top of page

Addressing the Complex Driving Needs of an Aging Population

The older adult (≥65 years) population in the US will increase to 88 million by 2050 and represent at least 25% of all drivers. Compared with younger cohorts, older adults experience decline in visual (acuity, contrast sensitivity, processing), cognitive (divided and selective attention), and motor functions (grip strength, range of motion) that are required for driving. These age-related changes can work in combination (eg, slower reaction time) to diminish reserve and increase the risk for driving impairment (unable to safely operate a vehicle) when superimposed on chronic diseases such as dementia, stroke, traumatic brain injury, untreated sleep apnea, and Parkinson disease. However, older adults tend to drive less during rush hour, avoid nighttime driving and inclement weather, and make fewer trips and drive shorter distances.1

Deaths associated with driving increase among older drivers compared with younger drivers, with a 2.5-times higher risk of a crash death for those aged 75 to 79 years and 5-times higher risk for those 80 years or older.2 Specifically, annual fatality rates among licensed drivers increase across advanced age groups with rates of 7.4 fatalities per 100 000 licensed drivers aged 65 to 69 years and 16.9 per 100 000 drivers 85 years or older.3 This increase may be due to the association between aging and frailty, driving older vehicles, and/or being overrepresented in left-turn (incoming cross-traffic) crashes where drivers are more vulnerable and less protected than in rear-end crashes.

Medically Impaired Older Drivers

Numerous age-associated conditions increase the risk of driving impairment among older adults. Multidimensional impairments (eg, poor reaction time, mistaking brake/gas pedals) can occur in combination to impact behavior such as failing to locate familiar routes/places, spatiotemporal navigation (driving a specific route), not observing traffic signs, and maintaining lane control or appropriate speed. Neurological conditions (eg, stroke, Parkinson disease) may impact motor function with paresis, limited range of motion, tremors, or impaired coordination. Conditions such as seizures, syncope, arrhythmia, sleep apnea, and orthostatic hypotension have associated symptoms that can abruptly impact the ability to safely operate a vehicle.

In case-control studies (cases were crashes in which the driver was responsible and controls were crashes in which the driver was not responsible), medication classes associated with increased odds of crash responsibility included benzodiazepines, opioids, anticholinergic drugs, anticonvulsants, and antipsychotics.4 Determining causality between specific medication classes and motor vehicle crashes is challenging given the contributions of sedating effects of the drug/dose, the underlying medical disease/condition, or other comorbid medical conditions. Polypharmacy is a growing concern for the aging cohort, and numerous drug combinations may be additive and contribute to lethargy, fatigue, and drowsiness.

Fitness to Drive, Legal Issues, and Driving Cessation

There are no validated criterion standard measures to determine driving competency, especially given variations in licensing requirements across US states. Primary care clinicians can take several steps to assess fitness to drive among older adults. Family members should be encouraged to accompany an older driver when appropriate and acceptable to identify potentially unsafe driving behaviors. Reducing possible distractions when driving (eg, turning off cell phones) may reduce crash risk. Reversible causes of driving impairment should be investigated and corrected, including eliminating sedating medications, treating sleep apnea, and referring for cataract surgery to address visual deficits.

In many cases, the patient will stop driving when there are no reversible medical causes contributing to a decline in driving fitness. In equivocal cases, clinicians can refer patients for a performance-based road test. Comprehensive driving evaluations may be performed by an occupational therapist and/or driving instructor or through state licensing agencies. Medicare, Medicaid, and private insurance typically do not cover evaluations purely to assess driving ability, but vocational rehabilitation or workers’ compensation may cover part of the cost.5 Ethically, given the “duty to warn,” it is appropriate for clinicians to refer to their state licensing agencies for patients who drive despite increased risk of driving impairment that could put themselves or others in danger. Some states provide civil immunity for clinicians to report unsafe drivers, while others have mandatory reporting laws; however, reporting is voluntary in most states. Clinicians should be familiar with their state laws concerning reporting and seek legal counsel when developing clinic policies to refer to state licensing authorities. Driving cessation discussions should be conducted sensitively, with family support (if available) and considerations made for alternative transportation or referrals to social services to maintain social connectedness along with the ability to access important locations/destinations. Ride-sharing may preserve independence, but cost, availability, reliability, trust, and access can limit its use.

Driving in an artificial environment on a driving simulator is an alternative to conventional road examinations and tests the performance limits of how a driver responds to situations not replicable on the road (eg, pedestrian walking into traffic, vehicle cutting into a lane). Driving simulators can be used as part of driving assessments conducted by occupational therapists who have access to simulators. Training on the simulator may be offered in some inpatient rehabilitation settings. Simulation is standardized across patients and maintains driver safety, and measurement of driver performance is objective. However, some participants may develop simulator sickness (a form of motion sickness) and the high cost associated with equipment and maintenance limits availability.6 There are also concerns with validity and reliability across simulator brands, models (high-fidelity [real vehicle] vs low-fidelity [computer screen]), and scenario parameters of programmed courses. There have been very few studies correlating impaired performance on the simulator with real-world crash data and/or naturalistic driving.

Technological advances in the long-distance transmission of data and an automobile’s computer infrastructure now allow tracking of real-life driving behaviors using global positioning system (GPS) data loggers. These data loggers, typically equipped with an accelerometer, gyroscope, GPS module, and sim card, plug into the on-board diagnostic port under the dashboard of vehicles manufactured after 1995. Spatiotemporal data (speed, date, time, latitude, longitude) are instantaneously extracted from the vehicle’s computers and wirelessly transmitted to servers. Summative metrics such as miles, trips, start/end time, and starting/ending locations are captured along with speeding, hard braking, sudden acceleration, forceful cornering actions, jerk, and impacts longitudinally in real time.7 These naturalistic driving data that are collected during real-life driving can facilitate risk modeling of the likelihood of a crash, provide a longer period to capture unsafe driving behavior, serve as a proxy for decline in overall health or development of preclinical dementia,8 and inform conversations with behavioral data between clinicians and patients about a nondriving future.

