Study design
The study had an ethnographic design with the main aim to understand the subjective realities of the participants and to get an insiders’ perspective of the folk illness
kimeo and uvulectomy, in other words the “locally valid representations of illness” [
41]. In medical anthropology, the term
illness refers to the patients’ experience of their condition, in contrast to the term
disease, which refers to professional health providers’ understanding of a given medical condition [
9].
The study included ten focus group discussions (FGDs) [
42‐
44] with caregivers and individual in-depth interviews (IDIs) [
45,
46] with four
kimeo cutters. The IDIs and FGDs were conducted by the two authors. We both took part in introducing the study to the participants and asking the introductory questions and follow-up questions. When we introduced ourselves to the participants, we explained that we were health researchers and that we had heard about the folk illness
kimeo in another part of the country where we had worked previously, but that we did not fully understand this illness and thus wanted to learn more. We took care to communicate that we were unbiased and emphasised that we were interested in the participants’ own ideas and experiences [
41]. None of the people we invited to participate refused, and none dropped out.
Both the FGDs and individual interviews were carried out solely in Swahili. This facilitated an atmosphere where the conversations and discussion could have a natural flow since there were no interruptions for translation purposes. In addition to notes during the interviews and FGDs, the first author took ethnographic field notes throughout the study [
45].
FGDs with caregivers We used purposive sampling to recruit participants to the FGDs with the aim of obtaining variation with respect to geographical location within the city, gender, age, occupation, and level of education. The aim was to detect whether there were any differences in the perceptions of
kimeo between these social groups. Because the public health authorities disapprove of uvulectomy, we decided that recruiting participants who know each other well would enhance trust and openness [
32,
35]. We recruited participants who were either neighbours or members of a local association. Seven of the ten groups were recruited through our personal networks with the help of a gate keeper who asked his/her friends to participate in the study. The inclusion criteria were that the participants should have one or more children (of any age), know the other participants in the group, and be of the same sex as the gate keeper who was recruiting them. There were no exclusion criteria. The FDGs took place at one of the group members’ homes (in most cases a backyard or a veranda). The three other groups were mutual support associations whom we approached face to face and asked for their participation. The participants of these FDGs were self-employed and the FDGs were held at their place of work at a time that they found convenient. No members were excluded, since they all had children, and the groups were single sex, thus fulfilling the inclusion criteria. Apart from some young children who were with their mothers, no outsiders were present during the FDGs.
We purposefully recruited people who were part of networks because we believed this would facilitate openness and trust. In rotating saving clubs, women entrust each other with their savings and they meet regularly. In mutual support associations, a group of self-employed people, usually within the same sector, agree to support each other in times of extreme financial stress (e.g. the death of a family member).
We believe that by recruiting members to FDGs who knew each other well, and who trusted each other, we avoided some of the problems that other researchers who have carried out FGDs in Tanzania have encountered [
44]. At the same time, there is the possibility that some participants would have been more open with strangers. Upon recruitment, we had no information about the participants’ experience with, or perceptions of, uvulectomy.
Caregivers’ characteristics Ten FGDs were carried out in three of the five administrative districts of Dar es Salaam: Ilala, Kinondoni, and Temeke. We recruited groups that were already established, so the number of participants varied, with the majority of the groups having four or five members (see Tables
1,
2). In each FGD, the participants were of the same sex, and were relatively equal in age and educational level. All the groups were ethnically mixed, but in two the participants hailed from the same region. The majority of the groups had both Christian and Muslim participants, but there were more Muslim participants compared to Christian. FDGs with males included members of three different mutual support associations (tailors, water sellers and taxi drivers), as well as two FDGs with neighbours (who had different occupations, mainly self-employed, but also some working in, or retired from, the formal sector). FDGs with women included self-employed members of saving clubs, as well as neighbours (two groups with home makers, one group with women who had retired from formal sector jobs).
Table 1
Focus groups discussion with female caregivers (n = 22)
01 | Self-employed | Saving club | 25–35 | Primary | 5 |
02 | Self-employed | Saving club | 25–35 | Primary | 5 |
03 | Homemakers | Neighbours | 25–35 | Primary | 4 |
04 | Homemakers | Neighbours | 35–45 | Primary | 5 |
05 | Retired from formal sector jobs | Neighbours | 60–70 | Secondary and tertiary | 3 |
Table 2
Focus group discussion with male caregivers (n = 21)
06 | Self-employed | Mutual support association | 25–45 | Primary | 4 |
07 | Self-employed | Mutual support association | 25–45 | Primary and secondary | 5 |
08 | Self-employed | Mutual support association | 25–45 | Primary and secondary | 4 |
09 | Self-employed and formal sector jobs | Neighbours | 25–45 | Primary, secondary and tertiary | 4 |
10 | Retired from formal sector jobs | Neighbours | 60–70 | Secondary and tertiary | 4 |
The FGDs were conducted in January 2012. We used an interview guide with open ended questions, focusing on perceptions of kimeo and uvulectomy. Many of the participants shared their own accounts of illness experiences, and this was often followed by a lively discussion. The FGDs lasted between 1 and 1.5 h, and the atmosphere was open and relaxed. The FDGs were recorded, and the authors took notes during and after the FGDs.
