Could scale-up of parenting programmes improve child disruptive behaviour and reduce social inequalities? Using individual participant data meta-analysis to establish for whom programmes are effective and cost-effective

Background Child disruptive behavioural problems are a large and costly public health problem. The Incredible Years® (IY) parenting programme has been disseminated across the UK to prevent this problem and shown to be effective in several trials. It is vital for policy to know for which families IY is most effective, to be sure that it helps reduce, rather than widen, socioeconomic inequalities. Individual trials lack power and generalisability to examine differential effects; conventional meta-analysis lacks information about within-trial variability in effects. Objectives To overcome these limitations by pooling individual-level data from the IY parenting trials in Europe to examine to what extent it benefits socially disadvantaged families. Secondary objectives examine (1) additional moderators of effects on child behaviour, (2) wider health benefits and potential harms and (3) costs, cost-effectiveness and potential long-term savings. Design Individual participant data meta-analysis of 14 randomised trials of the IY parenting intervention. Settings UK (eight trials), the Netherlands, Ireland, Norway, Sweden and Portugal. Participants Data were from 1799 families, with children aged 2–10 years (mean 5.1 years; 63% boys). Interventions IY Basic parenting programme. Main outcome measures Primary outcome was disruptive child behaviour, determined by the Eyberg Child Behavior Inventory Intensity scale (ECBI-I). Secondary outcomes included self-reported parenting practices, parenting stress, mental health, children’s attention deficit hyperactivity disorder (ADHD) and emotional symptoms. Results There were no differential effects of IY on disruptive behaviour in families with different levels of social/socioeconomic disadvantage or differential effects for ethnic minority families, families with different parenting styles, or for children with comorbid ADHD or emotional problems or of different ages. Some moderators were found: intervention effects were strongest in children with more severe baseline disruptive behaviour, in boys, and in children with parents who were more depressed. Wider health benefits included reduced child ADHD symptoms, greater parental use of praise, and reduced harsh and inconsistent discipline. The intervention did not improve parental depression, stress, self-efficacy or children’s emotional problems. Economic data were available for five UK and Ireland trials (maximum n = 608). The average cost per person of the IY intervention was £2414. The probability that the IY intervention is considered cost-effective is 99% at a willingness to pay of £145 per 1-point improvement on the ECBI-I. Estimated longer-term savings over 20 years range from £1000 to £8400 per child, probably offsetting the cost of the intervention. Limitations Limitations include a focus on one parenting programme; the need to make assumptions in harmonising data; and the fact that data addressed equalities in the effectiveness of, not access to, the intervention. Conclusions There is no evidence that the benefits of the IY parenting intervention are reduced in disadvantaged or minority families; benefits are greater in the most distressed families, including parents who are depressed. Thus, the intervention is unlikely to widen socioeconomic inequalities in disruptive behaviour and may have effects in narrowing inequalities due to parent depression. It was as likely to be effective for older as for younger children. It has wider benefits for ADHD and parenting and is likely to be considered to be cost-effective. Researchers/funders should encourage data sharing to test equity and other moderator questions for other interventions; further research is needed on enhancing equality of access to interventions.