20] for Panama – an upper-middle income country like Malaysia with comparable GDP per capita (US$ 6717 for Panama, US$ 7605 for Malaysia), with similar costs per case (US$ 1256 and US$ 391 respectively for hospitalized and ambulatory cases in Panama and US$ 1259 and US$ 422 respectively for hospitalized and ambulatory cases in Malaysia), and comparable number of reported annual cases per thousand people (1.7 for Panama, 1.5 for Malaysia) for the study periods. It should be borne in mind that this paper and Armien et al.[20] are not strictly comparable because Armien et al. did not consider several parameters such as annual variation in reported cases, variation in hospitalization rates, and spending on household insecticides. However, they did consider the economic loss due to fatal dengue which is beyond the scope of this paper. We can also compare our estimates for Malaysia with Brazil – another upper-middle income country with comparable GDP per capita (US$ 8450), and a slightly higher number of reported annual cases (2.2 during 2002–2007) per thousand people.[21] Brazil’s 2002 federal dengue prevention budget was reportedly BRL 1034 million,[22] and the estimate for 2007 is at least BRL 1527 millions. [21] These figures translate to an annual federal spending of US$ 4.3 per capita for dengue prevention in Brazil during 2002–2007. Separate state and municipal budgets for dengue prevention and illness costs are not included in this Brazilian estimate. In Malaysia, illness costs due to dengue are typically 11 times (range 5 to 28 times) the government spending on Aedes vector control (see Table 5 and Figure 1). Similarly in Thailand, illness costs due to dengue are typically 13 times (range 1 to 106 times) the government spending on Aedes vector control (see Table 5 and Figure 2). This result shows that increased investments in prevention could potentially generate large offsets in illness costs. Additional considerations There are three additional considerations worth mentioning, but are beyond the scope of the immediate costs. First, this study has not considered chikungunya, another Aedes- borne viral infection, which has affected Malaysia with increasingly bigger outbreaks in recent years: 27 cases in 1999, 227 cases in 2006, and 4271 cases in 2008.[23,24] Thailand experienced a severe ongoing outbreak with 20 541 reported cases from 23 provinces for the period 1 January 2009 to 20 May 2009. [25] The potential threat of chikungunya can be clearly seen from the experience of Indian Ocean countries such as India[1,26] which had around 1.5 million suspected cases during 2006–2008, and Réunion Island[27] where one third of the population was affected during the 2005-2006 outbreak. Although dengue and chikungunya are virologically distinct, cases are indeed comparable from disease prevention and management perspectives, therefore policy-makers should consider their socioeconomic impact together. 74 Dengue Bulletin – Volume 34, 2010 Immediate cost of dengue to Malaysia and Thailand: An estimate References [1] Mavalankar D, Puwar T, Govil D, Murtola T, Vasan S. A preliminary estimate of immediate cost of chikungunya and dengue to Gujarat, India. W.P.No. 2009-01-01. Ahmedabad, India. Indian Institute of Management Working Paper Series 2009. Available at: http://www. iimahd.ernet.in/publications/data/2009-01- 01Mavalankar.pdf. Accessed 30 July 2009. [2] Mavalankar D, Puwar T, Murtola T, Vasan S. Quantifying the impact of chikungunya and dengue on tourism revenues. W.P.No. 2009- 02-03. Ahmedabad, India: Indian Institute of Management Working Paper Series. 2009. Available at: http://www.iimahd.ernet.in/ publications/data/2009-02-03Mavalankar.pdf. Accessed 30 July 2009. [3] Cummings DAT, Irizarry RA, Huang NE, Endy TP, Nisalak A, Ungchusak K, Burke DS. Travelling waves in the occurrence of dengue haemorrhagic fever in Thailand. Nature, 2004, Jan 22;427(6972):344-347. Secondly, it has been shown elsewhere[2] that a 4% decline in tourists from non-endemic countries would result in a substantial loss of tourism revenue: US$ 65 million for Malaysia and US$ 363 million for Thailand. While all these potential revenue losses do not constitute economic costs, it is nevertheless striking that the impact on tourism is comparable to the total cost of illness estimated by following the traditional approach, limited strictly to economic costs. In other words, the impact on tourism revenues should not be ignored when calculating the burden of infectious diseases. Finally, we note that the emotional and long-term burden of illness and deaths due to dengue are beyond the scope of this study, therefore we recommend reading this working paper in conjunction with studies that report impact on quality of life,[11] lost DALYs (disability adjusted life years), long- term effects of illness on education and economic growth,[28,29] per capita income,[30,31] foreign direct investment,[32,33] etc. Such long-term effects are likely to be important for chikungunya, as some cases experience persistent joint pain for several months, even years.[1,34] We hope that our estimate of the immediate cost of dengue will trigger more refined studies on cost of illness, as well as analyses of the cost-effectiveness of vaccines and other interventions to combat these neglected tropical diseases. Acknowledgements Funded in part by a grant to the Regents of the University of California from the Foundation for the National Institutes of Health through the Grand Challenges in Global Health initiative. The authors are grateful to the Vector-Borne Disease Control Section (Ministry of Health, Malaysia) and the Malaysian CropLife and Public Health Association for providing us with data. The authors also thank L.S. Alphey (University of Oxford), J.P. Leao (Oxitec Limited), R.S. Lees (University of Malaya), A. Mandaltsi (University of Oxford), N. de Mel (University of Oxford), P. Vasan (University of Reading), and C.L. Hurley (Brandeis University) for their comments. Dengue Bulletin – Volume 34, 2010 75 Immediate cost of dengue to Malaysia and Thailand: An estimate [4] Haddix A, Corso P, Gorsky R. Costs (chapter 4, pp 53-78). In: Haddix A, Teutsch S, Corso P, editors. Prevention Effectiveness: A Guide to Decision Analysis and Economic Evaluation (ISBN 0-1951-4897-5, 2nd ed). New York: Oxford University Press, 2003. [5] Rafael M, Taylor T, Magill A, Lim Y, Girosi F, Allan R. Reducing the burden of childhood malaria in Africa: the role of improved. Nature, 2006, 23: 39-48. Available at: http://www. nature.com/nature/journal/v444/n1s/full/ nature05445.html. Accessed 30 July 2009. [6] Vose D. Risk Analysis - A Quantitative Guide. Chichester: John Wiley, 2000. [7] Shepard DS, Suaya J, Caram M, Lum L, Kongsin S. Dengue Cost of Illness Studies: Status Report. Seoul, Korea: Semi-Annual Meeting of Board of Counselors, 2006. Available at: http://www.pdvi.org/PDFs/BoC-Burden-4-17- 06f.ppt. Accessed 25 July 2008. [8] Anderson KB, Chunsuttiwat S, Nisalak A, Mammen MP, Libraty DH, Rothman AL, Green S, Vaughn DW, Ennis FA, Endy TP. Burden of symptomatic dengue infection in children at primary school in Thailand: a prospective study. Lancet, 2007,369:1452-1459. [9] Clark DV, Mammen Jr MP, Nisalak A, Puthimethee V, Endy TP. Economic impact of dengue fever/dengue hemorrhagic fever in Thailand at the family and population levels. American Journal of Tropical Medicine and Hygiene, 2005,72(6):786-791. [10] Ministry of Health Malaysia. Malaysia National Health Accounts Project. Health Expenditure Report 1997-2002. Malaysia: Putrajaya, Ministry of Health, 2006. [11] Lum LCS, Suaya JA, Tan LH, Sah BK, Shepard DS. Quality of life of dengue patients. American Journal of Tropical Medicine and Hygiene, 2008, Jun;78(6):862-867. [12] Suaya J, Shepard D, Armien B, Caram M, Castillo L, Chantha N, Garrido F, Kongsin S, Lum L, Montoya R, Sah BK, Siqueira JB, Sughayyar R, Tyo, K. Multi-country study of costs of dengue among ambulatory and hospitalized patients. Presented at the American Society of Tropical Medicine and Health 56th Annual Meeting, Philadelphia, PA, Nov. 5, 2007, Abstract 28, p. 9. [13] Suaya JA, Shepard DS, Chang M-S, Caram M, Hoyer S, Socheat D, Chantha N, Nathan MB. Cost-effectiveness of annual targeted larviciding campaigns in Cambodia against the dengue vector Aedes aegypti. Tropical Medicine and International Health, 2007,12(9):1026– 1036. [14] Garg P, Nagpal J, Kairnar P, Sinveratne SL. Economic burden of dengue infections in India. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2008,102(6):570- 577. [15] Cohen J. Statistical power analysis for the behavioral sciences (2nd edition). Hillsdale, NJ: Hove: Erbaum Associates, 1988. [16] Kongsin S, Jaimton S, Suaya J, Vasanawathana S, Sirisuvan P, Shepard DS. Cost of dengue in Thailand. Dengue Bulletin, 2010, 34:77-88. [17] Okanurak K, Sornmani S, Indaratna K. The cost of dengue hemorrhagic fever in Thailand. Southeast Asian Journal of Tropical Medicine and Public Health, 1997, 28(4):711-717. [18] Suaya JA, Shepard DS, Siqueira JB, Martelli CT, Lum LCS, Tan LH, Kongsin S, Jiamton S, Garrido F, Montoya R, Armien B, Huy R, Castillo L, Caram M, Sah BK, Sughayyar R, Tyo KR, Halstead SB. Cost of dengue cases in eight countries in Americas and Asia, a prospective study. American Journal of Tropical Medicine and Hygiene, 2009, 80(5):846-855. [19] World Health Organization (WHO). Core Health Indicators. WHO Statistical Information System (WHOSIS), 2006. Available at: http://www.who.int/whosis/database/core/ core_select_process.cfm. Accessed 18 August 2008. 76 Dengue Bulletin – Volume 34, 2010 Immediate cost of dengue to Malaysia and Thailand: An estimate [20] Armien B, Suaya JA, Quiroz E, Sah B, Bayard V, Marchena L, Campos C, Shepard DS. Clinical characteristics and national economic cost of the 2005 dengue epidemic in Panama. American Journal of Tropical Medicine and Hygiene, 2008, 79(3):364-371. [21] Ministério da Saúde. Dengue. 2008. Available at: http://www.senado.gov.br/web/ comissoes/cas/ap/AP20080401_MinSaude_ FabianoPimenta.pdf. Accessed 15 August 2008. [22] Programa Nacional de Control da Dengue. Instituído em 24 julho de 2002. Ministério da Saúde, Brasil. Available at: http://portal. saude.gov.br/portal/arquivos/ pdf/pncd_2002. pdf. Accessed 15 August 2008. [23] Liow T. Liow orders action on chikungunya. 2008. Available at: http://www.nst.com. my/Current_News/NST/Monday/National/ 2298988/Article/index_html. Accessed 1 August 2008. [24] ProMED-mail post. Chikungunya (25): Malaysia, India (GOA). 2009. Available at: http://www.promedmail.org/pls/otn/f?p=2 400:1202:2241869329467759::NO::F24 00_P1202_CHECK_DISPLAY,F2400_P1202_ PUB_MAIL_ID:X,78212. Accessed 27 July 2009. [25] ProMED-mail post. Chikungunya (17): Thailand. 2009. Available at: http://www. promedmail.org/pls/otn/f?p=2400:1202:2 241869329467759::NO::F2400_P1202_ CHECK_DISPLAY,F2400_P1202_PUB_MAIL_ ID:X,77691. Accessed 27 July 2009. [26] Mavalankar D, Shastri P, Bandyopadhyay T, Parmar J, Ramani K. Increased mortality rate associated with chikungunya epidemic, Ahmedabad, India. 