Prior research on cognition and everyday functioning in older adults has focused on cognitive predictors of everyday functioning. However, there is some evidence that subtle changes in self-reported everyday functioning may precede cognitive decline. The present study focused on one specific aspect of everyday functioning, driving, in relation to multiple domains of cognition in a longitudinal study. The aim of the present study was to determine whether changes in different aspects of self-reported driving predict cognitive decline.

A sample of 2,802 older adults who were cognitively healthy at the start of the Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) Study aged 65 to 94 years (Mage = 73.6, SD = 5.9, 75.9% female) were included in the current analysis. Participants took part in follow-up assessments at 1, 2, 3, 5, and 10 years post-intervention which included the completion of questionnaires and neuropsychological tests. Surveys were administered to assess self-reported driving (e.g., avoidance of specific driving situations, difficulty experienced in specific driving situations). Neuropsychological tests of memory [Auditory Verbal Learning Test (AVLT), Hopkins Verbal Learning Test (HVLT), Rivermead Behavioral Memory Test (RBMT)], processing speed [Useful Field of View (UFOV)], and reasoning [Letter Sets, Letter Series, Word Series] were administered. Confirmatory factor analysis on the neuropsychological tests administered yielded factor scores for each cognitive domain assessed: memory, reasoning, and processing speed. Bivariate latent change score models with coupling were used with each driving variable as a predictor of each cognitive functioning factor over time in order to examine the directionality of the relationship between these two variables.

In terms of memory, greater driver avoidance (estimate = -4.74, p < .001) and cognitive decline (estimate = 7.1, p < .05) had a bidirectional association (i.e., both variables predicted changes in the other variable over time). In terms of reasoning, greater driving difficulty (estimate = -0.27, p < .05) and cognitive decline (estimate = 1.68, p < .05) also had a bidirectional association. In terms of processing speed, cognitive decline (estimate = 0.47, p < .001) predicted greater driver avoidance over time.

Our research extends prior research on cognition and everyday functioning in showing that changes in specific aspects of self-reported driving both precede and follow changes in different neuropsychological domains. Future research might work to examine objective naturalistic driving behavior via GPS as an indicator of future cognitive decline and greater risk for neurodegenerative conditions.