Predictors of Motor Vehicle Crashes in a Dementia Clinic Population: The Role of Gender and Arthritis

To the Editor: We recently reviewed driving records for patients seen in a dementia Clinic1 to determine which demographic variables or medical conditions contributed to the occurrence of motor vehicle crashes. Records were examined for 249 subjects referred by physicians to the Clinic for Alzheimer Disease and Related Disorders at the Vancouver Hospital and Health Sciences Centre, Vancouver, Canada, a tertiary referral diagnostic clinic. We found no difference between the motor vehicle crash rates of persons with dementia and those who did not meet dementia criteria. Characterization of the groups revealed a variety of medical conditions other than dementia that may have affected driving skills. Subjects were selected for the study if they had been driving within 2 years of attending the Clinic. Each subject underwent a comprehensive multidisciplinary assessment, including physical, neurological and psychiatric examinations, neuropsychological assessment, social work and genetics interviews, and laboratory investigations. Clinical diagnoses were made by the multidisciplinary team using DSM-III-R2 criteria for dementia. Of the 249 subjects, 165 met criteria for dementia and 84 did not. The complaints and identified deficits for the not demented group were highly heterogeneous and included neurological disorders not accompanied by dementia, psychiatric disorders, other medical conditions, and very mild memory problems.1 We elicited information about medical conditions for all Clinic patients through physician-assisted self-report and caregiver report as well as by physical examination. Official motor vehicle crash records were obtained for each individual from the provincial Motor Vehicle Branch Drivers File, the Motor Vehicle Branch Accident Database, and the Insurance Corporation of British Columbia (single automobile insurance carrier in the province) Claims database.3 Thus, both police-reported and non-police-reported crashes resulting in an insurance claim were included. Because the demented and not demented samples showed no significant differences with respect to the number of drivers with one or more crashes (Demented 43/165, Not Demented 19/84; chi-square = 1.03, P < .50),1 they were combined for the present analyses. When the conditions listed in Table 1 and a sum of total number of medical conditions were entered into a multiple logistic regression predicting the occurrence of a motor vehicle crash, the presence of arthritis (Odds Ratio = 3.23, P = .009) and female gender (Odds Ratio = 0.53, P = .04) were the only significant factors emerging. Our finding that more men than women have accidents is not uncommon. In a geriatric medical screening program, Campbell and Bush4 found that women were twice as likely as men to have stopped driving. In an older cohort where the man of the household is the only driver, continuation of driving may be seen as a means to maximize the autonomy of the couple, thereby avoiding dependence on other family members. Alternatively, male drivers may drive more than women drivers, and this greater driving exposure may result in more frequent motor vehicle crashes.5 Of the 249 cases, 23 were diagnosed with arthritis, and 11 of these individuals had had motor vehicle crashes. We examined the issue of medication status in relation to presence of arthritis. The percentages with motor vehicle crashes for persons with arthritis and taking nonsteroidal anti-inflammatory medication (NSAIDS), with arthritis but not taking NSAIDS, without arthritis but taking NSAIDS, and without arthritis and not taking NSAIDS were 55 (5/9), 43 (6/14), 42 (5/12), and 21 (46/214), respectively. The first group differed significantly from the last group with respect to the percent of cases with accidents (P = .0086) but did not differ from the second (P = .28) or third (P = .26) groups. Even though the sample sizes for the affected (i.e., arthritis and/or NSAIDS) groups are small and the sample is limited to individuals referred to a dementia diagnostic clinic, further investigation of this finding appears warranted. The combination of arthritis and NSAIDS use yielded the greatest number of drivers with one or more motor vehicle crashes. Either condition in isolation yielded more drivers with crashes than when neither condition was present, though not significantly so. It is unclear from this data whether this reflects a compounding of risk associated with mobility problems and medication side effects or is solely indicative of the severity of the arthritis as manifest in the use of NSAIDS for pain and in restriction of mobility. Clarification of the association between driving crashes and these medical factors is of particular concern as the prevalence of arthritis in older people is high6 and NSAIDS are commonly prescribed. Despite the recent emphasis on the relations between cognitive impairment and motor vehicle crash rates in older individuals,7 our findings underscore the need to address other medical conditions8 and use of pharmaceutical agents9 when evaluating driving status in older adults.