Abstract
Background: This study examined the associations between measures of alcohol abuse/dependence (AAD) and several classifications of offending behaviour to age 30 in a New Zealand birth cohort. Methods: Outcomes included: assault; use of a weapon; theft/burglary/vehicle conversion; property damage/vandalism/arson; and fraud/embezzlement/misappropriation of funds. The study also used measures of AAD symptoms; and time-dynamic covariate factors including life stress, other substance use, mental health status, peer and partner substance use and offending, and unemployment. Data were analysed using conditional fixed effects regression modelling augmented by time-dynamic covariate factors to control for confounding. Results: Those with five or more AAD symptoms had unadjusted odds of offending that ranged from 6.23 to 21.25 times higher than those with no symptoms, with little evidence to suggest these associations varied with age. Adjustment for both unobserved fixed effects and time-dynamic covariate factors reduced the magnitude of the associations between AAD and offending, with those with five or more AAD symptoms having odds of offending that were 0.88-4.10 times higher than those with no symptoms. After adjustment, only the associations between AAD and: a) assault (OR. =. 4.10; 95% CI. =. 1.91-8.62; p<. 0.0001); and b) property damage/vandalism/arson (OR. =. 3.87; 95% CI. =. 1.30-11.39; p<. 0.0001); remained statistically significant. Conclusions: The results suggest a causal association between alcohol misuse and " impulsive" crimes such as assault and property damage/vandalism/arson, with estimates suggesting that AAD accounted for approximately 9.6-9.9% of these types of reported offending in the cohort.