Errors versus speed on the trail making test: Relevance to driving performance

Abstract Background/Objectives Many studies have demonstrated that speed to complete items on the Trail Making Tests (TMT A and TMT B) is useful in the prediction of driving safety. However, there is no consensus regarding optimal “cut scores” to discriminate between safe and unsafe drivers. In this study, we examine TMT speed and errors in drivers referred for a road test. Design Retrospective analysis. Setting Patients referred for a DriveWise® evaluation at Beth Israel Deaconess Medical Center in Boston, Massachusetts. Participants Drivers age 65 or older were included (total n  = 373). Forty-five percent of the sample had been diagnosed with Cognitive Impairment (CI) whereas the remaining participants were in the No Cognitive Impairment (NCI) group. Measurements TMT Parts A & B, Folstein Mini Mental Status Examination, Washington University Road Test. Results CI drivers with TMT A speed exceeding 46 s were more likely to fail the road test whereas TMT B speed was not a sensitive metric in this group. In the No Cognitive Impairment (NCI) group, TMT B speed exceeding 131 s predicted driving impairment whereas TMT A speed was not sensitive. Error scores were not useful in the determination of driving fitness for either group. Conclusions This study provides useful criteria for health providers working with older people in the determination of driving fitness. Results suggest that TMT speed, but not error rate, is associated with road test performance. Based on our work, we advocate that pre-existing dementia should be taken into consideration when using TMT performance as a screen for driving.