2008: Emerging Infectious Diseases [serial on the Internet]. Available at: http://www.cdc.gov/EID/content/14/3/412. htm. Accessed 30 July 2009. [27] Reiter P, Fontenille D, Paupy C. Aedes albopictus as an epidemic vector of chikungunya virus: another emerging problem? Lancet Infectious Diseases, 2006, 6(8):463-464. [28] Bloom DE, Canning D. Epidemics and Economics . Program on the Global Demography of Aging, Working Paper No. 9. Harvard Initiative for global Health, 2006. Available at: http://www.hsph.harvard.edu/ pgda/Working%20Papers/2006/BLOOM_ CANNINGWP9.2006.pdf. Accessed 30 July 2009. [29] Bloom DE, Canning D, Sevilla J. The effect of health on economic growth: A production function approach. World Development, 2004,32(1):1-13. [30] Barro R, Sala-I-Martin X. Economic growth. New York: McGraw-Hill, 1995. [31] Bhagava A, Jamison DT, Lau LJ, Murray CJL. Modeling the effects of health on economic growth. Journal of Health Economics, 2001, 20(3):423-440. [32] Alsan M, Bloom DE, Canning D. The effect of population health on foreign direct investment inflows to low- and middle-income countries. World Development, 2006, 34(4):613-630. [33] Jones T. The Panama Canal: A Brief History. 1990. Available at: http://www.ilovelanguages. com/tyler/nonfiction/pan2.html. Accessed 30 July 2009. [34] World Heal th Organizat ion (WHO). Chikungunya Fact sheet No. 327. 2008. Available at: http://www.who.int/mediacentre/ factsheets/fs327/en/print.html. Accessed 30 July 2009. Dengue Bulletin – Volume 34, 2010 77 Cost of dengue in Thailand# Sukhontha Kongsina, Sukhum Jiamtonb, Jose A. Suayac, Sirijitt Vasanawathanad, Petcherut Sirisuvana and Donald S. Shepardc* aResearch Centre for Health Economics and Evaluation, Faculty of Public Health, Mahidol University, Bangkok, Thailand bFaculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand cSchneider Institutes for Health Policy, Heller School, Brandeis University, Waltham, MA, USA dMedical Staff of Khon Kaen Hospital, Ministry of Public Health, Thailand Abstract The burden of hospitalized dengue cases in Thailand is high. In Khon Kaen Hospital, a provincial hospital, the cost per non-fatal case in 2005 averaged (±standard deviation) US$ 573 (±351). The majority of this cost is incurred by the government, indicating the importance of dengue to public policy. The projected national cost is US$ 158 (±33) million or US$ 3.55 (±$0.53) per capita per year. In international dollars, which adjust for purchasing power parity, the annual cost is 485 (±106) million international dollars. Of this total, 28% is for vector control while 72% represents the cost of dengue illness. The substantial cost of dengue illness indicates the potential economic value of effective control measures. Keywords: Dengue; economic burden; costing; Thailand. #Supported by the authors’ institutions and the Paediatric Dengue Vaccine Initiative *E-mail: shepard@brandeis.edu, Tel: +1-781-736-3975; Fax: +1-888-428-2672 Introduction The first epidemic of dengue in Thailand occurred in 1958.[1] Dengue is regarded as Thailand’s most important arbovirus infection of the 21st century,[2] and is the country’s largest single cause of paediatric hospitalizations.[3] Khon Kaen province, the site of the present study, has been particularly affected by dengue in recent epidemics.[4] Clinically, a dengue infection is usually classified as dengue fever, dengue haemorrhagic fever (DHF) or dengue shock syndrome (DSS) according to severity. Dengue illness entails considerable cost for medical treatment; school-age children also lose school days and working parents and caregivers lose income while caring for patients. Caregivers also incur expenses related to food, transportation and lodging for those visiting and caring for hospitalized patients. In addition, central, 78 Dengue Bulletin – Volume 34, 2010 Cost of dengue in Thailand regional and local governments incur costs for vector control activities, risk of dengue may reduce revenue from tourism, and dengue is sometimes fatal. Previous studies have examined several components of the annual economic burden due to DHF in Thailand, but they excluded items such as the value of lost school days, government subsidies to public hospitals, and costs of vector control.[5-7] A 1995 report estimated that the annual economic burden due to DHF in Thailand ranged from US$ 19 million to US$ 51 million per year.[7] The number of disability-adjusted life years (DALYs) lost to dengue in Thailand was 45 214 (731.7 per million population).[6,8] Anderson et al. estimated out-of-pocket costs to patients for treatment, but did not estimate other components.[5] In 2009, Suaya et al. published a comprehensive study of dengue illness costs in eight countries, including Thailand.[9] Nevertheless, many gaps remain. Cost-related studies are needed by vaccine developers, donors, the Ministry of Health and local governments in Thailand, NGOs and other stakeholders in developing policies around dengue prevention and treatment. Prevention decisions include public support and awareness around vector control programmes, commitments to recommend and use a dengue vaccine should it prove effective and be licensed, and funding for dengue treatment through national insurance systems and other mechanisms. This study is the first to conduct a comprehensive examination of costs of dengue to Thai society, including medical care costs for care at a public hospital, indirect costs to households for lost income and school absence, and for vector control. Materials and methods Study design This study on disease burden and cost of hospitalized dengue cases was conducted in Khon Kaen Hospital, Khon Kaen, Thailand, from October 2004 to June 2006. The study population consisted of hospitalized patients aged 2 to 15 years, who had a final clinical diagnosis of dengue at discharge. Patient-level data came from a patient questionnaire and abstracts of patients’ medical records (e.g., clinical characteristics, length of hospital stay and laboratory data). We did not influence treatment in any way nor order any laboratory tests specifically for this study. All data were entered into two identical standardized Microsoft Access databases created by one of the authors (SJ) for the multicountry study[9]–one for interview data and the other for medical record information. Study methods and databases are described further elsewhere.[9] Data collected from interviews The patient questionnaire was developed in English, translated into Thai, and then piloted. The questionnaire covered demographic and socioeconomic characteristics of the patient and households, characteristics of the illness, impacts on health, use of medical services, absence from school and work, time for patient care, and spending and income lost by households while caring for the dengue patient. A well-trained interviewer conducted two half-hour interviews, one at the time of admission using face-to-face communication and the second at the time of follow-up approximately two weeks later, usually by Dengue Bulletin – Volume 34, 2010 79 Cost of dengue in Thailand telephone. The majority of the interviewees were caregivers instead of patients, but a few patients spoke directly about their symptoms in the presence of the legal surrogate. Three workshops were conducted in Thailand (two also included investigators from Malaysia and Cambodia) to ensure consistency and accuracy in data collection, cleaning and analysis. When data were found to be inconsistent, data records were reviewed and imputations were made after a consistency check. Cost of a dengue case To derive the economic cost of each case in the study cohort, we measured the direct medical costs borne by the government, clients and households, and direct non-medical costs and indirect costs borne exclusively by households. Costs were expressed both in United States dollars (US$) based on the official 2005 exchange rate of 39.0 Thai Baht per US dollar, and international dollars (I$) based on purchasing power parity. The figure for I$ was 3.07 times the number of US$.[9] To quantify the societal cost of absence from school, we made the conservative assumption that it was equal to the cost per day of primary schooling, which was I$ 5.80 (US$ 1.90). This assumption was based on the fact that schooling is publicly funded, and hence society values it at least as highly as its cost. We then calculated the economic loss attributed to school days lost as the product of the daily cost and the number of school days lost. We valued a day of work lost to the worker or to the employer as the higher of the reported daily loss or Thailand’s minimum daily wage (I$ 11.00 or US$ 3.60) and then calculated the total economic costs of workdays lost as the product of this average daily loss times and the number of workdays lost. To value “other” days (caregiver and patient days lost other than for school or work) we also used Thailand’s daily minimum wage for patients or household members 15 years of age or above. Household total days affected are the sum of school, work and other days lost. As there were no deaths in our cohorts, the economic costs of premature deaths were incorporated subsequently under aggregate national projected costs. Macrocosting was the primary approach to estimate unit costs of inpatient and ambulatory care, with activity-based costing explored for validation. The facility’s annual operating budget, outpatient visits and inpatient visits were extracted from the hospital database for calculating unit costs. The unit cost for an average hospital day was derived from total hospital expenditure, and numbers of inpatients and outpatients during the fiscal year 2005. Using a relationship from international literature, we assumed the unit cost of an outpatient hospital visit was 25% of the cost of an inpatient bed day.[10] As many outpatient visits occurred below the level of a provincial hospital (e.g., at a physician’s office, in a health centre, or a secondary hospital clinic), we assumed the average cost of an ambulatory visit was only 60% a hospital outpatient visit. The average cost of an inpatient bed day was I$ 263.90 (US$ 86.00) and of an ambulatory visit was I$ 39.60 (US$ 12.90). Out-of-pocket payments for medical services were calculated from information on the costs of pre-admission, admission and follow-up periods obtained by interviews. Costs of vector control Vector control programmes consist of educating households about the reduction of breeding sites by eliminating containers that can hold 80 Dengue Bulletin – Volume 34, 2010 Cost of dengue in Thailand multiplying its average annual reported cases by its cost per case. In Thailand, as in most countries, the number of officially reported dengue cases is lower than the true number.