Fitness-to-Drive in Neurologically Impaired Older Adults

David B Carr, M.D.

Project Overview:
In elderly persons, neurological diseases (e.g., dementia or stroke) impair driving ability, which can result in a motor vehicle accident.  Moreover, older adult driver crashes tend to involve multiple vehicles, so the older driver and individuals of all ages are at risk of serious injury or death. Appropriate evaluations to determine fitness-to-drive in these patients often include performance-based road tests. However, there are many barriers to such tests, including cost, patient acceptability, and the availability of examination centers. There are also valid concerns regarding the safety of the driving instructor, the patient, and the public when administering such procedures in a real world setting.

The primary objectives of the proposed research are determining the feasibility and the clinical utility of assessing impaired driving behaviors and traffic safety knowledge along with novel functional abilities that are important when operating a motor vehicle.  The central hypothesis is that older adults with neurological disease will be able to perform these tests and demonstrate a wide range of impairments based on their disease severity, and that a combination of specific tests will predict who will fail a standardized road exam and/or experience a motor vehicle collision. 

To test this hypothesis, 100 elderly (55+ years old) men and women with dementia and 100 with a history of a cerebrovascular accident will be evaluated using both interview/history questions and measurements of key driving abilities that will include both computerized and paper and pencil psychometric tests. A standard driving questionnaire that focuses on the presence of impaired driving behaviors will be administered to an informant or collateral source, while knowledge of traffic safety rules of the participant will be assessed.  Measures of visual function (far and near visual acuity, visual fields, color detection, phoria, etc), cognitive function (sustained and divided attention, visual processing speed, visuospatial skills, executive function, road sign recognition) and physical function (performance-based tests of strength, flexibility, balance, motor speed, and coordination) will be determined before administering a performance-based road test.

The results of the history (questionnaires) and physical examination (functional abilities) will be utilized to derive a model that will predict failure on a standardized road test and/or a motor vehicle collision.  Encouraging results from other pilot studies suggest the feasibility and utility of utilizing multiple predictors in creating fitness-to-drive models. The data generated from this pilot study will ultimately contribute toward establishing appropriate evaluation guidelines for physicians, occupational therapists, and highway patrol examiners who are faced with evaluating patients with neurological impairment. Making a decision to drive without an actual evaluation in a real traffic setting would allow additional patients to be examined across the country due to a decrease in cost for such testing, wider acceptance, and increased availability. This could have a major impact on reducing the number of unsafe drivers on the road and subsequently decrease the number of motor vehicle collisions and injuries on our roadways in an ever-expanding population of neurologically impaired drivers.

Progress Report:
As of September 28th, 2008, 66 potential participants were screened and 49 have been enrolled and completed the driving evaluations. In our protocol we utilize the Functional Activity Questionnaire (FAQ) which is completed separately by informants and participants. The range of scores has been wide (0 -27), with an average score of 3.5+6.1 (higher scores indicating more functional impairment).  Informants have been able to report abnormal driving behaviors that could indicate at risk driving on the participants prior to the assessment.  Thus for, informants have documented a range of 0 to 8 unsafe driving behaviors in our participants, with an average of 1.3+2.1 behaviors.

Two measures that the Missouri State Highway Patrol currently utilizes for testing not only novice drivers but also those referred for fitness-to-drive evaluations (e.g. drivers with dementia or stroke), are road sign recognition tests and written examinations that test traffic safety knowledge. Our driving clinic has also created a written test that features common driving scenarios and situations that require decision-making and intact judgment. To date, our participants only have been able to name correctly 9.3+ 1.9 road signs (out of 12), and obtain an average score of 9.9+ 2.1 out of 13 questions on our written examination. This data will be compared to on road performance as our sample size increases.  These tests have been easy to administer, take minimal time by the examiner, and have some face validity and are acceptable to the older adult as a proxy for safe driving.

We have adopted important off-road tests that not only tap into these key functional abilities, but also are often utilized by occupational therapists in driving clinic settings and by the American Medical Association (AMA) for physicians in the office setting. To date, the majority of our off-road tests have been acceptable to our participants and feasible to administer.  All of these tests have exhibited a wide range of scores and significant portions have fallen into areas that have been associated with driving impairment.

As part of this study, we are revising the Washington University Road Test (WURT), and during the initial stages it has undergone several revisions in an attempt to make the assessment more functional, usable, and generalizable to other clinics.  Our participants thus far have failed the road test (WURT) on 13 out of 40 occasions, representing 32.5% of the sample to date. If this percentage continues, this should more than satisfy our power calculation assumptions to create a model that can predict our qualitative outcome measures for stroke patients. As a measure of inter-rater reliabilty, 50% of our evaluations were scored by two instructors, one that is blinded to the test result. So far, these evaluators have reached a 95% concordance in classification on the qualitative scoring of the WURT (pass/fail). In addition, there will be a quantitative measure (which quantifies the number/type of errors made on the road) to be included at a later date.   

The STARS crash data has just been made available to us and will be incorporated into the database during the 2nd year of the study. Our current sample size is too small to provide any preliminary data on a fitness-to-drive model for our stroke patients, which would utilize a logistic regression approach based on our qualitative outcomes (e.g. pass-fail on the road test and crash-no crash based on data from the Missouri Statewide Traffic Accident Recording (STARS) system).

The findings to date indicate that our chosen testing measures are acceptable, feasible, and provide a good dispersion of scores across ranges of neurological impairment for our stroke patients. Recruiting is on schedule to test another 50 subjects during the 2nd year of funding. We do not anticipate difficulties with recruitment and can increase visibility in other institutions outside Washington University if needed. We will also begin our follow up of our first year clients as to their level of independence and functioning, care needs, depression, and crash risk.  Our clinical tests have potential applicability to identify those neurologically impaired individuals that should no longer be operating a motor vehicle as well as those that are safe to continue or resume driving. The potential cost savings of not having to obtain an on-road test would be a potential cost savings to these patients and families, and the public safety benefit from saved lives and prevention of injuries could be substantial.