Emerging Available Automotive Technologies

More than 90% of crashes are due to human error. Advanced driver-assistance systems (ADAS) are electronic technologies that assist drivers in safely operating a vehicle by detecting crash risk stimuli and reducing driver error. Technological growth has evolved from features such as antilock braking system and cruise control to heads-up displays that project dashboard information to the direct line of sight, along with continuous monitoring alerts (warnings for lane departure, blind spot vehicles, impending crashes, pedestrian movement), task support (parking, lane centering, emergency braking), and environmental monitoring (inclement weather, night, obstructed traffic signs). The system may strategically respond (eg, engage the brake) even before the driver has time to respond. However, the variety in automotive manufacturers, makes, models, and year leads to a lack of standardization. Vehicles may include or exclude different systems, and even those with similar functions may have different names. ADAS are not errorless and may function when not needed or fail to function when needed. Older drivers’ willingness to use ADAS depends on individual perceived usefulness, safety, ease of learning, and anxiety.9 No comprehensive studies have been conducted to prove the effectiveness of ADAS in reducing crash risk among older drivers.

Autonomous Vehicles

Autonomous vehicles are driverless cars that can detect, evaluate, and respond to the vehicle environment without human involvement. The Society of Automotive Engineers’ Levels of Driving Automation (level 0-5) is a taxonomy that depicts the progression from human to automated features monitoring the driving environment. Level 0 classifies vehicles with no automation (most vehicles on the road), while level 5 is full automation (concept vehicles only). Contemporary manufacturers of autonomous vehicles are only at level 2, in which a vehicle can maintain lane-keeping and safe-following behaviors. Automotive engineers predict that most vehicles will be heavily dependent on human interaction for many years before level 4 and higher automation are marketable. If the cost is quite high, older adults and individuals from low socioeconomic groups, 2 groups already vulnerable to transportation disability, might have less access to autonomous vehicles when the technologies come to market. The industry still needs to reduce the high cost of sensors, determine liability that could transform the insurance industry, refine decision-making artificial intelligence algorithms, conduct extensive testing, improve reliability in inclement weather, ensure security systems are tamper-proof, and work with federal driving regulations.10


Technology has the potential to absolve problems of the human condition, such as the inability to drive in older age. However, technology cannot yet completely supplant all human driving and it cannot independently determine crash risk and safety, but it can be used to supplement decision-making. Objective data from technologies, coupled with a clinical assessment, can help clinicians determine fitness to drive.

Article Information

Corresponding Author: David Brian Carr, MD, Washington University School of Medicine, 600 S Euclid, Box 8303, St Louis, MO 63110 (

Published Online: September 1, 2023. doi:10.1001/jama.2023.16093

Conflict of Interest Disclosures: Dr Babulal reported receiving grants from the National Institute on Aging and BrightFocus Foundation during the completion of this work. No other disclosures were reported.



Molnar LJ, Eby DW, Charlton JL, et al. Driving avoidance by older adults: is it always self-regulation?  Accid Anal Prev. 2013;57:96-104. doi:10.1016/j.aap.2013.04.010PubMedGoogle ScholarCrossref


Cox AE, Cicchino JB. Continued trends in older driver crash involvement rates in the United States: data through 2017-2018.  J Safety Res. 2021;77:288-295. doi:10.1016/j.jsr.2021.03.013PubMedGoogle ScholarCrossref


Ratnapradipa KL, Pope CN, Nwosu A, Zhu M. Older driver crash involvement and fatalities, by age and sex, 2000-2017.  J Appl Gerontol. 2021;40(10):1314-1319. doi:10.1177/0733464820956507PubMedGoogle ScholarCrossref


Brubacher JR, Chan H, Erdelyi S, et al. Medications and risk of motor vehicle collision responsibility in British Columbia, Canada.  Lancet Public Health. 2021;6(6):e374-e385. doi:10.1016/S2468-2667(21)00027-XPubMedGoogle ScholarCrossref


Betz ME, Dickerson A, Coolman T, et al. Driving rehabilitation programs for older drivers in the United States.  Occup Ther Health Care. 2014;28(3):306-317. doi:10.3109/07380577.2014.908336PubMedGoogle ScholarCrossref


de Winter JCF, van Leeuwen PM, Happee R. Advantages and disadvantages of driving simulators.  Measuring Behavior. 2012:47.Google Scholar


Babulal GM, Traub CM, Webb M, et al. Creating a driving profile for older adults using GPS devices and naturalistic driving methodology.  F1000Res. 2016;5:2376. doi:10.12688/f1000research.9608.2PubMedGoogle ScholarCrossref


Babulal GM, Johnson A, Fagan AM, et al. Identifying preclinical Alzheimer’s disease using everyday driving behavior.  J Alzheimers Dis. 2021;79(3):1009-1014. doi:10.3233/JAD-201294PubMedGoogle ScholarCrossref


Motamedi S, Masrahi A, Bopp T, Wang JH. Different level automation technology acceptance: older adult driver opinion.  Transp Res, Part F Traffic Psychol Behav. 2021;80:1. doi:10.1016/j.trf.2021.03.010Google ScholarCrossref


Fagnant DJ, Kockelman K. Preparing a nation for autonomous vehicles: opportunities, barriers and policy recommendations.  Transp Res Part A Policy Pract. 2015;77:167-181. doi:10.1016/j.tra.2015.04.003Google ScholarCrossref



bottom of page