Equity effects of parenting interventions for child conduct problems: a pan-European individual participant data meta-analysis

Childhood conduct problems are a costly public health problem and are five times more common in socially disadvantaged groups than they are in advantaged groups. Untreated, conduct problems have a poor prognosis, with increasing gaps between socioeconomic groups, and high rates of subsequent criminality. Incredible Years is a high quality parenting programme for reducing conduct problems and is widely disseminated in Europe. Many trials have shown Incredible Years to be effective but the potential effects of parenting interventions on social inequality are unknown. Some behavioural interventions (eg, smoking cessation programmes), although beneficial overall, can widen inequality gaps. Because single trials and aggregate-level meta-analyses are ill equipped for examining differential intervention (moderator) effects, we pooled individual-level trial data to assess the effects of Incredible Years on social equity. We did a systematic review and individual participant data meta-analysis by searching CINAHL, Embase, Global Health, Medline, and PsycINFO, for studies published from inception to March 15, 2019. We also searched the Incredible Years website library and consulted with experts, including the European Incredible Years mentors' network. We included data from all completed randomised trials of the Incredible Years parenting intervention in Europe that included children aged 1–12 years, including unpublished trials, without restriction on publication year or outcome measures. We included prevention (selective or universal) and treatment or indicated prevention trials (for children diagnosed or above the clinical cutoff for conduct problems). We excluded trials or conditions within trials that were not randomised, included additional non-parenting material (eg, child-focused interventions), or were abbreviated, non-standard versions of the usual Incredible Years intervention of 12–14 weekly sessions. We requested individual participant data from the study authors. The primary outcome was child conduct problems, assessed using the Eyberg Child Behavior Inventory Intensity (ECBI-I) scale. Moderators were analysed using multilevel modelling with multiple imputation. Of 15 European trials of Incredible Years parenting programmes (n=1696 children), individual participant data were unavailable for one trial and one trial did not assess the primary outcome. Children were aged 2–10 years (median 5·1), 492 (30%) of 1651 children were from an ethnic minority and