Citation

Kim JH, Mooney SJ. Int. J. Epidemiol. 2016; 45(5): 1668-1675.

Affiliation

Department of Epidemiology, Columbia University, New York, NY, USA sjm2186@cumc.columbia.edu.

Copyright

(Copyright © 2016, International Epidemiological Association, Publisher Oxford University Press)

DOI

10.1093/ije/dyw172

PMID

27524819

Abstract

This article describes the epidemiological principles underlying four observational study designs commonly used to assess traffic safety: the case-control, case-crossover, culpability and quasi-induced exposure designs. We focus in particular on the specific challenges for preventing bias using each design. Whereas recruiting controls representative of the source population poses a special challenge in case-control traffic safety studies, case-crossover designs are prone to recall bias, and culpability and quasi-induced exposure studies can be undermined by difficulties assigning crash responsibility. Using causal diagrams and worked examples, we provide a simple way to teach traffic safety designs to epidemiologists and to encourage proper application of epidemiological principles among researchers designing traffic safety studies.

© The Author 2016; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association.

Language: en