In this study we examine the effect of several alcohol-related measures on self-reported collision involvement within the previous 12 months while controlling for demographic and driving exposure factors based on a large representative sample of adults in Ontario. Data are based on the 2002-2006 Centre for Addiction and Mental Health Monitor, an ongoing cross-sectional telephone survey of Ontario adults aged 18 and older (n=8542). Three logistic regressions of self-reported collision involvement in the past 12 months were implemented, each consisting of 3 steps: (1) demographic factors and driving exposure entered, (2) driving after drinking within the last 12 months entered, and (3) one of three alcohol-related measures (AUDIT subscales of alcohol consumption, dependence and problems) entered. In each step, measures from the preceding step were included in order to control for those variables. In Step 1, age (OR=0.989), region overall, Central East region (OR=0.71), West region (OR=0.67), and North region (OR=0.67), income overall and those who did not state income (OR=0.64), marital status overall and those married or living common law (OR=0.60), and number of kilometers driven in a typical week (OR=1.00) were found to be significant predictors of collision involvement. The analyses revealed that driving after drinking was a significant predictor of collision involvement in Step 2 (OR=1.51) and each of the Step 3 models (ORs=1.52, 1.37, 1.34). The AUDIT Consumption subscale was not a significant factor in collision risk. Both the AUDIT Dependence and AUDIT Problems subscales were significantly related to collision risk (ORs=1.13 and 1.10, respectively). These findings suggest that alcohol, in addition to its effects on collision risk through its acute impairment of driving skills, may also affect collision risk through processes involved when individuals develop alcohol problems or alcohol dependence.
2010 Elsevier Ltd. All rights reserved.