Variables Predicting Clinical Decision-Making to Drive: A Retrospective Analysis

Introduction.Driving is a dynamic activity involving physical, visuo-perceptual, and cognitive skills.There are multiple domains to assess in persons returning to drive -visual, cognitive, motor, and on-road assessments.It is recommended to perform an on-road driving assessment based on a battery of off-road tests.The objective was to identify the most important tests from the battery of tests that best predict the recommendation for an on-road driving assessment in our Driving and Mobility Services (DMS) Clinic. Methods.A retrospective analysis of our driving service data was gathered from 2017 to 2019.We analyzed data from 98 patients (65 men; mean age 68.8±14.2years).The patients were referred to the DMS by clinical departments in the University of Kansas Health System.Four key functions that were extracted from the dataset were: vision, motor function, cognition, and simulated driving assessments.Results.A backward linear regression identified possible predictors of the outcome, the clinical decision to drive.The analysis showed that the Montreal Cognitive Test (OR = 1.17, p = 0.01), break reaction time (OR = 0.12, p = 0.002), history of at-fault collision in the past five years (OR = 0.16, p = <0.001),Trail Making Test A (OR = 0.96, p = 0.01), Road Sign Recognition Test (1.42, p = 0.005), Dot Cancellation Test (OR = 0.97, p = 0.03) had the most influence on our decision to recommend a practical driving assessment or not; 52.9% of variance in the decision was explained by the model.Conclusions.Among several physical, visual, cognitive, simulator-based assessments, we were able to identify the top six variables that were predictive of clinical decision-making to permit driving.