The most important impact of age-related illness on transportation is likely to be a reduction of personal mobility. This has been demonstrated for people with dementia,13 but also happens with other illnesses. Older people report that impaired health is the most common reason for driving cessation.14, 15 However, patients rarely discuss this radical decision with a healthcare provider.16 Physicians dealing with older adults need to be aware of these limitations and to be able to support their patients to maintain their independence.

The issue of crash risk has been overstated but sadly forms a negative public backdrop to our professional practice: a study of British and Irish media showed an overwhelmingly negative portrayal of older drivers, despite their excellent safety record.17 Physicians must not allow a negative but inaccurate popular perception to interfere with their task of assessing, treating and advising older people in relation to their independence. This extends to the interpretation of studies on crashes: for example, certain illnesses are more common in older people who have crashes.18 However, since older people have lower crash rates, the likelihood of effective public health interventions to reduce this low rate further are unrealistic and may cause further problems (e.g. driving retirement) recognized with screening of older drivers.

Clearly, for individual patients the maintenance of autonomy must be balanced with public safety to an extent consistent with that applied to the rest of the population. Age-related visual and cognitive diseases, in particular macular degeneration, Alzheimer’s disease (AD), stroke and Parkinson’s disease,19 are likely to be the conditions most often associated with mobility and safety problems. Our main ethical prerogative is to preserve a sense of dealing with the issue in a hierarchical fashion common to good practice for all healthcare conditions. This emphasizes in turn the World Health Organization approach of health gain, health maintenance, compensation and finally palliation.20

The Older Drivers Project, an initiative between the American Medical Association and the US National Highway Traffic Safety Administration, has reaffirmed this principle, stating that a primary objective of its approach involves helping older drivers stay on the road safely to preserve their mobility and independence.21 The Older Drivers Project recommends that this can be accomplished through three methods: (1) optimizing the driver, (2) optimizing the driving environment and (3) optimizing the vehicle. In this approach, driving cessation is recommended only after the safety of the driver cannot be secured through any other means.

Clinicians may be reluctant to address the driving issue in the office practice setting. There is a clear need for education in the assessment of mobility and driving: the good news is that such education appears to be effective.22 Evaluating driving skills should not be viewed differently from the evaluation of risk for falls or other risk for injuries in older adults. Clinicians should consider the recommendation for driving retirement in older adults in a similar way to the decision that a previously ambulatory patient is now wheelchair bound for life: efforts should be made to preserve mobility when possible.

Definitive guidelines on how clinicians can intervene effectively to ensure adequate mobility, driving safety or effect driving cessation in impaired older adults are still needed, but current evidence and available resources indicate a general approach to this issue. There is an increasing body of evidence on the subject and some helpful guides.23 In addition, the relatively broad approach to Comprehensive Geriatric Assessment means that geriatricians will better understand the limitations of a predominantly cognitive approach to driving assessment, even in conditions characterized by cognitive decline. A good proxy is entry to nursing home care, which is poorly predicted by individual’s neurocognitive testing, but better matched by behavioural and functional limitations.24

Physicians assessing older people for transport/driving capability need to know:

Taking a Driving/Transportation History

The interested clinician can check static visual acuity (Snellen chart), hearing (whisper test or hand-held audiometry), attention and reaction time (Trail Making Test A or B), visual spatial skills (clock drawing task), judgment, insight, joint range of motion and muscle strength.40–43 Many of these tests were recently included in an American Medical Association resource on older drivers as reasonable for assessing and counselling older adult drivers and are available on their web site along with evidence-based medicine references. These tests are probably more important for gaining an overall perspective on the patient’s abilities and disabilities, rather than relying overly on the performance of any one component.

Although it may seem obvious that transportation should figure in a comprehensive assessment, this is not necessarily the case. An extreme example of this is the failure of the referring physician to advise on driving restriction for a significant number of people referred to a syncope clinic.28 It is also likely that many patients do not obtain formal advice or assessment about driving after stroke.29 As there is potential to improve driving and transportation options, there is also a need to discuss restrictions or planned withdrawal from driving for many patients. There are data to suggest that license restriction is associated with lower crash risks.30 Thus, a graduated driving reduction may be a viable option rather than driving cessation.

The patient’s own assessment of driving should be assessed and a promising approach in this regard is the Adelaide Self-Efficacy Scale.31 It is encouraging that self-assessed driving skills in mild cognitive impairment seem preserved.32 A collateral (witness) history of driving abilities is important, given the often collaborative nature of driving in later life,33 but cognizant of the conflict of interest of a spouse who does not drive.34 Recent data indicate that informants are able to recognize impaired driving behaviour in some older adults with medical illness.35

What Factors are Important in Driving Assessment?

