‘If he sees it with his own eyes, he will understand’: How gender informed the content and delivery of a maternal nutrition intervention in Burkina Faso

A growing body of literature urges policymakers, practitioners and scientists to consider gender in the design and evaluation of health interventions. We report findings from formative research to develop and refine an mHealth maternal nutrition intervention in Nouna, Burkina Faso, one of the world’s most resource-poor settings. Gender was not an initial research focus, but emerged as highly salient during data collection, and thus guided lines of inquiry as the study progressed. We collected data in two stages, first using focus group discussions (FGD; n = 8) and later using FGDs (n = 2), interviews (n = 30) and observations of intervention delivery (n = 30). Respondents included pregnant women, breastfeeding mothers and Close-to-Community (CTC) providers, who execute preventative and curative tasks at the community level. We applied Morgan et al.’s gender framework to examine intervention content (what a gender-sensitive nutrition programme should entail) and delivery (how a gender-sensitive programme should be administered). Mothers emphasized that although they are often the focus of nutrition interventions, they are not empowered to make nutrition-based decisions that incur costs. They do, however, wield some control over nutrition-related tasks such as farming and cooking. Mothers described how difficult it is to consider only one’s own children during meal preparation (which is communal), and all respondents described how nutrition-related requests can spark marital strife. Many respondents agreed that involving men in nutrition interventions is vital, despite men’s perceived disinterest. CTC providers and others described how social norms and gender roles underpin perceptions of CTC providers and dictate with whom they can speak within homes. Mothers often prefer female CTC providers, but these health workers require spousal permission to work and need to balance professional and domestic demands. We recommend involving male partners in maternal nutrition interventions and engaging and supporting a broader cadre of female CTC providers in Burkina Faso.

