AI Take the Wheel: Revolutionising Elderly Driver Assessments
The 90-year-old man sits confidently across the desk, a retired carpenter with weathered hands and an independent streak. Harold has been a patient for eight years. Today's appointment is his mandatory annual driving assessment, required by NSW law since he turned 75. A man who doesn't own a mobile phone shouldn't make you think of artificial intelligence, but he is someone who might be brilliantly placed to benefit from AI
"Doctor, I've been driving longer than you've been alive," he says with a dismissive wave. "Seventy-three years behind the wheel without a single accident that was my fault. I don't need some bureaucrat telling me I can't drive."
His chart tells a different story—type 2 diabetes with peripheral neuropathy, moderate hearing loss he refuses to acknowledge, ten different medications, and suspected early cognitive impairment. Just last week, the pharmacist mentioned Harold had tried to pay for his prescriptions three times, forgetting each previous attempt.
His family physician is in an unenviable position. On the desk lie the standard assessment tools: A MOCA paper based test which will involve asking a veteran to name animals, maze tests that resemble those on the back of children's restaurant menus, and a brochure for occupational therapy driving assessment—$1,000, with a four-month wait. Harold lives on the pension.
But what if instead of arguments about memory and the cost of testing, Harold's assessment relied on months of objective data from his actual driving behavior? What if location tracking from his car had already revealed concerning patterns—shorter trips, confusion in familiar areas, harsh braking incidents—that neither Harold nor his doctor could dispute? What if this conversation could be collaborative rather than adversarial?
This technological revolution is already beginning. For family physicians across NSW, these conversations represent some of the most emotionally challenging aspects of modern medicine. Research reveals that physicians find elderly driver assessments more difficult than end-of-life care discussions, considering driving conversations among the most contentious and emotionally challenging encounters in clinical practice (1). They are tired of getting yelled at for trying to do their job with inadequate tools. But emerging technologies promise to transform this broken system entirely.
The digital revolution in driving assessment
Recent breakthrough research demonstrates that vehicle tracking and artificial intelligence could eliminate the guesswork from elderly driver assessment, replacing subjective office consultations with objective, real-world driving behavior analysis.
Location tracking data as a digital biomarker
Studies published in Alzheimer's Research & Therapy show that driving location patterns can predict cognitive decline with remarkable accuracy (2). Machine learning analysis of one year's driving location data can distinguish cognitively normal older drivers with preclinical Alzheimer's disease from those without, achieving an F1 score of 0.82—indicating high robustness and precision (2).
The technology monitors comprehensive driving patterns including trip frequency, route complexity, hard braking events, speed variations, and spatial navigation patterns. Drivers with early cognitive changes show distinctive digital signatures: shorter distances, fewer unique destinations, smaller driving areas, and increased difficulty with complex navigation.
A landmark study of 2,990 drivers aged 65-79 used simple location tracking loggers connected to vehicles' diagnostic ports, collecting detailed data on speed, location, and driving patterns over months (3). The results were revelatory—the technology could identify at-risk drivers before clinical symptoms became apparent.
Real-world driving behavior trumps clinical guesswork
This approach addresses the core frustration family physicians face—making life-altering safety decisions based on 15-minute consultations rather than actual driving performance. Real-time data on harsh braking, route confusion, speed inconsistencies, and near-misses would provide objective evidence that neither doctors nor patients could dispute.
Consider Harold's case with location tracking monitoring. The data might reveal that over the past six months, his driving radius has shrunk by 40%, he's had 23 hard braking incidents in familiar areas, and he's repeatedly gotten lost on routes he's driven for decades. This objective evidence would transform the conversation from accusatory assessment to collaborative problem-solving.
Commercial vehicle tracking devices can now monitor dozens of safety-relevant driving behaviors: sudden acceleration and braking patterns, speed variations, route complexity errors, night driving patterns, and frequency of trips to unfamiliar locations.
AI algorithms that see what humans miss
Machine learning algorithms can identify subtle patterns in driving data that human observers would miss. Research shows these systems can detect early cognitive decline markers through navigation difficulties before clinical symptoms emerge (2), medication effects through changes in reaction times (4), vision problems through speed reductions (5), and physical limitations through difficulty with complex maneuvers (6).
