Menu
Menu
Menu
Menu
Background: The implementation of mobile health (mHealth) projects in low- and middle-income countries raises high and well-documented expectations among development agencies, policymakers and researchers. By contrast, the expectations of direct and indirect mHealth users are not often examined. In preparation for a proposed intervention in the Nouna Health District, in rural Burkina Faso, this study investigates the expected benefits, challenges and limitations associated with mHealth, approaching these expectations as a form of situated knowledge, inseparable from local conditions, practices and experiences. Methods: The study was conducted within the Nouna Health District. We used a qualitative approach, and conducted individual semi-structured interviews and group interviews (n = 10). Participants included healthcare workers (n = 19), godmothers (n = 24), pregnant women (n = 19), women with children aged 12-24 months (n = 33), and women of childbearing age (n = 92). Thematic and content qualitative analyses were conducted. Results: Participants expect mHealth to help retrieve patients lost to follow-up, improve maternal care monitoring, and build stronger relationships between pregnant women and primary health centres. Expected benefits are not reducible to a technological realisation (sending messages), but rather point towards a wider network of support. mHealth implementation is expected to present considerable challenges, including technological barriers, organisational challenges, gender issues, confidentiality concerns and unplanned aftereffects. mHealth is also expected to come with intrinsic limitations, to be found as obstacles to maternal care access with which pregnant women are confronted and on which mHealth is not expected to have any significant impact. Conclusions: mHealth expectations appear as situated knowledges, inseparable from local health-related experiences, practices and constraints. This problematises universalistic approaches to mHealth knowledge, while nevertheless hinting at concrete, expected benefits. Findings from this study will help guide the design and implementation of mHealth initiatives, thus optimising their chances for success.
The study was conducted within the NHD in Burkina Faso. The NHD is located approximately 300 kilometres to the northwest of Ouagadougou, the capital of Burkina Faso. It is one of the six districts of Boucle du Mouhoun Health Region and covers the geographical area of the Kossi Province in the western part of the country. The NHD comprises the town of Nouna with a total population of 29,297 inhabitants and a rural area of approximately 331,020 inhabitants. The health infrastructure of the NHD consists of one District Hospital (DH) in Nouna and 43 PHCs, out of which 10 PHCs are included in MOS@N. MOS@N is a 36-month project that includes both qualitative and quantitative components. In this paper, we focus on the qualitative component, and specifically on data collected at the beginning of the project. By focusing on this initial phase of data collection, we aim to examine mHealth expectations prior to the full implementation of MOS@N. To collect data, semi-structured interviews and group interviews were conducted. The research was designed and implemented by a team of researchers from the CRSN, McGill University and Université de Montréal. Ethical approval for the study was granted by the ethics committee of the Ministry of Health of Burkina Faso and by the Institutional Research Ethics Board of the CRSN. Qualitative data presented in this paper was collected over 2 months, in May and June 2014. Mixed purposive sampling methods were used to select participants [53]. As is usually the case with qualitative research, the aim was not to obtain a representative sample of the various categories of participants, but to gather a substantial body of information from them [54]. Participants can be divided into five different groups, namely (1) health workers in PHCs of the NHD; (2) godmothers participating in the MOS@N project; (3) pregnant women; (4) women with children aged 12–24 months; and (5) women of childbearing age. Table 1 presents the distribution of participants. Distribution of participants Following a purposive, expert sampling method, every health worker (n = 19) in participating PHCs (n = 10) was interviewed. Health workers belong to two subgroups, namely head nurses (infirmiers chef de poste, or ICP; n = 8) and midwives (n = 11). ICPs supervise the daily medical operations of the PHC. All the ICPs working in the participating PHCs were male. The midwives oversee maternal health services at the level of the PHC. All the midwives working in the participating PHCs were female. Data collection also involved individual and group interviews with women from the local population of the NHD. First, semi-structured interviews were conducted with every godmother recently selected to participate in MOS@N. At the time of the interviews, 48 godmothers living in 26 villages had just been selected. Following a purposive sampling method, half of them (n = 24) were interviewed during this first phase of qualitative data collection. Interviews with godmothers did not focus on their actual experiences of MOS@N, which was just starting, but rather on their overall perceptions and expectations. Secondly, semi-structured interviews were conducted with pregnant women (n = 20) enrolled in MOS@N. Participants were selected following a purposive, non-proportionate quota sampling method, with at least one respondent in every participating PHC, with an average of two per PHC. Thirdly, semi-structured interviews were also conducted with women with children aged 12–24 months (n = 33). Participants were selected following a purposive, non-proportionate quota sampling method in which the sole criteria for inclusion was attending any of the 10 participant PHCs. At least two women were interviewed in every PHC. Finally, women of childbearing age (n = 92) were recruited to participate in group interviews (n = 10). These participants were selected following a purposive, non-proportionate quota sampling method in which the main inclusion criteria was geographical, since there was one group interview (with an average of nine participants per PHC) in every participating PHC. We conducted semi-structured interviews with health workers, godmothers, pregnant women and women with children aged 12–24 months. Interviews lasted on average 30 minutes. Pregnant women and women with children aged 12–24 months were approached when coming to the PHC, either for an ANC visit or for a consultation with one or more children. Health workers and godmothers were approached as part of their broader participation in MOS@N and were met at their local PHC. The interviews were conducted by trained researchers from the CRSN. Interviews with health workers were conducted in French and interviews with godmothers, pregnant women and women with children aged 12–24 months were conducted in Dioula. Interviews were digitally recorded and transcribed. Those conducted in Dioula were transcribed into French. Interviews followed a pre-established interview guide, addressing various topics related to mHealth, access to maternal healthcare and mobile phones in general. As is usual with qualitative, semi-structured interviews, the main aim was to ask open-ended questions, which leave room to unexpected answers and are particularly adapted to discussing sensitive, health-related topics [55]. Semi-structured interviews were chosen over unstructured interviews, since while they allow the interviewer to depart from the interview guide, they are better suited to address specific issues when the research already has a fairly clear focus [56]. They also provide more consistency when there is more than one researcher involved in data collection, as was the case here. Qualitative interviews aim at gathering descriptions of the life-world of the interviewee, while remaining open for ambiguities and changes [57]. Group interviews, which were conducted with women of childbearing age, lasted between 60 and 90 minutes. They were also conducted in Dioula, digitally recorded and transcribed into French. Group interviews are particularly useful as part of such a multi-method design to clarify, extend, qualify or challenge data collected through other semi-directed interviews [58]. Each group interview included between 8 and 10 respondents. Interviews were moderated by trained researchers from the CRSN, whose role was to lead the discussion and elicit participation from all members [59]. Data analysis followed common qualitative data analysis guidelines. The first analytic step taken was data organisation and indexing. Recorded interviews were transcribed into French, and read repeatedly by three of the authors, while noting down initial ideas. We then used content and thematic analysis methods [60]. First, two of the authors proceeded to content analysis, by doing an in vivo/emergent, open coding of the relevant data. ATLAS.ti qualitative data software was used for coding. This allowed the researchers to create categories, to group codes under higher order headings and to formulate a general description of the research topic [61]. Then, another author organised data into thematic categories, first by searching for themes, and then reviewing, defining and naming them [62]. Compelling extract examples were selected, analysed and related back to the research questions and literature. The authors then compared thematic analysis and content analysis, moving on to more focused coding, with particular emphasis on concepts related to mHealth expectations. As usual with qualitative analysis, the goal was not to achieve representativeness, but rather to identify meaningful patterns and variations [63].
N/A
