The Candrive Older Driver Risk Stratification Tool

Author(s): Marshall, Bedard, Vrkljan, Tuokko, Porter, Naglie, Rapoport, Mazer, Glinas, Gagnon, Charlton, Koppel, MacLeay, Myers, Mallick, Ramsay, Stiell, Wells, Man-Son-Hing

Slidedeck Presentation:

Candrive RST DEvelopment CARSP 2022_Marshall

Abstract:

Background:

Driving is important to maintain mobility, independence, and quality of life. Unfortunately, many older drivers are no longer able to drive safely and must cease driving. Yet, screening older drivers for fitness-to-drive is challenging. It is accepted that age alone and medical diagnoses are not accurate predictors of fitness-to-drive, and current screening tools lack sufficient accuracy to make sound determinations. Furthermore, many approaches to screen for fitness-to-drive are not necessarily straightforward, nor easily conducted in a clinician’s office.

Aims:

The primary aim of the Candrive Older Driver study was to develop a practical risk stratification tool (RST) to screen for fitness-to-drive that can be easily administered by health care providers in their offices and without the need for a computer.

Methods:

We used a prospective multicenter cohort design. Participants were recruited in seven Canadian sites starting in June 2009; data collection ended December 2016. Participants had to be active drivers aged 70 years and older. They underwent in-person comprehensive assessments annually and shorter follow-up assessments every four months. The primary outcome measure was police-reported, expert-validated, at-fault collisions adjusted per annual kilometers driven. A total of 601 variables potentially related to fitness-to-drive were documented. Through a reduction process we identified 52 potential objective variables that could feasibly be collected in an office assessment and could contribute to the RST. We used Poisson regression modelling, in combination with Generalized Estimating Equations (GEE), to develop models where Nomograms provided the relative contributions of each variable to the level of estimated risk of an at-fault collision and produce a parsimonious tool.

Results:

We recruited 928 older drivers (62% male). The average age at enrolment was 76.2 years (SD = 4.8). The participants combined for a total of 4,483 person-years of driving and were involved in 231 collisions of which 112 (48%) were at-fault collisions. The derived Candrive RST includes four predictors. The majority of the person-years (74.8%) fell within the lowest risk category, whereas 9.3% fell in the low-medium category, and 13.0% and 2.9%, respectively, fell within the medium-high risk and high-risk categories. The relative risk for an at-fault collision per 10,000km driven was 5.26 (95% CI = 2.84-9.81) for the highest risk category compared to the lowest.

Discussion:

The risk of an at-fault collision can be estimated based on the Candrive RST. It is noteworthy that the vast majority of person-years of driving were in the lowest category, illustrating that as a whole, older drivers are not a high-risk group. However, for a small proportion of drivers, further evaluation and/or discussion regarding driving cessation may be warranted. While our sample appeared representative of the Canadian older driver population, further validation of the Candrive RST is required, including demonstration of its applicability within the clinical setting.

Conclusions:

While the Candrive RST does not replace sound clinical judgement, it may, for older drivers whose functional limitations create uncertainty regarding their fitness-to-drive, assist primary health care providers when initiating a conversation about driving and help guide further evaluation. Further validation of the tool is underway.