Date Published: April 24, 2018
Publisher: Public Library of Science
Author(s): Mark Tomlinson
Abstract: In a Perspective, Mark Tomlinson discusses research on early interventions to support child development in developing countries.
Partial Text: Peter Rockers and colleagues  assessed the impact on ECD of community-based home visiting incorporating health screening and parenting groups. The study was a 2-year follow-up of a cluster-randomised controlled trial conducted in Zambia with the caregivers of children between the ages of 6 and 12 months. The study included fortnightly home visits conducted by child development agents (CDAs) for the first year as well as parenting groups every 2 weeks for 2 years. The intervention significantly reduced stunting (odds ratio [OR] 0.45, 95% CI 0.22–0.92; p = 0.028) and was associated with an improvement in child language (β 0.14, 95% CI 0.01–0.27; p = 0.039), but there was no impact on other child-development outcomes. While Rockers and colleagues’ study is important and speaks to the difficulties of improving child development in low-resource contexts, one of the key conclusions of the paper is that parenting groups may be a promising avenue for improving physical growth and child development. This of course may be true, but it is more likely that the improvements in stunting were a result of the home visits of the CDAs, and not the parenting groups, given that their visits focused on screening for infections and acute malnutrition and encouraging caregivers to attend routine health services. Without an understanding of the mechanisms involved, drawing conclusions about what component of a complex intervention is the likely agent of change is difficult.
Despite significant current global health focus on scaling up interventions, knowledge is limited about scaling up programmes in ECD. In the Zambia study, one of the conclusions is that scale-up efforts would likely require a delivery platform integrated into existing structures . Unfortunately, in a low-resource setting such as Zambia, the intervention described by Rockers and colleagues is simply not scalable. The intensive nature of the intervention, including home visits and parenting groups, is beyond the means of all LMICs. The intervention in the Colombia study, on the other hand, was integrated from the start within a national programme . One of the explanations the authors proffer for the null results are concerns with extrapolating findings from efficacy trials (from which much of the current evidence comes) to interventions implemented at scale. This intervention was, however, always an integrated one implemented at scale that had positive outcomes (albeit small), and any compromises were likely there from start. Having said that, the acknowledgment that it may have been worthwhile to hire local supervisors and to increase the frequency of supervision  is illustrative of the urgent need for research that attempts to understand what is needed for successful scaling up—above and beyond programme content. When scaling up programmes, the ‘soft’ elements, such as recruitment, training, supervision, and accountability, are often the first to be dropped or reduced in frequency .