Assessing the cost-effectiveness of economic strengthening and parenting support for preventing violence against adolescents in Mpumalanga Province, South Africa: An economic modelling study using non-randomised data

Abstract

There is limited evidence around the cost-effectiveness of interventions to reduce violence against children in low- and middle-income countries. We used a decision-analytic model to evaluate the cost-effectiveness of three intervention scenarios for reducing adolescent emotional, physical, and sexual abuse in Mpumalanga Province, South Africa. The intervention scenarios were: 1) Community grant outreach to link households to South Africa’s Child Support Grant (CSG) if they are eligible, but not receiving it; 2) Group-based parenting support; and 3) Group-based parenting support ‘plus’ linkage to the CSG. We estimated average cost-effectiveness ratios (ACERs) for intervention scenarios over a ten-year time horizon, and compared them to a South Africa-specific willingness-to-pay (WTP) threshold (USD3390). Health effects were expressed in disability-adjusted life years (DALYs) averted. Our model considered four combinations of routine service versus trial-based costing, and population-average versus high prevalence of violence. Under routine service costing, ACERs for grant outreach and parenting support were below the WTP threshold when considering a population-average prevalence of violence USD2850 (Lower: USD1840-Upper: USD10,500) and USD2620 (USD1520-USD9800) per DALY averted, respectively; and a high prevalence of violence USD1320 (USD908-USD5180) and USD1340 (USD758-USD4910) per DALY averted, respectively. The incremental cost-effectiveness of parenting support plus grant linkage relative to parenting support alone was USD462 (USD346-USD1610) and USD225 (USD150-USD811) per DALY averted at a population-average and high prevalence of violence, respectively. Under trial-based costing, only the ACER for grant outreach was below the WTP threshold when considering a high prevalence of