The impact of community-based childhood obesity prevention interventions in Australia by socio-economic position: An individual participant data meta-analysis

Abstract

Introduction: Community-based interventions (CBIs) to prevent childhood obesity show promise in reducing body mass index z-scores (zBMI). Assessing whether this approach produces equitable outcomes across socio-economic sub-groups is typically limited by inadequate sample size. This study aimed to assess the effectiveness of CBIs on zBMI and weight-related behaviours by socio-economic position (SEP).

Methods: Individual participant data meta-analysis using seven Australian childhood obesity prevention CBIs conducted between 2003 and 2022. Trials used consistent intervention approaches, objectively measured height and weight, and self- or parent-reported behaviours, and lasted 2-4 years. Area-level SEP was estimated using Australian Index of Relative Socio-economic Advantage and Disadvantage tertiles based on home or school postcodes. Multi-level linear and logistic models estimated the intervention effect on zBMI and behaviours across SEP levels.

Results: While zBMI scores increased across all groups, the increase was significantly greater in control compared to intervention students (difference: -0.036 [95%CI -0.065, -0.007]), among the 25 346 observations analysed. The effect of CBIs was greater in low compared to high SEP students (intervention effect difference = -0.10 [95%CI -0.18, -0.02]).

Discussion: Obesity prevention CBIs may have greater impacts among children from low SEP areas, potentially contributing to reducing health inequities. Further research is needed to understand barriers to improving weight-related behaviours across socio-economic groups.

Keywords: childhood obesity; community‐based interventions; dietary intake; physical activity; socio‐economic position.

FIGURE 1

Intervention effect on weight‐related behaviours. Intervention effect (change in proportion of participants meeting recommendations/guidelines between baseline and endpoint) on weight‐related behaviours from mixed effect logistic regression models. Negative results indicate greater change in control group; positive results indicate greater change in intervention group. ^p < 0.05 for intervention effect. *p < 0.05 for difference in intervention effect with high SEP.