Cognitive screening tools for predicting unsafe driving behavior among senior drivers

Abstract

As Canada’s elderly continue to represent the fastest growing population in Canada, there has been an increasing need for effective and efficient screening tools for senior drivers, especially ones which identify possible cognitive impairments. Thus, the objective of this study was to conduct a meta-analysis of the available research surrounding the predictive value of pencil-and-paper cognitive screening tools. A systematic review of existing literature was conducted, with a final sample of 15 evaluation outcomes that identified 10 different pencil-and-paper tools. Multiple techniques were used to evaluate the data, including random effects modeling, meta-regression analysis and tests for bias, including publication bias. Finally, a multilevel meta-regression model was used to account for dependence of evaluation outcomes coming from the same study. A small to medium-sized significant pooled effect of 1.94 was found, indicating that when pencil-and-paper cognitive screening tools predict a driver is unsafe, there is a 94% greater chance that this driver will exhibit unsafe driving behaviors. Results, however, only provide partial evidence to inform the selection of pencil-and-paper cognitive screening tools, as it was not possible to unequivocally identify which tool performed best.