INT J TUBERC LUNG DIS 11(10):11081113 2007 The Union

Perceptions and beliefs about cough and tuberculosis and implications for TB control in rural Rwanda
P. N. Ngang,* J. Ntaganira, A. Kalk, S. Wolter, S. Ecks
* Central Hospital Yaounde, Yaounde, Cameroon; School of Public Health, National University of Rwanda, Kigali, Gesellschaft fur Technische Zusammenarbeit, Kigali, Rwanda; Department of Tropical Hygiene and Public Health, University of Heidelberg, Heidelberg, Department of Anthropology, South Asian Institute, University of Heidelberg, Heidelberg, Germany SUMMARY
S E T T I N G : Two southern provinces of Rwanda, Butare and Gikongoro. O B J E C T I V E S : To identify beliefs and popular perceptions on cough and tuberculosis (TB) in rural Rwanda and determine how they shape health-seeking behaviour. M E T H O D S : Eight focus group discussions, 21 key informant interviews and 12 illness narratives were conducted between May and June 2004. S T U D Y P O P U L A T I O N : TB patients, community members, traditional healers and health workers. R E S U L T S : There is wide use of herbal treatment for chronic cough in Rwanda. Patients seek conventional care when alternative treatment options fail or when severe symptoms such as shortness of breath, bloody sputum and weight loss appear. There are several local illnesses associated with chronic cough, with different alternative

treatments. TB symptoms are often mistaken for the acquired immune-deciency syndrome (AIDS). Identied causes for cough-related illnesses can be classied as biomedical (germs, internal body dysfunction and worms), environmental (seasonal changes and dust), cultural (inheritance), socio-economic (hard work, malnutrition and tobacco), and supernatural (witchcraft). Three healthseeking end points emerge for chronic cough: home care, health facility and the traditional healer. Healers in some areas, however, believe TB due to witchcraft can only be treated traditionally. C O N C L U S I O N : This study unveils beliefs and treatment options for chronic cough in Rwanda, with important implications for TB control that should be addressed. K E Y W O R D S : cough; tuberculosis; health-seeking behaviour

A NATIONAL TUBERCULOSIS (TB) Programme (NTP) has existed in Rwanda since 1990. This DOTSbased programme aims at eliminating TB as a public health problem. The programme was re-launched in 1995 following the 1994 genocide, but was extended to all health districts only in 2001. Health centres diagnose and treat smear-positive TB and follow up treatment for smear-negative and extra-pulmonary disease diagnosed at district hospitals.1 Between 2001 and 2003, a drop in case detection was observed. This was unexpected given a human immunodeciency virus (HIV) prevalence of 11% and an annual risk of TB infection of 1.8%.24 There was therefore a need to identify factors that contribute to low case detection. Studies from different parts of the world have shown that both the TB epidemic and NTPs are inuenced by multiple factors.58 The present study was conducted between May and June 2004 in two southern provinces of RwandaButare and Gikongoro. The aim was to identify popular beliefs associated with cough and determine how these beliefs affect healthseeking behaviour for TB.

METHODOLOGY
Setting In 2003, Butare had four health districts and a population of 722 616, while Gikongoro had three districts and a population of 492 607. Gikongoro is a mountainous region with little economic activity and poor geographic access to health care.4 The presence of key national institutions in Butare has favoured social and economic growth, and its non-mountainous terrain provides better geographic access to health services. Both provinces were seriously affected by the 1994 genocide and are still recovering from its effects. Material and methods Focus group discussions (FGDs), key informant interviews and illness narratives were used to obtain data. Two eld assistants helped the principal investigator. Tape recordings and notes were used to record information. Interviews were conducted in French and Kinyarwanda, a national language. Original recordings were transcribed in Kinyarwanda or French and then

Correspondence to: Ngang Peter Nchotu, BP 25622 Yaound, Cameroon. Tel: (237) 7 707 9480. Fax: (237) 2 221 0989. e-mail: ngangbmvr@yahoo.co.uk; awahbless@yahoo.com Article submitted 9 April 2006. Final version accepted 6 July 2007.

