Healer shopping in Africa: new evidence from rural-urban qualitative study of Ghanaian diabetes experiences

Data collection

I used three methods to gather data: individual interviews, group interviews, and ethnographies (table 1). Individual interviews tap into individual biographies and thus “provide[s] the basic data for the development of an understanding of the relations between social actors and their situation.”17 Individual interviews were conducted with 20 participants to gather subjective accounts of illness experiences and practices.

Whereas the individual interview relies on the subjective account, the group interview examines the structure, process, and outcomes of social communication. This provides insight into “how knowledge, and more importantly, ideas, develop and operate within a given cultural context.”18 Six focus groups were carried out with 44 rural and urban people to gather accounts and meanings of shared experience. Table 2 shows the demographics of the people who took part in individual or group interviews.

Both individual and group interviews are inadequate as tools to examine the social interactions and practices that shape subjective worlds. Ethnographies facilitate powerful analysis of these processes, as their object of study is locally produced “through the activities of particular people in particular settings.”19 I did six month ethnographies with three people with diabetes and significant others in their life worlds (table 3). These constituted multiple interviews (at least three) with each primary interviewee, observation of participants, and reflexive field notes, during monthly home visits.

Initial interview guides were informed by key empirical and policy themes within local and regional discussions on diabetes, as well as health psychology perspectives on chronic illness. I further refined the guides after interviews with 11 Ghanaian healthcare and policy experts. Guides for rural interviews, conducted in the local Akan language, Bono; two urban interviews conducted in the Akan languages Fanti and Twi; and one urban focus group discussion, conducted partially in Ga language, were further refined to reflect culturally appropriate meanings, after consultation with local field and translation experts. I am British, of Ghanaian (Akan) parentage. I can understand, read, and speak Fante and Twi fairly well but have less intricate knowledge of Ga and Bono. I did all urban English, Twi, and Fanti interviews and ethnography myself. A local field assistant did the urban Ga focus group, and three field assistants did rural individual and group interviews and ethnography interviews with myself in attendance as facilitator and note taker. Group discussions lasted from an hour to two hours. One rural group discussion lasted six hours over a three day period. Individual interviews lasted between 40 and 90 minutes. Permission was sought and granted to record all interviews and discussions. All participants were remunerated for participation. English interviews were transcribed verbatim. Bono, Twi, Fanti, and Ga transcripts were translated and transcribed jointly by bilingual research assistants and myself.

Analysis

I used the qualitative analysis package Atlas-ti to code all transcripts. Coding aimed to capture the range of views on the empirical categories of health, chronic illness and diabetes knowledge, knowledge and use of pluralistic medical systems, diabetes experiences, and illness practices (box 1). Informed by social representations theory, I paid attention to consensus, conflict, and absences across group and individual narratives.

With individual interview transcripts, I contextualised consensus, conflict, and absence within a singular unfolding narrative. I used a two stage process for focus group transcripts: firstly, I identified broad areas of group consensus, conflict, and absences; secondly, I followed through individual narratives systematically, where appropriate, to examine the progression and nature of interpersonal consensus, conflict, or changes in views within the group context. For each ethnography, I cross referenced individual transcripts, observation material, and field notes to identify consensus and divergences within a family unit. This systematic process facilitated the identification of converging and diverging discourses, motivations, relations, and practices between and within groups (such as urban and rural groups, people with type 1 v type 2 diabetes).

I used an intra-coder reliability test—a process that measures the consistency of a single coder coding the same material twice with a time interval—for the local language transcripts, given their complex multi-ethnic and language dimensions.20 A second coder subjected a selection of English transcripts to an inter-coder reliability test.

Biomedical management

Most participants sought biomedical care in the first instance and endorsed biomedical management as ideal self care practice. This stemmed from two factors. Firstly, biomedical, ethnomedical, and faith healing systems were subjected to public critique in terms of technical or practical knowledge of health problems, technological expertise, accessibility, and ethics; all three had strengths and weaknesses across these criteria, depending on the health problem. Chronic illnesses belonged to the category of complex contemporary illnesses; here, doctors were endorsed as diagnostic and pharmacological experts, and biomedical knowledge systems and practices were privileged over ethnomedical and faith healing systems, both described as diagnostically and pharmacologically inept and lacking professional ethics.

Secondly, biomedical explanations of and solutions to causes of diabetes were publicly legitimised. Diabetes was attributed to an array of interchangeable causes: high sugar diets, heredity, physiological imbalance, toxic foods, and spiritual disruption. High sugar diets and physiological imbalance were deemed primary causes, for which biomedicine offered the most successful solutions. Although the level of sophistication of biomedical explanations of diabetes increased with education, both uneducated and educated participants held sufficient biomedical knowledge on diabetes to prioritise drug and dietary management over alternative medical and self care practices.

