Are parenting programmes effective at scale? Associations with violence against adolescent girls, parenting and mental health in real-world delivery across eight African countries: a meta-analysis of pre-post surveys

INTRODUCTION: Evidence-based parenting programmes are widely used to prevent violence against children and improve parenting and mental health. Despite hundreds of randomised trials, little is known about their outcomes when delivered at scale within routine delivery. This study assesses the WHO-endorsed and UNICEF-endorsed Parenting for Lifelong Health programme for caregivers and adolescents, delivered through non-governmental organisation and government in Botswana, the Democratic Republic of the Congo, Eswatini, South Africa, South Sudan, Tanzania, Zambia and Zimbabwe, with support from the President's Emergency Plan for AIDS Relief (PEPFAR), the United States Agency for International Development (USAID) and the European Union. METHODS: Pre-post surveys for caregivers and adolescents were integrated into service data collection between 2016 and 2022. Abbreviated standardised measures of physical abuse, emotional abuse, approval of corporal punishment, positive involved parenting, monitoring/supervision, caregiver depressive symptoms, parenting stress and adolescent depressive symptoms and externalising behaviour were used. Individual country scores were analysed separately for caregivers and adolescents using generalised linear mixed-effects models, and cross-country data were combined using a random-effects meta-analytic model. RESULTS: 123 050 participants were included (93% retention, 57 908 adolescents (96% female), 56 423 caregivers at follow-up). In all-country meta-analyses, estimates showed reduced physical abuse (-65%; 95% CI 51% to 74%), emotional abuse (-59%; 95% CI 48% to 68%) and approval of corporal punishment (-55%; 95% CI 48% to 60%). Positive involved parenting increased (+52%; 95% CI 24% to 87%) and poor supervision/monitoring decreased (-48%; 95% CI 34% to 58%). Caregiver depressive symptoms (-25%; 95% CI 8% to 48%), parenting stress (-46%; 95% CI 41% to 52%), adolescent depressive symptoms (-22%; 95% CI 1% to 38%) and adolescent externalising behaviour problems (-43%; 95% CI 29% to 54%) all declined. There was heterogeneity in pre-intervention scores and extent of change between humanitarian and development settings, and between different target groups, but strong consistency across caregiver and adolescent reports. CONCLUSION: In eight African countries, including humanitarian and pandemic-affected contexts, an evidence-based parenting programme showed consistent associations with reduced violence against adolescent girls and improved parenting and mental health.