Cough and TB in rural Rwanda

1109

translated into English. The protocol was reviewed and approved by the ethics committee of the University of Heidelberg, Germany. The NTP of Rwanda authorised eld implementation. All participants provided informed verbal consent. Study population In Butare, data were collected from the Kabutare District hospital and the Kibilizi and Mugombwa health centres; in Gikongoro it was obtained from the Kigeme district hospital and the Kirambi and Cyanika health centres. The study population included community informants, traditional healers, TB patients and district health staff. Eight FGDs were conducted: four with TB patients and four with comparison groups of people who had never been diagnosed for TB. The comparison groups were selected to match the TB groups for age and sex. The number of participants for FGDs ranged from four to eight. Twelve other TB patients not involved in FGDs were asked to narrate their illness from onset of symptoms to the time of interview. Four came from Butare and eight from Gikongoro. As the number of TB patients in Cyanika was too small for an FGD, only illness narratives were conducted. Twenty-one key informants were interviewed: 10 traditional healers, six community members and ve TB health workers. Purposive sampling was used to select participants from amongst patients

attending the health institutions and from surrounding communities.9 Limitations Due to time and resource limitations, the study was restricted to two provinces and made use of only qualitative methods. Most interviews were conducted within the premises of health facilities. This may have introduced some information bias.

RESULTS
Illnesses associated with cough All participants were aware of TB (igituntu) and asthma (asima). The other major types of cough-related illnesses were identied only by traditional healers (Table). Other categories of participants turn to healers for treatment without explicitly linking symptoms to local illnesses. The main local illnesses presenting with cough included umugozi, agasema and sema. Umugozi is associated with hard work, genetic heritage and disorders of the veins, and presents with chronic cough, fever and haemoptysis. Agasema presents with cough, difcult breathing and blood in sputum, while sema involves difculty in breathing, bloody sputum and swollen limbs. Agasema is believed to result from dirt and dust and sema from seasonal changes and hard work. Healers in Butare reported only the former,

Table

Characteristics of some illnesses associated with cough

Main symptoms Cough Fever or sweating Weight loss Chest pain Blood in sputum Difficult breathing Poor appetite Abdominal distension Cough and vomiting Cough Fever or sweating Chest pain Blood in sputum Difficult breathing Swollen limbs/joint pain Cough Difficult breathing Skin rash Cough Swollen limbs/joint pain Blood in sputum Difficult breathing Cough Blood in sputum Difficult breathing Perceived cause(s) Hard work Tobacco Malnutrition Germs Genetic Poison and witchcraft Worms Main treatment option(s) Modern treatment (for TB not associated with witchcraft) Traditional, with concoctions that induce vomiting (for TB associated with witchcraft) Participant group* TB patients Community key informants Health workers Traditional healers

Illness (local name) Tuberculosis (igituntu)

Umugozi

Hard work Nerve disorder Genetic

Traditional herbal treatment

Traditional healers

Agasema

Dust Dirt Related to seasons Hard work

Traditional herbal treatment

Traditional healers

Sema

Referred to health facilities by some healers for modern care Traditional

Traditional healers

Asthma (asima)

Tobacco

TB patients Community key informants Traditional healers

* Main group(s) of participants that reported this type of cough. TB associated with witchcraft. The same word in Kinyarwanda seems to apply to veins. TB tuberculosis.

1110

The International Journal of Tuberculosis and Lung Disease

while those in Gikongoro knew only of the latter. It was not clear whether agasema and sema refer to the same entity. Belief in the supernatural cause of some forms of TB was more prevalent in Gikongoro than in Butare. Perceptions of cough All but one TB patient had pulmonary disease, which is the main source of human-to-human infection. Cough only became a matter of concern when other symptoms, such as breathing difculties, weight loss and bloody sputum appeared. The latter was sometimes considered a fatal sign and made some patients too afraid to seek early care. According to a male FGD participant:
The rst time I saw blood (in my sputum), I had gone to work in the elds, I hid the sputum, I will not tell a lie. Then I went home. People told me to go to the hospital and I refused, I hoped that it would go away. The following day I went back to the farm and could not work . . . People who came to visit me are the ones who told me to tell my wife . . . I dont know how they carried me to the hospital. I was in bad shape . . .

