Abstract
Background: Public health attempts to prevent obesity in children and young people should aim to minimize health inequalities. Two Cochrane reviews examining interventions aiming to prevent childhood obesity found that interventions promoting (only) physical activity have a small beneficial effect on BMI for people aged 5–18 years, as do interventions promoting physical activity alongside healthy eating for 5–11 year olds. We examined whether the effectiveness of the interventions included in these reviews differed according to eight factors associated with inequity: place, race/ethnicity, occupation, gender/sex, religion, education, socio-economic status, and social capital (the PROGRESS framework).
Methods: We collected data on change in BMI (standardized or unstandardized), subgrouped by baseline measures of PROGRESS factors, for intervention and control groups, from trial authors. We calculated the intervention effect per subgroup (mean difference), then contrasted these to estimate interactions between intervention and the baseline factors. We combined interaction estimates for each factor across trials using meta-analyses.
Findings: We collected subgrouped data from 81 trials that took place between 2001 and 2020, involving 84,713 participants. We found no substantial differences in effectiveness of interventions for PROGRESS subgroups in most scenarios. However, in the younger age group (5–11 years), the effect of interventions on standardized BMI appeared to be higher in boys (average difference in mean differences 0.03 (95% CI 0.01 to 0.06; 45 studies, n = 44,740), which was consistent in direction with the BMI effect (average difference in mean differences 0.06 kg/m2; 95% CI −0.02 to 0.13; 31 studies, n = 27,083).
Interpretation: Our findings suggest that those responsible for public health can promote these beneficial interventions without major concerns about increasing inequalities but should be mindful that these interventions may work better in boys aged 5–11 years than girls. More data are needed, so we encourage future trialists to perform subgroup analyses on PROGRESS factors.