By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
Corresponding Author Peter J. Winch, Department of International Health, Bloomberg School of Public Health, Johns Hopkins University, 615 North Wolfe Street, Baltimore, MD 21205, USA. Tel.: +1-410-9559854; Fax: +1-410-6147553; E-mail: firstname.lastname@example.org
Objective To explore Bambara language terminology and classification for locally available antimicrobial medicines in order to better target promotional messages and improve evaluation measures in Bougouni District, Mali.
Methods Mothers (n = 20) and drug vendors (n = 15) were asked to freelist medicines used to treat childhood illnesses, and to identify all medicines that corresponded to each of the listed terms from an array of medicines displayed with their packaging.
Results Each Bambara language medicine term can refer to numerous modern medicines, and each modern medicine has several Bambara names. The term nivakini (Nivaquine), often translated as ‘chloroquine’, refers to a wide range of medicines commonly used to treat malaria, many with no antimalarial effect. Antibiotics were also identified as common treatments for malaria. Mothers and vendors used slightly different terminology when discussing treatments for malaria, and sometimes employed the same term to refer to different medicines. Neither mothers nor vendors clearly differentiated between antimalarial medicines. Colour, shape and packaging play a large role in their recognition, classification and use.
Conclusions Current household survey methods are likely to provide inaccurate estimates of appropriate treatment of febrile illness, and thus alternative approaches are recommended. In introducing new malaria treatments, malaria control programmes should differentiate recommended treatments from other medications through distinctive packaging, drug appearance and appropriate Bambara language terms.
Objectif Explorer la terminologie et la classification en langue Bambara des médicaments antimicrobiens disponibles localement dans le but de mieux cibler les messages promotionnels et d'améliorer les mesures d’évaluations dans le district de Bougouni, Mali.
Méthodes Les mères (n = 20) et les vendeurs de médicaments (n = 15) devaient faire la liste des médicaments utilisés dans les maladies infantiles et identifier les médicaments correspondant à chacun des termes listes dans un étalage de médicaments présentés dans leur emballages.
Résultats Chaque nom de médicament en langue Bambara pouvait correspondre à plusieurs médicaments modernes, et chaque médicament moderne avait plusieurs noms Bambara. Le terme nivakini (Nivaquine), souvent traduit par «chloroquine», faisait référence à un large éventail de médicaments habituellement utilisés pour le traitement de paludisme, mais dont beaucoup d'entre eux sont dépourvus d'effet anti-paludéen. Les antibiotiques étaient également identifiés comme traitement classique du paludisme. Les mères et les vendeurs de médicaments utilisaient des terminologies légèrement différentes lorsqu'ils discutaient à propos des traitements contre le paludisme, et ils utilisaient parfois le même terme à propos de médicaments différents. Ni les mères ni les vendeurs ne faisaient clairement la différence entre les médicaments anti-paludéens. La couleur, la forme et l'emballage jouaient un rôle important dans leur reconnaissance, classification et utilisation.
Conclusions Les méthodes actuelles d'enquêtes de ménages fournissent probablement des évaluations imprécises des traitements appropriés des maladies fébriles et ainsi des approches alternatives sont nécessaires. En introduisant des nouveaux traitements anti-paludéens, les programmes de contrôles anti-paludéens devraient différencier les traitements recommandés des autres traitements grâce à des emballages distincts, à l'aspect du médicament et à des appellations appropriées en langue Bambara.
Mots clefs paludisme
Objetivo Explorar la terminología de la lengua Bambara y la clasificación de los medicamentos antimicrobianos disponibles localmente, con el fin de mejorar el objetivo final de los mensajes promocionales y las medidas de evaluación en el distrito de Bougouni, en Mali.
Métodos Se solicitó a madres (n = 20) y a vendedores de medicinas (n = 15) que elaboracen listas de medicinas utilizadas para tratar enfermedades infantiles, e identificar todos los medicamentos que correspondían con cada uno de los términos listados en un conjuntos de medicamentos mostrados con sus embalajes.
Resultados Cada término de medicamento en lengua Bambara puede referir a numerosas medicinas modernas, y cada moderno medicamento tiene varios nombres en Bambara. El término nivakini (Nivaquina), muchas veces traducido como cloroquina, se refiere a un amplio espectro de medicinas comúnmente utilizadas para tratar la malaria, muchas de ellas sin efecto antimalaria. También se identificaron antibióticos como tratamientos comunes para la malaria. Las madres y los vendedores utilizaban terminología ligeramente diferente cuando discutían acerca del tratamiento de la malaria, y muchas veces empleaban el mismo término para referirse a diferentes medicamentos. Ni las madres ni los vendedores diferenciaban claramente los medicamentos antimalaria. El color, la forma y el envoltorio juegan un importante papel en el reconocimiento, clasificación y uso.
