Abstract
OBJECTIVES: We examined whether effects of an agricultural livelihood intervention on food insecurity and psychosocial outcomes remained robust in pregnant compared to nonpregnant women living with HIV, and whether potentially negative associations between pregnancy and these outcomes were alleviated by the intervention. DESIGN: Secondary analysis of the Shamba Maisha cluster-randomized controlled trial ( N = 396 women; NCT02815579). The intervention included agribusiness training and supplies. METHODS: Food insecure women with HIV in Kenya were followed for 24 months between 2016 and 2019. Food insecurity, empowerment, social support, depression, HIV stigma, and intimate partner violence were collected at all visits. We estimated the effect of the intervention on trends for each outcome via mixed-effects regression, separately for women who did and did not become pregnant during follow-up, and whether trends differed by pregnancy status, separately by arm and adjusted for demographic factors. RESULTS: In comparison to controls, the intervention was associated with a greater decline in food insecurity among women who became pregnant (3.35 points, 95% confidence interval (CI): -5.63, -1.06) and who did not become pregnant (3.43 points, 95% CI: -4.34, -2.52). Effects on psychosocial outcomes were also comparable in pregnant and nonpregnant women. Having an incident pregnancy was associated with disempowerment among controls (difference in trend -0.22, 95% CI -0.44, -0.00) but not in the intervention arm. CONCLUSIONS: We observed comparable benefits of an agricultural livelihood intervention on food security and psychosocial outcomes regardless of pregnancy status. Agricultural livelihood interventions may hold promise for improving pregnancy outcomes through improved maternal food security.