Moderators of treatment effects in a child maltreatment prevention programme in South Africa

Previous research has found mixed results on whether the most disadvantaged families benefit as much as less disadvantaged families from parenting interventions designed to reduce child maltreatment, and little in known in low-income settings.

In this study, we test the effects of child, caregiver, household, and community characteristics as treatment moderators of intervention outcomes - child maltreatment and parenting practices. We test characteristics previously examined elsewhere as well as factors relevant to the South African context.

This analysis includes adolescents (ages 10-18) and their caregivers (N = 552 pairs) who participated in a randomised trial of a parenting programme in the Eastern Cape Province of South Africa.

Data from the caregiver and adolescent standardised questionnaires collected at baseline, post-test (1-month post-intervention), and follow-up (5-9 months) were analysed using longitudinal multilevel analyses. We tested seven hypothesised moderators for each of the primary outcomes through interactions of treatment effect with baseline moderators.

No moderator effects were statistically significant after correcting for multiple comparisons testing. Hence, in line with several recent studies examining moderation effects in parenting programmes, our study suggests that parenting interventions aiming to reduce child maltreatment and promote parenting skills in low- and middle-income countries may be similarly effective for families facing various levels of economic, social, and health risk factors.

It may be useful to explicitly power trials for testing moderator effects, study different types of moderators and use person-centred analyses to further understand variations in treatment effects.