True status of smear-positive pulmonary tuberculosis defaulters in Malawi.

True status of smear-positive pulmonary tuberculosis defaulters in Malawi M.L. Kruyt,1 N.D. Kruyt,2 M.J. Boeree,3 A.D. Harries,4 F.M. Salaniponi,5 & P.A. van Noord6 The article reports the results of a study to determine the true outcome of 8 months of treatment received by smear- positive pulmonary tuberculosis (PTB) patients who had been registered as defaulters in the Queen Elizabeth Central Hospital (QECH) and Mlambe Mission Hospital (MMH), Blantyre, Malawi. The treatment outcomes were documented from the tuberculosis registers of all patients registered between 1 October 1994 and 30 September 1995. The true treatment outcome for patients who had been registered as defaulters was determined by making personal inquiries at the treatment units and the residences of patients or relatives and, in a few cases, by writing to the appropriate postal address. Interviews were carried out with patients who had defaulted and were still alive and with matched, fully compliant PTB patients who had successfully completed the treatment to determine the factors associated with defaulter status. Of the 1099 patients, 126 (11.5%) had been registered as defaulters, and the true treatment outcome was determined for 101 (80%) of the latter; only 22 were true defaulters, 31 had completed the treatment, 31 had died during the treatment period, and 17 had left the area. A total of 8 of the 22 true defaulters were still alive and were compared with the compliant patients. Two significant characteristics were associated with the defaulters: they were unmarried; and they did not know the correct duration of antituberculosis treatment. Many of the smear-positive tuberculosis patients who had been registered as defaulters in the Blantyre district were found to have different treatment outcomes, without defaulting. The quality of reporting in the health facilities must therefore be improved in order to exclude individuals who are not true defaulters. Voir page 389 le re´sume´ en franc¸ais. En la pa´gina 390 figura un resumen en espan˜ol. Introduction Malawi has had a National Tuberculosis Control Programme (NTP) since the country gained its independence in 1964 but, like many countries in sub-Saharan Africa, is now burdened with a large and growing tuberculosis (TB) problem, mainly as a result of the epidemic caused by human immunodeficiency virus (HIV). In 1984, the TB programme adopted the DOTS (directly observed treatment, short course) strategy and was supported in this endeavour by the International Union against Tuberculosis and Lung Disease (IUATLD); all districts in the country implemented this strategy over the following 1–2 years. The good recording and reporting system, which is inherent in the DOTS strategy, allowed the TB epidemic to be reliably monitored, and the number of notified cases rose from 5334 in 1985 to 20 630 in 1996 (Malawi NTP, 1996). As the NTP struggled to cope with the increasing number of cases and a deteriorating economic situation, national cure rates for smear-positive TB patients decreased from 86% in 1986 to 63% in 1992, and increased slightly to 68% in 1994 (Malawi NTP). In Malawi, the TB registers are maintained by the district TB officers (DTOs). Each registered TB patient is given a unique registration number, and the name, age, sex, address, date of diagnosis, type and category of TB are recorded. Patients are treated with standardized antituberculosis regimens, depending on the type and category of TB. Newly diagnosed smear-positive pulmonary tuberculosis (PTB) cases receive 8 months of short-course chemotherapy, comprising 2 months of initial intensive treatment in hospital with daily supervised doses of streptomycin, rifampicin, isoniazid and pyrazinamide, followed by 6 months of unsupervised continuation therapy at home with isoniazid and thiacetazone or isoniazid and ethambutol. The drugs used in the continuation phase may be supplied to patients by the hospital or health centre. Sputum smears are examined after 2 months, 5 months and 8 months of treatment, and the outcome at the end of treatment is recorded according to the guidelines (see Table 1) established by the IUATLD and WHO (1, 2). In the Blantyre district, which has 15 govern- ment health centres providing a TB service to the community, the cure rates in 1991 were less than 50%, with high default and transfer rates among 1 College of Medicine, University of Utrecht, Utrecht, Netherlands. 2 Department of Medicine, University of Amsterdam, Amsterdam, Netherlands. 