Citation

Staples JA, Redelmeier DA. Inj. Prev. 2019; ePub(ePub): ePub.

Affiliation

Evaluative Clinical Sciences, Sunnybrook Research Institute, Toronto, Ontario, Canada.

Copyright

(Copyright © 2019, BMJ Publishing Group)

DOI

10.1136/injuryprev-2019-043262

PMID

31302609

Abstract

We thank Harper & Palayew for replicating our analysis of traffic risks on April 20.1 2 We agree that the absolute risks on April 20 must be modest because the majority of Americans do not celebrate the ‘high holiday’ and because comparison days are not devoid of impaired driving. Similarly, secular trends in relative risks for April 20 reflect evolving driving norms, fluctuating traffic enforcement, changing baseline rates of cannabis use, variable celebration behaviours and rapid growth of the cannabis industry. We also agree that crashes are already recognised to be more frequent on traditional holidays such as Independence Day and Thanksgiving.3 Beyond this shared understanding, however, we disagree with Harper & Palayew on a key assumption in their analysis.

Distant or unmatched control days can be biased comparators because prevailing weather, road conditions, daylight hours, driver fatigue and travel patterns vary substantially across time. More crashes, for example, occur on weekends relative to weekdays, in summer than winter and during 2016 than 2010.4 This was our rationale for performing a matched analysis by comparing April 20 to control days precisely 1 week earlier and 1 week later, thereby exactly controlling for weekday, month and year. In addition to reducing bias ...


Keyword: Cannabis impaired driving

Language: en

Keywords

driver; drugs; epidemiology; motor vehicle occupant