Keywords: physical activity/inactivity, exercise, sports, adolescence, body weight, body mass index, obesity, longitudinal studies, randomized trials, matching
Evidence of associations between physical inactivity in adolescence and obesity in adulthood has been inconsistent; past studies used regression methods which may lead to residual confounding. This study evaluates whether adolescent physical activity predicts obesity risk among adults aged 47-55 years, using full matching in the National Longitudinal Study of Youth 1979 with the R MatchIt library. Matching balanced the physically active and inactive groups on 10 potential confounders. Post-matching, our sample comprised 6331 physically active and 3307 physically inactive respondents. Bivariate analysis and multivariate regressions were conducted, controlling for 10 potential confounding variables. Adults who had been physically active in adolescence had 41% lower odds of obesity (adjusted odds ratio = 0.59, 95% CI=0.54,0.64), in multivariate regression in the matched sample. Interventions that help adolescents maintain physical activity may improve mid-life health. Subsequent analysis will use sensitivity analysis to evaluate the extent to which these differences can be attributed to an unobserved confounder.