3 Head, Department of Medicine, College of Medicine, Private Bag 360, Chichiri, Blantyre 3, Malawi. Requests for reprints should be sent to Dr M.J. Boeree at this address. 4 Adviser, National Tuberculosis Programme, Lilongwe, Malawi. 5 Programme Manager, National Tuberculosis Programme, Lilongwe, Malawi. 6 Senior Lecturer, Department of Epidemiology, College of Medicine, University of Utrecht, Utrecht, Netherlands. Reprint No. 5792 Research 386 # World Health Organization 1999 Bulletin of the World Health Organization, 1999, 77 (5) patients (3). Although cure rates have improved since 1991, the recorded default rates are still unacceptably high at around 10–15%. Defaulters are an extra risk for the population because of possible contagion and the development of drug resistance by tubercle bacilli. The reasons for such defaulting in Malawi and other parts of sub-Saharan Africa are not known, but may include failure to report or record the cure or death of a patient, failure to collect the necessary drugs from the hospital or health centre because the patient was too ill, discontinuation of treatment by the patient after experiencing an initial improvement, and failure to notify the health care system that the patient had transferred to another health centre. In the present study we investigated the status of patients with smear-positive PTB who had been registered as defaulters in the two principal hospitals which manage TB in the Blantyre district, and the possible reasons for their defaulting. Patients and methods The outcomes of 8 months of treatment of all smear- positive PTB patients who were registered between 1 October 1994 and 30 September 1995 in the TB registers at the Queen Elizabeth Central Hospital and the Mlambe Mission Hospital, Blantyre district, were studied. The name, age, sex, home address and treatment unit of patients who had been registered as defaulters were recorded, and a search was initiated to determine what had happened to them. Between 31 December 1996 and 31 March 1997 the mortality records of patients admitted to the hospital for the intensive phase of treatment were inspected to determine whether those registered as defaulters had in fact died in hospital. The peripheral unit (usually a health centre) where the patients received the continuation phase of treatment was then visited to determine whether there were any records of the treatment outcome. If there was no such record or the patient was still registered as a defaulter in the health centre, attempts were made to trace the patient or a relative in the home or workplace. This required making journeys to villages by motorcycle. If there was still no trace of the patient, a letter was written if a post- office box number was present in his/her treatment unit records or the TB register. In this way we investigated and documented the true 8-month treatment outcome of patients who had been registered as defaulters. According to the IUATLD and WHO definition (1, 2), true defaulters are patients who at any time during the course of treatment had not collected their drugs for two or more consecutive months. All defaulting patients who were still alive, or a near relative, were interviewed (using a structured questionnaire) to determine the following: the current health status of the patient or, if deceased, the date of death; any financial or other difficulties which prevented collection of antituberculosis medication; the dis- tance from the home to the health facility and the time taken to make this journey; and the presence of any other disease besides PTB. In order to determine the factors that could have been associated with defaulter status, a second interview, using another questionnaire, was carried out with the true defaulters and with a group of cured, fully compliant TB patients (matched for age and sex) who had been registered during the same period. This second interview, which was based on the health belief model (4–6), gathered information about demographic and socioeconomic variables, the patient’s health-seeking behaviour and knowledge about antituberculosis treatment, the continuing availability of such treat- ment, the relations between the health care worker and the patient, the health status of the patient during the continuation phase of treatment, as well as ease of access to the health unit. Results A total of 1099 new smear-positive PTB patients were registered at the Queen Elizabeth Central Hospital and the Mlambe Mission Hospital during the 12-month study period; 126 (11.5%) of them had been registered as defaulters. Information about treatment outcomes