Conclusiones Los actuales métodos de encuestas en los hogares probablemente proporcionarán estimaciones poco confiables sobre los apropiados tratamientos de enfermedades febriles, por lo que se recomiendan estrategias alternativas. Al introducir nuevos tratamientos contra la malaria, los programas de control de la enfermedad deberán diferenciar tratamientos recomendados de otros medicamentos, a través de embalajes distintivos, apariencia del medicamento, y terminología apropiada en lengua Bambara.
In the context of increasing resistance to chloroquine (CQ) and other drugs, most African countries are adopting new malaria treatment policies, in many cases making the transition to artemisinin-based combination therapy (ACT) as first-line treatment (Bloland et al. 2003). The introduction of these treatment policies, however, occurs in a context in which antimalarials such as CQ have become both commoditized (commodities actively traded in both the formal and informal sectors) (van der Geest & Whyte 1989) and indigenized (integrated into local understandings of health and illness) (Etkin et al. 1990). In Mali and other African countries, people are highly familiar with currently available modern and traditional malaria treatments that are significantly less expensive than the more effective ACTs (Williams et al. 1999; Tarimo et al. 2001). When combination therapy is formally introduced in authorized health facilities and pharmacies, it is likely that the less expensive drugs such as CQ will continue to be widely available from unauthorized points of sale such as markets, ambulatory vendors and shops (Baume et al. 2000; Bloland et al. 2003). Consumers are likely to continue to purchase this familiar medicine, not only because of the lower price but also because of characteristics such as its antipyretic and anti-inflammatory effects, its rapid action and the bitter taste that many people have come to associate with efficacy (McCombie 1996; Williams et al. 1999). Despite its decreasing biomedical efficacy, CQ has come to be endowed with high social efficacy (Whyte et al. 2002): it is still widely recognized as the treatment one should use for malaria (Williams et al. 1999; Tarimo et al. 2001). Health officials and health educators must find ways to differentiate the new drugs from existing malaria treatments and explain the advantages of using the new treatments exclusively.
Efforts to monitor the implementation of interventions to promote correct management of fever in young children face the challenge of distinguishing between cases in which effective antimalarial treatments have been administered and those in which less effective or ineffective antimalarial treatments have been used. One of the indicators commonly used by Roll Back Malaria (RBM) to monitor the outcomes and impacts of RBM efforts is the proportion of children less than 5 years old with fever who received an appropriate antimalarial treatment with 24 h (Remme et al. 2001). When this indicator is measured in household surveys, it is assumed that it is possible to distinguish between appropriate and inappropriate medications used to treat a sick individual, to distinguish, for instance, between CQ, sulphadoxine–pyrimethamine (SP), ACTs and other medications. However, studies of local terminology for modern medicines suggest that there is not always a one-to-one correlation between a local medicine term and the brand name or even generic name for a pharmaceutical product (Etkin et al. 1990). Therefore, such distinctions may be difficult to make, particularly when the majority of survey respondents are illiterate.
The current research was conducted to better understand how people classify available antimalarials and examine local terminology for these medicines in order to better target promotional messages and improve evaluation measures. Past education and evaluation efforts conducted by Save the Children, USA, and organizations in Mali have translated the Bambara term nivakini as CQ. One aim of this study was to assess the accuracy of this translation and to develop a list of all terms used to describe malaria treatments in order to inform and improve future education and research projects. The goals of this article are to describe Bambara terminology for antimalarial and antipyretic medicines in the Bougouni District, as well as to explore the validity of the assumptions inherent in the use of methods currently used to assess levels of appropriate treatment of febrile illness in this region. Additionally, the article presents suggestions for the introduction and marketing of new approved antimalarials in the region, as well as for monitoring and evaluation efforts that must differentiate between appropriate and inappropriate treatments for malaria.
