From Connor and Colleagues

Connor, Jennie; Ameratunga, Shanthi; Jackson, Rod

doi: 10.1097/01.ede.0000123328.08930.cb
Commentary: Author Response
Author Information

School of Population Health University of Auckland Auckland, New Zealand

Article Outline

It is indisputable that the injury burden due to road traffic is a huge public health problem and that epidemiologists have made a contribution to control of road traffic injury, largely in rich countries.

However in discussing the lack of specialized injury epidemiology training programs, Dr. Robertson1 has failed to acknowledge the enormous advances in the understanding of epidemiologic methods in general in the last 20 years or so, and their increasing application to injury prevention. Whether or not one sees injury epidemiology as a separate discipline, this line of research requires a thorough understanding of epidemiologic principles and study design. In the area of case-control studies in particular, our understanding has been revolutionized by the work of “modern” epidemiologists. This has provided opportunities for the more rigorous study of component causes of injury in populations, as the case-control design and its variants are particularly well suited to studying uncommon events where at least some contributing factors are transient.

The commentary suggests that the use of unmatched controls is a major weakness of our study. Dr. Robertson misses the point of our research.2 The controls are a good representation of the underlying study population, selected in proportion to their driving exposure. While matching may improve efficiency, this approach would make it impossible to estimate the population attributable risks or the effects of the matching factors. Our use of representative controls does not preclude anything that could be done in a matched study.

Robertson’s description of modern methods as “intellectual fun” also misses the point. Improving methodology is central to improving the reliability of results. Descriptive studies serve a range of useful functions, as all epidemiologists know, but reliably quantifying contributing causes of injury is not one of them.

Back to Top | Article Outline


1.Robertson LS. Causes and prevention of motor vehicle injuries [commentary]. Epidemiology. 2004;15:350–351.
2.Conner J, Norton R, Ameratunga S, Jackson R. The contribution of alcohol to serious car crash injuries. Epidemiology. 2004;15:337–344.
© 2004 Lippincott Williams & Wilkins, Inc.