was obtained for 101 (80%) of these defaulters (Table 1) through a home or village visit (100 cases) and by writing to one patient. Among patients whose treatment outcome was not a true default, we discovered a failure in communication between the district TB officer and the health centre in 70% of cases and between the patient and the health centre in the remainder. Table 1. True 8-month treatment outcomes for 101 study patients who had been registered as defaulters True outcome No. of patients Defaulted from treatment 22 (i.e. patients who during their treatment had not collected drugs for two consecutive months) Cured 27 (i.e. patients who had completed treatment and whose sputum smear results were negative at the end of treatment) Treatment completed 4 (i.e. patients who had completed treatment but whose sputum smear results were not known at the end of treatment) Died 31 (i.e. patients who had died during the 8-month treatment regimen) Transferred 17 (i.e. patients who had transferred to another district and whose treatment outcome was unknown) Total 101 Pulmonary tuberculosis defaulters in Malawi 387Bulletin of the World Health Organization, 1999, 77 (5) Only 22 patients were true defaulters according to the IUATLD and WHO definition (1,2). At the time of the first interview, we found that 12 out of the 22 true defaulters had died since the end of their 8-month treatment period, 1 patient had moved out of the area, 1 patient had been wrongly registered in another unit and had defaulted again from treatment, and 8 patients were alive and well. In our first interviews we investigated 20 cases (the 8 who were alive and the relatives of the 12 patients who had died). The most important findings from these interviews are shown in Table 2. We also compared the 8 defaulters who were still alive with a group of age- and sex-matched cured patients who had been fully compliant with their treatment. The main differences between the two groups were as follows: all 8 compliant patients were married, compared with none of the defaulters; and 7 of the compliant patients knew the correct duration of antituberculosis treat- ment, compared with none among the defaulters. Discussion The true outcome of 8 months of treatment was determined for 80% of the patients who had been registered as defaulters. The remaining 20% could not be traced because either their addresses in the TB register were incorrect and our inquiries in their villages did not help or our letters sent to their post- office box numbers were not answered. Of the 101 patients who were traced, only 22 were true defaulters. The true default rate was therefore considerably lower than that given in the TB register. The possible causes for the different treatment outcomes and suggestions for corrective measures are summarized in Table 3. About one-third of the registered defaulters had in fact completed their treatment and many of them were cured. The DTO was not aware of this because either the officer had not visited and checked the results in the treatment unit registers, or the patient’s transfer to another health centre had not been communicated to the DTO (one-fifth of the cases). Failure in communication between districts was also a problem, and this accounted for a small number of defaulters who should have been registered as ‘‘transferred’’ to another health centre. About one-third of the defaulting patients had died. Over 70% of tuberculosis patients registered for treatment at the Queen Elizabeth Central Hospital were HIV-seropositive (7); therefore, since the mortality rates among HIV-infected TB patients in sub-Saharan Africa are high (8–11) it is not surprising that many ‘‘defaulters’’ were in fact patients who had died. In the rural areas, with few telephones and difficulties in communication, it is unlikely that relatives would travel to the nearest health centre to report the death of a patient when there is no incentive to do so. Health centre staff are supposed to follow up all defaulters, but owing to transport difficulties, increasing workload, lack of motivation, concerns about safety in remote areas, and poor record-keeping, few of them do so. Ways to improve communication between health facilities within a district and between different districts must therefore be found. We investigated the reasons why patients defaulted. Although only a small number of our defaulting patients could be compared with compli- ant cured patients, we found that unmarried status and ignorance of the duration of antituberculosis treatment were characteristic of defaulting beha- viour. The public should be given more information about tuberculosis, especially the total duration of treatment and the need to complete the full course, Table 2. Results of interviews with 8 study patients and relatives of 12 patients who had died, concerning 20 defaulters No. of defaulters 20 No. of males/females 14/6 Mean age (years) 38.9+13.0 Mean time to obtain the medicine (minutes) 147+118 Mean distance to hospital/health centre (km) 7.7+4.7 Mean survival time (months) 11.2+6. 