The research presented here was conducted in the Bougouni District of the Sikasso region of southern Mali. The region is predominantly rural, and the majority of people engage in subsistence agriculture. The dominant ethnic group is the Bambara, with a small minority of Fulani residents, and the culture is largely polygynous and patrilineal. The primary language spoken is Bambara, although formal education is conducted in French. Literacy rates are extremely low; close to 90% of mothers in the Sikasso region cannot read at all (Demographic and Health Surveys 2001a). Access to the media is relatively limited; in 2001, 4.2% of women read a newspaper at least once a week, 20.2% watched television at least once a week and 65.3% listened to the radio at least once a week; yet, 31.4% of women reported no exposure to the media (Demographic and Health Surveys 2001a).
Within this region, malaria transmission is seasonal; 6 months of high transmission occur between June and November every year (Tanser et al. 2003). A palpable spleen rate of 58.1% (indicative of high prevalence of malarial disease) was found among children randomly sampled in the community during the rainy season (Thèra et al. 2000). Resistance of malaria parasites to CQ had been relatively low in Mali for many years, with CQ-sensitive parasitological response between 83% and 87% and adequate therapeutic responses to CQ between 86% and 90% (Plowe et al. 2001). As recently as 2002, an adequate clinical response to CQ of 83% was found in Bancoumana (Doumbo et al. 2005). However, a study conducted in the Bougouni District between October 2003 and March 2004 documented far higher rates of CQ resistance. Early therapeutic failure was estimated at 30%, controlling for reinfections; and the combined parasitological and therapeutic adequate response was estimated to be only 10% (de Radigues 2006). These findings prompted authorities to speed discussions regarding changes to malaria treatment policies in Mali.
In this setting, antimalarial medicines are available from a variety of authorized sources: government health centres, pharmacies and community health workers. Despite the variety of authorized sources, a large percentage of modern medicines are purchased from unauthorized sources: shops that sell CQ along with candy, powdered milk and laundry soap; market vendors who display their brightly coloured wares in mayonnaise jars arranged on low wooden tables and ambulatory vendors who wander around town with baskets of medicines and sundries perched atop their head (Djimde et al. 1998). A wide range of pharmaceutical products are sold by these vendors, from vitamins and menthol candy to codeine and antibiotics.
This qualitative research project was conducted in three villages of the Zantiebougou health zone during a 7-week period in October and November of 2004. In order to maximize the chances that study participants would have encountered modern medicines from a variety of sources, the selected villages were characterized by the presence of a village drug kit and a weekly market, and were located relatively close to a paved road and less than 20 km from the community health centre (Centre de Santé Communautaire). The members of the qualitative research team were local residents who were able to read and write Bambara, had prior qualitative research experience and had participated in a 5-day training in qualitative research methods prior to the start of data collection.
Ethical approval was granted by the Committee on Human Research at the Johns Hopkins Bloomberg School of Public Health and by the Institutional Ethics Committee of the University of Bamako, Faculty of Medicine, Pharmacy and Dentistry.
Freelisting involves asking respondents ‘what are all the kinds of X’ and repeatedly probing until the respondent can think of no additional terms (Weller & Romney 1988). Mothers (n = 20) and drug vendors (n = 15) were asked to freelist all modern medicines used to treat childhood illnesses in general, and to describe the specific indications for each medicine. They then were asked to freelist all modern medicines used to treat specific local illnesses associated with fever and respiratory symptoms. The decision to ask about medicines used to treat local illnesses compatible with both malaria and respiratory illness was based upon the fact that there is extensive overlap between symptoms of malaria and pneumonia, a fact that sometimes makes it difficult for health professionals, let alone parents, to distinguish between the two illnesses (O'Dempsey et al. 1993; English et al. 1996; Kallander et al. 2004). Next an array of locally available modern medicines was displayed, and the mothers and vendors were asked to identify all of the medicines that corresponded to each of the previously listed terms (Figure 1). Subsequently, participants were asked to indicate the most typical or ‘best’ example of a medicine corresponding to that term. Finally, participants were asked to give the name for any medicines in the array that they had not previously identified, and to describe the corresponding indications. If participants indicated that they did not know the name of a given medicine, that response was recorded. In order to reduce the possibility that participants would feel pressure to guess or fabricate the names of unfamiliar medicines, interviewers were instructed not to probe after a respondent indicated that he or she did not recognize a given medicine.
