<Back    TITLE PAGE    Forward>

DISCUSSION

In this study, we have determined the rates of citations, crashes and at-fault crashes of drivers licensed in the medical conditions program and compared them to the rates of demographically similar drivers. Our study describes the driving performance of drivers who voluntarily report their medical conditions to the licensing agency. It does not describe the direct influence of medical conditions on driving performance.

This study is a continuation and expansion of our previous work in this area: Evaluating Drivers Licensed with Medical Conditions in Utah, 1992-1996, which was completed in June 1999 (DOT HS 809 023, available from the National Technical Information Service, Springfield, Virginia 22161). The Executive Summary for that first report is included as Appendix D. As in that first study, we evaluated the rates of various adverse driving events experienced by all 68,770 drivers licensed in the state with medical conditions. In that study, we reported analysis of each functional ability category (medical condition) by restriction status; that is, the various functional ability levels were collapsed into unrestricted (levels 3-5) and restricted (levels 6-11) groups for analysis. (Recall that levels 1-2 are irrelevant for the general population of drivers, and that level 12 signifies no driving permitted.) Also in the first study, those individuals who were in more than one functional ability category (reported more than one medical condition) did not undergo subgroup analysis for specific combinations of conditions. Rather, they were analyzed together as a single group of “drivers with multiple medical conditions”.

In the present study, we performed more detailed analyses, in order to expand on existing knowledge of the effects of specialized licensing programs that regulate such drivers . Rather than being collapsed into unrestricted and restricted groups, each functional ability category (medical condition) was analyzed for each functional ability level. Further, drivers in more than one functional ability category were analyzed by specific combinations of medical conditions.

Approximately 80% (54,825) of the study population reported a single medical condition for the study period. When these drivers were analyzed for adverse driving events by specific medical condition and functional ability level, some patterns could be discerned. The rate of citations for medical conditions drivers did not differ in any consistent way from that of comparison drivers. The rates of crashes and at-fault crashes, however, were significantly higher than the comparison group for some functional ability levels for most functional ability categories (medical conditions). Further, this effect was generally seen at the numerically lowest functional ability levels, where driving privileges were not restricted (recall that the numerically lowest functional ability levels are the least restrictive). Specifically, categories showing significantly higher rates of citations included psychiatric conditions and visual acuity, levels 3-5. Categories showing significantly higher rates for crashes and at-fault crashes included all medical conditions except cardiovascular, for levels 3-5 depending on the specific condition (but always including level 3). Note that “significantly” is used here in its statistical sense. It does not necessarily follow that these differences have major import for public safety or for the medical conditions program.

Approximately 20% (13,408) of the study population were in more than one functional ability category (reported more than one medical condition) for the study period. Most (10,595) of these individuals reported two such conditions, although combinations of as many as 7 did occur. Although an extremely large number of different combinations were mathematically possible, only a limited number of two-way combinations (and no combinations of three or more functional ability categories) contained sufficient numbers of individuals for analysis. The rates of citations for these drivers with combinations of medical conditions did not generally differ from the comparison groups. Rates of crashes and at-fault crashes were higher than the comparison groups for approximately half of the combinations analyzed. In these cases, unrestricted (functional ability levels 3-5) and restricted (levels 6-11) did not appear to differ markedly. Specifically, higher rates of crashes and at-fault crashes, compared to comparison drivers, were found for drivers with the following combinations of medical conditions:

Unrestricted Drivers
Restricted Drivers
diabetes and cardiovascular
diabetes and cardiovascular
diabetes and neurological
neurological and musculoskeletal
diabetes and psychiatric
neurological and functional motor
diabetes and visual acuity
musculoskeletal and functional motor
neurological and musculoskeletal
cardiovascular and visual acuity (at-fault crashes only)
neurological and functional motor
 
musculoskeletal and functional motor
 

We have presented a degree of risk for each category as a relative risk (odds ratio). This is a ratio of the rates of events that compare medical condition drivers to the rates of comparison drivers. While this ratio compares rates of adverse driving events for medical condition populations to those of demographically similar drivers, the numbers of events themselves are also of interest. The absolute numbers of adverse driving events for many subgroups described in this study are very small, often less than 100 and sometimes less than 10. Medical condition and functional ability level subgroups with high relative risks, but low numbers of occurrences, probably do not have a major adverse impact on public safety. Thus, the risk caused by these groups may not warrant changes to this safety program.

As noted elsewhere, small numbers in subgroups was a major obstacle to data analysis in this study. In order to compensate for this, it might seem logical to analyze all the functional ability levels in aggregate, by combining all the various medical conditions. That is, drivers from functional ability level 3 for each functional ability category (medical condition) could be combined into one group containing all drivers at functional ability level 3, and similarly for every other level. We considered conducting such an analysis, but have elected not to do so. Primarily, this is because there is no particular commonality between functional ability levels for different medical conditions. For instance, functional level 3 is defined differently for diabetes than it is for visual acuity. There is no basis for concluding that functional ability level 6 assigned for diabetes, for instance, implies the same degree of impairment as the same level assigned for psychiatric conditions. A global analysis by functional ability level might lead to a conclusion that the levels should be modified globally, but such a conclusion probably would not be warranted since there is no fundamental similarity amongst functional ability levels across medical conditions categories.

As with our first study of this program, there are two major limitations that must be considered when evaluating these results. First, accurate measurements of exposure (miles driven) and other factors that affect the risk of citation or crash are not available. In the absence of exposure data, we reasoned that the amount and conditions of driving for persons with medical conditions could be estimated by selecting comparison drivers using age group, county of residence, and sex. This may in fact be incorrect, as other factors influence the number of miles that people drive. They include marital and economic status, employment, higher education, being a member of a social or religious organization, residential demographics, and the condition itself .

The second major limitation of our study is that the medical conditions program relies on self-reporting through a general questionnaire administered by the Utah Driver License Division. This is also a major limitation of the medical conditions program itself. There is actually a disincentive for applicants to report a medical condition, since doing so may require a longer wait for a driver license or a visit to a health care professional or both. Medgyesi and Koch showed that for every driver with a cardiovascular disease known to the licensing division in Saskatchewan, Canada through its medical review program, there were 94 such drivers who were unknown . The proportion of drivers who have medical conditions and actually report their conditions is unknown in Utah. The Utah Department of Health estimates that the prevalence of diabetes was 2.9% (57,900) of the general population in 1996 ; however less than half (26,458, 46%) of these persons, although not all would be licensed drivers, reported their condition to the driver license division.

In addition to these two major limitations, other potential problems with the system may exist. For instance, drivers who were required to have an evaluation by a physician might have “doctor-shopped” to acquire a more favorable rating and thus avoid restrictions on their driving privileges. Further, it is unknown how consistently health care professionals assign functional ability levels according to the guidelines. Compliance with the program for restricted drivers (e.g., time, area or speed) was also assumed. For example, while some restricted drivers are not supposed to drive after dark because of the restrictions placed upon their driver licenses, we did not verify that these drivers were following their restrictions at the time of the crash or citation.

<Back    TITLE PAGE    Forward>