5 Were there difficulties in obtaining medicine in general? Yes 89.5% No 10.5% Were financial problems incurred to obtain medicine? Yes 55.6% No 44.4% Were there difficulties in obtaining medicine due to patients’ physical problems? Yes 36.8% No 63.2% Table 3. Possible causes and treatment outcomes for 79 patients who had incorrectly been registered as defaulters, and possible corrective measures Treatment outcome Possible causes Corrective measures Cured or treatment completed (n = 31) DTO not visiting health centresa Regular DTO visits to health centresa Health centres not communicating with DTO Educate health centre staff about communicating the results of treatment outcome Transferred (n = 17) DTOs failing to communicate with each other Quarterly DTO meetings at regional level Died (n = 31) Difficulties in relatives providing information to health centres Better follow-up by health centre staff would result from having adequate transport, staff numbers and training a DTO: district tuberculosis officer. Research 388 Bulletin of the World Health Organization, 1999, 77 (5) via intensive health education through health care workers and the use of posters, leaflets and flyers in the local language. Malawi is a poor country, and it is therefore unlikely that substantial socioeconomic and demographic improvements can be made in the near future. This study may raise questions about the reliability of data for other treatment outcomes. Cure rates and treatment completion rates can be verified from the patients’ treatment cards, and sputum smear results from the laboratory records. Details of patients who move to another district should be recorded in the TB registers. These treatment outcomes are likely to be correct, although we have not specially investigated this. Deaths cannot always be verified if the patient died in the village, and further research to assess whether death rates are accurate is probably warranted. n Acknowledgements This study was funded partly by a British Govern- ment grant from the Department for International Development (DFID) to the Malawi National Tuberculosis Control Programme. We thank DFID for this support, and are grateful to the Royal Dutch Tuberculosis Association (KNCV) for advice, and the District TB Officer at the Queen Elizabeth Central Hospital for his help. Ethical approval was granted by the Malawi National Health Science Research Committee of the Ministry of Health and Population. Re´sume´ La ve´ritable situation des tuberculeux non observants a` frottis positif au Malawi Pour faire face a` la charge de plus en plus lourde que repre´sente la tuberculose, de nombreux pays d’Afrique subsaharienne comptent sur leurs programmes natio- naux de lutte antituberculeuse. L’un des e´le´ments essentiels de la lutte antituberculeuse consiste dans la notification et le suivi des non-observants au cours du traitement. Au Malawi, le taux de non-observance varie de 10 a` 15%. Au cours de la pre´sente e´tude, nous avons cherche´ a` de´terminer quelle e´tait la situation ve´ritable des malades atteints de tuberculose pulmonaire avec frottis positif qui avaient e´te´ enregistre´s comme non- observants dans les deux hoˆpitaux antituberculeux du district de Blantyre, en nous efforc¸ant de trouver des raisons pouvant expliquer leur comportement. Nous avons examine´ les re´sultats de 8 mois de traitement chez tous les cas de tuberculose pulmonaire avec frottis positif consigne´s entre octobre 1994 et septembre 1995 sur les registres de la tuberculose de ces hoˆpitaux. Nous avons releve´ l’identite´ de ceux qui e´taient note´s comme non-observants et nous les avons suivis par des visites a` domicile ou, a` de´faut, en consultant les statistiques de mortalite´ des hoˆpitaux ou les dossiers des centres de sante´. Nous avons interroge´ les malades qui e´taient de ve´ritables non-observants selon la de´finition de l’OMS ainsi que les proches des malades qui avaient de´me´nage´ ou e´taient de´ce´de´s, afin de nous faire une ide´e de leur e´tat de sante´ et de connaıˆtre les raisons qui les avaient amene´s a` ne pas observer le traitement. Nous avons