The limits of clinical intuition
Family physicians sometimes report having a "gut feeling" about certain patients' driving safety—a sense of concern based on years of knowing a patient, observing subtle changes in cognition or physical ability, or hearing worrying stories from family members. These instincts shouldn't be dismissed entirely; they often reflect pattern recognition from extensive clinical experience.
However, the evidence shows that physician clinical judgment alone is limited for driving assessment. Small international studies suggest that most doctors' clinical judgment alone has limited ability to predict driving assessment results. One study highlighted doctors could only predict driving safety in 62-78% of cases when compared to actual on-road testing results (7). When physicians base predictions primarily on office cognitive tests, the correlation with actual driving ability is limited(8). Further Australian data in larger studies would be useful.
Perhaps most tellingly, only 33% of non-GP specialists report confidence in their ability to assess fitness to drive, and 73% feel they would benefit from further education (9). These aren't the statistics of a reliable assessment method. That said, there is probably most utility for the GP to filter out low risk cases, leaving the moderate to high risk for further assessment by on roads testing.
Occupational therapy driving assessments remain the gold standard, providing comprehensive evaluation of actual driving performance. But at $800-1200 with no Medicare rebate and months-long waiting lists, they're inaccessible to most patients. The question isn't whether a family physician's clinical intuition has value—it's whether we should continue forcing doctors to make high-stakes decisions with inadequate tools when better technology exists.
Satellite navigation as a mobility aid
Paradoxically, navigation technology also helps older drivers stay on the road longer and safer. A comprehensive study of 895 older drivers found that those using satellite navigation systems drove more frequently than non-users, particularly among drivers with declining spatial orientation abilities (10).
Navigation systems compensate for age-related wayfinding difficulties (11), allowing drivers with cognitive changes to maintain mobility safely for longer periods. Modern systems provide turn-by-turn voice guidance, real-time traffic alerts, automatic route recalculation, and emergency assistance features.
The patients who break your heart
The current system creates heartbreaking scenarios that technology could prevent. Take "Louis," a 67-year-old with vascular dementia who insisted all his accidents were "other people's fault" (12). Location tracking data would have revealed objective evidence of unsafe driving patterns that family members and medical professionals could address collaboratively.
Then there's 75-year-old "Olive," whose dementia meant she lacked insight into her deteriorating abilities. Continuous vehicle monitoring would have identified declining performance before dangerous incidents occurred.
These scenarios highlight how objective data could transform adversarial relationships into supportive interventions. Instead of family physicians becoming the "bad guys" who take away independence, technology could provide early warning systems that help maintain safe mobility longer.
When therapeutic relationships shatter
The emotional toll on family physicians is profound under the current system. Research reveals that 75% believe reporting patients negatively impacts the doctor-patient relationship (13). Even more concerning, 46% experience undue pressure from patients to reconsider their reporting decisions, while 23% have had patients leave their practice over license revocation.
Location tracking-based assessment could eliminate this role conflict entirely. Instead of making subjective judgments that damage relationships, family physicians could review objective data collaboratively with patients. The conversation shifts from "I think you're unsafe" to "the data shows some concerning patterns—let's work together to address them."
Technology could restore family physicians to their proper role as healers and advocates rather than reluctant gatekeepers.
The promise of autonomous vehicles
Perhaps the ultimate solution lies not in better assessment systems but in technology that eliminates the need for them entirely. Self-driving cars could revolutionize elderly mobility, removing difficult decisions about when to stop driving while maintaining independence and safety.
Since 94% of crashes result from human error, full automation would virtually eliminate driver-related accident risk (14). Current Advanced Driver Assistance Systems (ADAS) already help elderly drivers stay safer longer: automatic emergency braking compensates for slower reaction times, lane-keeping assistance helps with vision changes, and blind-spot monitoring addresses physical limitations (15).
Research consistently shows that driving cessation leads to depression, social isolation, and higher mortality risk, with older adults four to six times more likely to die within three years of stopping driving (16,17). Autonomous vehicles could maintain mobility without requiring active driving capability.
However, elderly drivers show mixed attitudes toward fully autonomous vehicles, preferring lower levels of automation that maintain some driver control (18). This suggests gradual implementation may be more acceptable than immediate transition to full automation.
