The present project aimed to develop and validate a methodology for driving simulators to assess and diagnose driving ability of elderly drivers. A driving simulation course has been developed which covered a representative selection of driving tasks of moderate difficulty as well as scenarios which are particularly difficult for elderly drivers. Driving errors were semi-automatically registered and classified by a tablet PC application. Based on the registered driving errors, the driving competence of each driver was rated on an 11-point fitness-to-drive (FtD) rating scale by specifically trained raters. The driving course was validated on the basis of a 60-minute standardized driving test in real traffic. By including similar driving tasks, it was ensured that it was structurally comparable to the simulated course. 30 elderly drivers (> 70 years) and 30 control drivers (25-50 years) were assessed in the simulation and in real traffic. During the driving tests, more driving errors were registered for the elderly drivers than for the controls both in the simulator and in real traffic. FtD-ratings and total number of driving errors during the driving tests in the simulation correlated up to r = .80 with the FtD-ratings of the driving tests in real traffic. ROC-Analyses revealed at Sensitivity-Specificity Ratio of 85.71 : 82.61 at best. Overall, driving simulation was well accepted by the subjects. The findings of the study confirm the validity of driving simulation as a tool to diagnose driving ability and argue for its introduction as a diagnostic tool. (Author/publisher)
Samenvatting