Apart from the fear caused by the unusual nature of the symptoms, patients were also apprehensive of having the acquired immune-deciency syndrome (AIDS). Although the majority had only TB, the AIDS stigma seemed to outweigh the fear of having TB. According to one health worker, patients were slow in seeking help for TB symptoms, but once the disease was diagnosed they became reluctant to leave the health facility. Knowledge that TB can be cured seemed to reassure patients, who consequently felt more comfortable staying at the health facility to receive treatment. Symptoms identied for TB were very similar to those dened in modern medicine. Identied causes could be classied as biomedical (germs, internal body dysfunction and worms), environmental (seasonal changes and dust), cultural (genetic heritage), socioeconomic (hard work, malnutrition and tobacco) and supernatural (witchcraft). The main symptoms for TB associated with witchcraft were cough with blood in sputum and abdominal distension. When reporting on deaths that occur in the community, a healer said:
When the person is going to die of TB (caused by witchcraft), he vomits blood,* vomits the inner parts of the body, the lungs, the heart. . . . He dies after vomiting all those inner body parts. Thats how TB from witchcraft kills a person. (Interview, female healer)

Home care targeted symptoms. The popular local plant, umuramvumba, was reported as the main treatment for cough regardless of duration. It is considered a broad-spectrum cough medicine (Figure). Some TB patients reported having had symptoms for over 6 months before seeking care away from home. Most abandoned home care only when severe weight loss, breathing difculties and haemoptysis appeared, or when attempts at home remedies had failed. Traditional healers identied two categories of patients. The rst are those who presented with conditions such as umugozi, agasema, asthma or TB due to witchcraft. The second category comprises cases they would refer to the health facilities, often after attempts at traditional treatment had failed. Most healers claimed they could treat umugozi. Treatment lasted several weeks. Some believed sema could only be treated in health facilities. Several healers said they used herbal medications to treat and cure asthma. Another less popular plant identied for treating cough was umunyinya. Healers in Gikongoro strongly believed that TB due to witchcraft could only be treated traditionally. Treatment was by administering concoctions that resulted in vomiting. For this to be effective, patients

Care seeking Three care-seeking end points emerged from this study: home care, treatment at health facilities and treatment by traditional healers.
* Literally means sends out blood.

Figure The umuramvumba plant (in the front and central part of the picture, with bigger leaves).

Cough and TB in rural Rwanda

1111

have to present early after the onset of symptoms or the infection forms roots in the body and becomes difcult to treat. Some healers felt it was dangerous to send patients with TB from witchcraft to the hospital, because if . . . they inject the person he will die immediately. Most TB patients reported several trips to the health facility before being screened for TB. This was a demanding task in the mountainous areas of Gikongoro, where there is poor geographic access to health care. Traditional healers claimed to refer cases to health facilities, but no TB patient acknowledged receiving advice from a healer to seek modern care. One healer reported that some patients he saw heard about him during discussions with other patients in the waiting areas of health facilities. Health staff in Gikongoro reported heavy expenditures (payment in kind) incurred when visiting healers that rendered patients very poor before they decided to seek conventional care. This was the rationale for starting an education programme for various categories of traditional healers at the Kirambi health centre. According to observations made by health staff, the populations in the mountain regions of Gikongoro seemed to rely more on herbal treatment for cough and other chest symptoms than was the case in Butare. The reported number of TB cases in one health centre in Gikongoro had steadily dropped over time. The health staff did not provide a reason for this drop. A female participant believed her TB was related to the HIV infection she contracted when she was raped during the 1994 genocide.

