OBJECTIVES: This study investigates the impact of Attention-Deficit/Hyperactivity Disorder ("ADHD") on the prevalence and costs of accidents. The unique contribution of this research is that it estimates a model that predicts the probability of having an accident for individuals with and without ADHD. METHODS: Administrative claims data from a large employer (n>100,000) were analyzed. ADHD patients were identified as beneficiaries with one or more claims for ADHD in the study period (1996-98). The ADHD patient sample was compared to a matched control sample. In addition to descriptive analysis, multivariate analysis involving logistic regression was used to estimate a model of the probability of having an accident. RESULTS: ADHD patients have a greater incidence of accidents than their controls (32% v. 20%). However, given that a patient has an accident, the average number of accident claims is the same between both groups (3.6 v. 3.5). Accident specific costs are greater for ADHD patients than controls ($209 vs. $131) because patients with this condition are more likely to have accidents. Among patients with accident claims, ADHD patients and controls had statistically similar costs. The multivariate analysis confirms the utilization patterns found in the descriptive analysis. The odds of having an accident for ADHD patients are 1.7 times as great as that for controls after taking account of all the other variables. CONCLUSIONS: ADHD is a significant predictor of having an accident. However, given that a person has an accident, ADHD patients generally have a similar number of accident claims and costs as their controls.