Burkina Faso is a landlocked country in Sub-Saharan Africa that ranks 182 of 189 countries in the Human Development Index (United Nations Development Programme, 2019). Income is low at US$731 per person in 2018, placing Burkina Faso among the poorest 10% of countries globally (The World Bank, 2018). Approximately 21% of Burkinabe children under five are malnourished (United Nations Development Programme, 2019), and the country ranks among the lowest 10% of countries globally in terms of gender equity across health, education, economic status and political representation. The average years of education completed are 1.6 years, which is exceptionally low even when compared with other low income, African countries (∼3–6 years)(United Nations Development Programme, 2019). The level of formal education is low for everyone but worse for women: in 2010, 57% of 15- to 19-year-old women had not received any formal education compared with 47% of men (INSD, 2012). Furthermore, educational attainment decreases with age, older women are the least likely to have received any formal education (7% of 50- to 54-year-old women have any formal education) (INSD, 2012). Nearly half of married Burkinabe women (42%) live in polygamy, and 44% of Burkinabe women think that a man has the right to punch his wife (INSD, 2012). A majority (88%) of married women who earn money can decide for themselves how to spend their income (INSD, 2012). Only 8% of Burkinabe women have the principal decision-making power in terms of their own healthcare (INSD, 2012). Our study site, Nouna town and its surrounding villages, lies in the northwest of Burkina Faso within the Boucle du Mouhon region. About 30 000 people live in Nouna town and about 100 000 people live in the Health and Demographic Surveillance System (HDSS) site, within which we conducted data collection (Sie et al., 2010). In the region, women’s educational level and decision-making power concerning their income are near the country’s average (INSD, 2012). However, more than half of women (56%) living in Boucle du Mouhon accept domestic violence, which is higher than the national average (44%) (INSD, 2012). This research stems from a formative study conducted in preparation for a nutrition-promotion trial. We adapted South African maternal nutrition videos that female community health workers (CHWs) originally showed to mothers during home visits via tablets (Rotheram-Borus et al., 2011). The South African research team developed the video intervention using a human-centred design approach grounded on community feedback and iteration (Adam et al., 2019). We chose to follow their approach when adapting the video intervention: to maximize engagement of mothers and CTC providers throughout the adaptation process to ensure internal relevance particularly as several intervention components stem from a different cultural background. The Burkinabe intervention involves CTC providers visiting pregnant and breastfeeding mothers at home to show a set of maternal nutrition videos on a tablet. The videos cover different food groups and emphasize the importance of a varied diet. They feature a mother of a small child making choices about her own and her family’s nutrition regimen (Isler et al., 2020). CTC providers include CHWs, who are predominantly male, and Mentor Mothers (MM) who are exclusively female. CHWs are the government-installed personnel linking a community to its local health centre (Centre de Santé et de Promotion Social, CSPS), and they engage in all areas of prevention. MMs are older women who voluntarily accompany pregnant mothers to health centres, sharing practical advice and sometimes assisting during labour. Respondents included pregnant and breastfeeding mothers (78 women), CHWs (5 males, 3 females) and MMs (exclusively female). Respondents were purposively selected because they could either receive or deliver the intervention, and they were able to speak and understand Dioula, the most commonly spoken local language. As a means of reflecting the population distribution, we sampled from the catchment areas of two urban and four rural health centres and were guided by CTC providers in terms of identifying pregnant and breastfeeding mothers. The study consisted of two phases. In the first phase, we gathered data on how to adapt video content. In the second phase, we sought feedback on the adapted videos and preferences for their distribution. Data collectors were female, bilingual (Dioula and French) and came from Nouna town or surrounding villages. They held high school diplomas and had previously conducted research. We worked with only female interviewers to ensure that mothers would be comfortable. We trained data collectors for 3 days on research ethics, maternal nutrition, video interventions, qualitative research and audiotaping techniques. We piloted in-depth interviews (IDIs) and focus group discussions (FGDs) and refined them in close collaboration with the data collection team. Data collection took place between April and June 2018. We conducted FGDs in quiet areas located in or near health centres. For IDIs and observations, we visited eligible participants at home and delivered the video intervention using tablets. We interviewed 30 mothers and conducted FGDs with 48 mothers, 8 CHWs and 8 MMs. Initially, we intended to work exclusively with CHWs to deliver the video intervention. CHWs are predominantly male; only one health centre in the vicinity employs female CHWs. The reason for this gender imbalance is not clear. However, it became clear that some mothers were uncomfortable interacting with a male CHW. We, therefore, included MMs, and consequently conducted two additional FGDs with MMs and 15 observations of MMs, as well as 15 observations of CHWs. CHWs and MMs were thus involved to a similar extent. For each observation, two research team members joined a CTC provider for a home visit of a pregnant or breastfeeding woman. After initial greetings and explanations, the CTC provider sat with the woman and showed her the videos on a tablet. Research team members sat nearby to observe the CTC provider’s approach to video presentation, the woman’s reactions, how the CTC provider and woman interacted generally, other family members’ involvement and any other pertinent details of the setting in which video viewing took place. After the observation, the mother and CTC provider exchanged thanks and good wishes, the research team extended their own thanks and the team left. For a detailed list of data collection activities see Table 1. Data collection activities We regularly debriefed the data collection team (McMahon and Winch, 2018). Bilingual research assistants transcribed and translated the audio-recorded data from Dioula to French. A member of the research team checked the transcripts for consistency and quality. We developed a codebook grounded on debriefing notes, and structured codes into principal and secondary categories. During initial coding, we refined the codebook and agreed on a final version, which two researchers applied to all transcripts. We incorporated data triangulation by comparing FGDs, IDIs and observations for consistency. Incongruities were discussed with a senior researcher within the study team. We used existing gender analysis frameworks because they addressed our research questions and we expected them to provide a meaningful basis for our work. We began with the work of Deshmukh and Mechael (2013) because it focuses on gender in mHealth within maternal, newborn and child health. However, this framework was too focused on the intersection of technology and gender to be helpful for analysis of our nutrition-related data. Ultimately, Morgan et al. (2016) informed our analysis, because their categorization scheme allowed for a more holistic analysis. Morgan et al.’s (2016) framework developed out of a review of existing gender frameworks and argues that gender is a power relation that is negotiated through (1) access to resources, (2) division of labour, (3) social norms and (4) decision-making. Where men are responsible to provide for the family, they are typically favoured in terms of access to resources, both within and beyond the household (March et al., 1999). Types of work are rewarded differently and typical female tasks like household maintenance and childcare are at the lower end of the hierarchy of rewards because they are unpaid and invisible (March et al., 1999). Social norms and rules help to decide in everyday life what behaviour is acceptable, but they can seem set and unchangeable, thus sustaining and justifying gender inequalities (March et al., 1999). As unequal access to resources and a set division of labour are justified by social rules, some individuals gain power over others, thus becoming key decision-makers who can then make decisions that reinforce their own power (March et al., 1999). By grouping categories of our codebook as subcategories within the categorization scheme outlined by Morgan et al., we were able to apply this framework to our research process (see Table 2) and content (see Results section). A process of merging, ordering and renaming those subcategories followed. The subcategories presented in the results thus emerged from our own analysis. How gender as a power relationship influences research process domainsa a Morgan et al. (2016) encourages researchers to ask the following questions to ensure sensitivity to gender throughout the data collection process. We conducted this research with the approval of the ethics committee of the medical faculty of Heidelberg University (S-140/2018) and the ethics committee of the Burkinabe Health Ministry in Nouna (N°2018-07-/CIE/CRSN). We obtained written consent before all IDIs and FGDs.

N/A

• isler, jasmin
• sawadogo, nathalie
• harling, guy
• bärnighausen, till
• adam, maya
• sié, ali
• mcmahon, shannon a.