[11] A secondary analysis of data from Kampaeng Phet, where schoolchildren are under active surveillance for dengue, found that the estimated number of dengue cases is four times the officially reported number of hospitalized cases. Since 83% of the reported cases are from hospitals, the estimated number of dengue cases is 3.3 times the total number of reported cases. That is, the estimated expansion factor is 3.3. Statistical analysis The unit of analysis was a clinically confirmed hospitalized dengue episode with or without serological confirmation. We compared costs between patients without laboratory confirmation and those with laboratory confirmation. Statistical significance between continuous variables was examined by t-tests (for normally distributed variables) or by the Wilcoxon rank sum test (for skewed data) and the significance between categorical variables was examined by the Fisher’s exact test. Ethical considerations The study protocol was approved by the ethics committees of Khon Kaen Hospital, Faculty of Public Health of Mahidol University, Ministry of Public Health, Royal Government of Thailand, and Institutional Review Boards (IRB) at Brandeis University, and the initial funding organization, the International Vaccine Institute, on behalf of the Pediatric Dengue Vaccine Initiative (PDVI). standing water; when possible treating those that cannot be eliminated and spraying insecticide to reduce the abundance of adult mosquitoes. These activities are implemented at all levels of the administration from the national to sub-district. To ascertain their costs, we obtained data on costs (both personnel and other costs) at the national levels, and from representative examples of lower-level administrative units from the province to the sub-district. At each administrative unit, we divided by the population served to get per capita costs. At the local levels, we computed a weighted average of rural and urban costs. We then summed the levels to get the total per capita costs. National costs of dengue To project national dengue costs, we used the preliminary economic model and the Monte Carlo simulation from the multicountry study.[9] Since the number of dengue cases varies substantially by year, we used the five-year (2001–2005) average of dengue cases and deaths that was officially reported to WHO, with supplementary data to show the breakdown by setting. As Thailand’s surveillance system is based primarily on hospitals, the share of cases (± standard error of the mean, SEM) from hospitals from 2002–2005 averaged 83.0% (± 1.7%). The average age of death as a result of dengue in Thailand was 7.6 years (± 2.0). As a provincial hospital, Khon Kaen Hospital was organizationally located midway in a health- care system with national hospitals at the tier above and district hospitals and health centres at the tier below, and we assumed that the cost per case was nationally representative. We estimated Thailand’s aggregate cost by Dengue Bulletin – Volume 34, 2010 81 Cost of dengue in Thailand remaining 131 patients, who had only clinical diagnoses for dengue. Except for the facts that laboratory-confirmed children were more likely to come from urban areas and were slightly older, there was no significant difference in demographic characteristics between the serologically confirmed and the clinically diagnosed groups. Health status before illness in almost all patients in both groups was good or very good. All patients experienced fever for an average number of six days and had 11-day duration of illness. Half of the patients from Results Demographic and clinical characteristics As shown in Table 1, our subjects consisted of 173 patients classified into two cohorts. The laboratory-confirmed cohort consisted of 42 patients who were serologically confirmed for dengue by polymerase chain reaction (PCR). These laboratory confirmations were done under another clinical research investigation simultaneously underway in the same hospital. The clinical-only cohort consisted of the Table 1: Demographic characteristics of the study cohorts Characteristic Laboratory-confirmed Clinical only Entire cohort p-value Number of patients 42 131 173 n.a. Gender (% female) 55 43 46 0.174 Males 55 43 46 0.174 Age distribution (%) 0–4 yrs. 5–9 yrs. 10–14 yrs.* Mean age, yrs. (SD) 0 33 67 10.0(2.4) 5 51 44 9.0(3.0) 4 47 49 9.2(2.9) 0.098 0.048 Parents' highest level of education (%) Primary school or less Secondary school Vocational school or more 48 31 21 49 31 20 49 31 20 0.975 Residence (% urban) 48 24 30 0.004 Interview completion (%) Both interviews completed Mother as patient's proxy 95 88 91 73 92 77 0.363 0.048 Interview timing, days, mean (SD) Onset to first interview First to second interview 4.4(2.2) 14.7(3.6) 5.4(2.4) 13.7(2.5) 5.2(2.4) 13.9(2.9) 0.017 0.093 * Includes 1 patient aged 15 years; notation: SD denotes standard deviation; n.a. denotes not applicable. 82 Dengue Bulletin – Volume 34, 2010 Cost of dengue in Thailand both groups sought medical care within 24 hours of the onset of symptoms, suggesting that the symptoms were rather severe. Almost all the patients received ambulatory care at a public facility prior to admission to Khon Kaen Hospital. Tab le 2 shows that the c l in ica l characteristics of the patients were similar in both cohorts apart from eye or retro-orbital pain, low platelet count and low white cells count, which were significantly more frequent in serologically confirmed patients. Thirteen per cent of patients took aspirin before hospitalization, a drug that can increase the risk of bleeding in dengue. Total cost of a dengue case Table 3 shows that there were no significant differences in overall cost between the Table 2: Clinical characteristics of the study cohorts Characteristic Laboratory-confirmed Clinical only Entire cohort p-value Number of patients 42 131 173 n.a. Health status before illness Good or very good 81% 87% 86% 0.330 Seeking health care promptly Within 24 hours from onset 45% 52% 50% 0.452 Symptoms and signs during illness Fever Headache Eye or retro-orbital pain Muscle or joint pain Abdominal pain Vomiting Diarrhea Dizziness Excessive thirst Reduced urine output Sore throat and/or running nose Skin rash Bleeding Plasma leakage Low platelets count Low white cells count High white cells count 100% 88% 50% 79% 76% 81% 60% 74% 60% 33% 36% 71% 95% 76% 74% 93% 12% 100% 82% 26% 69% 79% 77% 43% 76% 52% 46% 40% 56% 87% 63% 56% 56% 13% 100% 84% 32% 71% 79% 78% 47% 75% 54% 43% 39% 60% 89% 66% 61% 65% 13% 1.000 0.387 0.004 0.220 0.660 0.600 0.058 0.818 0.389 0.155 0.645 0.085 0.138 0.125 0.046 0.000 0.856 Quality of life during illness Mean, 0 to 100 (SD) 68.3(14.5) 65.9(16.7) 66.5(16.2) 0.408 Dengue Bulletin – Volume 34, 2010 83 Cost of dengue in Thailand Characteristic Laboratory-confirmed Clinical only Entire cohort p-value Patient impact, days, mean (SD) Fever Feeling bad or very bad Illness Onset to hospitalization 6.1(2.3) 4.2(2.4) 11.8(6.2) 2.2(1.3) 6.0(3.8) 4.2(2.5) 10.5(4.5) 2.9(1.5) 6.0(3.5) 4.2(2.5) 10.8(5.0) 2.7(1.5) 0.947 0.971 0.199 0.017 Utilization of health services, mean(SD) Ambulatory visits % in public sector Hospital stay (days) % in public sector 3.6(1.8) 86 5.5(4.1) 100 4.4(2.1) 87 4.6(3.0) 100 4.2(2.0) 87 4.9(3.3) 100 0.028 0.123 Treatment, % of cohort Self-medication Aspirin Antibiotics 2 12 74 4 14 73 3 13 73 1.000 0.760 0.946 Household impact, days, mean (SD) School lost Work lost Period affected 5.9(3.9) 4.5(5.9) 20.9(13.0) 5.4(3.2) 3.7(4.7) 17.9(9.4) 5.5(3.4) 3.9(5.0) 18.6(10.4) 0.340 0.324 0.108 Notation: SD denotes standard deviation; n.a. denotes not applicable. Table 3: Cost of a dengue episode in Khon Kaen Hospital by laboratory confirmation status (I$) Cost component Laboratory-confirmed (n=42) Clinical only (n=131) p-value M (SD) M (SD) Direct costs Ambulatory care Doctor visits, private Doctor visits, public Prescriptions, lab. tests, etc. Inpatient care Non-medical care 1799 129 123 4 1 1464 206 1186 60 61 8 5 1083 146 1564 159 151 5 2 1226 179 928 82 79 12 21 805 123 0.185 0.034 0.129 0.238 Indirect cost 157 120 130 90 0.124 Total cost 1956 1293 1694 998 0.171 Notation: M denotes mean; SD denotes standard deviation; I$ denotes international dollars. 84 Dengue Bulletin – Volume 34, 2010 Cost of dengue in Thailand laboratory-confirmed and clinically diagnosed dengue cases. For the cohort as a whole, Table 4 shows that direct costs were the higher of the two broad cost components, with a value of I$ 1621 or US$ 528 per case. Of these direct costs, inpatient care comprised 79%, ambulatory care 9%, and non-medical care (e.g. transportation) 11%. In this study, indirect costs were only 11% of the total cost of a dengue case (I$ 137, US$ 45). Household out-of-pocket payments (for direct non- medical costs plus the value of time lost from work, school and other household activities) constituted only a small portion of the cost per case. Most of the cost of a dengue case was borne by the government primarily through its subsidy of services provided at Khon Kaen Hospital. The sum of these components (I$ 1758 or US$ 573) was the average cost to society of a dengue case. Table 5 shows the financial challenges faced by families of hospitalized children with dengue. Although patients were hospitalized for an average (±standard deviation) of Table 4: Cost of a dengue episode in Khon Kaen Hospital by component for entire cohort Cost component I$ US $ M (SD) M (SD) Direct costs Ambulatory care Doctor visits, private Doctor visits, public Prescriptions, laboratory tests, etc. Inpatient care Non-medical care 1621 152 145 5 2 1284 186 998 79 75 11 18 883 129 528 49 145 5 2 418 60 325 26 75 11 18 288 42 Indirect cost 137 98 45 32 Total cost 1758 1079 573 351 Notation: M denotes mean; SD denotes standard deviation; I$ denotes international dollars. 