The medicines were displayed along with their packaging, so as to allow for the possibility that medicines are identified by the appearance of packaging and not only by the appearance of the medicine itself. While parents often buy single capsules or tablets without their original packaging, most medicines in the local markets are displayed in the original packaging, and thus it seemed important to allow for package recognition, particularly because many of the medicines themselves are similar in appearance. The decision to include packaging was also based upon the fact that the vendors interviewed were likely to depend, at least in part, upon package recognition when purchasing and selling their wares. The medicines used in the array were purchased from market vendors by the interviewers who, using local illness terms identified in earlier semi-structured interviews (data not presented here), requested all medicines used to treat cough, fever, malaria and respiratory illness. The interviewers also purchased all medicines specifically mentioned by mothers during previous interviews. Medicines sold by the government health centre and/or by the community health workers for the treatment of malaria and pneumonia were also included in the array. A similar medicine array was used by Caroline Bledsoe during her research on western pharmaceuticals among the Mende of Sierra Leone, although she did not include samples of the packaging along with the medicines, and she asked about the uses but not the names for each medicine (Bledsoe & Goubaud 1985).
For a list of all medicines included in the medicine array, as well as their contents, see Table A1, Appendix A. Given the fact that one aim of this research was to learn how respondents classify the wide range available medicines and to investigate questions such as ‘Are different brands of medicines with similar contents referred to by the same name?’ and ‘Do parents differentiate between medicines with different contents?’, brand names will be used throughout this article. Each brand name, however, will be followed by the generic drug name or drug category (i.e. ‘cold medicine’) in parentheses.
Data management and analysis
The modern medicine naming array and pile sorting exercises were recorded using handheld tape recorders. The interviewers completed detailed data collection sheets during the interview, and the data were subsequently entered into Microsoft Excel for organization and analysis. After the exercises, interviewers listened to the tapes and amplified the notes they had written. During the course of the research, the research team met periodically to discuss the data collected prior to that point, to come to a consensus regarding the spelling of each Bambara term, to discuss variations and discrepancies in the data and to modify the data collection tools as needed.
Bambara medicine terms
In the remainder of this article, words in italics are Bambara or French language terms as they were mentioned by the study participants. The freelisting exercises generated a far greater number and variety of terms than anticipated, eliciting 88 medicine terms from mothers and 113 medicine terms from vendors. Very few terms, however, were mentioned by a large proportion of respondents. The majority of terms were mentioned by fewer than 25% of mothers and less than 30% of vendors. Table 1 illustrates the number of terms that were mentioned by varying numbers of respondents, presenting the number of terms listed by 1–5, 6–10, 11–15 and 16–20 respondents, respectively. Of the terms listed, 26 were mentioned as treatments for local illnesses identified in a previous series of interviews as those with symptoms compatible with those of clinical malaria. Table 2 presents the treatments mentioned by mothers and vendors; and mothers alone or vendors alone for the treatment of local illnesses associated with fever, convulsions and anaemia. The nine terms most frequently mentioned as treatments for these illnesses, with approximate translations in parentheses, were aspirini (Aspirin), daga/dakan (brand name for Nigerian paracetamol), fatôkèni (little crazy man), kunbilèni (pill with a red head), nivakini (Nivaquine), paracetamoli (Paracetamol), pikiri (injection), sosoni (little mosquito) and sumayafura (drug for ‘malaria’). Of these, nivakini was the term most commonly mentioned by both mothers and vendors, and it was the only term mentioned by both groups as a treatment for sumaya, the local illness term most commonly associated with clinical malaria, and used by health workers as a translation for malaria.
Table 1. Number and percentage of elicited Bambara medicine terms mentioned by varying numbers of mothers and vendors respondents
Mothers (n = 20)
Vendors (n = 15)
Number of terms mentioned by 1–5 respondents
Number of terms mentioned by 6–10 respondents
Number of terms mentioned by 11–15 respondents
Number of terms mentioned by 16–20 respondents
Total number of terms mentioned
Table 2. Bambara terms mentioned by more than one respondent as treatments for local illnesses compatible with clinical malaria
Bambara terms for malaria treatments mentioned by both mothers and vendors
Bambara terms for malaria treatments mentioned by mothers only
Bambara terms for malaria treatments mentioned by vendors only
Pikiri (pikiri ji, pikiri mugu, sumaya pikiri, chèmanchèpikiri)
Each of these nine terms was associated with several different modern pharmaceuticals, often with very different biomedical indications. These terms refer primarily, but not exclusively, to antimalarial or antipyretic drugs. This trend is illustrated by Figure 2, which illustrates the full range of medicines identified as examples of nivakini as well as those indicated as the ‘most typical’ examples, by mothers and vendors. For example, 19 mothers (100%) used the term nivakini to refer to Nirupquin (CQ), but as many as 13 mothers (68%) used this term to refer to diaminodiphenylsulphone (DDS), a drug that has limited antimalarial effects when used alone, although it has been used in combination with chlorproguanil as a treatment for malaria (Bukirwa et al. 2004). Additionally, a multivitamin and a zinc supplement were among the medicines associated with this term (Figure 2). Mothers often failed to distinguish between different types of white tablets, even when these medicines with similar appearances were displayed with their distinctive packaging. In general, a wide variety of small white tablets were indicated whenever mothers were asked to point to examples of nivakini.
