Validity and reliability of a driving simulator for evaluating the influence of medicinal drugs on driving performance.

Abstract

Although driving simulators (DS) are receiving increasing attention due to concern over traffic accidents under the influences of drugs, few DS are recognized for their reliability and validity. Therefore, the development of an evaluation system using DS for driving performance is urgently needed. To investigate whether the standard deviation of lateral position (SDLP) increases with blood alcohol concentration (BAC) using a DS with reliability and calculate the SDLP threshold from the difference between BAC levels of 0 and 0.05%. Twenty healthy Japanese men performed the DS tasks up to 60min in Study 1 and DS tasks twice at 1-week intervals in Study 2. Twenty-six healthy men conducted the same DS tasks under BAC level (0, 0.025, 0.05, and 0.09%) in double-blind, randomized, crossover trial in Study 3. The primary outcome was SDLP in a road-tracking test. The test-retest reliability of DS data was assessed, and the estimated difference in SDLP between BAC levels of 0 and 0.05% was calculated using a linear regression model. The cumulative SDLP values at 5-min intervals were stable, and the intraclass correlation coefficient for its values was 0.93. SDLP increased with BAC in a concentration-dependent manner. The predicted ΔSDLP value for the difference between BAC levels of 0 and 0.05% was 9.23cm. No participants dropped out because of simulator sickness. The new DS used in these studies has reliability, validity, and tolerability and is considered suitable for evaluating the influence of drugs on driving performance.