A community-based child health and parenting intervention to improve child HIV testing, health, and development in rural Lesotho (Early Morning Star): a cluster-randomised, controlled trial.

Background: When caregivers live in remote settings characterised by extreme poverty, poor access to health services, and high rates of HIV/AIDS, their caregiving ability and children's development might be compromised. We aimed to test the effectiveness of a community-based child health and parenting intervention to improve child HIV testing, health, and development in rural Lesotho.

Methods: We implemented a matched cluster-randomised, controlled trial in the Mokhotlong district in northeastern Lesotho with 34 community clusters randomly assigned to intervention or wait-list control groups within a pair. Eligible clusters were villages with non-governmental organisation partner presence and an active preschool. Participants were caregiver-child dyads, where the child was 12-60 months old at baseline. The intervention consisted of eight group sessions delivered at informal preschools to all children in each village. Mobile health events were hosted for all intervention (n=17) and control (n=17) clusters, offering HIV testing and other health services to all community members. Primary outcomes were caregiver-reported child HIV testing, child language development, and child attention. Assessments were done at baseline, immediately post-intervention (3 months post-baseline), and 12 months post-intervention. We assessed child language by means of one caregiver-report measure (MacArthur-Bates Communicative Development Inventory [CDI]) and used two observational assessments of receptive language (the Mullen Scales of Early Learning receptive language subscale, and the Peabody Picture Vocabulary Test 4th edn). Child attention was assessed by means of the Early Childhood Vigilance Task. Assessors were masked to group assignment. Analysis was by intention to treat. This trial was registered with ISRCTN.com, ISRCTN16654287 and is completed.

Findings: Between Aug 8, 2015, and Dec 10, 2017, 1040 children (531 intervention; 509 control) and their caregivers were enrolled in 34 clusters (17 intervention; 17 control). Compared with controls, the intervention group reported significantly higher child HIV testing at the 12-month follow-up (relative risk [RR] 1·46, 95% CI 1·29 to 1·65, p<0·0001), but not immediately post-intervention. The intervention group showed significantly higher child receptive language on the caregiver report (CDI) at immediate (effect size 3·79, 95% CI 0·78 to 6·79, p=0·028) but not at 12-month follow-up (effect size 2·96, 95% CI -0·10 to 5·98, p=0·056). There were no significant group differences for the direct assessments of receptive language. Child expressive language and child attention did not differ significantly between groups.

Interpretation: Integrated child health and parenting interventions, delivered by trained and supervised lay health workers, can improve both child HIV testing and child development.

Funding: United States Agency for International Development (USAID) and the President's Emergency Plan for AIDS Relief (PEPFAR).