Older adults who stop driving face 5-year mortality 68% higher than those who continue, per one cohort in Chihuri et al.’s 2016 systematic review and meta-analysis of 16 studies (12 in the US, plus Australia, Finland, and Kuwait) — also posted on the AAA Foundation for Traffic Safety’s website. No citable survey reports what share of former drivers wish they still drove, so this entry uses that mortality elevation as the best available proxy for the action side’s regret rate. The same review found driving cessation almost doubled the risk of depressive symptoms (summary odds ratio 1.91) and made former drivers nearly five times as likely to enter long-term care (HR 4.85). The underlying cost reflects what driving represents in car-dependent societies: not a leisure activity but a primary mechanism for social participation, medical appointments, and daily errands.
The comparison is not between a safe choice and a risky one. IRTAD data across OECD countries show that drivers aged 80 and older have a per-mile fatal crash rate roughly double that of drivers aged 35 to 64, and IIHS finds per-mile fatal crash involvement begins climbing at ages 70 to 74 and is highest among drivers 85 and older — elevated largely by injury susceptibility (fragility) rather than by a greater tendency to crash. About 15 to 25% of older drivers report at least one near-miss or driving difficulty in the past year — a signal that a meaningful minority are aware of declining capability even while continuing to drive. The asymmetry in the data is that continuing generates fewer subjective regrets while producing more objective risk; stopping generates more subjective regrets while producing a set of health consequences that are real but partly confounded by the health deterioration that prompted cessation.
Reverse causation is the dominant methodological concern. Drivers who stop usually do so because of a health crisis or family pressure following a crash or cognitive screen, meaning the depression and mortality outcomes measured in follow-up studies may partly reflect the pre-existing decline that triggered cessation rather than the cessation itself. This does not eliminate the causation — social isolation and loss of autonomy plausibly do worsen health independently — but it makes the direction of effect difficult to quantify. The practical implication is that the timing and circumstances of stopping matter: planned gradual reduction under the person’s own agency produces better outcomes than abrupt cessation following a crash. Country context also matters substantially: in dense urban areas with accessible public transit, the isolation effect of stopping driving is far smaller than in the US suburbs and rural areas where most American elders live.