DISCUSSION
Findings from this study unveil community views and treatment options for cough and TB that are different from the conventional biomedical approach. Although herbal remedies such as umuramvumba might be appropriate for mild coughs and colds, indiscriminate use in cases of chronic cough will likely create a pool of TB infection in the community. Local illnesses such as umugozi, agasema and sema, identied by healers, represent possible cases of TB that are not brought to the attention of the health services. Umugozi, in particular, presents with cough, fever, chest pain, blood in sputum and dyspnoea, which should lead to screening for TB. Similar ndings from Indonesia showed that conditions dened as one disease by Western medicine were represented by a network of local illnesses by ethnic groups.6 From this study, healers consistently identify and recommend treatment for local illnesses that could be TB. The other participants turn to them for care without clearly associating symptoms to the illnesses. This might partly reect the poor nancial and geographic accessibility of modern services, especially in the mountainous terrain of Gikongoro, where patients travel long distances

to reach health facilities. This concords with the view, held by some health workers, that patients seek modern care only after making large payments in kind to healers. The predominantly farming population, who have little free time, might have easier access to healers living in their communities than to modern health services. Patients did not acknowledge referrals to health facilities by healers. This could be interpreted as fear of discussing experiences at the hands of traditional healers with investigators. In Malawi, it was observed that although two thirds of traditional healers claimed to refer patients to hospitals, most decisions to seek help from health facilities were made by the patient, the family or following advice from health workers.10 Belief in some rural areas that TB caused by witchcraft can only be treated traditionally is a further barrier to modern care. The diagnosis was based primarily on the presence of abdominal distension and haemoptysis. About 70% of the forms of TB diagnosed in HIVinfected patients affect organs other than the lungs.11 Peritoneal TB, characterised by abdominal distension, is a common nding. The reported cases of deaths from TB related to witchcraft could have been cases of massive haemoptysis, a known and often dramatic occurrence in advanced pulmonary disease, with bleeding from lung cavities. The observation that illnesses caused by witchcraft are treated traditionally is not new. An earlier study in Rwanda had revealed that 90% of mothers of children aged 6 years with diarrhoea associated with witchcraft turned to traditional treatment.12 Contrary to the situation in Malawi, where healers claim they could cure TB,10 those in the present study limited themselves to witchcraft-related TB. Although TB patients did not specically mention fear of AIDS as a reason for delay in seeking help, several patients reported undergoing HIV testing before sputum examination for TB, which shows that these two diseases are intertwined in their perceptions. Similar ndings have been reported from various sites worldwide, such as Malawi13 and Thailand.7 The key here is to raise awareness that TB is not always associated with HIV,14 and that it can be cured whether the victim is HIV-positive or HIV-negative. The fact that patients were reluctant to return to the community once the diagnosis of TB was made could mean that, apart from seeking cure, another goal of patients might simply be to rid themselves of stigmatising symptoms before going back home. This differs from NTP objectives that focus on controlling TB as a disease, often with little or no attention given to the experience patients go through when living with the illness in society. It has been observed that patients have a broader focus of both dealing with the experience of ill health and trying to cope with daily life, while NTPs focus more on the disease.15 The differences between Gikongoro and Butare with regard to TB treatment options might result from

1112

The International Journal of Tuberculosis and Lung Disease

social and economic inequality. Local and global inequalities have been identied as an obstacle to basic health care.16 Attention has also been drawn to the need to create a balance between provision of treatment through targeted public health interventions such as DOTS, and appropriate action to address barriers to effective chemotherapy.17 Infectious diseases such as TB have been shown to be concentrated among the poor who live under the weight of major social inequalities.18 Although health education in the short term can help people come to grips with some biomedical challenges, there is a need also to address the social, economic and environmental factors that predispose and contribute to maintaining preventable infectious diseases.19 The possible case of HIV infection from rape during the 1994 genocide also highlights the role of social and political instability in the spread of infection. However, in seeking to address these issues, it is still important to pay attention to the universal provision of high quality TB control, given that no global strategies have been adopted as yet to alleviate poverty and other causes of inequality.20

CONCLUSION
This study unveils beliefs and treatment options for chronic cough in Rwanda, with important implications for TB control that should be duly addressed. The ght against infectious diseases such as TB needs to go beyond the biomedical and address social, economic, environmental and political factors as well. Left unattended, these factors, alongside popular local beliefs, may create important barriers between the community and health services. Finally, while integrating care, it is important to educate the public about the curability of TB, irrespective of HIV status. Acknowledgements
This study was conducted within the framework of an MSc thesis on Community Health and Health Management (CHHM) in Developing Countries at the University of Heidelberg, Germany, 200304. Collaborating institutions were GTZ-Sant in Kigali, the National TB Control Programme of Rwanda and the School of Public Health in Butare, Rwanda.