When families become allies instead of adversaries
Current family dynamics add complexity to already difficult situations. 82% of family physicians receive indirect contact from families expressing concerns, but these conversations often create more problems than they solve (19).
Vehicle tracking monitoring would transform family dynamics from secretive reporting to collaborative safety monitoring. Instead of adult children calling doctors behind their parents' backs, families could review objective driving data together, making decisions based on evidence rather than accusations. Technology provides transparent, non-judgmental safety assessment that protects family relationships while addressing legitimate safety concerns.
A call for urgent technological implementation
The professional medical community should embrace technological solutions that could transform elderly driver assessment.
Make vehicle tracking monitoring standard for elderly drivers
Mandatory vehicle tracking for drivers over 75 would provide continuous, objective assessment of driving ability. Unlike annual clinical evaluations, tracking systems provide 365 days of real-world performance data.
The technology exists and costs are minimal—basic vehicle tracking devices cost under $200 with minimal monthly fees. Government subsidy could make this technology universally accessible while providing far superior safety assessment than current clinical methods.
Accelerate autonomous vehicle deployment for elderly populations
Targeted deployment of autonomous vehicles in elderly communities could begin immediately with current technology. Government investment in autonomous shuttle services for elderly populations could provide immediate mobility solutions.
The path forward: embracing the digital revolution
The research is clear: technology offers comprehensive solutions to problems that have plagued the medical community for decades. Vehicle tracking provides objective assessment, AI algorithms identify subtle changes before they become dangerous, and autonomous vehicles promise to eliminate the problem entirely.
Pilot programs could begin immediately in NSW communities: vehicle tracking trials for willing elderly drivers, AI assessment validation studies comparing tracking data to clinical evaluations, and autonomous shuttle services in retirement communities. These programs would provide real-world data on technology effectiveness while beginning the transformation away from the current broken system.
Conclusion: Stop the mucking around, embrace the future
The elderly driver assessment crisis in NSW represents more than a clinical challenge—it's a fundamental threat to the therapeutic relationships that underpin quality healthcare. But unlike many healthcare crises, this one has a technological solution that's already emerging.
It's time to stop mucking about with maze tests and Trail Making assessments that we know have limited ability to predict real-world driving performance (20,21). We need better tools to be making life-altering decisions. We know occupational therapy driving assessments are the gold standard, but they're expensive and inaccessible (22). We know vehicle tracking monitoring and AI assessment provide objective, continuous evaluation of actual driving behavior.
The technology is here. The evidence is overwhelming. It's time to get on with AI-guided assessments.
Vehicle tracking and AI assessment eliminate the guesswork that forces family physicians into adversarial relationships with their patients. Autonomous vehicles eliminate the problem entirely by maintaining mobility without requiring driving ability. This isn't just better for patients—it's better for family physicians, better for families, and better for community safety.
Harold wouldn't need to sit defensively across from his doctor, denying obvious safety concerns. His doctor wouldn't need to get yelled at when he simply wants to protect the community. Instead, they could review months of objective driving data together, exploring how technology might help Harold maintain safe mobility for years to come—whether through enhanced driver assistance, satellite navigation support, or eventually, autonomous vehicles.
The current system serves no one well. Patients feel attacked and lose trust in their doctors (23). Family physicians are forced into roles they're ethically uncomfortable with (13). Families are caught in impossible situations. Communities remain at risk because subjective assessments miss dangerous drivers while restricting safe ones.
We have the technology to fix this. We have the evidence to support it. We have the moral imperative to implement it. The question isn't whether change is needed—it's whether we'll have the courage to embrace solutions that could transform this broken system into one that serves everyone well.
Stop the bureaucratic inertia. Stop the technological timidity. Stop forcing doctors to be the bad guys when technology could be the solution. The tools exist to preserve therapeutic relationships, maintain elderly independence, and actually improve road safety (2,3). The time for implementation is now.
References
Betz, M. E., Jones, J., Petroff, E., & Schwartz, R. (2013). "I wish we could normalize driving health:" a qualitative study of clinician discussions with older drivers. Journal of General Internal Medicine, 28(12), 1573-1580.
Bayat, S., Babulal, G. M., Schindler, S. E., et al. (2021). GPS driving: a digital biomarker for preclinical Alzheimer disease. Alzheimer's Research & Therapy, 13(1), 115.
