Abstract
INTRODUCTION: Children who are HIV-exposed and uninfected (CHEU) are a growing population at potential risk of poor neurocognitive development. We tested a nurturing care intervention on children's neurocognitive development and maternal depressive symptoms (primary) with mediation through caregiving activities (secondary). METHODS: This study was conducted among six intervention and nine comparison antenatal-care/prevention of vertical transmission (ANC/PVT) HIV clinics in eSwatini. We enrolled pregnant women and measured infant development at 9 and 18 months. mothers2mothers (m2m) designed and implemented the clinic-home-community-based intervention. We measured infants' neurodevelopment, maternal depressive symptoms and caregiving activities with the Mullen Scales of Early Learning (MSEL), Edinburgh Postnatal Depression Scale, HOME Inventory and Family Care Indicators. We fitted linear mixed effects regression models with clinic random effects to compare intervention versus comparison arms, and generalised structural equation models to evaluate mediation, adjusting for confounders. RESULTS: Mother-infant pairs (n = 429) participated between January 2016 through May 2018. Socio-demographic characteristics were balanced between arms except for higher rates of peri-urban versus rural residence and single versus married mothers in the comparison group. The 18 month retention was 82% (180/220) intervention, 79% (166/209) comparison arm, with 25 infant deaths. Intervention MSEL scores were significantly, and modestly, higher in receptive language (55.7 [95% CI 54.6, 56.9] vs. 53.7 [95% CI 52.6, 54.8]), expressive language (42.5 [95% CI 41.6, 39.8] vs. 40.8 [95% CI 39.8, 41.7]) and composite MSEL (85.4 [95% CI 83.7, 84.5] vs. 82.7 [95% CI 81.0, 84.5]), with no difference in maternal depressive symptoms or in observations of mother-child interactions. Intervention book-sharing scores were higher (0.63 vs. 0.41) and mediated the effect on MSEL scores (indirect effect, p-values ≤ 0.024). The direct effects on visual reception and expressive language scores were significantly higher in the intervention compared to the comparison arm (coefficients 1.93 [95% CI 0.26, 3.60] and 1.66 [95% CI 0.51, 2.79, respectively]). CONCLUSIONS: Nurturing care interventions can be integrated into ANC/PVT clinic-home-community programmes. The intervention, mediated through interactive caregiving activities, increased language development scores among CHEU. Partnering with a local team, m2m, to design and implement a culturally relevant intervention illustrates the ability to impact parent-child play and learning activities that are associated with children's neurodevelopment.