The case of sumayafura (medicine for malaria), another term frequently used by mothers to refer to treatments for malaria indicates, however, that while all small white tablets might be considered treatments for malaria, all treatments for malaria are not considered to be equivalent. No antimalarials were included in the range of small white medicines indicated by mothers as examples of sumayafura. This suggests that for mothers, contrary to what one might expect, the term nivakini (derived from the brand name Nivaquine) is inclusive of medicines classified as sumayafura (medicine for malaria), but that the inverse is not true. Nivakini seems to be a cover term for all medicines used to treat malaria (including many white tablets with no proved antimalarial or antipyretic effects), while sumayafura refers primarily to antipyretics and not to antimalarials.
When asked to identify medicines for the treatment of local illnesses resembling malaria, vendors were more likely than mothers to point only to medicines with antimalarial or antipyretic effects. Vendors used the term nivakini to refer to a variety of different available antimalarials, without differentiating between Maladrine (CQ), Nirupquin (CQ), CQ phosphate (generic brand), Melaxime (SP), and quinine (Figure 2). However, three vendors used this term to refer to codeine. Multivitamins, DDS and the cough medicines Carbetux and phenylbutazone were also associated with this term. The term sosoni (little mosquito), mentioned by all but two vendors as a treatment for local illnesses resembling malaria, was much more specifically applied. With one exception, all vendors used this term to refer exclusively to antimalarial medications, but again, they did not clearly differentiate between CQ and SP.
Of particular note is the fact that two of the terms included in the list of treatments for local illnesses resembling malaria –kunbilèni (pill with a red head) and kunbilèni jèman (pill with a red head that is white) – are commonly associated with antibiotics. Kunbilèni generally refers to tetracycline (sold as an yellow and red capsule), but is also used as a name for cotrimoxazole, amoxicillin and ampicillin. This is similar to the way in which the Hausa term farin kapso, originally used to refer to chloramphenicol, expanded in usage to incorporate penicillins, sulphonamides and vitamins (Etkin et al. 1990). Kunbilèni jèman is used interchangeably to refer to both CQ phosphate and chloramphenicol; both are sold in the form of white capsules but they have very different biomedical indications.
Differences in terminology use between mothers and vendors
There was a great degree of overlap in the list of terms used by mothers and vendors to describe treatments for local illnesses compatible with clinical malaria, as illustrated by Table 2; however the overlap was not complete. Vendors were more likely to use terms derived from the brand or generic names of medicines, such as melaxini (Melaxime) and metro (metronidazole). In contrast, mothers tended to use names referring to the symptoms or illnesses the medicine is commonly used to treat, or to the physical appearance of the medicine. Additionally, as illustrated by Figure 2, even when mothers and vendors use the same term, they do not always use it to refer to the same pharmaceutical.
There was very little agreement between mothers and vendors when it came to Bambara terms for medicines used to treat sumaya, the local illness term most commonly translated as ‘malaria’. Nivakini was the only term mentioned by both mothers and vendors as a treatment for sumaya. Sumayafura (medicine for ‘malaria’) was mentioned by 100% of mothers as a treatment for sumaya, but by none of the vendors. All of the vendors mentioned daga, melaxini and sosoni, as treatments for sumaya, but these terms were mentioned by, at most, one mother.
The naming of antimalarial medicines
The Bambara terms for modern pharmaceuticals appear to evolve from specific characteristics of the medicines they name: some refer to colour, some to shape or form, some to a picture or word on the packaging, and still others to the symptom or illness treated with the medicine or to its effect on the body. These different ways of naming modern medicines are illustrated through examples presented in Table 3.