The greatest advance in this area has been the understanding that a purely cognitive model of driving ability does not adequately reflect the complexity and hierarchical nature of the driving task.36 Psychometric approaches have generally been disappointing for a number of reasons37 and efforts to find a best cognitive battery resemble the alchemist’s search for transforming base metal into gold rather than a carefully thought out scientific endeavour. However, recent efforts have pushed correct classification rates in demented samples to 80%, which is encouraging.38

Currently, many prediction efforts have focused on heterogeneous groups of older adults during licence renewal. However, the crash rate of this group is already low (e.g. like the odds ratios for predicting crashes) and it is unlikely that any set of tests in this arena will be useful for the prospective prediction of driving safety. Further success at developing fitness-to-drive models will likely need to focus on specific groups of medically impaired older adults that are homogenous in their specific cognitive or visual domains (e.g. dementia with a specific subtype such as AD). In addition, these models will likely need to incorporate additional measures or proxies such as lifelong driving habits (e.g. history of tickets, crashes, abnormal driving behaviours), personality characteristics (e.g. too aggressive or too passive), traffic density and perhaps other measurement concerns such as test anxiety or confidence issues.

The most common model for driver assessment would involve a combination of physician, occupational therapist, neuropsychologist, specialist driving assessor and/or social worker. Not all disciplines will be needed by all patients: a patient with severe dementia clearly cannot drive and simply a referral to the social worker to plan alternative transportation is appropriate. Equally, a mild cognitive defect may only require a review by the physician and occupational therapist. The overall interdisciplinary assessment should attempt to provide solutions to both maintaining activities and exploring transport needs. However, even in a skilled rehabilitation setting, the predictive value of team assessments may be low for diseases such as stroke39 and the on-road test is the current best assessment available.

The on-road test may be helpful, as it may demonstrate impairments to a patient or caregiver who is ambiguous about the patient stopping driving. At a therapeutic level, members of the team may be able to assist the patients in coming to terms with the losses associated with stopping driving. The occupational therapist may be able to maximize activities and function and help focus on preserved areas of achievement, while the social worker can advise on alternative methods of transport. This approach should save time and valuable resources for occupational therapy, neuropsychology and on-road driver assessors.

In addition to the usual work-up, the medical assessment should include a driving history from patient and ideally (with the patient’s permission) also from a caregiver or informant. The physician needs to weigh judiciously the collateral history, taking into account whether or not the carer is also dependent on the patient operating a motor vehicle. Physicians should also inquire about new unsafe driving behaviours. These behaviours can be apparent in mild dementia and would raise concerns about continuing driving privileges. It is important to recognize that these behaviours represent a change from baseline. They include becoming lost in familiar areas, driving too fast, reacting too slowly, consistently making poor judgments, failure to notice street signs, having more accidents, receiving indecent gestures from other drivers, miscalculating speed and distances, new dents on the car, knocking off rear-view mirrors, showing poor judgment when making turns or impaired ability to recognize or understand road or traffic signs.

The next stage of testing includes evaluations from occupational therapists and/or neuropsychologists. None of the studies have been sufficiently large to have a reasonable predictive value or to determine cut-off points on neuropsychological test batteries. This situation is paralleled in memory clinics where there is a wide variation in test batteries used: it is likely that the important elements of successful assessment are choice of key domains, familiarity with a test battery and the development of an understanding and close liaison between the physician and the occupational therapist and/or neuropsychologist. In addition, a recent review indicated that to date, there is little evidence to support the use of performance based road tests.

A wide range of tests have been correlated with driving behaviour but few have been sufficiently robust to calculate cut-off points for risky driving. All of these tests can be criticized for taking an over-cognitive view of the driving task.44 A comprehensive review of tests is available from the US National Highway Transportation Safety Administration45 and a recent meta-analysis limited to traditional neuropsychological tests indicated that visuospatial skill impairment was the cognitive ability with the strongest association with impaired driving in studies with dementia.46

The other interesting aspect is that there may be a disparity between scores on a test battery and the clinical assessment of the neuropsychologist. In a short paper by Fox et al., the neuropsychology test scores and the neuropsychology prediction were found not to be significantly associated, suggesting that the clinicians made their decisions on items not formally measured in the neuropsychology test battery.47 In conjunction with the clinical assessment and collateral history, these tests will guide the physician as to which patients require on-road testing, and also those who are likely to be dangerous to test!

At present, simulators of sufficient sophistication are not widely available but may represent opportunities for both driver rehabilitation48 (analogous to training aeroplane pilots) and assessment. The main benefit of large, sophisticated simulators such as the Iowa simulator has been to try to develop and understand neuropsychological and behavioural test batteries in a safe and reliable method and to correlate them with unsafe driving behaviour and crashes. The classic paper by Rizzo et al. in 1997 revealed that 29% of AD patients experienced crashes in the simulator versus none of 18 control participants.49 The drivers with AD were also more than twice as likely to experience ‘close calls’. There was also evidence that some drivers with mild AD did not crash and showed fair control of their vehicles compatible with the idea that some patients with mild dementia should be allowed to continue to drive.