4.9 (±3.3) days, the inpatient care itself was provided in a public hospital. Therefore, out- of-pocket payments for inpatient care were generally nominal. Furthermore, medications were generally available in the hospital, so families faced few out-of-pocket payments for that item as well. However, patients reported an average of 4.2 (±2.0) medical visits before and after hospitalization. Forty-seven per cent of patients named payment for these services as a financial challenge. There were no significant differences in the prevalence of financial challenges between laboratory- confirmed and clinically diagnosed dengue cases. National cost of dengue The costs of vector control, shown in Table 6, aggregate to US$ 1 per capita. Most of the costs are incurred at the regional and provincial levels. From a per capita perspective, the national-level Ministry of Health costs are trivial. With a population of 62 million, Dengue Bulletin – Volume 34, 2010 85 Cost of dengue in Thailand Table 5: Challenges faced by households during the patient’s illness episode Challenge Laboratory- confirmed (n=42) Clinical only (n=131) Entire cohort (n=173) p Specific financial challenges, % Could not afford cost of visits to medical providers Could not afford drugs Could not afford cost of transportation Savings (e.g. bank account) Sold items (e.g. furniture, animals, jewelry) Borrowed from family members or friends from outside the household Borrowed from someone other than a friend or family Other (e.g. from the church) 48 7 5 7 2 12 7 0 47 16 7 8 1 18 7 0 47 14 6 8 1 17 7 0 1.00 0.20 1.00 1.00 0.43 0.48 1.00 1.00 Number of financial challenges, mean (standard deviation) 0.88 (1.02) 1.04 (1.15) 1.00 (1.12) 0.20 Table 6: Estimated spending on vector control in Thailand, by administrative level Administrative level Role Name b Population Per capita amount, US$ National Strategic mgmt. Country 62 000 000 0.0003 Regional Train, supply, supervise Region 6 7 506 340 0.68 Provincial Implement Roi Ed Province 1 505 314 0.20 Locala Implement 0.12 – Municipality Chiang Mai 151 011 0.01 – Subdistrict Padad 13 793 0.12 Total 62 000 000 1.00 aMunicipalities cover urban areas and subdistricts rural areas. The original per capita amounts have been multiplied by 20% and 80%, respectively, to represent the shares of Thailand’s population covered by these local administrative units. bThe specific administrative units are representative examples for estimating national costs. Region 6 contains Khon Kaen province. 86 Dengue Bulletin – Volume 34, 2010 Cost of dengue in Thailand Thailand’s national cost is US$ 62 million. The mean (±standard error of the mean) reported numbers of dengue cases and deaths were 81 000 (±20 000) and 93 (±25), respectively. The mean annual costs (±standard deviation) of reported dengue illness was I$ 147 (±32) million or US$ 48 (±10) million. Applying the expansion factor of 3.3 gives projected national costs of I$ 485 (±106) million or US$ 158 (±33) million. Spread over the entire population, the projected cost of dengue in Thailand is I$ 10.89 or US$ 3.55 per capita per year. Of this total, 72% represents the cost of illness and the remaining 28% is vector control. Discussion The burden of each hospitalized dengue case is high. The average of I$ 1758 or US$ 573 is about one fifth of the per capita GDP of Thailand (I$ 8440 or US$ 2750)[9]. The majority of this cost is incurred by the government, indicating the importance of dengue to public policy. We validated the macro-cost analysis for direct medical costs with activity-based costing (ABC). ABC, developed by Kaplan and Bruns in 1987,[12] is a type of cost accounting that seeks to allocate direct and indirect costs to the services provided. Originally developed to improve the efficiency of manufacturing in the private sector, the technique has been extended to some hospitals in both industrialized and developing countries.[12,13] It was used on a pilot basis in Khon Kaen Hospital to for the first 112 dengue cases, in which costs were allocated based on the number of hours that patients spent in each unit. The results showed that labour costs represented 66% of total costs while non-labour costs (including capital, utilities, and materials) comprised the remaining 33% and validated the overall costs from macro-costing. It is reassuring that the total annual cost of dengue in the present paper of US$ 158 million or US$ 3.55 per capita is reasonably similar to a largely independent estimate in another paper in Dengue Bulletin of US$ 135 million or about US$ 2.20 per capita.[14] While that study also included estimated costs of ambulatory only cases, its estimate of the cost per case, based on a lower estimated government subsidy for hospital care, was smaller. Elsewhere in South-East Asia, vector control programmes and potential vaccines have been shown to be cost-effective strategies to reduce the cost of dengue.[15] Among the eight countries in the multicountry study, Thailand’s cost per inpatient case (I$ 1758) was 26% higher than the average of I$ 1394 for all inpatient cases. Thailand’s per capita cost of dengue (US$ 3.55) is 32% lower than that of the other country in which the total cost of dengue has been estimated with a similar approach—Panama (US$ 5.22 per capita).[16] This relationship is consistent with Thailand’s 31% lower 2005 per capita GDP (US$ 2750 in Thailand and US$ 4630 in Panama) and the exclusion of costs of ambulatory-only cases in Thailand. Similarly, the share of the per capita cost represented by vector control in Thailand (28%) was very similar to that in Panama (30%). In public reporting of dengue, some countries report only laboratory confirmed cases, while others such as Thailand base the reporting on clinical criteria. In Khon Kaen Hospital and many other facilities in the region, the absence of laboratory confirmation does not mean that the case was not dengue. Rather, clinicians in Thailand, who are experienced in Dengue Bulletin – Volume 34, 2010 87 Cost of dengue in Thailand managing dengue, often did not feel that the laboratory test and associated delay were sufficiently useful to justify the discomfort to the patient and the cost involved. The similarity in demographics, clinical characteristics, and costs between cases that were laboratory confirmed with those not confirmed suggests that the majority of the clinically diagnosed but non-confirmed cases were, in fact, dengue. This fact shows the utility of reporting clinically co-diagnosed cases alongside those with laboratory confirmation. Since illness costs are 2.6 times greater than vector control costs, moderate increases in vector control costs would be economically justified even if they achieved only a small reduction in dengue illness. For example, a 26% increase in vector control cost would be justified provided it could reduce dengue burden by at least 10%. In this example, this 26% increase in vector control costs would add US$ 0.26 per capita while the 10% reduction in illness costs would save US$ 0.26 from that component, leaving the total unchanged at US$ 3.55 per capita. These findings suggest investing more money in vector control (or in a vaccine when available) would be worthwhile due to their offsets in the economic cost of dengue. Acknowledgements The authors are indebted to Binod Sah, MD, MS, and Rana Sughayyar, MS, for assistance in data cleaning and analysis, to Clare L. Hurley, MM (also at Brandeis University) for editorial assistance, and to the physicians, nurses and other staff of 16 departments at Khon Kaen Hospital for their helpful cooperation in abstracting financial and medical data. Our special thanks go to patients with dengue and their family members who enrolled in the study for providing such useful information. References [1] Ka l a yana roo j S , Chan s i r iwong s V, Nimmannitya S. Dengue Patients at the Children’s Hospital, Bangkok: 1995-1999 Review. Dengue Bulletin, 2002, 26:33-43. [2] Gubler DJ. Epidemic dengue/dengue hemorrhagic fever as a public health, social and economic problem in the 21st century. Trends in Microbiology, 2002,10:100-3. [3] W i c h m a n n O , H o n g s i r i w o n S , Bowonwatanuwon C, Chotivanich K, Sukthana Y, Pukrittayakamee S. Risk factors and clinical features associated with severe dengue infection in adults and children during the 2001 epidemic in Chonburi, Thailand. Tropical Medicine and Interational Health, 2004, 9:I022-9. [4] Epidemiology Bureau. Annual Epidemiological Surveillance Report. Bangkok: Ministry of Public Health, 2006. [5] Anderson KB, Chunsuttiwat S, Nisalak A, Mammen MP, Libraty DH, Rothman AL, Green S, Vaughn DW, Ennis FA, Endy TP. Burden of symptomatic dengue infection in children at primary school in Thailand: a prospective study. Lancet, 2007, 369:1452-9. [6] Clark DV, Mammen MP Jr, Nisalak A, Puthimethee V, Endy TP. Economic impact of dengue fever/dengue hemorrhagic fever in Thailand at the family and population levels. American Journal of Tropical Medicine and Hygiene, 2005, 72:786-91. 88 Dengue Bulletin – Volume 34, 2010 Cost of dengue in Thailand [7] Sornmani S, Okanurak K. Social and Economic Impact of Dengue Haemorrhagic Fever in Thailand. Unpublished report, Faculty of Tropical Medicine. Bangkok: Mahidol University, 1995. [8] Clark D. Quantification of Disease Burden: Measuring the Impact of Dengue Fever in Thailand, Worcester Polytechnic Institute, Armed Forces Research Institute of Medical Sciences, 2001. [9] Suaya JA, Shepard DS, Siqueira JB, Martelli CT, Lum LCS, Tan LH, Kongsin S, Jiamton S, Garrido F, Montoya R, Armien B, Huy R, Castillo L, Caram M, Sah BK., Sughayyar R, Tyo KR, Halstead SB. Cost of dengue cases in eight countries in the Americas and Asia: a prospective study. American Journal of Tropical Medicine and Hygiene. 2009; 80:846-55. [10] Shepard DS, Hodgkin D, Anthony Y. Analysis of Hospital Costs: A Manual for Managers. Geneva, Switzerland: World Health Organization, 2000. [11] Suaya JA, Shepard DS, Beatty M, Farrar J. Disease burden of dengue fever and dengue hemorrhagic fever. In: Preedy VR, Watson RR, eds. Handbook of Disease Burdens and Quality of Life Measures. New York: Springer, 2010:pp.1263-79. [12] Kaplan RS, Bruns WJ. Accounting and Management: A Field Study Perspective. Boston: Harvard Business School Press, 1987. [13] Waters H, Abdallah V, Santillán D, Richardson P. Application of activity-based costing (ABC) in a Peruvian NGO healthcare system. Operations Research Results 1(3, Revised). Bethesda, Maryland: Published for the U.S. Agency for International Development (USAID) by the Quality Assurance Project (QAP), 2003. [14] Lee LH, Vasan SS, Birgelen L, Murtola TM, Gong H-F, Field RW, Mavalankar DV, Ahmad NW, Hakim LS, Murad S, Ng CW, Lum LCS, Suaya JA, Shepard DS. Immediate cost of dengue to Malaysia and Thailand: An estimate. Dengue Bulletin, 2010, 34:65-76. [15] Shepard DS, Suaya JA. Economic evaluation of dengue prevention. In: Ungar W, ed. Economic Evaluation in Child Health Oxford: Oxford University Press, 2009:Chapter 12, pp. 225-37. [16] Armien B, Suaya J, Quiroz E, Sah B, Bayard V, Marchena L, Campos C, Shepard D. Clinical characteristics and national economic cost of the 2005 dengue epidemic in Panama. American Journal of Tropical Medicine and Hygiene, 2008, 79:364-71. Dengue Bulletin – Volume 34, 2010 89 Clinical characterization, diagnosis and socioeconomic impact of hospitalized dengue in Cambodia# Jose A. Suayaa, Ngan Chanthab, Rekol Huyb, Binod K. Saha, Moh-Seng Changc, Duong Socheatb, Philippe Buchyd, Te Vanthae, Ong Sivuthd, Elizabeth Haileselassief and Donald S. Sheparda* aSchneider Institutes for Health Policy, Heller School, Brandeis University, Waltham, MA, USA bNational Center for Parasitology, Entomology and Malaria Control, Ministry of Health, Phnom Penh, Cambodia cWHO, Phnom Penh, Cambodia dInstitut Pasteur du Cambodge, Phnom Penh, Cambodia eTakeo Provincial Hospital, Cambodia fIndependent Consultant, Richmond, VA USA Abstract