WITHDRAWN: Young Adult Substance Use and the Timing of Parenthood

Introduction: Substance use is prevalent during late adolescence and early adulthood (Johnston et al., 2015), a time in which most social role transitions also take place (Shanahan, 2000). Adult social roles are related to subsequent declines in substance use (Bachman et al., 1997, 2002; Weitzman & Chen, 2005), although it is less clear how substance use leads to variation in the timing of adult role acquisition. In the current study, we therefore investigate the associations between substance use of young adults and the timing of parenthood.

Methods: We used data from the Monitoring the Future (MTF) study, an ongoing series of surveys on a nationally representative sample of American 12th-graders, including an oversample of drug users. Our analysis involved panel data from 47,309 individuals who were high school seniors in 1976-2002 and followed subsequently until ages 29/30. Our exposures of interest were substance use behaviors (i.e., cigarette use, binge drinking, marijuana use, and other illicit drug use) at ages 19/20. Due to inherent differences between substance users and non-users, we employed a propensity score weighting (PSW) method to adjust for potential baseline confounders. We estimated a separate propensity score model for each substance use behavior that included gender, race/ethnicity, substance use indicators at age18, religiosity, cohort, high school characteristics (i.e., grade, truancy, evenings out, high school type), and family characteristics (i.e., two parent family, parental education). The absolute standardized mean difference (ASMD) revealed that sufficient balance was achieved after weighting between users and non-users of each substance (ASMD < .20). We then investigated the effect of each substance use on the timing of parenthood using a separate Cox proportional hazard model. Survival models were weighted using a composite weight comprised of the product of PSW and the MTF sampling weight.

Results: Preliminary results showed that (after balancing for potential confounders) binge drinking, marijuana and other illicit drug use at ages 19/20 were significantly related to delayed parenthood (hazard ratio [HR]=0.84, p<.001; HR=0.86, p<.001; and HR=0.94, p<.05 for binge drinking, marijuana, and other illicit drugs, respectively) although cigarette use was not related to the timing of parenthood (HR=1.03, p=.18).   

Conclusions: Findings of this study suggest that, among those who are not parents by age 20, young adults involved in substance use may delay parenthood, relative to non-substance users. However, no significant association between smoking and the timing of parenthood was found, which may reflect substantial confounding effects of adolescence characteristics on cigarette use at ages 19/20. The discussion will focus on the intervention implications of the effects of substance use on young adults.