|Date added||July 10, 2018|
|Category||2018 CARSP XXVIII Victoria|
|Tags||Research and Evaluation, Session 4B|
|Author/Auteur||Crizzle, Gilfoyle, Mychael|
|Stream/Volet||Research and Evaluation|
The increasing number of older drivers presents a major public health concern given that older adults are more likely to develop functional impairments associated with age-related medical conditions that can impair their ability to drive safely. One particular screening tool, the SIMARD (Screen for the Identification of Cognitively Impaired Medically At-Risk Drivers), is proposed as a valid and reliable screening tool for identifying cognitively impaired drivers whose driving skills may have declined to an unsafe level. However, recent articles have reported poor sensitivity and specificity of the SIMARD, questioning its usefulness as a stand-alone screening tool. The purpose was to examine whether 1) the SIMARD is correlated with other clinical tests; and 2) to determine the SIMARD's sensitivity and specificity for predicting pass/fail on the road test. Data was collected from one driving assessment center in South-Western Ontario. Data was collected retrospectively from 2012-2015 and prospectively from 2015 to October 2016. In total, there were 175 client records containing information concerning demographics, clinical test scores (SIMARD, Montreal Cognitive Assessment [MoCA] Trails A & B, Useful Field of View [UFOV]) and on-road pass/fail outcomes. Descriptive statistics (mean and standard deviation) and frequencies (valid percent) were used to describe the sample. Chi-square and independent t-tests were used to compare categorical and continuous variables, respectively. A binary logistic regression was performed to examine whether the SIMARD was predictive of failing a road test. All statistical tests were considered significant at p < .05.
Of the 175 assessments, clients were primarily referred due to concerns related to mild cognitive impairment (26%), dementia (17%), stroke (19%) and traumatic brain injury (5%). The average age of clients was 69±14.9 years; 73% were male. On the SIMARD, the mean score was 49.7±22.8 (range of 2 to 114) with clients being classified as low risk (18%), requiring further testing (61%) and high risk (21%). Clients scored on average 21.9±4.6 on the MOCA, 69±80 and 240±147 seconds on Trails A and B, respectively. Clients were classified on the UFOV as being very low to low (55%), low/moderate (18%), or moderate to high risk (26%). Approximately 42% failed the road test. The mean score on the SIMARD was significantly correlated with age (r=.57, p<.001), Trails A (r=.39, p<.001), Trails B (r=.59, p<.001) and the MoCA (r=.65, p<.001). The SIMARD high risk index was also more likely to result in a higher risk index on the UFOV (χ2=36.63, p<.001). A logistic regression with only the SIMARD found the test to be predictive of pass/fail outcomes on the road test (N = 151; -2 Log Likelihood = 174.56; Nagelkerke R = .247), however, the SIMARD misclassified a large number of participants. When the SIMARD was combined with other clinical tests (i.e. Age, MoCa, Trails A & B, UFOV), the SIMARD was not a significant predictor of pass/fail outcomes on the road test.
The findings show the SIMARD should not be used as a sole screening tool for referring clients for further driving assessment due to a degree of misclassifications. The SIMARD should not be used as a stand-alone driver screening tool.