Dengue infection is endemic among children in Cambodia. Few studies have described the clinical characteristics, diagnoses and the socioeconomic impact of hospitalized dengue cases. As part of a multicountry prospective dengue study, we analysed interviews and record reviews of 123 consecutive cases of children hospitalized with confirmed dengue in a major public provincial hospital. About 80% of cases evidenced plasma leak. On average, a hospitalized dengue case lasted 6.3 days, with 3.3 days of inpatient care, and cost US$116 (Singapore dollar 48). The majority (88%) of these cases’ households reported a substantial adverse economic impact, perhaps explained by their inability to afford medical care (39%), or on account of loss of five or more days of income (91%), need to borrow money beyond family or friends (39%), being compelled to sell or sale of property (53%) to pay for treatment. Hospitalized dengue has major clinical and socioeconomic consequences in Cambodia. Keywords: Dengue; socioeconomic burden; Cambodia. # This research was supported by research agreements from the Paediatric Dengue Vaccine Initiative (PDVI) to the National Center for Parasitology, Entomology and Malaria Control of the Ministry of Health of Cambodia, the Institut Pasteur du Cambodge, and Brandeis University in the USA, and by the endowment of the Schneider Institutes for Health Policy, Brandeis University. * E-mail: shepard@brandeis.edu 90 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia Introduction Dengue illness, a mosquito-borne viral disease, is a major global public health problem in tropical and subtropical countries and causes more morbidity and deaths than any other arbovirus disease in humans.[1] In Cambodia, dengue is endemic and, as in other South-East Asian countries, represents one of the primary causes of hospitalization and death among children.[2] Official reports, based only on hospitalizations of children, showed an annual average of 15 614 cases and 189 deaths during the period 2000–2007. In 2007 alone, 40 000 cases and 407 deaths were reported: the highest number of children affected in one year during this period.[3] Cambodia, with a population of 14 million inhabitants and a life expectancy at birth of 58 years, had a per capita gross domestic product (GDP) of US$ 380 and public spending on health of US$ 6.46 (1.7% of GDP) per person in 2005.[4] Only a few studies have examined economic aspects of dengue in Cambodia.[5,6,7,8] During the 2001 epidemic in Thmar Pouk district of Cambodia, a study was conducted on 162 dengue patients.[5] The mean household out-of-pocket expenditure per dengue case was estimated at US$ 8 for public hospitals and US$ 89 for private services. A remarkable finding of this study was that 70% and 33% of households of sick children receiving care in the private and public sector, respectively, went into debt in order to pay for medical services. Similarly, another study reported average out-of-pocket spending of up to US$ 49 for a single dengue episode in a public facility in Cambodia.[7] Yet another study found that the use of annual targeted larviciding campaigns as the major component of dengue vector control were actually cost-effective.[9] A more recent study found that dengue patients incurred costs similar to those with other febrile illnesses, both overall ($ 31.50 vs. $ 27.20, respectively) and among the hospitalized subset of both groups ($ 40.10 vs. $ 36.20, respectively).[6] Using data from a cohort of children who had been hospitalized for dengue in a major public provincial hospital in Cambodia, this study analyses the clinical characteristics and diagnosis of laboratory- confirmed dengue cases, the associated use and cost of health services received, and the socioeconomic impact of the illness episode on households. This study was part of a prospective multicountry study to measure dengue disease and the economic burden in eight dengue-endemic countries.[8] Methodology Study population Our study population consisted of children aged 0–14 years who were admitted into Takeo Provincial Hospital with clinical suspicion of dengue between December 2004 and December 2005. The Institutional Review Boards of the collaborating institutions and the funding agency approved the study protocol. All consecutive patients admitted during this period were candidates for this study. Their legal surrogates were informed of the purpose, invited to participate, and asked to sign an informed consent once they agreed to. All the patients had agreed to participate. Survey of parents or other legal surrogates The hospitalized child’s parent or legal surrogate was interviewed twice by a trained physician using a standardized survey instrument that Dengue Bulletin – Volume 34, 2010 91 Socioeconomic impact of hospitalized dengue in Cambodia included sections from the World Health Survey[10] and EuroQol (visual scale).[11] The survey ascertained clinical characteristics of the child’s illness as well as an assessment of the socioeconomic impact on the child’s household. All the interviews were conducted in person and took place at the hospital. The initial and follow-up interviews were timed at two days after hospital admission and one week after discharge (as part of a follow-up medical visit), respectively. Each interview took about 45–60 minutes. When inquiring about symptoms and health status, legal surrogates were invited to consult with the patient if necessary. Clinical characterization of a dengue case The study ascertained the clinical characteristics of a dengue case. Both the above-mentioned interviews of legal surrogates and a review of medical records were used as data sources. Parameters of interest included days of fever, days of overall illness, symptoms and signs of disease, perceived severity and quality of life (QoL), care-seeking behaviour, and use of aspirin and antibiotics. To measure QoL during the illness, the study used the EuroQol’s thermometer-like visual analog scale[11] where 100 corresponded to perfect health and zero to death, and asked parents or other legal surrogates to assess their child’s QoL. To complement the clinical information obtained from surveys, we reviewed medical records and extracted relevant clinical data (e.g. days of fever, clinical manifestations such as vomiting, diarrhoea, evidence of plasma effusion and bleeding) and laboratory data (e.g. platelet and white cell count, haematocrit, radiological results, etc.) related to the patient’s illness during the hospitalization. Classification of dengue cases by plasma leakage In order to evaluate the severity of dengue, we grouped patients based on evidence of plasma leakage, which is the most important pathophysiologic characteristic of dengue severity. Plasma leakage was defined by the verification in the medical records of plasma effusion or haemoconcentration. Plasma effusion was based on clinical or radiological signs of pleural/pericardial effusion, hepatomegaly, or ascites. Haemoconcentration was defined by evidence of a 20% increase in haematocrit (compared with the stabilized haematocrit at hospital discharge) or by a haematocrit value 20% above that considered normal for children (haematocrit >42).[12] Cost of a dengue case For each dengue case, we performed a full economic cost analysis consisting of the estimation and subsequent combination of three major cost categories: direct medical, direct non-medical, and indirect costs, all estimated from a societal perspective. We used the methodology from the multicountry study, with the interviews of parents or other legal surrogates and the annual hospital expenses as main data sources.[8] Cost-related domains surveyed included: health impact, use of medical services, schooling, work productivity, leisure time, out-of-pocket spending, and income lost. The time horizon was the acute dengue episode. To calculate direct medical costs for each patient, we summed the products of the type and amount of services received by setting (ambulatory or inpatient) and by provider (public or private), and multiplied this by their respective unit costs. We used 92 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia lost was calculated by multiplying the daily cost and the number of school days lost. The societal value of a day of work lost was considered the maximum of the worker’s reported income lost per day or Cambodia’s daily minimum wage (US$ 1.9).[15,16] The total economic costs of work days lost were calculated as the product of this average daily loss and the number of work days lost. “Other” days were days caregivers and patients lost for reasons other than school or work, such as days a non-working parent cared for an ill child. To value “other” days for those aged 15 or above, we multiplied the country’s daily minimum wage by the “other” days lost. In addition, the total days the household was affected was then calculated as the sum of school, work and “other” days lost. Finally, the total cost of a dengue case was calculated for each patient as the sum of all his or her direct (medical and non-medical) and indirect costs. Costs were expressed in 2005 US dollars. Economic and financial impact of a dengue case on households Parents or other legal surrogates were also interviewed about the type and degree of financial challenges resulting from their child’s illness (e.g., ability to afford medical care and days of income lost), their coping mechanisms adopted (e.g., borrowing money and/or selling valuables such as family silver or livestock), and their perception of the overall economic impact on their households. Analytical framework The unit of analysis was a dengue case defined by a documented acute febrile illness episode with clinical suspicion of dengue and each patient’s actual out-of-pocket payments for ascertaining the cost of private medical services. To estimate the economic cost of the medical care provided by public facilities, we used a macro-costing approach.[13] Three steps were used in this approach. First, using admissions, length of stay, and numbers of ambulatory visits in Takeo hospital as data, we estimated the hospital’s annual number of hospital-day equivalents. This estimation was computed by multiplying the annual number of admissions by the average length of stay and the number of hospital outpatient visits by 0.25 (based on the observation that the cost of a hospital outpatient visit was one fourth of a hospital day).