Some participants remained committed to biomedical management. This group had three distinct or interlinked experiences: people who had lived with diabetes in the long term; people who had experienced few or no life disruptions from diabetes; and people who lived with other serious conditions, such as hypertension and prostate cancer. Practical daily routines were geared towards controlling symptoms through drug and dietary management and routine lifestyle change (drinking less, taking up exercise).

However, for most participants, biomedical management was undermined by two key factors. Firstly, daily drug and dietary management routines imposed a considerable psychosocial burden on all participants. Secondly, the high cost of biomedical drugs and recommended foods undermined the commitment of low income and financially destitute groups to long term engagement in biomedical care. For example, rural participants paid 60 000-90 000 cedi (£6-9; $11-16.5; €8.8-13.2 at the time of the study) for a monthly supply of insulin. This constituted 60% of the monthly income of those on a minimum daily wage of 5000 cedi (£0.50).

Spiritual action

There were two dimensions to spiritual action. Within rural communities, diabetes was partially attributed to two spiritual causes: witchcraft (malevolent action caused by “witches”—people possessing mystical powers used to harmful ends) and sorcery (malevolent action caused by ordinary people through knowledge and use of spells or rituals). For some, these constituted primary theories and demanded traditional religious intervention. For instance, five participants sought traditional religious healing initially because the suddenness and mysteriousness of their symptoms were attributed to witchcraft. Consultation was short lived, and all turned to biomedical intervention as their physical health deteriorated. For most, spiritual causal theories were secondary to biological, lifestyle, and heredity theories and were drawn on only if a discrepancy existed between their pre-diagnosis knowledge of diabetes and their subjective experiences. Crucially, this group made a distinction between legitimate and illegitimate spiritual action: the first occurred within the Christian sphere and the second in the traditional religious sphere. This distinction resonated with broader rural and urban discussions of general illness practices.

The second kind of spiritual action, Christian prayer, was more widespread and had a consensual motivational basis. Most rural and urban participants were practising Christians, who stressed that as everyday experiences and practices required spiritual mediation—prayer and Christian communion—diabetes treatment required spiritual mediation. Spiritual intervention offered by faith healers was rejected because of empirical failures of faith healing. People who had sought faith healing in the past drew attention to faith healers' inability to diagnose and their emphasis on fasting as a primary healing technique. Several participants attributed past complications to ineffectual faith healing practices.

Cure seeking—passive and active

Although participants understood and stressed that diabetes was incurable, many hoped for a cure. Discourses on “hope for a cure” centred predominantly on a collective need for relief from the physical and psychological burden of drug and dietary management and was synonymous with hope for future biomedical breakthroughs. This need shaped passive cure seeking, which constituted intermittent healer shopping, particularly during acute illness phases (see box 2). Socioeconomic status and geographical location shaped distinct healer shopping choices. Wealthy urban people were more likely to doctor shop. When they healer shopped for ethnomedical treatment, a persistent preference was for scientifically approved ethnomedical drugs. Low income urban and rural groups, with fewer material resources and restricted access to biomedical services and regulated ethnomedical services, healer shopped within mainly unregulated ethnomedical services; some augmented ethnomedical care with biomedical dietary advice. Most participants reported negative experiences from unregulated ethnomedical treatments. Problems included unspecified dosages, unclear information about drug action, and side effects of drugs. By contrast, wealthy urban groups endorsed the pharmacological benefits of scientifically approved ethnomedical drugs. Although participants remained highly critical of unregulated ethnomedicine, few had recourse to desired alternatives.

A minority of participants believed that diabetes could be cured; this group constituted mainly rural, and low income, people recently diagnosed with diabetes or experiencing acute physical disruption. Their belief was fuelled by unwavering faith in God's ability to heal all health problems. This group engaged in active cure seeking, which constituted persistent and committed healer shopping between ethnomedical and faith healing systems. Some combined ethnomedical drug treatment with biomedical dietary management.

Medical inaction

Medical inaction constituted passive withdrawal from drug and dietary management. This, according to the narratives of low income groups, was underpinned by chronic suffering characterised by severe financial difficulties, escalating complications, and a lack of family and social support. During fieldwork, five rural and urban participants were medically inactive. They made frequent references to depression and suicidal thoughts. They all emphasised that death was a desirable alternative to chronic suffering. Participants in Nkoranza received psychosocial support from their self help group, but this did not address the critical financial support needed for diabetes management.

Although this group had passively disengaged from medical care, they were spiritually active. They engaged in constant prayer, simultaneously seeking death as an end to chronic suffering and a miracle (such as a change in socioeconomic circumstances, a benefactor) to have their health restored. Additionally, and despite awareness of the shortcomings of faith healers, this group relied on faith healing (see box 3).

Healer shopping: three findings on choices and motivation

Firstly, healer shopping for ethnomedical treatment was not the primary choice for most participants. Biomedical management was a preferred persistent choice. However, the high cost of biomedical care, coupled with the psychosocial burden of diabetes, set most people on a passive cure seeking course; a minority either actively sought cures or became medically inactive. Crucially, as the processes of cure seeking and medical inaction showed, healer shopping did not occur only within ethnomedical systems but across biomedical, ethnomedical, and faith healing systems. Different choices depended on people's socioeconomic status and access to these pluralistic services, as well as on whether participants hoped for a biomedical breakthrough or believed in a (Christian) spiritual cure for diabetes.

