Multiple-domain predictors of problematic alcohol use in young adulthood

The goal of this study was to identify predictors of problematic young adult alcohol use.

The sample consisted of 141 subjects (81 females) participating in a national study of genetic risk factors for alcoholism. All subjects were evaluated first as children or adolescents, then approximately 5 years later as young adults. Outcome consisted of the number of alcohol symptoms (0-10) endorsed at this second time point. Predictors of outcome were drawn from five domains representing: (1) Demographic Characteristics, (2) Child/Adolescent Problematic Alcohol Use, (3) Biological Risk, (4) Externalizing Behaviors, and (5) Family Environment. A two-stage analytic strategy was used in which (1) separate multiple regression analyses were conducted within each of the five domains and (2) statistically significant predictors of problematic alcohol use from each domain were combined into one regression model to determine which remained significant.

In the final model, 31% of the variance in the number of alcohol symptoms in young adulthood was predicted by a high number of alcohol symptoms in childhood and adolescence, low initial sensitivity to alcohol, and a negative child/adolescent relationship with the father.

These results demonstrated that GABRA2--originally associated with a diagnosis of alcohol dependence in adults--also predicted the onset of symptoms among subjects in their 20s, confirmed specific hypotheses about three other predictors in the fi nal model, and suggested the utility of incorporating biological and nonbiological predictors to optimally predict young adult alcohol problems.