ensuite proce´de´ a` un deuxie`me interrogatoire, base´ sur une mode´lisation des croyances en matie`re de sante´, des ve´ritables non-observants et d’un groupe te´moin de la meˆme pe´riode, constitue´ de malades gue´ris (apparie´s par sexe et par aˆge). Sur les 1099 malades que nous avons identifie´s au cours de la pe´riode e´tudie´e, 126 (11,5%) avaient e´te´ enregistre´s comme non-observants, mais 22 d’entre eux (dont 8 encore en vie) ont pu ve´ritablement eˆtre qualifie´s comme tels. De fait, environ un tiers des non-observants enregistre´s e´taient alle´s jusqu’au bout de leur traitement et un grand nombre avait gue´ri. La principale raison des erreurs d’enregistrement tenait a` une mauvaise commu- nication entre le Service antituberculeux districal concerne´ et les centres de sante´. L’insuffisance de la communication entre les districts e´tait e´galement en cause, de meˆme que l’absence de suivi des malades par les centres de sante´. Au de´part, nous avons interroge´ 20 personnes: les 8 non-observants encore vivants et 12 proches des malades de´ce´de´s. Au cours d’un second interrogatoire, nous avons compare´ les 8 malades non observants a` des patients gue´ris et parfaitement observants. Environ 90% des malades ou de leurs proches ont de´clare´ avoir eu des proble`mes financiers ou diverses difficulte´s pour obtenir leurs me´dicaments. Comparativement aux patients gue´ris, les ve´ritables non- observants se caracte´risaient par le fait qu’ils e´taient ce´libataires et n’avaient aucune ide´e de la dure´e normale de leur traitement. Parmi les non-observants re´pertorie´s, nombreux e´taient les malades de´ce´de´s. Dans des zones rurales avec des proble`mes de communication dus notamment au sous-e´quipement te´le´phonique, on ne pouvait gue`re s’attendre a` ce qu’un parent se rende jusqu’au centre de sante´ le plus proche pour y de´clarer le de´ce`s d’un malade, s’il n’e´tait pas pousse´ a` le faire. L’absence de suivi des malades non observants par les centres de sante´ s’explique par les proble`mes de transport, l’accroisse- ment de la charge de travail que cette activite´ aurait produit, un manque de motivation et la mauvaise tenue des dossiers. Il importe donc de trouver le moyen d’ame´liorer la communication entre les e´tablissements de soins, tant entre districts qu’a` l’inte´rieur d’un meˆme district. En ce qui concerne les difficulte´s financie`res et les proble`mes pratiques auxquels les non-observants ont eu Pulmonary tuberculosis defaulters in Malawi 389Bulletin of the World Health Organization, 1999, 77 (5) a` faire face, on peut remarquer que le Malawi est un pays pauvre, dont la situation socio-e´conomique et de´mo- graphique n’a gue`re de chance de s’ame´liorer dans un proche avenir. Toutefois, en pratiquant une e´ducation sanitaire intensive, il devrait eˆtre possible de corriger l’ignorance qu’ont les non-observants de la dure´e normale du traitement antituberculeux. Resumen Situacio´n real de los pacientes con tuberculosis pulmonar y frotis positivo remisos a cumplir el tratamiento en Malawi Muchos paı´ses del A´frica subsahariana tienen programas nacionales de lucha antituberculosa para hacer frente a la creciente carga de tuberculosis (TB). Una parte esencial del control de la TB consiste en la notificacio´n y el seguimiento de los pacientes remisos durante la terapia; en Malawi, la tasa de abandonos varı´a entre un 10% y un 15%. En el presente estudio investigamos la situacio´n real de los pacientes con tuberculosis pulmonar (TBP) y frotis positivo que habı´an sido registrados como remisos a cumplir el tratamiento en los dos hospitales principales para casos de tuberculosis del distrito de Blantyre; se analizaron asimismo las posibles razones del abandono. Examinamos los resultados de ocho meses de tratamiento de todos los pacientes con TBP y frotis positivo que figuraban en los registros de TB en esos hospitales entre octubre de 1994 y septiembre de 1995. Tras identificar a los pacientes remisos registrados, procedimos a hacer un seguimiento de los mismos, bien acudiendo a visitarlos a su domicilio, o bien, cuando ello no era posible, examinando los registros de mortalidad de los hospitales y los registros de los centros de salud. Entrevistamos tanto a pacientes que eran verdaderos remisos, segu´n la definicio´n de la OMS, como a familiares de pacientes que habı´an cambiado de domicilio o habı´an muerto, al objeto de conseguir informacio´n sobre su estado de salud y sobre las posibles razones del abandono del tratamiento. Se llevo´ a cabo una segunda entrevista, basada en el modelo de creencias de salud, entre los