Table 3. Ways of naming modern medicines in Bougouni District, Mali
Type of medicine term
Terms originating from generic or brand names
Nivakini (Nivaquine brand chloroquine)
Terms that refer to the colour of the medicine
Kunbileni (‘pill with a red head,’ referring to the red end of bivariate capsules)
Terms that refer to the packaging of the medicine
Sosoni (‘small mosquito’, referring to antimalarial medicines with pictures of mosquitoes on the packaging)
Terms that refer to the symptom or illness treated
Sumayafura (malaria medicine)
Terms referring to the effect of the medicine on the body
Fatokeni (‘little crazy man’)
Terms referring to a cultural or historical figure or event
Yaro (named for a nurse who distributed large quantities of this drug in the region)
Of the various defining drug characteristics, colour and packaging seem to play the largest role in the naming, classification and recognition of modern pharmaceuticals. The medicines whose names were almost universally agreed upon, and which were least often conflated with other medicines, were Ibumol (ibuprofen, paracetamol and caffeine) and Sudrek (paracetamol and caffeine), both pain medicines with very distinctive colouring. Ibunol is a dark blue and light blue dichromatic capsule, while Sudrek is a dichromatic tablet with one pink side and one white side. Melaxime (SP) and Maladrine (CQ), which were also generally recognized as being distinct from other medicines (although not from each other), have pictures of mosquitoes on their packaging. Respondents had difficulty distinguishing between medicines with similar appearances, as illustrated by the fact that the terms both nivakini and paracetamoli were used to refer to a wide range of white tablets, and white tablets were generally given a variety of different Bambara names. All dichromatic capsules, regardless of colour or contents, were called kunbilèni (red head).
Improving strategies for antimalarial drug deployment
Mothers and vendors in the Bougouni District use a large variety of terms to refer to modern medicines; however, a given term may be used to refer to many different modern medicines. Mothers, in particular, have difficulty distinguishing between different medicines with similar appearance. Hence, it is likely that mothers are highly dependent upon the counsels of vendors when purchasing medicines in the market, although additional research must be conducted to test this hypothesis. It is likely that the vendors’ drug recognition and classification practises have a larger impact on actual drug treatment practises than do those of the mothers. For this reason, the vendors’ inability to differentiate between drugs with distinct biomedical indications is particularly important to highlight. While vendors in our study frequently identified the correct biomedical indications for a given modern pharmaceutical, they were not always able to do so. Of particular concern is the fact that vendors, as well as mothers, indicated that kunbilèni is used to treat malaria in children. This term is inclusive of a wide range of antibiotics, none of which are recommended for the treatment of malaria, and one (tetracycline) which is not recommended for children under 5 years. Additionally, even if drug vendors can identify correct indications and dosages for a given medicine, they may not put this knowledge into practice (Igun 1994). Evidence suggests that market vendors often make informal ‘prescriptions’ for a given illness, but rarely offer adequate instructions regarding dosage, indications, side effects and counterindications when medicines are purchased (Abosede 1984; Djimde et al. 1998; Kamat & Nichter 1998; Brieger et al. 2004). For these reasons, interventions to promote appropriate treatment of childhood malaria should emphasize care-seeking to authorized health facilities or trained community-based providers, where the medicines sold for the treatment of malaria can be assumed to actually have antimalarial and antipyretic qualities.
The ideal would be for all cases of malaria to be treated at authorized health facilities with recommended drugs, and then only after consultation, akin to the ‘limitation of drug resistance’ paradigm for antimalarial drug deployment described by Bloland et al. (2003). However, the reality is that barriers of distance, cost and cultural attitudes are often prohibitive and market and ambulatory vendors remain major sources of malaria medications. Medicines purchased from private vendors may be used to treat as many as 87% of cases of presumptive malaria in Africa (Deming et al. 1989; Ruebush et al. 1995; McCombie 1996), and unauthorized drug sellers account for as much as half of all sales of antimalarials sold in developing countries (Foster 1991). In Mali, the majority of consumers choose non-recommended treatments for malaria, even when sanctioned providers limit their first choices to nationally recommended antimalarials (Djimde et al. 1998). Therefore, the more applicable of the paradigms for antimalarial drug deployment of Bloland et al. may be that of ‘improved access’ (Bloland et al. 2003). It is difficult to ensure an acceptable quality of care when implementing through multiple sources of care in the formal and informal sectors. Nevertheless, a number of different interventions focusing on the informal sector have been shown to improve malaria treatment practises. These interventions include training shop keepers about appropriate drug choice and dosage (Le Grand et al. 1999; Marsh et al. 1999, 2004); developing vendor-to-vendor education programmes (Tavrow et al. 2003); training mothers to teach other mothers about the need to promptly administer CQ for malaria (Kidane & Morrow 2000); training community health workers to treat malaria, pneumonia and other conditions (Winch et al. 2005); and public education campaigns (Le Grand et al. 1999), including some that focus on educating school children (van der Geest & Geissler 2003).
