The AAN Quality Standards Sub-committee updated its 2000 practice parameter on the evaluation and management of driving risk in patients with dementia in an attempt to determine the usefulness of patient demographic characteristics, driving history, and cognitive testing in predicting driving capability in patients with dementia. It also tried to identify interventions that help reduce driving risk.
The guideline authors reviewed 422 out of 6,000 evidence-based studies published between 1970 and 2006. The findings were released online April 12 ahead of the April 20 print edition of Neurology®.
Lead guideline author Donald J. Iverson, MD, a neurologist with the Humboldt Neurological Medical Group, Inc., responded by e-mail to our questions about the parameters.
Figure. THE NEW GUID...Image Tools
WHAT IS NEW IN THE UPDATED GUIDELINES?
The new practice parameter identifies a number of historical features, not present in the original, that are associated with increased driving risk. Similar to the 2000 guideline, the updated version confirms that patients with dementia are at an increased risk for unsafe driving; however, while the previous guideline concluded that this means patients with mild dementia categorically should not drive, the update softens the message to “should strongly consider discontinuing driving.”
WHY DID THE COMMITTEE DECIDE TO DO ANOTHER REVIEW JUST 10 YEARS LATER?
The Quality Standards Subcommittee schedules a review of guidelines for updates every three years. Since the publication of the 2000 practice parameter, our search strategies are more comprehensive, and the evidence-rating process has become more stringent.
WHAT WAS THE OVERALL FINDING OF THIS REVIEW?
We found that patients with mild dementia are at an increased risk for unsafe driving but many are still safe drivers. Overall, the following historical factors are associated with increased risk: caregiver concerns, a history of accidents or tickets, restrictive driving, a Mini-Mental State Exam (MMSE) score of 24 or lower, and an aggressive or impulsive personality.
HOW USEFUL WAS THE MMSE IN PREDICTING RISK?
An MMSE score of 24 or lower is possibly useful in identifying patients at increased risk for unsafe driving. Otherwise the correlation between MMSE scores and driving performance is unclear, as data are conflicting.
CAN NEUROLOGISTS RELY ON GLOBAL MEASURES OF DEMENTIA TO IDENTIFY DECLINING DRIVING ABILITY?
The point estimates of relative risk for patients with a Clinical Dementia Rating (CDR) of 1 failing a driving test — when compared to those of drivers without dementia — ranges from 2.68 to 88.76. Some believe that this association is strong enough to categorically recommend that patients with mild dementia should not drive. Conversely, others look at the wide confidence intervals (1.4 – 1444) and high driving-test pass rates for patients with mild dementia — as high as 76 percent – and conclude that such a categorical recommendation is too restrictive.
WHICH NEUROPSYCHO-LOGICAL TESTS PROVIDE ADDITIONAL PROGNOSTIC INFORMATION?
Figure. SAMPLE ALGOR...Image Tools
While neuropsychological testing may define dementia severity, we did not find sufficient evidence to support or refute its benefit in evaluating driving risk. The neuropsychology subsection of the author panel did not identify a specific test that contributed additional prognostic information.
CAN NEUROLOGISTS RELY ON PATIENTS AND THEIR CAREGIVERS TO ASSESS DRIVING CAPABILITY AND RISK?
If the caregiver or patient has concerns, those concerns are usually validated with driving tests. However, if the patient or caregiver believes everything is fine, there may or may not be a problem. According to one Class I study, a caregiver's rating of “marginal” or “unsafe” is probably useful in identifying unsafe drivers; however, patients' self-ratings of “safe” are not useful, according to three Class I studies.
Many patients with dementia lack insight into their deficits, including their driving ability. Interestingly, some patients may not explicitly recognize their deteriorating driving skills, yet they will begin to restrict their driving, which suggests implicit recognition of their deficits.
WHAT ASPECTS OF THE DRIVING HISTORY WERE ASSOCIATED WITH DECREASED DRIVING ABILITY?
A history of a crash one to five years earlier or a traffic citation in the last two to three years is possibly useful in identifying patients with decreased driving ability. Everyone who purchases car insurance somewhat anticipates crashes and driving citations, which correlate with increased risk. In addition, patients who have reduced their driving mileage or report always avoiding driving at night or in the rain also are possibly high-risk drivers. In addition, there is a weak association with aggressive or impulsive personality characteristics.
WHAT PERSONALITY FACTORS WERE IDENTIFIED AS PREDICT-ING MORE DRIVING RISK?
One Class II study found that aggressive or impulsive personality characteristics were found to be possibly useful to identify patients at increased driving risk.
DO ANY INTERVENTIONS REDUCE DRIVING RISK?
We did not identify any high-level evidence of interventions that reduced driving risk.
WHICH METHODS ARE MOST EFFECTIVE IN IDENTIFYING PATIENTS AT INCREASED RISK FOR UNSAFE DRIVING?
With the understanding that mild dementia is itself a risk factor for unsafe driving, we recommend considering the historical factors we cited earlier to develop a qualitative estimate of risk.
WHAT SUGGESTIONS DO YOU HAVE FOR NEUROLOGISTS TREATING PATIENTS WITH DEMENTIA WHO MAY BE AT RISK FOR UNSAFE DRIVING?
Neurologists are, in some states, legally obligated to identify conditions that may put their patients' or the public's health at risk. Because there is no test result or historical feature that accurately quantifies driving risk, clinicians have to make qualitative estimates of driving risk.
My suggestions are dependent upon the role of the neurologist and in which state he or she practices.
HOW CAN NEUROLOGISTS USE THESE NEW GUIDELINES?
We suggest that clinicians present patients and their caregivers the data that show, for example, that as a group, patients with mild dementia — as reflected by a CDR score of 1 — are at a substantially higher risk for unsafe driving and therefore should strongly consider not driving.
As a general neurologist who practices in a medically underserved area of a state that has mandatory reporting laws, this parameter is probably more relevant to me than it would be for a behavioral neurologist providing a second opinion from a tertiary care center in a state without reporting laws. Even still, driving can be considered the same type of end-of-life issue as financial conservatorship, transition to assisted living, or advance health. This whole process is a discussion or negotiation among the clinician, patient, and caregivers or family.
In my experience, primary care practitioners are more likely to delegate the driving issue to the neurologist. The updated practice parameter suggests sharing the information on driving risk with the patient and family. If they are willing and able to make alternative transportation agreements — great. If not, the clinician should use his or her judgment as to how hard to push the issue. In that instance state reporting laws may come into effect. We recommend a follow-up evaluation every six months to determine whether the risk has increased and to re-approach the issue. A Class I study reported that drivers with a CDR of 1 were about 2 to 2.5 times more likely than drivers without dementia to fail a driving test at a six-month follow-up.
WHAT RECOMMENDATIONS DID THE COMMITTEE MAKE FOR FUTURE RESEARCH IN THIS AREA?
The quest for a single clinical test that will dichotomize drivers into “safe” and “unsafe” will probably go unrealized. I believe accurate estimates of driving risk will require a composite score that uses results from high-quality studies of bedside tests, which control for the presence and severity of dementia. Also, several researchers have created discriminant functions that accurately predict driving performance. •