Study setting

The Gambia is the smallest country in continental Africa and is surrounded by Senegal except at its western border which is the Atlantic Ocean. It has a population of almost 1.9 million citizens of whom 96% are Muslim ([24]).

The Medical Research Council Unit The Gambia (MRCG) at the London School of Hygiene & Tropical Medicine runs three HDSS which cover almost 15% of The Gambia’s population. This study was carried out in Farafenni HDSS in Kerewan region on the North Bank of the river Gambia. There are three main ethnic groups–Wollof 41%, Mandinka (31%) and Fula (22%), whose primary occupation is farming [14]. As at 31^st December 2015, there were 55,209 individuals being followed up [25]. In this region, under-five mortality is the third highest nationally at 72/1000 livebirths, adult mortality the second lowest and fertility second highest with a total fertility rate (TFR) of 6.8 children per woman.

The Basse HDSS established in 2007 is on the south bank of the river Gambia [26]. The Serahule (39%), Fula (30%) and Mandinka (29%) are the major ethnic groups in Basse (2013 census spatial distribution). As at 31^st December 2015, a total of 179, 548 people were being followed up in the HDSS. Under-five mortality is 66/1000 livebirths, ranking fifth highest while the level of adult mortality is second highest nationally and has a TFR of 6.4 children per woman.

The qualitative study reported on here was embedded within two larger household surveys conducted from December 2015 to June 2016 in the two HDSS These were re-enumeration surveys of a sample of the HDSS population carried out with several objectives, one of which was to use pregnancy histories to re-estimate under-five mortality estimates including perinatal mortality. For this objective, a pregnancy history was obtained from resident women of reproductive age by female only interviewers as opposed to routine surveillance conducted by male interviewers (S1 File). Other study protocols included conditions that women were to be interviewed in private using their preferred language.

Study participants

The first group of study participants of the qualitative study comprised 13 of the 31 female fieldworkers who conducted the pregnancy history survey (see Table 1). For the six months of the survey, the female fieldworkers were fed and accommodated in the community by the village chiefs at no cost, as is customary in The Gambia. They participated in daily activities of the community alongside the local women, including cooking, fetching water, grinding of cereals and washing, in addition to interviewing the women. The female fieldworkers each interviewed 332 women on average, although one of them interviewed 1700 women. The female fieldworkers were mostly Mandinka and Fula, which are two of three most common ethnic groups in The Gambia, the third being Wollof [27]. All fieldworkers spoke Wollof as it is the lingua franca in The Gambia.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic profile of the study participants.

https://doi.org/10.1371/journal.pone.0216924.t001

All the female fieldworkers had completed secondary school education, and two were in university while four were enrolled in college. Most had no fieldwork experience except for three who had worked with MRC or conducted nationwide surveys before (Table 1). Although they had been born in the study areas, most of them no longer lived in the rural areas as they had left to study or work.

AJR interviewed three key informants who formed the second group of study participants, two of whom were long-serving community liaison officers and one who was a field worker having over twenty years of research experience all over The Gambia. The three key informants hail from three distinct regions in The Gambia and were selected to provide more information about cultural issues identified on analysing the in-depth interviews. The third group of study participants were six male fieldworkers and their supervisor, who live in the community and work in the HDSS, and who were responsible for follow up the residents in the HDSS both before and after the survey was completed.

Study design

Three research methods were used: in-depth interviews of the female interviewers, key informant interviews, and a focus group discussion (FGD) with a group of male HDSS field staff. AJR developed the question guides for all the interviews and FGD after a cumulative process of information gathering during the two years she had worked in the HDSSs and lived in The Gambia. This involved observation and informal conversations about ways of communicating in the community that might affect the pregnancy history data that the female interviewers collected.

Some of the field team we interviewed had worked in both Farafenni and Basse, and thus provided different viewpoints covering the north and south bank communities. AJR did not interview the female fieldworkers as the relationship was that of employer and employees, and this power dynamic would probably have constrained their responses. The interviews of the female fieldworkers were therefore conducted by JN, who is white, and were carried out in English as all the participants were fluent. JN spent a week observing the survey procedures in the field and noted the interactions of the fieldworkers with the community members and the community’s reaction to both male and female fieldworkers. She is female and similar in age to most of the fieldworkers, which aided in building rapport.