References
1 Ministre de la Sant, Rpublique Rwandaise. Politique nationale en matire de sant. Kigali, Rwanda: Ministre de la Sant, 2002.

2 Programme National Intgr de Lutte Contre la Lpre et la Tuberculeuse. Rapport annuel dactivits 2001. Cit de Kigali, Rwanda: PNLT, 2002. 3 Programme National Intgr de Lutte Contre la Lpre et la Tuberculeuse. Rapport annuel dactivits 2002. Cit de Kigali, Rwanda: PNLT, 2003. 4 Programme National Intgr de Lutte Contre la Lpre et la Tuberculeuse. Rapport annuel dactivits 2003. Cit de Kigali, Rwanda: PNLT, 2004. 5 Steen T W, Mazonde G N. Ngaka ya Setswana, ngaka ya segoa or both? Health seeking behaviour in Batswana with pulmonary tuberculosis. Soc Sci Med 1999; 48: 163172. 6 Caprara A, Abdulkadir N, Idawani C, Asmara H, Lever P, De Virgilio G. Cultural meanings of tuberculosis in Aceh Province, Sumatra. Med Anthropol 2000; 19: 6589. 7 Ngamvithayapong J, Winkvist A, Diwan V. High AIDS awareness may cause tuberculosis patient delay: results from an HIV epidemic area, Thailand. AIDS 2000; 14: 14131419. 8 Long N H, Johansson E, Lonnroth K, Eriksson B, Winkvist A, Diwan V K. Longer delays in tuberculosis diagnosis among women in Vietnam. Int J Tuberc Lung Dis 1999; 3: 388393. 9 Dawson S, Manderson L. Methods for social research in disease; a manual for the use of focus groups, WHO/UNDP/ World Bank/TDR, Boston, MA, USA: International Nutrition Foundation for Developing Countries (INFDC), 1993: p 23. 10 Brouwer J A, Boeree M J, Kager P, Varkevisser C M, Harries A D. Traditional healers and pulmonary tuberculosis in Malawi. Int J Tuberc Lung Dis 1998; 2: 231234. 11 McQuaid K R. Alimentary tract. In: Tierney L M, McPhee S T, Papadakis M A, eds. Current diagnosis and treatment, adult ambulatory and in-patient management. 41st ed. New York, NY, USA: Langue Medical Books/McGraw-Hill, 2002: p 596. 12 Csete J. Health seeking behaviour of Rwandan women. Soc Sci Med 1993; 37: 12851292. 13 Banerjee A, Harries A D, Nyirenda T, Salaniponi F M. Local perceptions of tuberculosis in a rural district in Malawi. Int J Tuberc Lung Dis 2000; 4: 10471051. 14 Sonnenberg P, Glynn J R, Fielding K, Murray J, Godfrey-Faussett P, Shearer S. HIV and pulmonary tuberculosis: the impact goes beyond those infected with HIV. AIDS 2004; 18: 657662. 15 Ogden J. The resurgence of tuberculosis in the tropics. Improving tuberculosis controlsocial science inputs. Royal Society of Tropical Medicine and Hygiene; Meeting at Manson House, London, 20 May 1999. Trans Roy Soc Trop Med Hyg 2000; 94: 135140. 16 Farmer P. Pathologies of power: rethinking health and human rights. Am J Public Health 1999; 89: 14861496. 17 Farmer P, Nardell E. Editorial: Nihilism and pragmatism in tuberculosis control. Am J Public Health 1998; 88: 7. 18 Castro A, Farmer P. Science and society. Infectious disease in Haiti. EMBO Rep 2003; 4 (Suppl): S20S23. 19 Porter J, Ogden J, Pronyk P. Infectious disease policy: towards the production of health. Health Policy Plann 1999; 14: 322328. 20 Grange J M, Grandy M, Farmer P, Zumla A. Historical declines in tuberculosis: nature, nurture and the biosocial model. Int J Tuberc Lung Dis 2001; 5: 208212.

