Citation

Staples JA, Redelmeier DA. JAMA Intern. Med. 2019; 179(3): 456.

Affiliation

Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada.

Comment On:

JAMA Intern Med 2018;178(4):569-572.

Copyright

(Copyright © 2019, American Medical Association)

DOI

10.1001/jamainternmed.2018.8094

PMID

30715145

Abstract

To the Editor In our Research Letter1 examining crash risks on the “4/20” counterculture holiday, we identified drivers (rather than crashes) as the unit of analysis because one or more drivers may contribute to a crash. This approach also helps optimize interpretability for clinicians and policymakers around driving risks despite breaching strict assumptions for statistical independence. Furthermore, as Aydelotte and colleagues2 point out, similar results were obtained whether the driver (OR, 1.12; 95% CI, 1.05-1.19) or the crash (incidence rate ratio, 1.10; 95% CI, 1.02-1.20) was the unit of analysis. The observed difference in calculated absolute risks simply indicates the average crash involves more than one driver.


Keyword: Cannabis impaired driving

Language: en