Abstract
PURPOSE: This study examined the associations between measures of alcohol abuse/dependence symptoms and a range of psychosocial outcomes from ages 21 to 30 in a New Zealand birth cohort. METHODS: Outcome measures included measures of: criminal offending, family violence and relationship instability, sexual risk-taking and consequences, mental health, and other adverse health and adjustment outcomes. Bivariate associations between a three-level classification of alcohol misuse (no symptoms, subclinical level of symptoms, met criteria for alcohol dependence) and each outcome during the period 21-30 years were computed using Generalised Estimating Equation models. These associations were then adjusted for non-observed sources of confounding using conditional fixed effects regression modelling, augmented by time-dynamic covariate factors. For both sets of models estimates of the attributable risk (AR) were computed. RESULTS: There were statistically significant (p<.05) bivariate associations between alcohol misuse and each of the fifteen outcome measures, with estimates of the AR ranging from 7.4% to 46.5%. Adjustment for non-observed fixed effects generally reduced the magnitude of these associations; however, after adjustment, 12 of the 15 associations remained statistically significant (p<.05). Estimates of the AR after adjustment for fixed effects ranged from 3.6% to 44.3%. CONCLUSIONS: The results suggest that there are pervasive and persistent linkages between alcohol misuse and a range of adverse psychosocial outcomes. A reduction in levels of alcohol misuse amongst individuals of this age group could reduce substantially the overall level of personal and societal cost of hazardous levels of alcohol consumption.