TORONTO — There’s no hard evidence that retesting older drivers makes roads noticeably safer, according to an expert who has made a career studying the issue.

Dr. Richard Tay, formerly of Calgary and now an associate dean at the Faculty of Law and Management at La Trobe University, Melbourne, Aus. Has spent years examining aging as it pertains to road safety.

He points out that a paper to be published in the journal "Ageing and Society" shows that among five provinces (Alberta, British Columbia, Saskatchewan, Manitoba, and Ontario), Ontario has the most onerous license renewal scheme, but does not have the lowest crash rates.

"Theoretically, retesting has some merits," Tay says, "but the 64-million-dollar question is how should we do it. At what age should mandatory retesting begin? Sixty-five, 70, 75, 80, or 85? Is age-based regulation fair?"

Choosing the correct criteria to justify the choice is tricky, he admits. Crash statistics vary significantly depending which yard stick you use (crashes per capita, crashes per licensed driver, crashes per km driven). Many road safety researchers and policy makers are fixated on the crashes per km driven and aging drivers have a higher crash rate than middle-aged drivers, but so do younger drivers.

"More importantly, is that the right criterion to use?" he says. "Compare Driver A who has a higher crash risk per km driven but rarely drives, with Driver B who has a lower crash rate-per km driven but drives a lot, who is more likely to be involved in a crash in the next two or three years, or however long the license is valid for? Whose license should you take away first?"

He says papers published in Journal of Gerontology, show there is no accurate, reliable and valid test available right now to accurately weed out people likely to be involved in crashes.

"There are many tests that have been developed that claim to be valid but do not stand up to the test when put under the microscope," he says.

"Health professionals are obsessed with odds ratio; if you fail this test, you have a 50-percent higher chance of being involved in a crash. On the surface, it sounds reasonable to use such a test; after all, we have been using this kind of decision making for almost all health and medical issues.

"However, it does not tell us the most important thing; if you fail this test, what is the likelihood that you will be involved in a crash?"

He says a driver who fails the test may be 50-percent more likely to be involved in a crash than a person who passed the test. But that information is useless unless we know how likely it is that the other person who passed the test will be involved in a crash.

"Do you want to take away someone’s licence if he has a one chance in two million of being involved in a crash, compared to someone who has a one chance in three million?"

Tay also points out that the reliability of exams varies widely in the various provinces.

"I have taken many of them," he says. "Sometimes I pass and sometimes I fail within a very short period. But I am the same driver."

In general, he says, he does not support mandatory re-testing but likes the idea of voluntarily self-assessment.

— For proof on how some drivers get better with age, check out this week’s online feature Grays Under Pressure by clicking here.