Depressive symptoms and violence exposure in a population-based sample of adult women in South Africa

Abstract

Depressive symptoms are a major burden of disease globally and is associated with violence and poverty. However, much of the research linking these conditions is from resource-rich settings and among smaller, clinical samples. Secondary data from a household survey in Gauteng Province of South Africa examines the cross-sectional association between adult women’s elevated depressive symptoms and markers of violence. Using tablet computers, participants self-completed interview modules to screen for depressive symptoms (Patient Health Questionnaire 2-item screener), childhood exposure to physical and sexual abuse (Childhood Trauma Questionnaire 4-item index), as well as past-year exposure to sexual or intimate partner violence (SIPV; WHO Multicountry Study instrument 4-item index). Socio-economic status, food security, education, and income were self-reported. Representative data at the ward level allows for modeling of results using survey commands and mixed-level modeling. Of the 7,276 adult women participating in the household survey, 42.1% reported elevated depressive symptoms. A total of 63.9% reported childhood violence exposure and 5.3% had past-year SIPV. Multi-level modeling suggests that violence is a strong predictor of depressive symptoms. Childhood abuse alone increases the odds of high depressive symptomology, after controlling for individual-level markers of poverty and neighborhood of residence (aOR 1.31, 95%, CI 1.17–1.37). Combined exposure to childhood abuse and past-year SIPV increased odds of reporting elevated depressive symptoms (aOR 2.05, 95%, CI 1.54–2.71). Ward characteristics account for 6% of the variance in depressive symptoms, over and above the contributions of household food security and socio-economic status. Exposure to violence in childhood and past-year SIPV were associated with depressive symptoms among women. These associations persist after controlling for socio-economic markers and latent neighborhood characteristics, which also had significant association with elevated depressive symptoms. These data suggest that efforts to reduce the burden of depressive symptoms may benefit from approaches that prevent violence against women and children.