“Generally the young mom suffers much:” Socio-cultural influences of maternal capabilities and nutrition care in Uganda

Abstract: We conducted 40 in-depth interviews and eight focus groups among mothers and fathers (n = 91) of diverse ages in western Uganda to define the relevant domains of maternal capabilities and their relationship to infant and young child feeding practices. This study was directed by a developing theory of maternal capabilities that posits that the impact of health-directed interventions may be limited by unmeasured and poorly understood maternal characteristics. Ugandan caregivers defined three major life events that constrain women’s capabilities for childcare: early pregnancy, close child spacing, and polygamous marriage. Women describe major constraints in their decision-making capabilities generally and specifically to procuring food for young children. Future nutrition programs may improve their impact through activities that model household decision-making scenarios, and that strengthen women’s social support networks. Findings suggest that efforts to transform gender norms may be one additional way to improve nutrition outcomes in communities with a generally low status of women relative to men. The willingness of younger fathers to challenge traditional gender norms suggests an opportunity in this context for continued work to strengthen resources for children’s nutritional care. Significance: Maternal factors such as autonomy are associated with child feeding practices and nutritional status, with varying degrees depending on the definition of maternal-level constructs and context. This study describes the events and processes that constrain maternal capabilities—intrapersonal factors that shape mother’s abilities to leverage resources to provide care to children—as they relate to nutrition and hygiene practices. We report community beliefs and understandings about which capabilities have meaning for child nutrition and hygiene, and develop a conceptual framework to describe how these capabilities are formed and describe implications for future nutrition programs in East Africa and similar settings. The study was conducted in the Bundibugyo District of western Uganda. Bordered to the north by Lake Albert and to the west by the Democratic Republic of Congo, the 261,000‐person district remains one of the most remote and impoverished in the country. In 2010, the region was partially electrified, and the first paved road was constructed in 2013. These developments have ushered in a slow trend towards peri‐urbanization, though the majority of the population relies on subsistence farming and a small cash crop market for cocoa and matooke. The national fertility rate of 6.2 live births per woman is the highest of any country in the East and Southern African region. In the western region, women give birth to a mean of 6.4 children; over a fifth (22.6%) of women begins childrearing before the age of 20 years. The stunting prevalence (height‐for‐age Z score < −2) is highest in the western region at 43.9%, compared with 32% nationally (UBOS & ICF International Inc., 2012). Women's access to education is poor: 20% of women receive no formal education, and 80% do not proceed past primary school (UBOS & ICF International Inc., 2012). Child feeding is traditionally done by women. Following an early diet or porridges and soft‐starchy foods, children are fed family foods by around 18 months. Meals consist of starch‐based “foods” such as maize porridge and matooke, and protein and lipid‐rich sauces such as groundnuts, bean, and dried fish (Ickes et al., 2015). Bundibugyo inhabited by the Babwisi and the Bakonjo people groups, who represent the Batooro and Bwamba Kingdoms, respectively. Both groups speak Bantu‐derived languages. The society is patrilineal and patrilocal: land is primarily owned by men and inherited through male lines. Following marriage, couples typically settle near men's families. A clan structure, led by men and based on location of residence and family origin, orders the society. Marriage is recognized culturally and religiously. Marriage traditionally involves a bride price paid to a woman's family. Traditionally, a garden hoe or digging stick is given at marriage ceremonies to signify the replacement of lost labor as a girl leaves her family. The process of marriage was done historically both through arrangement and elopement (called Kuhaiya Mukali), and negotiated between families through a middleman (Kibonabuko for the Bakonjo and Mukwenda for the Babwisi) (Uganda Travel Guide). Polygynous marriage is common. The majority of residents in Bundibugyo are Christians, mainly in the Anglican and Catholic churches. There is a small Muslim, and African tradition religious population (United States Department of State, 2009). We purposively sampled participants for in‐depth interviews and focus groups. We conducted four all‐female and four all‐male focus groups for a total of eight focus groups. To ensure that that the study sample had current or potential experience with providing child care, all participants were required to have a child under 3 years of age, or to be expecting child. To avoid potential conflict within households, we did not sample men and women from the same household. Participants were asked, if possible, to be interviewed in a private setting at their compounds or at the local health center (in the case of focus groups) to ensure protection. All participants were free to participate irrespective of their marital status. Participants were purposively sampled to achieve saturation of key themes (Ulin, Robinson, Tolley, & McNeil, 2002). As interviews progressed, we identified specific populations that were necessary to understand how cultural processes and factors affected capabilities (e.g., mothers who care for more than five children and younger mothers). Hence, we used the initial demographic questionnaire to ensure that study members reflected these groups. Our study sampled households from three of the seven sub‐counties in Bundibugyo: Nyahuka, Busaru, and Ndugutu. The research team visited 10 villages and two peri‐urban centers. Participants were approached at their homes and at health centers by study staff, which included two researchers from the United States and three Ugandan research assistants. Participants were asked if they would be willing to participate in a study about men's and women's roles in caring for children in Bundibugyo. We explained that the study purpose was to “better understand social issues in the community and how they relate to children's nutrition, and to identify ways to improve the programs