Babulal, G. M., Traub, C. M., Webb, M., et al. (2019). Creating a driving profile for older adults using GPS devices and naturalistic driving methodology. F1000Research, 8, 1-15.
Marshall, S. C. (2008). The role of reduced fitness to drive due to medical impairments in explaining crashes involving older drivers. Traffic Injury Prevention, 9(4), 291-298.
Freeman, E. E., Munoz, B., Turano, K. A., & West, S. K. (2006). Measures of visual function and time to driving cessation in older adults. Optometry and Vision Science, 83(10), 765-773.
Molnar, L. J., Eby, D. W., Charlton, J. L., et al. (2013). Driving avoidance by older adults: Is it always self-regulation? Accident Analysis & Prevention, 57, 96-104.
Ott, B. R., Anthony, D., Papandonatos, G. D., et al. (2005). Clinician assessment of the driving competence of patients with dementia. Journal of the American Geriatrics Society, 53(5), 829-833.
Iverson, D. J., Gronseth, G. S., Reger, M. A., et al. (2010). Practice parameter update: evaluation and management of driving risk in dementia. Neurology, 74(16), 1316-1324.
Marshall, S., Man-Son-Hing, M., Bédard, M., et al. (2013). Determining fitness to drive in older persons: a survey of medical and surgical specialists. Canadian Journal on Aging, 32(4), 304-321.
Jette, A. M., & Branch, L. G. (1992). A ten-year follow-up of driving patterns among the community-dwelling elderly. Human Factors, 34(1), 25-31.
Choi, M., Adams, K. B., & Kahana, E. (2012). Self-regulatory driving behaviors: Gender and transportation support effects. Journal of Women & Aging, 24(1), 33-53.
Yassuda, M. S., Wilson, J. J., & Von Mering, O. (1997). Driving cessation: The perspective of senior drivers. Educational Gerontology, 23(6), 525-538.
Kelly, R., Warke, T., & Steele, I. (1999). Medical restrictions to driving: the awareness of patients and doctors. Postgraduate Medical Journal, 75(887), 537-539.
Lyman, S., Ferguson, S. A., Braver, E. R., & Williams, A. F. (2002). Older driver involvements in police reported crashes and fatal crashes: trends and projections. Injury Prevention, 8(2), 116-120.
Dickerson, A. E., Molnar, L. J., Eby, D. W., et al. (2007). Transportation and aging: a research agenda for advancing safe mobility. The Gerontologist, 47(5), 578-590.
Marottoli, R. A., Mendes de Leon, C. F., Glass, T. A., et al. (1997). Driving cessation and increased depressive symptoms: prospective evidence from the New Haven EPESE. Journal of the American Geriatrics Society, 45(2), 202-206.
Ragland, D. R., Satariano, W. A., & MacLeod, K. E. (2005). Driving cessation and increased depressive symptoms. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 60(3), 399-403.
Siren, A., & Hakamies-Blomqvist, L. (2004). Private car as the grand equaliser? Demographic factors and mobility in Finnish men and women aged 65+. Transportation Research Part F: Traffic Psychology and Behaviour, 7(2), 107-118.
Johnson, J. E. (1998). Older rural adults and the decision to stop driving: The influence of family and friends. Journal of Community Health Nursing, 15(4), 205-216.
Dobbs, B. M., & Schopflocher, D. (2010). The introduction of a new screening tool for the identification of cognitively impaired medically at-risk drivers: The SIMARD a modification of the DemTect. Journal of Primary Care & Community Health, 1(2), 119-127.
Stutts, J. C., Stewart, J. R., & Martell, C. (1998). Cognitive test performance and crash risk in an older driver population. Accident Analysis & Prevention, 30(3), 337-346.
Dickerson, A. E., Reistetter, T., Davis, E. S., & Monahan, M. (2011). Evaluating driving as a valued instrumental activity of daily living. American Journal of Occupational Therapy, 65(1), 64-75.
Windsor, T. D., Anstey, K. J., Butterworth, P., et al. (2007). The role of perceived control in explaining depressive symptoms associated with driving cessation in a longitudinal study. The Gerontologist, 47(2), 215-223.