Fired from school: Evaluating whether models for the health effects of job loss explain young adults' health outcomes after high school suspension and expulsion

Stressful life events such as involuntary unemployment predict negative health outcomes years later.  High school suspension has doubled recently and disproportionately affects disadvantaged youth.  This paper evaluates whether high school suspension may increase health disparities.  We compared blood pressure, cholesterol, body mass index, waist size, depression, sleep problems, tobacco, and marijuana use in 2008 of young adults ages 25--32 initially suspended from high school in 1996 (n=438) and compared them with non-suspended youth matched on 20 socioeconomic, educational, and health factors measured pre-suspension (1995) (n=8200) in the nationally representative Add Health data.  We estimated relative risks using a multivariate Poisson working model in the matched sample, and separately in schools with strictest suspension policies. Bivariate analysis found differences in blood pressure, tobacco, and marijuana use.  The matching method, coarsened exact matching, balanced on 20 pre-suspension factors.  After matching, suspended students were 44\% more likely than non-suspended to smoke and 48\% more likely to use marijuana daily (IRR 1.44 (1.11, 1.86), p=0.006; IRR 1.48 (0.97, 2.25), p=0.07); used marijuana 29\% and 33\% more often in the past month and year (IRR 1.29 (1.15, 1.45), p=0.0001; 1.33 (1.30, 1.37), p=0.0001); and twice as likely to be dependent on cannabis (IRR 2.03 (1.13, 3.67), p=0.02). Students attending schools with harsh suspension policies were twice as likely to smoke daily (IRR 2.27 (1.22, 4.22)).  Matched sampling balanced groups on baseline smoking or marijuana, so differences are not attributable to these factors.