Road safety

Cover of Model Licence Re-Assessment Procedure for Older Drivers: Stage 2 Research
Model Licence Re-Assessment Procedure for Older Drivers: Stage 2 Research
  • Publication no: AP-R259-04
  • ISBN: 0 85588 722 2
  • Published: 1 December 2004
Stage 1 research for Austroads lead to the development of a model licence re-assessment procedure for use in licensing authorities in Australasia. Stage 2 research set out to undertake a pilot trial of the procedure in an Australian jurisdiction and to undertake an evaluation study of four promising off-road driver-screening tests for use as part of the model procedure. An initial study was undertaken in Tasmania in 2000 to evaluate the procedural aspects of the model. The results showed that the duration of administration for the screening test was within acceptable limits and that agents and participants indicated a high level of acceptance of the new procedure. A subsequent study involving 1042 drivers was carried out in New Zealand in 2001-2002 to evaluate the effectiveness of four off-road screening tests (GRIMPS, CALTEST, EDS and DriveABLE) for predicting on-road driving performance. EDS was discarded following a preliminary analysis of the results for approximately 100 participants. The final analyses involving data from approximately 300 participants performing each of the remaining three tests revealed strong relationships between on-road driving performance and test outcomes. Logistic regression analyses were undertaken to explore the goodness-of-fit of the models for various combinations of sub-tests and showed good predictive power, sensitivity and specificity measures for each test and test variant. A hypothetical licensing authority scenario was then introduced to examine the comparative performance of the three tests. These findings revealed a number of differences between the three screening tests and their variants in terms of their ability to satisfy the designated criteria and other important performance measures. While all three tests were found to have relatively good hit rates (76-80%) and reasonably good rates of false negatives, CALTEST (3 subtests, standard scoring) and GRIMPS (regressed 5 subtests) clearly out-performed the other test options. Recommendations from this research i