While analysing the content of the initial interviews, AJR changed the interview guides as themes emerged that needed further investigation (S2 File). The questions focused on the experience of the female fieldworkers and their understanding of the community members’ attitude to being interviewed by a female-only team and cultural beliefs that influenced the community’s response. Recruitment was stopped after no further emerging themes could be identified by JN and AJR. AJR and JN subsequently conducted the FGD with the male HDSS field staff.

The female fieldworker interviews were conducted in Dampha Kunda and Gambisara, which are in Basse on the south bank of the river Gambia. The key informants were interviewed in Basse, Farafenni and in Fajara on the coast, where the MRCG headquarters are located. The FGD comprising the male fieldworkers and their supervisors was conducted in the Armed Forces Provisional Ruling Council (AFPRC) General Hospital in Farafenni, located on the north bank of the river Gambia.

Sampling

Purposive selection of female interviewers was done to learn from their experience when conducting the community survey, after which the snowball technique was applied to identify key informants [28, 29]. The key informants were selected based on their knowledge of the Gambia as their main roles are to facilitate community entry for all projects by MRCG, the fact that they hailed from different regions in The Gambia, being multi-lingual, and their extensive experience doing research throughout the country. The FGD was held with all the male fieldworkers working on the Farafenni HDSS. All the in-depth and key informant interviews were digitally recorded and transcribed verbatim, using the software F4. AJR and JN made expanded field notes after the FGD and recorded their reflection on the discussion immediately after its conclusion.

Data analysis

The analytic phase for each interview and the FGD included reading each transcript and field note, summarizing, designing tables with the contents and encoding material to emerging themes [30]. Relevant passages from the transcripts were added to the coding tables under the corresponding codes and recurring themes identified. Verification and validation of the themes through data and investigator triangulation were conducted by AJR and JN who compared their independent coding frameworks derived from responses from the different data sources and discrepancies between them were discussed and resolved[31, 32]. Both manual coding and analysis using NVIVO 10 Qualitative Data analysis software (QSR International Pty Ltd. Cardigan UK) was done.

Ethical approval

The Gambia Government/ MRC Joint Ethics Committee and the London School of Hygiene & Tropical Medicine Ethics Committee approved the study. The authors were given the discretion when obtaining informed consent to get written or verbal consent depending on the local situation. The participants were informed of the study objectives and were reassured about anonymity and their right to withdraw at any time with no consequence. According to the participants preference and in order to obtain more frank responses, verbal consent was obtained which was audio-recorded. Permission to record the interview was also given and thus all interviews were audio-recorded. As per the AAA guidelines for informed consent, ongoing consent from the participants throughout the interviews was obtained [33]. Both AJR and JN had permits to work and conduct research in The Gambia in alignment with national regulations.

Community reluctance to be counted and to speak of the deceased

The female interviewers reported that the Fula, Wollof and Serahule exhibited a reluctance to be counted. According to one key informant (I14) of Fula origin (community liaison officer), who speaks six local languages and is familiar with the Gambian culture, this was his experience unlike with the Mandinka who did not share this reluctance to be counted.

He reported that these groups of people held on to the belief that it was not good to be counted as it had been passed down generations. He suggested that they did not know the origin of this belief but still avoided being counted even after exposure to Western education. He said:

“As far as counting is concerned, the Fula will not want to use it for humans. Instead of asking ‘how many children do you have?’ you ask, ‘how many sticks does your mother have?’ They will then respond with a number. Even animals you do not count. The history goes back even before my ancestors. They do not like to count (I14)”.

The female fieldworkers interviewed corroborated this view, saying that it was not culturally acceptable to count numbers of children because counting them would reduce their numbers, that is, increase their chances of dying (I1). It is said that, if you know how many children someone has, it enables you to curse them and cause the children’s death. A female fieldworker attributed this to the low literacy level in the study areas and thought that, over time, this concept would change as more people were educated and therefore came to understand that counting people was not directly responsible for causing their deaths (I1).

