Using Serial Trichotomization to Determine Fitness to Drive in Medically At-Risk Drivers

This study examines if serial trichotomization of cognitive tests (i.e., MoCA, Trails A and B, UFOV subtests 1-3) predicts passing or failing a comprehensive driving evaluation (CDE) with 100% sensitivity and 100% specificity in 142 medically at-risk drivers (72% male, mean age = 69.2 years, SD = 14.1) referred by physicians for suspicion of being unsafe to drive. Approximately 32.7% passed (i.e., considered safe to continue driving) and 27.5% failed (i.e., considered unsafe to continue driving) the CDE; 40.1% had indeterminate outcomes (i.e., required lessons and retesting). Serial trichotomization of cognitive tests predicted more pass (43.7%) and fail outcomes (35.4%), reducing the number of indeterminate outcomes to 21.8%. These findings show that serial trichotomization of cognitive tests can increase the accuracy of making evidence-informed decisions and reduce the number of unnecessary on-road assessments for medically at-risk drivers referred by physicians for CDEs.