Given the fact that informal medicine sales in Mali are unauthorized and unregulated, interventions designed to educate vendors are not particularly practical or feasible at this time, but interventions can and should be designed to help parents to become more informed consumers, empowering them to select a specific brand of medicine sold by vendors, rather than simply asking for sumayafura (malaria medicine) and receiving any one of a number of medicines that vendors recognize as malaria treatments. Education efforts to increase appropriate treatment of malaria must emphasize the distinctive visual characteristics that set appropriate medicines apart from others. This will help to decrease inappropriate use of antibiotics, and thus hopefully limit the development of antibiotic resistance. It will also increase the chances that children with malaria are treated with effective antimalarials as opposed to vitamin or mineral supplements.
When changes to malaria drug policy in Mali are finalized and new drugs are introduced to this region, efforts must be made to distinguish them from existing medicines, through colour, packaging or both. Additionally, marketing the drug with a Bambara name may help to establish that particular name in the local vocabulary and prevent the drug from becoming just one more type of nivakini or sumayafura. Improving the written content of information on package inserts is unlikely to have a great effect, given the low literacy levels in the region, but including graphic representations of both the proper indications and proper dosages of medicines could help to improve treatment practises (Bloland et al. 2003).
Implications for survey research
The results of this study challenge the assumption that the self-report of treatments administered during a febrile episode – either in response to survey questions or chosen from photo or medicine arrays – provides a valid measure of the actual treatments administered. There does not seem to be a one-to-one correlation between medicine terms and modern medicines, which is in agreement with findings from a previous study of Hausa terminology for pharmaceuticals (Etkin et al. 1990). Because distinct terms for specific malaria treatments often do not exist in the local language, self-reports of malaria treatments are unlikely to accurately estimate levels of appropriate treatment. It seems likely that self-reports will overestimate appropriate treatment because all of the terms used for malaria medications are also inclusive of medicines that are not even antimalarials, let alone recommended antimalarials.
The presentation of an array of typical antimalarial medications is often used to aid in the identification of treatments administered [i.e. the malaria module of the Demographic and Health Surveys (2001b)]. This approach seeks to diminish recall bias and to clarify ambiguous medicine terminology; however, it may also provide poor estimates of appropriate treatments, as mothers do not seem to distinguish between medicines with similar appearances even when the medicines are displayed with their distinctive packaging. A mother who had treated a child with codeine (a small white tablet similar in appearance to CQ phosphate tablets) might still point to the samples of CQ tablets when asked to identify the medicines used to treat this illness. However, asking mothers to identify samples in a display is still likely to be more accurate than freelisting alone; it is also possible that mothers who have actually purchased a specific treatment for a sick child are likely to be more able to distinguish it from other medicines with similar appearances than were the mothers in our sample. Finally, the examination of prescriptions, receipts or the actual medicines themselves is often considered the ‘gold standard’, but caretakers are not always able to produce actual physical evidence to indicate which medicines were obtained and used, especially if medications were obtained outside of the formal health sector.
Survey researchers and programme evaluators attempting to assess levels of appropriate treatment of malaria in this region should begin by asking caretakers to list all of the treatments administered for the given case of illness, and then asking to examine any receipts or packages still present in the home. For those medicines that still cannot be clearly identified, interviewers should ask mothers to describe both the medicine and its packaging. These descriptions can then be compared with locally available medicines in order to make an identification. Using this approach, rather than asking respondents to point to medicines displayed before them, minimizes the possibility that mothers will point to medicines at random in an attempt to satisfy an interviewer who is probing for a response. While these strategies are likely to improve the accuracy of survey research, it is important to recognize the limits to what can be learned about medicine use from household surveys, and to complement quantitative research with qualitative investigations.
