Operating a motor vehicle is an overlearned motor task involving aspects of divided attention and visual perception (surveillance of roadway, traffic, and regulatory and informational signage), and time-on-task, accomplished by multiple limb-coordinated motions. Driving is a privilege granted by the government based on age (reflecting driver maturity), minimal performance standards (corrected vision, passing a written and a performance test under low workload conditions), and restricted based on bad experience (moving violations and collisions). Society tolerates the increased risk of the immature driver, because the only way their driving skills improve and the risk of collisions decrease is with practice. A trained driving examiner makes the licensing decision for the immature driver on behalf of the State. Society, through the legislative process, has restricted driving for certain conditions or driving violations. For example, States have progressively tightened the permissible level of breath alcohol concentration of drivers to a uniform 0.08 mg/dL. Through the same mechanism, some jurisdictions have increased the frequency of testing of older drivers and required physician reporting of potentially impaired drivers.
Physicians are not trained to determine fitness-to-drive. Thus, when called upon by a patient or their family to answer the question “can they drive?” the physician has no basis on which to base a decision. Previously, we addressed this question for drivers with dementia, through a systematic literature review.1 This evidence-based medicine practice parameter reported that there was an increased risk of collisions for drivers with a clinical dementia rating scale score of 0.5 (possible Alzheimer's disease) and 1 (probable Alzheimer's disease) compared to nondemented drivers. The clinical dementia rating is a complex rating scale, based on a structured interview, encompassing many domains, and not easily adaptable to a brief clinical encounter. Of the many psychometric tests used in the studies of driving in people with dementia, only Trails A (as part of a regression analysis including age, sex, and driving history) and the Sternberg Test (a triarchic memory test) were partially predictive of future crash risk.2, 3 Thus far, no simple psychometric test has been found to have a threshold level robust enough to reliably divide drivers into high or low risk for future crashes. This is because driving is a complex task involving many cognitive domains, and collisions are rare events that involve both capable and impaired drivers.
Bajaj and colleagues present an interesting report of motor vehicle crashes in drivers with cirrhosis, with and without minimal hepatic encephalopathy (MHE) over a 2-year period.4 After measuring cognition with a brief psychometric test battery, and the inhibitory control test (ICT), the drivers were assigned to either the MHE cohort or the non-MHE cohort. Collision and violation data were collected by self-report and from the Department of Transportation for the year before and the year after the psychometric testing. MHE as diagnosed by ICT was more highly associated with past and future collisions than MHE diagnosed by standard psychometric tests. ICT appeared to be more sensitive than the standard psychometric battery in the diagnosis of MHE.4 Similar criteria were used as in a prior report for the case definition of MHE: >2 standard deviations for any two of the four psychometric tests or > 5 lures on the ICT.5 Yet, the psychometric battery was slightly different in this study than in their prior work. It appears that there has been a shift, and the ICT has become the preferred test for the diagnosis of MHE. Yet, the difference between ICT and other psychometric tests in the diagnosis of MHE is not the focus of this work.
Their study builds on prior work including an on-the-road test of a convenience sample of 274 consecutive patients with liver cirrhosis, only 50 of whom qualified for the driving test. Five of the 14 with MHE had the on-the-road performance test stopped by the driving instructor who was blinded to their cognitive test results.6 There were no significant differences for individual components of the driving test, only for driving cessation. In a desktop driving simulator test, subjects with MHE made more wrong turns when performing a route following task, missed more divided attention tasks, and had more collisions than drivers with cirrhosis without MHE and compared to normal controls.7
Many questions regarding driving safety are left unanswered. In this study, driving exposure was not reported. To be included, subjects had to drive a minimum of 20 miles per week. This is not enough to maintain driving performance. Collisions involving elder drivers were previously believed to be associated with aging. Based on work from the Netherlands, when corrected for driving exposure on a yearly basis, older drivers were actually safer than younger drivers with similar exposure.8 It is still uncertain if drivers lose their previously acquired abilities through lack of practice, or that drivers at the beginning of impairment are self-restricting their driving. Without data on driving exposure, we cannot compare the MHE and non-MHE groups, or compare either cohort to historical controls.
In this combined retrospective/prospective design, the duration of MHE was not reported. Did drivers convert from non-MHE to MHE during the year before the psychometric testing? What happened to the dropouts? Did they become too impaired to drive? Did they stop driving because of collisions?
Lastly, the comparison of the psychometric battery to the ICT to make the diagnosis of MHE may not be appropriate. The ICT measures vigilance and time-on-task during a ≈15-minute test. The psychometric battery involved variations of Trails A and B, digit symbol test, and the block design test. Different aspects of cognition are measured with these test batteries. Without a reference standard, it is difficult to argue that one test is superior to another in detecting MHE.
Bajaj and colleagues are to be applauded for their studies of drivers with MHE. Outside of alcohol intoxication, normal aging, and dementia, very few studies have been undertaken on other conditions that have the potential to impair driving performance. With the addition of driving exposure, their study will yield important information about driving performance and the extent of the driving safety problem seen in drivers with MHE.
Most drivers operate a motor vehicle in an acceptable fashion, most of the time. Collisions are rare, and fault cannot always be assessed. Good drivers are involved in collisions. Most of our current driving restrictions are based on what level of risk society, through the governmental process, is willing to tolerate. The form of impairment that we have the most information on is the effect of alcohol on driving. Based on the Grand Rapids study, the relative risk of a collision by an alcohol-impaired driver compared to an alcohol-free driver is 5 at a breath alcohol concentration of 0.08%.9 This is the level that is currently used in the United States, and stricter levels are used in other jurisdictions.
As physicians, we are entrusted to first do no harm. Except in certain jurisdictions, physicians are not required to report drivers impaired by physical etiologies such as cognitive impairment, sudden loss of consciousness, or excessive sleepiness, all of which are associated with a higher relative risk of collisions. Not being trained to determine fitness to drive, at best we can only detect the extreme cases of impaired drivers. Whereas caution is always prudent when asked “is it safe to drive?” more research is needed to determine the true relative risk of collisions by drivers with MHE based on exposure, and what type of accidents are typical for this population. The Holy Grail of fitness-to-drive testing is an office-based test that is easy to administer, reproducible across many disorders and over time, sensitive to driver impairment, yet specific enough to not falsely label good drivers as impaired. Only then can society, through the legislative process, determine who is allowed to drive and who should not drive based on a chronic medical condition.