Reducing child abuse amongst adolescents in low- and middle-income countries: A pre-post trial in South Africa.

Background: No known studies have tested the effectiveness of child abuse prevention programmes for adolescents in low- or middle-income countries. 'Parenting for Lifelong Health' ( http://tiny.cc/whoPLH ) is a collaborative project to develop and rigorously test abuse-prevention parenting programmes for free use in low-resource contexts. Research aims of this first pre-post trial in South Africa were: i) to identify indicative effects of the programme on child abuse and related outcomes; ii) to investigate programme safety for testing in a future randomised trial, and iii) to identify potential adaptations.

Methods: Two hundred thirty participants (adolescents and their primary caregivers) were recruited from schools, welfare services and community-sampling in rural, high-poverty South Africa (no exclusion criteria). All participated in a 12-week parenting programme, implemented by local NGO childcare workers to ensure real-world external validity. Standardised pre-post measures with adolescents and caregivers were used, and paired t-tests were conducted for primary outcomes: abuse (physical, emotional abuse and neglect), adolescent behaviour problems and parenting (positive and involved parenting, poor monitoring and inconsistent discipline), and secondary outcomes: mental health, social support and substance use.

Results: Participants reported high levels of socio-economic deprivation, e.g. 60 % of adolescents had either an HIV-positive caregiver or were orphaned by AIDS, and 50 % of caregivers experienced intimate partner violence. i) indicative effects: Primary outcomes comparing pre-test and post-test assessments showed reductions reported by adolescents and caregivers in child abuse (adolescent report 63.0 % pre-test to 29.5 % post-test, caregiver report 75.5 % pre-test to 36.5 % post-test, both p < 0.001) poor monitoring/inconsistent discipline (p < .001), adolescent delinquency/aggressive behaviour (both p < .001), and improvements in positive/involved parenting (p < .01 adolescent report, p < .001 caregiver report). Secondary outcomes showed improved social support (p < .001 adolescent and caregiver reports), reduced parental and adolescent depression (both p < .001), parenting stress (p < .001 caregiver report) and caregiver substance use (p < .002 caregiver report). There were no changes in adolescent substance use. No negative effects were detected. ii) Programme acceptability and attendance was high. There was unanticipated programme diffusion within some study villages, with families initiating parenting groups in churches, and diffusion through school assemblies and religious sermons. iii) potential adaptations identified included the need to strengthen components on adolescent substance use and to consider how to support spontaneous programme diffusion with fidelity.

Conclusions: The programme showed no signs of harm and initial evidence of reductions in child abuse and improved caregiver and adolescent outcomes. It showed high acceptability and unexpected community-level diffusion. Findings indicate needs for adaptations, and suitability for the next research step of more rigorous testing in randomised trials, using cluster randomization to allow for diffusion effects.