Clinician Effectiveness in Assessing Fitness to Drive of Medically At-Risk Older Adults

Objectives To model the relative contributions of driver data and clinical judgments to clinical ratings of driver capability for a state licensing authority and to compare ratings with on-road test results. Design Retrospective, logistic regression. Setting Missouri Driver License Bureau. Participants Adults aged 60 and older (N = 652; 52% male) evaluated by a physician of their choosing and a portion subsequently road tested (n = 286). Measurements Clinical data from an evidence-based physician statement (Form 1528). A three-level rating (likely capable, unclear, not capable) was collapsed into two outcomes (0 likely capable; 1 unclear, not capable) as the dependent variable. Independent variables (predictors) were age, sex, driving exposure, recent crash or police action, number of medical conditions, medication side effects, driver insight, and disease functional severity rating for driving. Results Three variables in the model (Nagelkerke coefficient of determination = 0.64; P <.001) were significant in the expected direction: disease functional severity for driving (odds ratio (OR = 6.65), insight (OR = 2.35), and age (OR = 1.06). Proportionately more drivers rated likely capable (73%) passed the road test than those rated unclear or not capable (62%). Conclusion Judgments of disease severity, decrements in driver insight, and older age influenced clinician ratings of driving capability. Correspondence of physician ratings to on-road test outcomes was imperfect, highlighting the complexities in translation of clinical judgments to on-road performance. Both means of assessment have important and additive roles in driver licensing.