[13] Second, to calculate the cost of a hospital day, we divided the hospital’s annual expenses by the number of hospital-day equivalents. Third, as public ambulatory care was provided not only by Takeo Hospital but also by other health centres and dispensaries, we assumed that the cost of a public ambulatory visit was 60% of the cost of a hospital outpatient visit. To calculate direct non-medical costs we aggregated all the out-of-pocket payments by the patient’s household for transportation, food, lodging and miscellaneous expenses related not only to the patient seeking and obtaining medical care but also to household members’ visits to the patient at the hospital. To calculate indirect costs for each patient, we assigned monetary values to three major categories of days: (a) days of school lost; (b) days of work for pay lost; and (c) “other” days lost by either the sick child and/or any other household member who provided care to the patient during the illness. As Cambodia funds primary education, we assumed that the economic value of a day of schooling was equal to the cost of providing a day of public primary school (US$ 0.10).[14] The economic loss attributed to school days Dengue Bulletin – Volume 34, 2010 93 Socioeconomic impact of hospitalized dengue in Cambodia All of the dengue serotypes were identified, with DENV-2 being the predominant type, and there were no fatalities reported in the study cohort. The majority of the patients had plasma leakage (80%). Only a minority of patients were from urban areas (12%). Females represented 54% of the participants. All the participants were children aged 14 or less, with 5–9 years being the predominant age group (59%). All of the children under the age of five had evidence of plasma leakage. Though the mean age of participants was eight, children with leakage were about two years younger (p<0.01) than those without leakage. Patients represented a wide range of household socioeconomic characteristics measured as the parent’s highest level of education. All the participants had two interviews (not shown) with the first and the second interviews on the 6th and 16th day from the onset of symptoms, on an average. Clinical characterization Based on a five-point scale, the majority of the participants (90%) felt good or very good (the two best points) before the beginning of symptoms, with children with plasma leakage reporting overall better health status prior to the illness (Table 2). About two thirds of the sick children sought medical care within 24 hours from the onset. The most commonly reported symptoms were fever (100%), abdominal pain (98%), vomiting (81%), headache (80%) and muscle or joint pain (74%). The majority of the patients had evidence of bleeding (55%) and thrombocytopenia (89%). Vomiting, bleeding and thrombocytopenia were found to be more frequent in cases with leakage (p<0.01). Overall, proxies reported serious impairment in the patient’s quality of life during the illness. The worst health status averaged only 7% of the equivalent of perfect health, indicating a loss subsequent laboratory confirmation. For every patient enrolled in the study, the Institut Pasteur du Cambodge performed two serologies: an in-house IgM capture enzyme-linked immunosorbent assay (MAC-ELISA) adapted from Rossi et al.[17], and haemagglutination inhibition (HI) tests.[18] Assays for dengue viruses and Japanese encephalitis virus antigens were used in parallel. When patient’s admission serum was positive for MAC-ELISA or had an HI titer over 2560, the laboratory diagnosis was of a recent dengue or other flavivirus infection. When the admission sample was negative for MAC ELISA and had HI titers ≤2560, a second serum sample (convalescent sample) was collected seven days after the first sample and a new run of MAC-ELISA and HI tests were performed on both admission and convalescent sera. The early specimen of serologically confirmed patients was also tested for RNA detection and dengue virus serotyping by semi-nested reverse transcriptase-polymerase chain reaction (RT-PCR).[19] For analytical purposes, participants were divided into two groups based on documentation of plasma leakage being present or absent. Data were entered into a Microsoft Access database, and standard routines for cleaning, consistency and analysis were performed. We report results as means and standard deviations for continuous variables and frequencies for categorical variables. T-tests and Chi square tests with alpha level of significance at 0.05 were performed for key analyses. Results The study recruited 127 patients with clinical suspicion of dengue, of which 97% (n=123) were dengue laboratory-confirmed and therefore included in the analysis (Table 1). 94 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia Table 1: Demographic characteristics of dengue-confirmed patients by cohort Item Plasma leakage Entire cohort No Yes P Cohort 25 98 123 Residence, % of cohort Urban 20 10 0.182 12 Sex, % in cohort Females 56 53 0.793 54 Age groups, % in cohort 0–4 yrs 5–9 yrs 10–14 yrs* 0 52 48 12 61 27 0.040 10 59 31 Age distribution, mean (SD) Age 9.1 (2.3) 7.5 (3.1) 0.018 7.9 (3.0) Parent highest level of education, % in cohort Primary school or less Secondary school Vocational school, college or more 32 64 4 41 54 5 0.672 39 56 5 Interviews Mother as legal surrogate, % Days from onset to first interview Days from first to econd interview 56 5.9 (1.6) 9.7 (1.6) 63 6.4 (1.2) 9.7 (2.1) 0.505 0.185 0.887 62 6.3 (1.3) 9.7 (2.0) * includes three 15-year-old patients of 93% of quality of life. On average, an illness episode lasted for 6.3 days, including 3.7 days of fever. Patients felt bad or very bad throughout the majority of the days of illness. Overall the disease was half a day longer in children with plasma leakage (p<0.05). One fifth of the patients received medication not prescribed by a medical practitioner, and nearly one fourth of the participants received antibiotics. Health-care use and productivity loss Each patient recorded an average of 3.7 ambulatory visits and 3.3 days of hospital stay. Overall, 59% of the ambulatory visits and all of the hospital days took place in public facilities (Table 3). There were no major differences in use of ambulatory services or length of hospital stay by plasma leakage Dengue Bulletin – Volume 34, 2010 95 Socioeconomic impact of hospitalized dengue in Cambodia Table 2: Clinical characteristics of dengue-confirmed patients by cohort Item Plasma leakage Entire cohortNo Yes P Cohort size (cases) 25 98 123 Health status before illness, % of cohort Good or very good 80 93 0.053 90 Seeking health care, % of cohort Within 24 hours from onset 76 68 0.458 70 Symptoms and signs during illness, % of cohort Fever Headache Muscle or joint pain Skin rash Retro-orbital pain Sore throat and/or running nose Diarrhoea Abdominal pain Vomiting Bleeding Low platelet count Low white cells count High white cells count 100 92 72 56 20 12 20 100 64 32 76 44 8 100 77 74 61 34 17 26 97 86 61 93 33 20 1.000 0.101 0.800 0.634 0.187 0.762 0.567 1.000 0.013 0.009 0.014 0.288 0.241 100 80 74 60 31 16 24 98 81 55 89 35 18 Quality of life during illness episode, mean (SD) Mean, quality in % 7.8 (4.3) 6.7 (3.9) 0.240 7.0 (4.0) Time indicators, mean (SD) Days of fever Days feeling bad or very bad Days of illness 4.1 (2.0) 5.7 (1.3) 5.8 (1.4) 3.6 (1.1) 6.2 (1.4) 6.4 (1.2) 0.199 0.118 0.025 3.7 (1.3) 6.1 (1.4) 6.3 (1.4) Treatment, % of cohort Self-medicated Used antibiotics Used aspirin 28 20 0 16 21 1 0.181 0.876 1.000 19 21 1 96 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia status (p<0.05). Dengue illness resulted in an average of 6.6 days of school absenteeism, 14.2 days of work lost and 2.6 other days lost, when including both the patient and other household members (Table 3). Among children without plasma leakage, almost all (92%) were in school at the time of the illness, whereas among children with plasma leakage, only 61% were in school. This discrepancy is explained by the overall younger age and larger proportion of children aged 0–4 years with plasma leakage who, therefore, were too young to be in school. This age difference resulted in an average of three fewer days of school absenteeism (p<0.05) in households of children with leakage than those without it. However, among children attending school, a supplementary analysis found that those with plasma leakage lost overall more days of school than children without leakage (7.0 vs. 6.6 days). Table 3: Use of health services and household impact by cohort Item Plasma leakage Entire cohortNo Yes P Cohort 25 98 123 Utilization of health services, mean (SD) Ambulatory visits % in public sector Hospital stay (days) % in public sector 3.4 (0.6) 62 3.5 (1.6) 100 3.8 (0.9) 58 3.2 (1.2) 100 0.034 0.389 3.7 (0.8) 59 3.3 (1.3) 100 Household impact, mean (SD) Days of school lost* % borne by patient Days of work lost Other days affected Total days affected 9.2 (7.6) 63 10.5 (5.1) 1.3 (2.4) 21.0 (9.4) 5.9 (5.9) 73 15.2 (8.4) 2.9 (3.5) 24.0 (9.6) 0.019 0.001 0.008 0.161 6.6 (6.4) 70 14.2 (8.0) 2.6 (3.4) 23.4 (9.6) *The large majority of children without plasma leakage were schooling at the time of the illness (92%) compared with children with plasma leakage (61%). The households of children with plasma leakage lost, on average, five more days of work compared with households of children without plasma leakage (15.2 versus 10.5 days, p<0.01, respectively). An average dengue episode was associated with a total loss of 23.4 household member days. The loss tended to be three days longer if the child had plasma leakage than if he did not, though the trend was not significant at p<0.05 (p equals 0.161). Dengue cost and household impact The total cost of a dengue episode averaged US$ 116 (Singapore dollars 48), and consisted of medical costs ($28 or 24%), non-medical costs ($33 or 28%) and indirect costs ($55 or 48%) (Figure 1). Indirect costs and total costs were significantly greater in cases with plasma leakage than without it (p<0.01). Dengue Bulletin – Volume 34, 2010 97 Socioeconomic impact of hospitalized dengue in Cambodia Figure 1: Full cost of a hospitalized dengue case by cohort 23 27 42 92 30 34 59 122 28 33 55 116 Without Leakage With Leakage Entire Cohort Medical care Non-medical cost Total costIndirect cost $0 $20 $40 $60 $80 $100 $120 $140 Ec o n o m ic C o st (U S$ ) Households faced financial challenges during the child’s illness (Figure 2). About 40% of the households reported difficulties in paying their medical providers. In addition, all but one household lost at least one day of work and the large majority (91%) of the households lost at least five days of work. Households addressed financial challenges by borrowing money from their family or friends (60%) or from other people (39%) and selling items (53%) such as livestock, family silver or furniture. Overall, the large majority (88%) of the households rated their child’s illness as having “quite a lot of substantial impact” on their family’s economic situation, corresponding to the two worst categories on a five-point scale. Discussion Dengue illness, an acute disease, was associated with diverse clinical manifestations and remarkable economic impact for our cohort of hospitalized cases. The severity of the disease among hospitalized cases is indicated by the fact that 80% had evidence of plasma leakage, an important risk factor for death. The large proportion of cases with plasma leakage may indicate the severity of dengue cases that warranted admission to the hospital. The observation that all of the children aged under five had evidence of plasma leakage suggests that dengue in this age group may be more severe than in older children. 