Rapid deceleration occurs when substantial force slows the speed of a vehicle. Rapid deceleration events (RDEs) have been proposed as a surrogate safety measure. As there is concern about crash involvement of older drivers and the effect of age-related declining visual and cognitive function on driving performance, we examined the relationship between RDEs and older driver's vision, cognitive function and driving confidence, using naturalistic driving measures. Participants aged 75 to 94 years had their vehicle instrumented for 12 months. To minimise the chance of identifying false positives, accelerometer data was processed to identify RDEs with a substantial deceleration of >750 milli-g (7.35m/s2). We examined the incidence of RDEs amongst older drivers, and how this behaviour is affected by differences in age; sex; visual function, cognitive function; driving confidence; and declines over the 12 months. Almost two-thirds (64%) of participants were involved in at least one RDE, and 22% of these participants experienced a meaningful decline in contrast sensitivity during the 12 months. We conducted regression modelling to examine associations between RDEs and predictive measures adjusted for (i) duration of monitoring and (ii) distance driven. We found the rate of RDEs per distance increased with age; although, this did not remain in the multivariate model. In the multivariate model, we found older drivers who experienced a decline in contrast sensitivity over the 12 months and those with lower baseline driving confidence were at increased risk of involvement in RDEs adjusted for distance driven. In other studies, contrast sensitivity has been associated with increased crash involvement for older drivers. These findings lend support for the use of RDEs as a surrogate safety measure, and demonstrate an association between a surrogate safety measure and a decline in contrast sensitivity of older drivers.
}, year = {2017}, journal = {Accident Analysis and Prevention}, volume = {98}, edition = {2016/11/05}, pages = {312-319}, isbn = {1879-2057 (Electronic)