Using cognitive status to predict crash risk: blazing new trails?

INTRODUCTION: A computer-based version of an established neuropsychological paper-and-pencil assessment tool, the Trail-Making Test, was applied with approximately 700 drivers aged 70 years and older in offices of the Maryland Motor Vehicle Administration.

METHOD: This was a volunteer sample that received a small compensation for study participation, with an assurance that their license status would not be affected by the results. Analyses revealed that the study sample was representative of Maryland older drivers with respect to age and indices of prior driving safety. The relationship between drivers' scores on the Trail-Making Test and prospective crash experience was analyzed using a new outcome measure that explicitly takes into account error responses as well as correct responses, the error-compensated completion time.

RESULTS: For the only reliable predictor of crash risk, Trail-Making Test Part B, this measure demonstrated a modest gain in specificity and was a more significant predictor of future safety risk than the simple time-to-completion measure.

IMPACT ON INDUSTRY: Improved specificity and the potential for autonomous test administration are particular advantages of this measure for use with large populations, in settings such as health care or driver licensing.