Food-insecure pregnant women in South Africa: a cross-sectional exploration of maternal depression as a mediator of violence and trauma risk factors

OBJECTIVES: Better understanding of psychosocial risk factors for food insecurity (FI) during pregnancy and how they interact is crucial, given long-term health implications for maternal and child health. We investigated the association between maternal childhood trauma as well as intimate partner violence (IPV) and FI among pregnant women in South Africa, in the Drakenstein Child Health Study, and whether maternal depression mediates these relationships. SETTING: Two primary care clinics in Paarl, South Africa. PARTICIPANTS: 992 pregnant women; inclusion criteria were clinic attendance and remaining in area for at least 1 year; women were excluded if a minor. METHODS: We examined psychosocial predictors of FI using multivariate regression. Mediation analyses investigated whether depression mediated the relationship between IPV and FI as well as between childhood trauma and FI, including disaggregation by two study communities. FI was assessed using an adapted US Department of Agriculture food security scale; households were coded as food insecure where 2 of 5 affirmative responses were recorded. RESULTS: Among 992 pregnant women, there were high rates of IPV (7%–27%), depression (24%) and childhood trauma (34%). In multivariate cross-sectional analysis, emotional IPV (adjusted OR [aOR] 1.60; 95% CI 1.04 to 2.46), depression (aOR 1.05; 95% CI 1.01 to 1.08) and childhood trauma (aOR 1.52; 95% CI 1.08 to 2.15) predicted FI. In mediation models, depression partially mediated the relationship between emotional IPV and FI as well as physical IPV and FI; depression partially mediated the relationship between childhood trauma and FI. Differing degrees of mediation were found when applied to communities. CONCLUSIONS: Antenatal maternal depression, IPV and childhood trauma were highly prevalent and associated with FI. Depression, IPV and trauma screening services should be considered within routine antenatal care and may offer an opportunity to identify and intervene. Community-level differences in risk and in mediation analyses indicate that contextual tailoring of interventions may be important.