Background: Older drivers exhibit elevated crash risk per distance driven, yet accurately identifying high-risk individuals without unjustly restricting mobility remains challenging. Whether older drivers can accurately assess their own driving-related risks remains controversial, and educational attainment-as a proxy for cognitive reserve-may influence this capacity for self-assessment.
Methods: This population-based prospective cohort study included 1673 male drivers aged 60 years and older from the Korean Longitudinal Study on Cognitive Aging and Dementia (KLOSCAD). Participants were classified into four risk phenotypes based on concordance between baseline self-reported driving concerns and subsequent crash occurrence: Drivers with inherent safety (DIS), drivers with regulated safety (DRS), drivers with recognized risk (DRR), and drivers with unrecognized risk (DUR). Self-reported and informant-reported driving problems were assessed using the driving risk questionnaire (DRQ). Crash records were obtained from the Korean National Police Agency (KNPA) over a 2-year follow-up.
Results: Of 290 drivers involved in crashes, 179 (61.7%) had reported baseline concerns (DRR), while 111 (38.3%) had not, representing 'silent' at-risk drivers-those who reported no driving concerns at baseline but subsequently experienced crashes (DUR). Lower baseline Self-Informant Discrepancy (mean SID: -0.05 in the unrecognized risk group vs. 3.15 in the recognized risk group; p < 0.001) characterized the unrecognized risk group, while informant reports showed a floor effect (94.3% zero) and did not predict crashes, necessitating objective screening. Among drivers without self-reported concerns, lower education (≤ 9 years) combined with slower TMT-B performance (≥ 180 s) identified a high-risk subgroup with a crash rate of 20.9% (OR, 2.04; 95% CI, 1.29-3.22; p = 0.002).
Conclusions: Self-reported driving concerns demonstrate predictive validity for future crashes, but a substantial minority of crash-involved drivers remain undetected by self-report. Education and TMT-B performance may serve as clinically actionable markers for identifying these "silent" at-risk drivers.
Keywords: driving safety; older adults; self‐report validity; trail making test; unrecognized risk.
© 2026 The Author(s). Journal of the American Geriatrics Society published by Wiley Periodicals LLC on behalf of The American Geriatrics Society.