The fieldworkers encountered further resistance when asking about the respondents’ siblings, especially if the siblings were deceased, as Gambians prefer not to dwell upon those who have died (I1,I6,I7, I8). Most Fula and Serahule women interviewed refused to name their siblings. If the fieldworker probed for names, they would sometimes be told to ask their (the respondents’) parents as a way of dismissing them. In some cases, their parents were not part of the study sample, or were resident elsewhere or deceased. Often, respondents would acknowledge that they knew the number of their siblings, but some reported that they would never reveal this information, far less list their names. However, by gentle probing and by breaking down the question into two, starting by counting those who were alive first and then asking about those who had passed away, the team occasionally managed to get this information (I2).

A key informant (I15) reported that even Fula people from the neighbouring country of Guinea living in Basse have the same reticence, and that their reluctance was not related to fears about witchcraft.

Concealment of pregnancy

Women from all ethnic groups in The Gambia, seldom talk about or discuss their pregnancies, even when they are plainly visible. A well-known proverb that justifies this was provided by a key informant as follows: “If a snake wants to grow big it has to hide.” (I14). He further explained that hiding the pregnancy was a way of protecting it from harm from others who might cause the woman to miscarry by means of the evil eye or witchcraft. For the primary survey, two questions were included to try to improve the detection of pregnancies (S1 File). The women were first asked directly by the female interviewers if they were pregnant. Regardless of the answer, they were then asked about their menstruation cycles. If a respondent reported that she had missed her period, she was further probed as to the reasons for the absence of her period while keeping in mind the sensitivity and cultural implications of revealing a pregnancy (I4). This was useful as many women readily reported missing their menstrual periods for months, but still said that they were not pregnant when asked directly. This was not always the case for older women however, as some of them felt offended at being asked about menstruation by our ‘younger’ fieldworkers.

The female fieldworkers reported that, in some instances, the woman’s husband, mother or in-laws confirmed the pregnancy, even if she would not, despite attempts at privacy by the team such as interviewing the woman under a tree out of earshot. The woman’s relatives knew that the survey was asking women about pregnancies and would tell the female interviewers that so-and-so was pregnant in spite of her reluctance to report this.

Although the women eventually responded, they found it difficult to report on dates of their last menstrual period, as asking about dates of menstruation is an unusual practice in The Gambia. The female fieldworkers said that they would sometimes encourage the respondents to speak about periods and pregnancy on the basis that these were things that all women went through (I13).

Stigma about pregnancy losses

The pregnancy history survey (primary study) included questions on miscarriages, abortions and stillbirths (S1 File). Perhaps as a continuum of the reluctance to be counted noted previously, some women did not want to talk about their reproductive health problems and considered this information private. This was particularly so when the women were asked about pregnancy losses (I6, I7, I9, I12). The fieldworkers believed that younger women between 15 and 18 years of age found it most difficult to answer questions about pregnancy losses. For example, their standard response was that they were not married and therefore could not have had this experience (I7). This reticence is probably compounded by the fact that the community frowned upon both women having children out of wedlock and upon them having an abortion.

Although respondents were hesitant to talk about pregnancy losses, there were a few exceptions to this. Some interviewers reported that although it took a lot of time to sit and listen to the respondents, the ‘release’ reported by the respondents after talking about their pregnancy losses was worth it. This was more so as they lived in the households that sometimes belonged to the respondents (I4, I11). Unfortunately, some may have answered with the misconception that by giving out this information they would be helped to find a cure (I10). It was also difficult to capture accurate dates for stillbirths and miscarriages but not for livebirths, which were usually recorded on the child’s health card (I1, I4, I11, I13).

Challenges with legal documentation

In the urban area in Basse, people could clearly remember the dates of demographic events, and readily produced their identification documents. In the more rural areas, gaining access to documents proved difficult.

This sentiment is echoed by (I4, I9) and (I10), who mention that refusal to produce identification documents occurred more frequently among the Serahule community, who were concerned that the interviewers would take them.

Women’s roles and status in the community

In some villages, in addition to obtaining consent from women study participants, fieldworkers needed to first seek permission from the heads of households, who wished to determine which questions could be answered. If the household head was not around, the interviewers would have to wait or reschedule as doing otherwise meant that they could be expelled from the compound (I1). We found that most refusals were by older women or male heads of households from the Serahule community.

