Predicting at-fault car accidents of older drivers

Considerable research shows car accidents are difficult to predict using sc reening tests. The objective of this exploratory study is to determine whet her detailed accident analysis taking into account the specific accident ty pe might enhance the predictive power of a standardised road test and a set of selected neuropsychological tests. Moreover, this study addresses the v alidity and reliability of performance-based driving evaluation. The sample consisted of 84 older drivers between 65 and 96 years of age who were refe rred for a fitness-to-drive evaluation. Using discriminant analyses, the su bjects were classified as drivers with and without at-fault accidents. We c ompared the accuracy of neuropsychological tests and a road test for postdi cting all accidents, accidents classified into two categories and accidents classified into four different categories. The percentages of correctly cl assified subject were highest at the level of the most detailed classificat ion. These results suggest that, although accident prediction is difficult, the predictability of car accidents by neurocognitive measurements and a r oad test increases when the kind of accident is specified. (C) 2001 Elsevie r Science Ltd. All rights reserved.