The Mediating Effect of Marriage on the Association Between Young Adult and Middle Adult Substance Use

Introduction:Although high rates of adolescent substance use generally decline in young adulthood, some individuals are at risk to progress into problematic alcohol use in adulthood (Brook et al., 2002; Chassin, Fora, & King, 2004). Indeed, about 6% of adults aged 26 or older had an alcohol use disorder in 2014 in the US (Center for Behavioral Health Statistics and Quality, 2015). The causal pathways regarding development of an alcohol disorder in adulthood are poorly understood. We investigate whether marriage mediates the association between alcohol use behaviors in early and middle adulthood, theorizing that marriage may confer a protective effect.

Methods:Data were from the Monitoring the Future (MTF) study, an ongoing series of surveys on nationally representative samples of 12th-graders in the US. Analyses included panel data from 9,052 individuals who were high school seniors in 1990-1998 and were followed subsequently throughout their 20s and at age 35. Inverse probability of treatment weighting (IPTW) was used to obtain unbiased estimates of the outcome (i.e., alcohol use at age 35 [no use, non-disordered use, disordered use]) in the presence of confounding of the exposure (i.e., alcohol use at age 19/20 [no use, infrequent use, frequent use]) and the mediator (i.e., marital status at age 29/30 [not married, married]). Different scales of coefficients from IPTW logistic regression models were standardized for the estimation of mediation (Kenny, 2013).

Results:Infrequent and frequent alcohol use at age 19/20 (compared to no use) were positively associated with the likelihood of disordered use (OR= 1.62 and 1.86, respectively, p <.001) and negatively associated with the likelihood of abstaining (OR=0.23 and 0.65, respectively, p<.001) at age 35, compared to non-disordered use. Furthermore, infrequent alcohol use at age 19/20 (compared to no use) was positively associated with being married at age 29/30 (OR=1.09, p <.001). Marriage at age 29/30 was negatively associated with the likelihood of disordered use at age 35 (OR= 0.61, p <.001) but positively associated with the likelihood of abstaining (OR= 1.16, p <.001), compared to non-disordered use. The mediated effects of marriage were significant, but small; 0.3% to 2.5 % of the association between age 19/20 alcohol use and age 35 alcohol use was explained by marriage at age 29/30.

Conclusions: Findings suggest that, while marriage is a significant mediator of early and later alcohol use, the magnitude of this effect is small; young adult alcohol use has stronger effects on the development of problematic alcohol use in middle adulthood. The discussion will focus on the intervention implications of the effects of young adult alcohol use behaviors.