Also, in households where there was divorce, some of the women would not disclose that their husbands, who had been listed as the household head, had left, leading the fieldworkers to believe the women were not members of the selected households, when in fact they were (I6).

The Gambian culture dictates that hospitality offered by the community members should be accepted. The female fieldworkers were hosted at no cost for the entire six-month duration of the survey. They felt that they were well integrated and received more information because they lived in the community. They participated in daily activities including food preparation and eating from the same pot (I5). They reported that doing chores with the women reduced their workload and time pressure, which was appreciated, and they believed that this resulted in their receipt of more frank and accurate responses during the interviews. The female fieldworkers however had to adhere to dress codes that were acceptable to the communities they visited to encourage their responses. Being young, modern, educated women, sometimes they were met with disapproval when they wore ‘western’ attire such as trousers.

In The Gambia, women are required to live with their parents until they get married. For the female fieldworkers, living away from their parents was a challenge as most of them had never stayed away from their families before. Although they found this experience positive (I6, I9), they faced harsh weather conditions, and found the sleeping arrangements and food inadequate. Basse experiences the hottest temperatures in The Gambia, reaching over 40 degrees Celsius. The research team mitigated some of their discomfort by providing mattresses, insecticide-treated bed-nets, and bags of rice to supplement what the villagers provided. The hosts were also given a token of appreciation for providing accommodation.

The male fieldworkers during the FGD discussed their thoughts on the employment of female-only interviewers and what they had heard in the community about the female fieldworkers during their routine follow up visits so as to enable us to gauge the community’s attitudes to having women interviewers. According to them, female fieldworkers are the preferred choice for asking questions relating to pregnancy and menstruation as it was not culturally suitable to have men ask these questions (FGD_I-FGD_VII). They confirmed their discomfort in asking about pregnancy, so much so that they had developed ingenious ways of doing so without compromising women’s modesty. This they did, for example, by asking if the woman had a ‘yellow card’, which is the card given to women attending the antenatal clinic, thus alluding to her pregnancy without having to ask directly or make the woman respond directly. They also noted on their ledgers if the woman was visibly pregnant, thus eliminating the need to ask questions about this (usually in the late stages when it was obvious). It is important to note that male fieldworkers do not ask about menstruation during their routine HDSS rounds and they expressed great reluctance to ask a question about this if it was to be introduced, although they would do so if it were required. The male fieldworkers also felt that although the women in the community would answer questions about pregnancy when asked by men, social constraints made it difficult for them, whereas the female fieldworkers had no such constraint. Their visits to the community after the pregnancy history survey revealed that the female fieldworkers had left a good impression and the women had been happy to talk to them, stating that they were welcome to continue their surveys in the future.

Strengths and limitations

For this study, data triangulation was applied using three approaches to qualitative data collection, comparing views from different age groups, sex and different regions in The Gambia as well as investigator triangulation. This increased our confidence in our findings, resulting in a more comprehensive viewpoint of the cultural factors affecting data collection on pregnancies, births and deaths.

Although the employment of local interviewers is often advocated, we found that having most of our interviewers from urban areas did not hamper rapport building. A limitation of this study is that community members were not directly interviewed after the female-only survey, but information was gleaned by proxy via the male interviewers who went back to talk to them. Equally, in some regards, this may be a strength of the study as the male fieldworkers are insiders in the community and more frank information might have been given to them than would have been to the researchers, if we had asked community members about their thoughts on the survey and the interviewers[34].

The female interviewers had all completed secondary education which enabled good communication between us. However, their educational level may have created a ‘space’ between them and the women in rural villages, who are unlikely to have had western education. This westernization may have affected the responses and attitudes of the community members, judging from the comments the interviewers received about their clothing. However, the fact that there were no other criticisms of, or sanctions against, the female fieldworkers from the community apart from the one occasioned by their ‘westernized’ dressing, which they subsequently corrected, suggests that it is unlikely that it resulted in significantly less information being given to them. However, inevitably, this paper represents the interviewers’ views as modulated by their own beliefs and social constructs, as well as the researchers’ interpretation based on their worldview, and may not fully depict the community’s real perspective.