DE-PASS Best Evidence Statement (BESt): Determinants of self-report physical activity and sedentary behaviours in children in settings: A systematic review and meta-analyses.

Abstract

Previous physical activity interventions for children (5-12yrs) have aimed to change determinants associated with self-report physical activity behaviour (PAB) and/or sedentary behaviour (SB), however, the associations between these determinants and PAB/SB in different settings are uncertain. The present study aimed to identify modifiable determinants targeted in previous PAB/SB interventions for children. Intervention effects on the determinants and their associations with self-report PAB/SB were assessed across settings. Search of relevant interventions from pre-defined databases was conducted up to July 2023. Randomized and non-randomized controlled trials with modifiable determinants were included. Data extraction and risk of bias assessments were conducted by two independent researchers. Where data could be pooled, we performed Robust Bayesian meta-analyses. Heterogeneity, publication bias and certainty of evidence were assessed. Fifteen studies were deemed eligible to be included. Thirty-seven unique determinants within four settings were identified-school, family, school with family/home, and community with(out) other settings. Ninety-eight percent of determinants belonged to individual/interpersonal determinant categories. Narratively, intervention effects on student perception of teachers' behaviour (school), self-management, perceived barriers, external motivation, exercise intention, parental modeling on SB (school with family/home) and MVPA expectations (community) were weak to strong, however, corresponding PAB/SB change was not evident. There were negligible effects for all other determinants and the corresponding PAB/SB. Meta-analyses on self-efficacy, attitude, subjective norm and parental practice and PAB/SB in two settings showed weak to strong evidence against intervention effect, while the effect on knowledge could not be determined. Similarly, publication bias and heterogeneity for most analyses could not be ascertained. We found no concrete evidence of association between the modifiable determinants and self-report PAB/SB in any settings. This is presumably due to intervention ineffectiveness. Design of future interventions should consider to follow the systems-based approach and identify determinants unique to the context of a setting, including policy and environmental determinants.