Keywords: health disparities, educational attainment, matched sampling, causal mediation, propensity score matching
This study evaluates suspension during adolescence predicts heart disease among middle-aged adults from the National Longitudinal Study of Youth 1979, and whether heart disease risk appears mediated by factors that predict educational attainment. We performed Mahalanobis matching within propensity score calipers with exact matching on test score tertile and evaluated potential mediation by standardized test score using stratification and causal mediation by test score tertiles and parent education. Matching identified 1501 non-suspended adults similar to the 1016 suspended adults. In multivariate analysis within the matched sample, suspended adults had greater odds of heart disease at age 40 (OR 1.52 (0.98, 2.38)) and age 50 (OR 1.80 (1.31, 2.48)) than non-suspended adults. Test scores mediated 21% (95% CI (8%, 61%)) of the effect of suspension on heart problems at age 40 and 9% (0%, 22%) at age 50, according to the causal mediation model. School suspension predicts worse adult health at mid-life. Higher standardized test scores attenuate the effect of school suspension on heart disease at age 40 but not at age 50, which may indicate earlier onset among suspended adults.