IDIs and observations with practitioners Four
kimeo practitioners were purposely selected to cover different parts of the city: One in the city centre, two in two different residential areas of larger Dar es Salaam, and one in a semi-urban area at the outskirts of the Dar es Salaam region (see Table
3). The interviews took place at their clinics.
Table 3
Interviews with folk practitioners (n = 4)
1 | Rashidi | 50–60 | City center | Uvulectomy Circumcision (males only) Herbal treatments for several illnesses, including asthma, peptic ulcer and cough that persists after uvulectomy |
2 | Khasim | 40–50 | Densely populated unplanned Area | Uvulectomy Herbal treatment for several illnesses |
3 | Omari | 30–40 | Densely populated unplanned Area | Uvulectomy Herbal treatment for several illnesses, including epigastric pain |
4 | Abasi | 25–35 | Semi-urban area (District 4) | Uvulectomy Herbal treatment for several illnesses |
The first practitioner whom we asked for an interview said that he did not want to be recorded. This was probably because of the unclear legality of uvulectomy in Tanzania. Rapid notetaking was therefore used during the interviews with all four participants, and quotes from these interviews are therefore not verbatim in the strict sense. In addition to interviews, we observed the practitioners’ interaction with their clients before and after our interviews and took ethnographic notes.
The interviews were conducted in October 2011 and January 2012 and lasted between one and two hours. The interview guide was open ended and focused on the practitioners’ backgrounds, their perceptions of kimeo, their experiences with carrying out uvulectomies, their relationship to the authorities, and their perceptions of the professional health system. In December 2017, we conducted follow-up interviews with two of the practitioners. The main aim was to learn whether there were any changes in their practice and the way that the health authorities related to them. We also wanted to see the actual surgery and observed two cases of uvulectomy: a girl aged approximately five years, and a man in his mid-twenties.
Kimeo cutters’ characteristics The four
kimeo cutters were between 25 and 65 years old. Two of them grew up in Dar es Salaam, while the two others grew up in rural areas and migrated to Dar es Salaam as adults. Three of them had completed primary school, while one had four years of secondary education. All four practitioners worked full time in their clinics and offered other services in addition to uvulectomy, like male circumcision and herbal treatments (see Table
3).
The clinics varied in appearance, but the structures looked simpler than clinics offering professional health services and three of them consisted of one room only and an outdoor waiting area. None of the clinics had paraphernalia commonly associated with traditional healers in Tanzania (whisks, calabashes, or textiles in the ritual colours red, white and black), nor offered divination, a service that is typically associated with Tanzanian healers. Three of the participants dressed in “Western” regular clothes: trousers and shirts. The fourth practitioner dressed in a stark white uniform, very similar to the uniforms worn by higher ranking health personnel at formal health institutions in Tanzania.
The four practitioners share some important characterises in common: they are male, they are Muslim, their ethnic background is from the Eastern part of Tanzania, and they are members of Chama cha Utabibu Asilia (The Association for Traditional Therapies). All four have learned their skill in apprenticeship with their father or older brother and carry out their practice from permanent clinics, which clients often refer to as hospitali (clinic, small hospital). Two of them are third generation kimeo specialists and operate clinics that have existed since the late colonial era. The two other clinics have existed for approximately five and twenty years respectively.
Data analysis
Throughout the data collection period, the authors discussed the findings after each FDG and IDI. The second author transcribed verbatim the recorded FGDs and translated them to English. The two authors compared their hand-written notes from the IDIs and agreed on one set of typed notes for each practitioner, which included ethnographic observations.
The first author then read all transcripts and notes multiple times for familiarisation and coded the transcripts manually. She manually designed two coding trees referencing the coded information from the FDGs and IDIs. These coding trees reflected the ethnographic design of the study, which aimed at understanding care givers' and practitioners' perceptions and experiences with
kimeo and uvulectomy, and the theoretical focus on the social acceptability of uvulectomy. The first author sorted data using the two coding trees to split data into related hierarchical conceptual themes [
47]. Two main themes in the coding tree for the FDGs with caregivers were predefined based on the research questions (perceived causes of
kimeo and pathways of care), while two new themes (trust in the practitioners and communication between caregivers and professional health workers) and one sub-theme (caregivers’ own experiences with going through uvulectomy) were identified during the analysis of the data. Likewise, the two main themes in the coding tree for the IDIs with the healers were predefined (perceived causes of
kimeo and strategies to win the trust of caregivers and patients), while sub-themes were identified during the analysis process (procedures of the surgery and membership in associations). Acknowledging that illness representations are often varied within a community [
41], we looked both for perceptions of
kimeo and uvulectomy that many of the participants shared, as well as perceptions that only one or two of the participants voiced (outliers in the data).