available for parents to care for children in this community.” Interviews ranged from 30 to 60 minutes; while the focus groups lasted approximately 75 minutes. To protect participant privacy, to build trust and rapport with study participants, confidentiality was provided to all participants (Baez, 2002). The interviewer ensured each caregiver that the information shared would be kept confidential, and that no names would be used in publishing the results, nor would anyone be able to determine who participated in the study based on the information shared and published. Caregivers in focus groups were encouraged to not to use individual names in their responses, and were instructed to not share the names of anyone in the focus group. Further participants were asked to share only the general topic of the discussion with others after the conclusion of the meeting. The consent process explained that if participants were uncomfortable at any time, we would stop the interview. We ensured all participants that their participation in the study would have no impact on future interactions with the health center, NGOs or any nutrition activity in the district. Each contributor was given a bag of dried beans (valued at $0.70 USD) for their participation. Interviews were conducted in English and translated to the preferred local language of the participant (Lubwisi or Lukonjo). Responses were back translated in English to the interviewer. All participants completed a demographic questionnaire, based on the Uganda Demographic and Health Survey (UBOS & ICF International Inc., 2012). Following the initial development of the interview and focus group guides, we spent 1 week revising and translating the guides in person as a research team in Uganda. This process involved discussions of ways to convey complex ideas and culturally‐constructed ideas such as empowerment, gender, and marriage. Semi‐structured interview questionnaires were iteratively revised following each day of interviews. The initial questions applied cognitive testing of terms used to understand the domains of maternal capacities of initial interest: women's social support; psychological capabilities; roles, priorities and time; decision‐making/agency; and empowerment. Following early findings, we began to use phrases to help describe concepts in future interviews and focus groups, such as “to take another wife” to describe polygyny and “to feel free” to introduce the topic of empowerment. These domains were identified with a working group that developed a framework to be tested and refined in multiple contexts (Matare et al., 2015). In later stage interviews, the interview guide examined men's and women's child caring practices and roles related to nutrition and hygiene. These individual interviews revealed cultural patterns and processes that were assessed through vignettes in focus groups. Table 1 presents a sample of representative in‐depth interview questions, and Table 2 presents the focus group scenarios. Sample questions and responses from semi‐structured interview guides, based on the theoretical framework constructs Moderator: Was he supporting you in using family planning? Mother: “It was my own decision; he was not supporting.” (Mother's Interviews #23) “Having many wives on many shambas (gardens) of cocoa does not mean you will now become rich. Because the more wives you have the more children you will produce and the more money you will need to spend to take them to school and pay for their school fees. So many people are now realizing that they are now changing from having many wives or multiple wives to keeping up with only one.” (Father's Focus Group #3) [because]: some women don't go where they say they are going!” (Father's Interview #8) Focus group discussion guide vignettes “In the second couple, a girl loses her father when she is 14. Since her mom cannot pay her school fees, an older man of 28 years in community offers to pay. This girl soon becomes pregnant with his child and she becomes married to the older man, and stops schooling. The women in both couples are both pregnant for the first time.” What impact does this have on the decisions that they can make about caring for their children? Three distinct processes and events were identified that routinely constrained caring capabilities and compromised child nutrition: (a) Becoming pregnant and getting married as an adolescent; (b) Becoming pregnant while still caring for a very young child; and (c) Becoming a co‐wife in a polygamous marriage. The focus group narrative (Table 2) was orally presented in focus groups to gain collective responses from parents regarding the likely community response to two hypothetical scenarios that incorporated these concepts. Focus group findings were coded into two main themes that all examined the differences between young mothers who married older men and mothers who are married to men their same age: (a) decision‐making around money and (b) treatment by others and support from the community. Because of the interconnectedness of themes, focus group and interview findings were incorporated. All interviews and focus groups were audio‐recorded, and then back‐translated to English and transcribed verbatim by translators who were trained regarding the study purpose and theoretical background. Revisions were then made to the English transcripts as needed to ensure clarity and accuracy. Transcripts were read and coded by two separate analysts using Dedoose (Dedoose Software, Los Angeles, CA, USA) qualitative research software. Interview codes were developed deductively from the interview guides and inductively from emergent themes from the interview data. After agreeing on preliminary codes, a minimum inter‐rater reliability of >0.9 was established by testing for mutual agreement of code assignment with a sample of 10 transcripts. Direct text quotes were extracted from transcripts with participant identification numbers in order to link quotes with demographic data. Data reduction was accomplished by organizing representative quotes within each theme into schematic matrices to provide a hierarchical visual display of themes. Related themes were grouped together under dimensions of the social ecological framework (Bronfenbrenner, 1979). Demographic data was tabulated into an electronic database and response frequencies were calculated in STATA version 12.0 (Stata Corp., College Station, TX, USA). The study was approved by The College of William and Mary Protection of Human Subjects Committee (PHSC‐2013‐03‐13‐8583‐sbickes) and by the Ministry of Health at the Mbale Regional Hospital Institutional Review Committee (UG‐IRC‐012).