Challenges and limitations
The identification of vendors willing to participate in the research was extremely challenging, and thus we failed to recruit as many vendors as our research and analysis plan called for. Because we did not want to ask vendors to give up time during profit-making market days, the decision was made to identify vendors at local markets and to schedule interviews at a later date rather than to interview the vendors on the spot. Given the fact that their work is illegal, however, vendors were reluctant to speak with the research team; on one occasion, a vendor went so far as to set up an interview using a false name and address – which was discovered by the interviewer only when he arrived at a village to be told that there was no such vendor there. Vendors also may not have used the same terms when speaking to the interview staff as they use when speaking with mothers, given the fact that the interviewers have a higher level of education than most mothers and are affiliated with Save the Children. In conversations between mothers and vendors, the differences in the two sets of vocabulary may not be as pronounced. While the use of ‘mystery clients’ would have addressed both of these challenges inherent in working with vendors, such a practise was rejected for ethical reasons. Instead, local interviewers were used with the hope of increasing vendors’ willingness to speak candidly about the medicines they sell.
The method employed did not allow us to determine the true degree of consensus regarding the names of modern medicines. The fact that a respondent attached a given name to a given medicine during the naming array exercises does not necessarily mean that it was the only name the respondent knew. It is possible that the respondents shared a common vocabulary for each medicine, but that each respondent only mentioned a few of all of the known terms during the course of the interview. However, the interviewers probed extensively during each stage of the interview with the aim of addressing this potential weakness.
This study elicited important information about the way that modern medicines are named and classified in this region, and illuminates potential strategies for improving the accuracy of survey research and the effectiveness of programmes designed to improve appropriate treatment seeking for malaria. To our knowledge, this is one of the first studies in Mali to examine the knowledge of unauthorized medicine vendors regarding the products they sell. These findings are very pertinent to the current debate about how best to make the transition to combination therapy for malaria in Mali, given the high volume of malaria medicines sold in the informal sector. It is hoped that this study will provide the foundation for future social science research regarding the sale and use of modern medicines in this region, thus allowing researchers to address some of the gaps in the literature, which have been identified in recent reviews (Williams & Jones 2004; Mwenesi 2005). In order to go beyond naming, and to better understand both how local medical knowledge is constructed, and what symbolic importance is attached to modern medicines, there is a need to employ the terminology elicited by this study in further research efforts. Increased knowledge about perceptions of modern medicines will be valuable in the development of marketing and education campaigns to accompany changes in national drug policy, allowing marketers and educators to emphasize medicine attributes that are highly valued by mothers, and to better distinguish new treatments from the existing ones.
The authors thank the following people for their time, energy and support: Maureen Cunningham, Zana Daou, Melisse Murray, Salleye Yattara Touré, Soungalo Koné, Fodié Maguiraga and Ousmane Diakité, as well as the community health workers who work for Save the Children in Bougouni District, Mali, and all of the mothers and drug vendors who volunteered to participate in this research. This project was funded by The Thomas and Carol McCann Innovative Research Fund for Asthma and Respiratory Disease.
Table A1. Names, contents and sources of medicines included in the array of locally available modern pharmaceuticals
Key for sources of medications: CHC, community health center; CHW, community health worker; DV, ambulatory or market drug vendor.
Amodiaquine oral suspension
Amodiaquine 10 mg/ml
Amoxicillin trihydrate BP equivalent to amoxicillin 500 mg
Ampicillin trihydrate BP equivalent to ampicillin 500 mg
Benzylpenicillin potassium BP 1 000 000 IU
Citrate of carbetapentane 100 ml
Chloroquine phosphate tablets
Chloroquine phosphate BP 250 mg
CHC, CHW, DV
Chloroquine phosphate syrup
Chloroquine phosphate BP 160 mg/10 ml
Chloroquine phosphate capsules
Chloroquine phosphate BP 250 mg
Codeine (Terpine brand)
Codeine 100 mg/10 mg
Cold Caps Cophinza
Paracetamol BP 500 mg, phenylpropanolamine HCl BP 12.5 mg, chlorpheniramine maleate BP 2 mg, caffeine 30 mg, orange flavour qs
Paracetamol BP 500 mg, caffeine BP 30 mg, pseudoephedrine HCl BP 30 mg, chlorpheniramine maleate BP 2 mg
Co-trimoxazole oral suspension
Trimethoprime BP 40 mg/5 ml, sulphamethoxozole BP 200 mg/5 ml
Paracetamol BP 500 mg
Dexamethasone 0.5 mg
Drastin analgesic tablets
Aspirin BP 420 mg, caffeine anhydrous BP 30 mg
Piroxicam USP 20 mg
Ibuprofen BP 200 mg, paracetamol BP 325 mg, caffeine BP 30 mg, excipients qs