Family car ownership and activity in young people: cross-sectional and longitudinal analyses using the International Children's Accelerometry Database

Abstract

BackgroundUbiquitous car ownership may be negatively affecting children's activity and health. We aimed to assess the associations between household car ownership and children's inactivity and body-mass index (BMI), and between change in household car ownership and changes in children's inactivity and BMI. MethodsWe used pooled data from five cohort studies (UK, Australia, and Brazil) with valid exposure or outcome data and from four studies (UK, Australia) with longitudinal data within the International Children's Accelerometry Database (ICAD). Main outcome measures were average daily children's sedentary time (SED) and moderate-to-vigorous physical activity (MVPA) during 4 or more days assessed with accelerometery, and BMI z-score (and changes in these measures). Main exposure was household car ownership (none [9·0%, reference], one [29·7%], two or more [61·3%]) and change in ownership (no change [84·2% reference]), decrease [7·7%], increase [8·1%]). Associations were examined using mixed-effects linear regression with random intercepts, adjusted for sex, maternal education, country, BMI z-score (where relevant), and (change in) age and wear time. FindingsMean age of participants in the cohort studies (n=4193, 53·4% female) was 10·4 years (SD 2·0) and mean follow-up of participants in the longitudinal studies (1333, 54·5%) was 3·3 years (SD 1·1). Cross-sectionally, household car ownership was associated with higher SED (vs none): (1 car β 14·1 min per day, p=0·0002; ≥2 cars 12·8, p=0·0008), lower MVPA (−8·9, p<0·0001; −9·0, p<0·0001), and lower BMI z-score (−0·1, p=0·63; −0·3, p=0·022). An interaction with country (p=0·050) was observed for SED, suggesting stronger associations with a clear dose–response in Brazilian participants; no differences were observed between UK and Australian participants. Longitudinal analyses showed no associations for increased car ownership. A decrease in car ownership (93·1% households with ≥2 cars at baseline) was associated with greater decreases in MVPA (vs no change −8·5 min per day, p=0·002), but not with change in SED or BMI z-score. There was no difference in effect by country. InterpretationChildren living in households with car access are more sedentary and less active than those without. However, decreasing car access was associated with a decrease in MVPA, potentially related to changes in economic or family circumstances or reduced access to activity spaces. Decreasing car access and use are important public health targets (eg, reducing air pollution), but its potential impact on children's activity opportunities should be explored. FundingThe pooling of the data was funded through a grant from the National Prevention Research Initiative (grant no G0701877).