Secondly, the findings challenged the claim that healer shopping was underpinned by problematic spiritual causal beliefs, the endorsement of traditional religious healers as spiritual experts, and the need for a cure. The motivations underpinning healer shopping drew on but transcended culture; the need for affordable and effective pharmacological treatment and relief from the physical and psychosocial burden of diabetes shaped medical care and self care. Spiritual causal theories held cultural legitimacy within the Ghanaian public sphere, but their structure and functions were complex. Two spiritual causal theories of diabetes—witchcraft and sorcery—emerged as primary or secondary theories for a minority of rural participants. As primary theories, witchcraft and sorcery took precedence over biological and lifestyle theories. As secondary theories, they were subordinate to alternative theories but assumed importance when the onset of diabetes seemed to have no biological or lifestyle basis. More importantly, spiritual causal theories did not shape illness practices as predictably as is reported in the literature. People who accessed traditional religious healing abandoned this for biomedical treatment as their physical health deteriorated. Others attributed their diabetes to spiritual causes but did not seek spiritual treatment. Evangelical Christianity has assumed increasing relevance as a mediator of solutions to everyday problems, including those of health and disease, in Ghana.21 Within this context, Christian faith healers wield legitimate and greater therapeutic power over traditional religious healers. Most participants distinguished between legitimate spiritual action, occurring within the Christian sphere, and illegitimate spiritual action within the traditional religious sphere; they were more likely—in the event of extreme physical and economic disruption—to seek faith healing rather than traditional religious healing.

Finally, biomedical knowledge and technical expertise on diabetes were privileged over ethnomedical and faith healing versions. Thus most educated and uneducated groups sought and held sufficient biomedical knowledge to understand that diabetes was incurable and to commit to biomedical management. For most people, hope for a cure was synonymous with hope for biomedical breakthroughs in diabetes treatment and distinct from minority belief in a cure that was subsumed within an all-encompassing faith in the healing power of a Christian God.

Challenges for diabetes care in Ghana

These findings may not be generalisable to other regional and ethnic contexts. For example, religious and policy responses might be different in northern Ghana, which is predominantly Muslim and served by a better equipped voluntary biomedical sector. However, the theory driven, multi-site, multi-experiential qualitative approach adopted has uncovered clear patterns of illness practice across key demographic arenas such as geographical location, socioeconomic status, diabetes status, and support systems; these can instruct new developments in diabetes policy and service delivery. For instance, wealthy urban people tend to doctor shop or healer shop for regulated ethnomedical services (highly concentrated in the urban south); they might find restricted diets oppressive, but they can afford these recommended alternatives. The health of this group is least compromised. For low income and rural people, and especially those on expensive insulin treatments, the consequences of living with diabetes are more severe; this group has least access to biomedical and regulated ethnomedical services, they cannot afford recommended foods, and they are likely to gravitate towards faith healers as a spiritual response to chronic suffering.

Challenges for health care and policy development lie in providing affordable drugs and food, prioritising psychosocial support, and regulating faith healing practices. Policy discussions suggest that the provision of affordable or subsidised pharmaceutical drugs is unlikely to occur in the short term.11 Two partial solutions exist, however. A recently implemented national health insurance scheme provides cover for chronic conditions and may alleviate the financial burden for people who can afford premium payments. The second solution lies in the expanded use of scientifically approved ethnomedical drugs, which have proved effective for urban people on oral treatment. The clinical and policy climate is conducive for such a development; ongoing toxicological tests and clinical trials indicate that effective herbal medicines for diabetes exist, and research suggests that ethnomedical practitioners are willing to submit their drugs to trials if costs are covered and that doctors are willing to refer patients to regulated ethnomedical services.22 23 This development may serve simultaneously as a regulatory exercise aimed at making ethnomedical diabetes care more transparent and the implementation of a referral system more sustainable.

The high cost of recommended foods is neglected in the local literature but deserves empirical work that engages with the agricultural and food export sectors. The lack of social and psychological services and a welfare system in countries such as Ghana places the burden of psychosocial and economic support on families. Low income families are unlikely to provide consistent, long term support to their diabetic relatives.24 Self help groups are becoming important providers of psychosocial support in Africa and other low income regions.25 26 In Nkoranza, the diabetes self help group provided valuable support, especially for people abandoned by their families and community. Lessons can be drawn from this group for the other study sites. Finally, the role of faith healers as healthcare providers for people with diabetes deserves research and policy attention. Future research will have to examine the feasibility of developing educational and regulatory interventions that aim, simultaneously, to enhance their psychosocial and spiritual role and minimise damaging aspects of their healing repertoires.