remisos verdaderos y un grupo testigo de pacientes curados (armonizado por edad y sexo) del mismo periodo. De los 1099 pacientes identificados durante el periodo de estudio, 126 (11,5%) habı´an sido registrados como remisos, pero so´lo 22 resultaron serlo realmente, y de e´stos ocho seguı´an con vida. Casi un tercio de los remisos segu´n los registros habı´an terminado en realidad su tratamiento y muchos estaban curados; otra tercera parte habı´a fallecido. La razo´n principal de los errores de registro era la deficiente comunicacio´n entre la Oficina de TB distrital responsable y los centros de salud. La mala comunicacio´n entre los distritos tambie´n era causa de problemas, al igual que la incapacidad para efectuar un seguimiento de los pacientes desde los centros de salud. Inicialmente llevamos a cabo 20 entrevistas: con los ocho pacientes remisos que seguı´an con vida, y con 12 familiares de pacientes fallecidos. En una segunda entrevista se procedio´ a comparar a los ocho pacientes remisos con pacientes curados que habı´an seguido fielmente el tratamiento. Casi un 90% de los pacientes o sus familiares sen˜alaron que habı´an tenido problemas econo´micos o de otro tipo para obtener los medica- mentos. En comparacio´n con el grupo curado, los remisos verdaderos eran en general personas solteras y que no sabı´an cua´nto debı´a durar el tratamiento. Muchas de las personas consideradas remisas segu´n los registros habı´an fallecido. En las zonas rurales, con pocos tele´fonos y malas comunicaciones, era improbable que un familiar viajase hasta el centro de salud ma´s cercano para notificar la muerte de un paciente, cuando no tenı´a ningu´n incentivo para ello. El personal de los centros de salud no podı´a realizar un seguimiento de los pacientes remisos debido a los problemas de transporte, la gran carga de trabajo, la falta de motivacio´n y los fallos del sistema de registro. Por consiguiente, hay que hallar la manera de mejorar las comunicaciones entre los establecimientos de salud de cada distrito y entre los diferentes distritos. En lo que respecta a los problemas financieros y pra´cticos afrontados por los pacientes remisos, cabe sen˜alar que Malawi es un paı´s pobre y que es improbable que su situacio´n socioecono´mica y demogra´fica mejore en un futuro pro´ximo. Sin embargo, la ignorancia constatada entre los pacientes remisos respecto a la duracio´n correcta del tratamiento podrı´a remediarse mediante una educacio´n sanitaria intensiva. References 1. Maher D et al. Treatment of tuberculosis: guidelines for national programmes, 2nd ed. Geneva, World Health Organization, 1997 (unpublished document WHO/TB/97.220 Rev. 1). 2. Enarson DA et al. Tuberculosis guide for low-income countries, 4th ed. Geneva, International Union against Tuberculosis and Lung Disease, 1996. 3. Harries AD et al. Tuberculosis programme changes and treatment outcomes in patients with smear-positive pulmonary tuberculosis in Blantyre, Malawi. Lancet, 1996, 347: 807–809. 4. Kelly GR, Mamon JA, Scott JE. Utility of the health belief model in examining medication compliance among psychiatric out- patients. Social science and medicine, 1987, 25 (11): 1205–1211. 5. Fincham JE, Wertheimer AI. Using the health belief model to predict initial drug therapy defaulting. Social science and medicine, 1985, 20: 101–105. 6. Barnhorn F, Adriaanse H. In search of factors responsible for noncompliance among tuberculosis patients in Wardha district, India. Social science and medicine, 1992, 34 (3): 291–306. 7. Harries AD et al. An audit of HIV testing and HIV serostatus in tuberculosis patients, Blantyre, Malawi. Tuberculosis and lung diseases, 1995, 76: 413–417. 8. Nunn P et al. Cohort study of human immunodeficiency virus infection in patients with tuberculosis in Nairobi, Kenya. American review of respiratory diseases, 1992, 146: 849–854. Research 390 Bulletin of the World Health Organization, 1999, 77 (5) 9. Ackah AN et al. Response to treatment, mortality, and CD4 lymphocyte counts in HIV-infected persons with tuberculosis in Abidjan, Coˆte d’Ivoire. Lancet, 1995, 345: 607–610. 10. Perriens JH et al. Pulmonary tuberculosis in HIV-infected patients in Zaire: a controlled trial of treatment for either 6 or 12 months. New England journal of medicine, 1995, 332:779–784. 11. Elliott AM et al. The impact of human immunodeficiency virus on mortality of patients treated for tuberculosis in a cohort study in Zambia. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1995, 89: 78–82. Pulmonary tuberculosis defaulters in Malawi 391Bulletin of the World Health Organization, 1999, 77 (5)