BACKGROUND: As the number of older adult drivers increases, distinguishing safe from unsafe older adult drivers will become an increasing public health concern. We report on the medical and functional factors associated with vehicle crashes in a cohort of Alabama drivers, 55 years old and older. METHODS: This prospective study involved 174 older adults, on whom demographic, medical, functional, and physical performance data were collected in 1991. Subjects were then followed through 1996 for incident vehicle crashes. RESULTS: Sixty-one subjects experienced between one and four police-reported vehicle crashes during the study period. Following adjustment for age, race, days driven per week, and gender, Cox proportional-hazards models showed the following variables to be associated with crash involvement: reported difficulty with yardwork or light housework (relative risk [RR] = 2.1; 95% confidence interval [CI] 1.1, 4.0; p = .02), or opening ajar (RR = 3. 1; 95% CI 1.4, 6.7; p = .004); at least one crash before 1991 (RR = 2.1; 95% CI 1.2, 3.7; p = .008); using hypnotic medication (RR = 2.9; 95% CI 1.3, 6.6; p = .01); self-reported stroke or transient ischemic attack (RR = 2.7; 95% CI 1.1, 6.6; p = .03); scoring within the depressed range on the Geriatric Depression Scale (RR = 2.5; 95% CI 1.1, 6.0; p = .03), and failing the useful field-of-view test (RR = 1.9; 95% CI 1.0, 3.5; p = .05). CONCLUSIONS: Variables related to function, medication use, affect, neurological disease, and visuocognitive skills were associated with vehicle crash involvement in this cohort. Our findings suggest that multifactorial assessments are warranted to identify at-risk older drivers.