98 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia The vast majority of cases (97%) admitted to the hospital with suspicion of dengue during the study period had laboratory confirmation, which may indicate the high prevalence of dengue among children, the absence of another cause of childhood febrile illness (e.g. influenza) at the time of the study, and the ability of parents to judge when their child may have dengue. The majority of patients in this study reported good health before the illness, but felt severely debilitated as evidenced by the substantial loss in quality of life in both patients with and without plasma leakage. The promptness in seeking care observed in our cohort suggests that dengue illness was associated with severe symptoms from the very onset of the disease. The average across each patient’s lowest quality of life was a level of only 7% of that of perfect health. This represents an average loss of 93% of the quality of life expected for perfectly healthy individuals. This loss was substantially more than the 60% loss from a comparable study in Malaysia.[20] As this study was the first time that the EuroQol thermometer visual scale was used in dengue patients in Cambodia, further validation would be required to rule out interviewer artifacts or the proxies’ difficulties in grading the quality of life of their child.[21] After these data are validated, they can be used to transform results from the thermometer visual scale into utilities for calculations of quality-adjusted life years (QALYs) and disability-adjusted life years (DALYs).[22] Figure 2: Household’s financial challenges, coping mechanisms and economic impact from a hospitalized dengue case 39% 91% 39% 53% 60% 88% Coping mechanisms Overall economic 5+ days of income lost Borrowed from other than family/friends Quite a lot or substantial Sold items Borrowed from family/friends Financial challenges 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Could not afford medical care P er ce n ta g e o f h o u se h o ld s Dengue Bulletin – Volume 34, 2010 99 Socioeconomic impact of hospitalized dengue in Cambodia The severity of a hospitalized dengue case is also marked by the large number of medical services obtained during the illness. A patient had on average 3.7 ambulatory visits and 3.3 days of hospital stay within six days of the illness episode. The greater total cost of a hospitalized dengue case with plasma leakage than without it is explained by the greater ambulatory care costs, non-medical costs and indirect costs. Overall, cases with plasma leakage were associated with a greater number of private, and more expensive, ambulatory visits. We were unable to document differences in the cost of inpatient care by plasma leakage status because of the limitation of the hospital costing methodology used, which consisted of measuring the cost of a hospital day for an average hospitalization. However, as the hospital did not have an intensive care unit but rather a paediatric medical unit for all patients, the use of length of stay and average cost per hospital day was appropriate. Dengue illness affected not only the sick child but also other household members who provided care to the patient. For example, about 30% of the household’s days of school lost were borne by family members other than the sick child. In addition, a dengue episode with an average duration of less than one week resulted in twice as many days of work-for-pay lost by household members. The high share of other household members’ involvement for hospitalized patients reflects the custom and need for a family member to stay with the patient virtually full time while hospitalized, and substantially during the rest of the illness. Another indication of this spill-over effect is the number of person days affected in the entire household. A hospitalized dengue episode, when including the patient’s and other household members’ time, resulted in a loss of 23.4 days. Of these days, the majority (71% or 16.6 days) were borne by household members other than the patient (not shown in table). The remarkable economic impact of the disease reported by most of the households may be explained by both the large out- of-pocket spending on medical and non- medical items and the income lost from lost days of work. Since no households reported sick leave, the lost days of work reported by nearly all households represented income lost. To cope with these financial challenges, the majority of households borrowed money or sold some of their valuables. A recent study has also shown that dengue often forces Cambodian households to sell property or goods or take loans to meet the catastrophic expenses.[7] These observations highlight the potentially devastating effect of this disease on a household’s economic situation in a country with limited social protection. The main limitation of this study is that it is based only on children admitted as hospital inpatients. By contrast, a community-based study would likely have included many non- hospitalized patients with less severe disease, and probably without plasma leakage. As normal values of haematocrit in Cambodian children were not available, we used those for children in the United States,[23] an approach followed by dengue researchers in a Nicaraguan study.[12] As children in Cambodia are more likely than their American counterparts to have lower haematocrit values due to a higher prevalence of malnutrition, the identification of haemoconcentration in our study may be conservative.[12] We initially attempted to classify dengue illness between dengue fever and dengue haemorrhagic fever. We discontinued this attempt, however, as the then existing WHO classification, which dated from 1997,[24] led 100 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia to problems in interpretation. At the time of the study WHO was developing newer guidelines, but these did not become available until afterwards.[25] For patients whose condition was severe enough to require hospitalization, dengue was a severe illness even in the absence of plasma leakage documented in the medical record. Patients under study without plasma leakage lost significantly more days of school and stayed slightly longer in hospital than patients with leakage. Thus, the study shows that hospitalized dengue cases are severe illnesses with substantial socioeconomic consequences, regardless of documentation of plasma leakage in the medical record. As this study was conducted in a single rural provincial hospital during only one year, generalizations for the entire country require caution. The cases involved in this study may be less severe than those during other years and in other locations, no deaths were reported. By contrast, two deaths would have been expected in the study cohort based on the death rate of 1.7% which was arrived at on the basis of the 2001-2005 notifications from Phnom Penh and Kandal Provinces.[8] However, as over 85% of the Cambodian population live in rural areas, and the demographic characteristics of our sample resembled that of the national reported cases in 2005, we believe some overall extrapolations to frame the problem of dengue at the national level are warranted. For example, in 2005, Cambodia reported 9040 hospitalized cases and 155 deaths, a fatality rate of 1.7%. Applying our cost estimates, the total economic burden of hospitalized, reported, non-fatal dengue was US$ 917 000, or US$ 0.07 per capita. However, this figure does not include estimates for under-reporting of hospitalizations, cost of ambulatory cases, economic value of dengue deaths, or cost of vector control efforts. In fact, 90% of the dengue deaths reported in Cambodia in 2005 occurred in children aged 0–9 years, with the average age of death being five years. Based on the 2005 country- specific life tables, for every death there was an average loss of 58 years of life.[26] The total years of life lost from reported dengue deaths is, therefore, an estimated 7480 person-years. These years lost are also associated with lost productivity, which implies a considerable additional economic loss. Although we are not aware of any other economic studies of dengue costs that distinguish costs according to plasma leakage, we can compare our results to those from other studies on the overall cost of dengue. The economic cost per hospitalized case in this study (US$ 116) is substantially higher than the average costs of previous studies of US$ 8[5], 40[6], or 49[7]. A major reason for difference is perhaps on account of this study’s more comprehensive estimate of the government subsidy for hospital costs and the quantification of time lost which is not directly related to lost income. When costs of fatal dengue are factored in, Cambodia’s cost per case rises to US$ 265.[27] Compared to costs in other Asian countries, however, the cost per case in Cambodia is still low. Other studies in Dengue Bulletin reported costs per hospitalized case of US$ 573 for Thailand[28] and US$ 575 for India,[29] and a previous publication pegged the figure at US$ 947 for Malaysia.[27] Cambodia’s per capita cost of dengue derived here (US$ 0.07 for non-fatal dengue) is substantially less than the estimate for Cambodia when fatal dengue is included (US$ 0.20)[27]. It is also below the national figures for Dengue Bulletin – Volume 34, 2010 101 Socioeconomic impact of hospitalized dengue in Cambodia Thailand (US$ 0.53) for non-fatal cases only[28] or US$ 0.77 including fatal cases,[27] or for that of Malaysia at $1.41.[27] Even when results are converted to international dollars, the cost per case and cost per capita in Cambodia are less than half of that for Thailand and Malaysia,[27] largely due to Cambodia’s lower per capita gross national income with correspondingly lower costs of medical care and valuation of lost time. During a more representative outbreak, the duration of hospitalization, the hospital t reatment resources and tota l costs would likely have been far greater than recorded here. Moreover, official reports may substantially underestimate the true number of dengue cases.