On the road to retirement: Predicting nighttime driving difficulty and cessation using self-reported health factors

Older drivers now expect to drive longer than previous cohorts and will make up about 25% of licensed U.S. drivers by 2050. Identifying early predictors of nighttime driving difficulty, a precursor to driving retirement, can inform screening procedures and timely linkage to interventions supporting driving or transitioning to driving cessation.

We examined self-reported physical and mental health baseline predictors of greater nighttime driving difficulty in five and ten years using weighted multivariate logistic analyses of 2261 drivers, aged 57 to 85, from the National Social Life, Health, and Aging Project (NSHAP). Transition matrix models describe probabilities of having greater, lesser, or the same nighttime driving difficulty after five years based on baseline driving conditions and the significant logistic model factors. We built a transition matrix tool that offers users the ability to calculate expected probabilities of change in nighttime driving difficulty based on the identified salient factors.

Five-year predictors of greater nighttime driving difficulty included perceived poor physical health (OR = 3.75), limitations to activities of daily living (ADLs; OR = 1.97), and clinical levels of depressive and anxiety symptoms (OR = 1.63; OR = 1.71). Excellent physical health (OR = 0.52), mental health (OR = 0.60), and any frequency of physical activity compared to 'never' were protective (OR = 0.37-0.51). Physical health, walking pain, and limitations to ADLs were predictive at ten-years. Transition models showed physical health and anxiety were most indicative of greater nighttime driving difficulty at 5-years for those reporting no difficulty at baseline, but limitations to ADLs were more predictive otherwise.

Lay practitioners could capitalize on the use of self-report screening measures to identify older adults who may experience near-term nighttime driving difficulty. Earlier identification may better guide long-term driving retirement planning or engagement in appropriate health interventions. The transition matrix modeling tool is freely available to facilitate development and validation of related measures.