RSUM
L I E U D T U D E : Deux provinces au sud du Rwanda, Butar et Gikongoro. O B J E C T I F : Identifier les croyances culturelles et les perceptions populaires de la toux et de la tuberculose (TB) en zone rurale au Rwanda et dterminer leurs influences sur les comportements lgard des soins de sant. M T H O D E S : Huit discussions de groupes, 21 interviews et 12 narrations sur la maladie ont t conduits entre mai et juin 2004. P A R T I C I P A T I O N A L T U D E : Les patients tuberculeux, les membres de la communaut, les praticiens traditionnels et le personnel de sant.

Cough and TB in rural Rwanda

1113

RSULTATS :

La toux chronique est souvent traite par des plantes mdicinales locales en Rwanda. Les symptmes qui amnent les patients chercher des soins sont une dyspne, des crachats hmoptoques et une chute de poids. Plusieurs pathologies locales sont associes la toux, et sont traites de faon alternative. Les symptmes de la TB sont trs souvent confondus avec ceux du SIDA. Les causes biomdicales, environnementales, socioconomiques et mystiques (sorcellerie et empoisonnement) sont associes la TB. Les comportements des malades

pour la recherche de soins en cas de toux les dirigent principalement vers : le traitement domicile, les structures sanitaires et les praticiens traditionnels. Certains praticiens traditionnels dans les zones rurales estiment quune TB dorigine mystique ne peut tre traite que traditionnellement. C O N C L U S I O N : Cette tude rvle des croyances et options de traitement de la toux au Rwanda qui ont des consquences importantes sur la lutte contre la TB et qui doivent ainsi tre prises en compte.
RESUMEN

Butare y Gikongoro, dos provincias del sur de Ruanda. O B J E T I V O S : Definir las creencias y las percepciones populares sobre la tos y la tuberculosis (TB) en zonas rurales de Ruanda y determinar en qu forma estas definen los comportamientos de bsqueda de atencin de salud. M T O D O S : Se realizaron ocho grupos de opinin, 21 entrevistas a informadores clave y 12 narrativas de la enfermedad, entre mayo y junio de 2004. P O B L A C I N D E L E S T U D I O : Pacientes con TB, miembros de la comunidad, curanderos y profesionales de la salud. R E S U L T A D O S : La fitoterapia tiene amplias aplicaciones contra la tos crnica. Los pacientes buscan atencin convencional cuando los tratamientos alternativos fallan o cuando aparecen sntomas graves como disnea, hemoptisis o prdida de peso. Existen algunas enfermedades locales que se acompaan de tos crnica y tienen diferentes
MARCO DE REFERENCIA :

tratamientos alternativos. Los sntomas de la TB se suelen confundir con el SIDA. Las causas de las enfermedades asociadas con tos se pueden clasificar como : biomdicas (microorganismos, disfuncin interna del cuerpo y gusanos), medioambientales (cambios estacionales y polvos), culturales (herencia), socioeconmicas (trabajo intenso, desnutricin y tabaquismo) y sobrenaturales (hechicera). Aparecen tres criterios de evaluacin de la bsqueda de atencin de salud en la tos crnica : atencin domiciliaria, centros de salud y curanderos. Sin embargo, los curanderos de algunas regiones atribuyen la TB a la hechicera y consideran su tratamiento posible solo mediante los mtodos tradicionales. C O N C L U S I N : En el presente estudio, se ponen en evidencia las creencias y las opciones de tratamiento de la tos crnica en Ruanda, con implicaciones importantes que se deben tener en cuenta en los programas de control de la TB.