[8,30] Hence a more comprehensive estimate would require incorporating expansion factors to adjust for this underreporting. In conclusion, households of patients hospitalized for dengue face substantial and tangible economic hardship. Acknowledgements The authors would like to thank the following organizations and colleagues for their assistance and support in study implementation, data management and editorial comments: The Cambodia Ministry of Health, staff at the Takeo Hospital; Mariana Caram, Clare L. Hurley, Erika Moldow, Chrisann Newransky, and Rana Sughayyar, of Brandeis University; and Scott Halstead of the Pediatric Dengue Vaccine Initiative. References [1] Farrar J, Focks D, Gubler D, Barrera R, Guzman MG, Simmons C, Kalayanarooj S, Lum L, McCall PJ, Lloyd L, Horstick O, Dayal-Drager R, Nathan MB, Kroeger A. Towards a global dengue research agenda. Tropical Medicine & International Health, 2007, 12: 695-699. [2] DeRoeck D, Deen J, Clemens JD. Policymakers’ views on dengue fever/dengue haemorragic fever and the need for dengue vaccines in four southeastern countries. Vaccine, 2003, 22: 121-129. [3] National Dengue Control Program, National Center for Parasitology, Entomology, and Malaria Control, Ministry of Health, Cambodia, Reported annual dengue cases and deaths from 2001-2007. [4] United Nations Development Programme. Human development report 2007/2008: fighting climate change : Human solidarity in a divided world New York: United Nations Development Programme; Palgrave Macmillan, 2007. [5] Damme WV, Leemput LV, Por I, Hardeman W, Meessen B. Out-of-pocket health expenditure and debt in poor households: evidence from Cambodia. Tropical Medicine and International Health, 2004, 9: 273-280. [6] Huy R, Wichmann O, Beatty M, Ngan C, Duong S, Margolis HS, Vong S. Cost of dengue and other febrile illnesses to households in rural Cambodia: a prospective community- based case-control study. BMC Public Health, 2009, 9: 155. 102 Dengue Bulletin – Volume 34, 2010 Socioeconomic impact of hospitalized dengue in Cambodia [7] Khun S, Manderson L. Poverty, user fees and ability to pay for health care for children with suspected dengue in rural Cambodia. International Journal for Equity in Health, 2008, 7: 10. [8] Suaya JA, Shepard DS, Armien B, Caram M, Castillo L, Chantha N, Garrido F, Kongsin S, Lum L, Montoya R, Sah BK, Siqueira JB, Sughayyar R, Tyo K. Multi-country study of costs of dengue among ambulatory and hospitalized patients. Presented at the American Society of Tropical Medicine and Health 56th Annual Meeting, Philadelphia, PA, Nov. 5, 2007, Abstract 28. American Journal of Tropical Medicine and Hygiene, 2007, 77: 9. [9] Suaya JA, Shepard DS, Chang MS, Caram M, Hoyer S, Socheat D, Chantha N, Nathan MB. Cost-effectiveness of annual targeted larviciding campaigns in Cambodia against the dengue vector Aedes aegypti. Journal of Tropical Medicine and International Health, 2007, 12: 1026-1036. [10] World Health Organization (WHO). World Health Survey instruments and related documents, 2002. http://www.who.int/ healthinfo/survey/instruments/en/index.html. Accessed: January 15, 2008. [11] EuroQol Group. EuroQol-A new facility for the measurement of health-related quality of life. Health Policy, 1990, 16: 199-208. [12] Harris E, Videa E, Perez L, Sandoval E, Tellez Y, Perez M, Cuadra R, Rocha J, Idiaquez W, Alonso R, Delgado M, Campo L, Acevedo F, Gonzalez A, Amador J, Balmaseda A. Clinical, epidemiologic, and virologic features of dengue in the 1998 epidemic in Nicaragua. American Journal of Tropical Medicine and Hygiene, 2000, 63: 5-11. [13] Shepard DS, Hodgkin D, Anthony Y. Analysis of Hospital Costs: A Manual for Managers. Geneva, Switzerland: World Health Organization, 2000. [14] The World Bank Group, EDSTATS. http:// devdata.worldbank.org/edstats/cd1.asp. [15] Economic Research Institute, HR Tax Codes and Laws. http://www.erieri.co.uk/freedata/ HRCodes/. [16] International Labour Organization, Minimum Wages Database. http:/ /www.i lo.org/ travaildatabase/servlet/minimumwages. Accessed 18 Mar 2011. [17] Rossi CA, Ksiazek TG, Enzyme-linked immunosorbent assay (ELISA). Lee CSC, ed. Manual of hemorrhagic fever renal syndrome and hantavirus pulmonary syndrome. Seoul, 1998, 87-91. [18] Clarke DH, Casals J. Techniques for hemagglutination and hemagglutination- inhibition with arthropod borne viruses. American Journal of Tropical Medicine and Hygiene, 1958, 7: 561-73. [19] Lanciotti RS, Calisher CH, Gubler DJ, Chang GJ, Vorndam AV. Rapid detection and typing of dengue viruses from clinical samples by using reverse transcriptase-polymerase chain reaction. Journal of Clinical Microbiology, 1992, 30: 545-51. [20] Lum LCS, Suaya JA, Tan LH, Sah BK, Shepard DS. Quality of life of dengue patients. American Journal of Tropical Medicine and Hygiene, 2008, 78: 862-867. [21] Prosser LA, Hammitt J, Keren R. Measuring health preferences for use in cost-utility and cost-benefit analyses of interventions in children: theoretical and methodological considerations. Pharmacoeconomics, 2007, 25: 713-26. [22] O’Leary JF, Fairclough DL, Jankowski MK, Weeks JC. Comparison of time-tradeoff utilities and rating scale values of cancer patients and their relatives: evidence for a possible plateau relationship. Medical Decision Making, 1995, 15: 132-7. [23] Soldin S, Brugnara C, Gunter K, Hicks J, eds,. Pediatric Reference Ranges. Washington DC: AAAC Press, 1997. Dengue Bulletin – Volume 34, 2010 103 Socioeconomic impact of hospitalized dengue in Cambodia [24] World Health Organization (WHO). Dengue haemorrhagic fever: diagnosis, treatment, prevention and control. 2nd edition. Geneva, 1997. [25] World Health Organization (WHO). Dengue: Guidelines for Diagnosis, Treatment, and Control. New Edition. World Health Organization and Special Program for Research and Training on Tropical Diseases, 2009. http://whqlibdoc. who.int/publications/2009/9789241547871_ eng.pdf. Accessed: 12 August 2010. [26] World Health Organization (WHO), Life tables for WHO member states for year 2005. http:// www.who.int/whosis/database/life_tables/ life_tables.cfm. [27] Suaya JA, Shepard DS, Siqueira JB, Martelli CT, Lum LCS, Tan LH, Kongsin S, Jiamton S, Garrido F, Montaya R, Armien B, Huy R, Castillo L, Sah BK, Sughayyar R, Tyo KR, Halsted SB. Cost of dengue cases in eight countries in Americas and Asia, a prospective study. American Journal of Tropical Medicine and Hygiene, 2009;80: 845-55. [28] Kongsin S, Jaimton S, Suaya J, Vasanawathana S, Sirisuvan P, Shepard DS. Cost of dengue in Thailand. Dengue Bulletin, 2010;34:77-88. [29] Bhavsar A, Shepard DS, Suaya JA, Mafowosofo M, Hurley CL, Howard M. A private hospital- based study assessing knowledge, attitude, practice and cost associated with dengue illness in Surat, India. Dengue Bulletin. 2010;34: 54-64. [30] Suaya JA, Shepard DS, Beatty M, Farrar J. Disease burden of dengue fever and dengue hemorrhagic fever. Preedy VR, Watson RR, eds. Handbook of Disease Burdens and Quality of Life Measures. New York: Springer, 2010. 104 Dengue Bulletin – Volume 34, 2010 Instructions for contributors Dengue Bulletin welcomes all original research papers, short notes, review articles, letters to the Editor and book reviews which have a direct or indirect bearing on dengue fever/ dengue haemorrhagic fever prevention and control, including case management. Papers should not contain any political statement or reference. Manuscripts should be typewritten in English in double space on one side of white A4-size paper, with a margin of at least one inch on either side of the text and should not exceed 15 pages. The title should be as short as possible. The name of the author(s) should appear after the title, followed by the name of the institution and complete address. The e-mail address of the corresponding author should also be included and indicated accordingly. References to published works should be listed on a separate page at the end of the paper. References to periodicals should include the following elements: name and initials of author(s); title of paper or book in its original language; complete name of the journal, publishing house or institution concerned; and volume and issue number, relevant pages and date of publication, and place of publication (city and country). References should appear in the text in the same numerical order (Arabic numbers in parenthesis) as at the end of the article. For example: (1) Nimmannitaya S. Clinical spectrum and management o f dengue haemorrhagic fever. The Proceedings of the International Conference on Dengue Haemorrhagic Fever, Kuala Lumpur, September 1-3, 1983:16-26. (2) Gubler DJ. Dengue and dengue haemorrhagic fever: Its history and resurgence as a global public health problem. In: Gubler DJ, Kuno G (ed.), Dengue and dengue haemorrhagic fever. CAB International, New York, NY, 1997, 1-22. (3) Nguyen Trong Lan, Nguyen Thanh Hung, Do Quang Ha, Bui Thi Mai Phuong, Le Bich Lien, Luong Anh Tuan, Vu Thi Que Huong, Lu Thi Minh Hieu, Tieu Ngoc Tran, Le Thi Cam and Nguyen Anh Tuan. Treatment of dengue haemorrhagic fever at Children’s Hospital N.1, Ho Chi Minh City, 1991-1996. Dengue Bulletin. 1997; 22: 150-161. Figures and tables (Arabic numerals), with appropriate captions and titles, should be included on separate pages, numbered consecutively, and included at the end of the text with instructions as to where they belong. Abbreviations should be avoided or explained at the first mention. Graphs or figures should be clearly drawn and properly labelled, preferably using MS Excel, and all data clearly identified. Articles should include a self-explanatory abstract at the beginning of the paper of not more than 300 words explaining the need/ gap in knowledge and stating very briefly the area and period of study. The outcome of the research should be complete, concise and focused, conveying the conclusions in totality. Appropriate keywords and a running title should also be provided. Articles submitted for publication should be accompanied by a statement that they have not already been published, and, if accepted Dengue Bulletin – Volume 34, 2010 105 Instructions for contributors for publication in the Bulletin, will not be submitted for publication elsewhere without the agreement of WHO, and that the right of republication in any form is reserved by the WHO Regional Offices for South-East Asia (SEARO) and the Western Pacific (WPRO). One hard copy of the manuscript with original and clear figures/tables and a computer diskette/CD-ROM indicating the name of the software should be submitted to: The Editor Dengue Bulletin WHO Regional Office for South-East Asia Indraprastha Estate Mahatma Gandhi Road New Delhi 110002, India Telephone: 91-11-23370804 Fax: 91-11-23379507, 23370972 E-mail: dengue@searo.who.int Manuscripts received for publication are subjected to in-house review by professional experts and are peer-reviewed by experts in the respective disciplines. Papers are accepted on the understanding that they are subject to editorial revision, including, where necessary, condensation of the text and omission of tabular and illustrative material. Original copies of articles submitted for publication will not be returned. The principal author will receive 10 reprints of the article published in the Bulletin. A pdf file can be supplied on request.