The traditional concept of confounding was previously introduced in this Key Research Skills series — a confounder is a variable that is associated with both the exposure and outcome of interest without being an intermediate on the causal pathway between them.1 Although the earlier article also outlined how a randomised controlled trial (RCT) can be used as a solution to handle confounding at the study design phase, it is not always feasible to use an RCT. For example, consider a study looking at the association between smoking (exposure) and cardiovascular disease (CVD; outcome). It is not ethical to randomly assign people to smoke! The alternative is to conduct an observational study in which confounders are carefully identified, measured without too much error, and adjusted for in the statistical analysis.
The full article is accessible to AMA members and paid subscribers. Login to read more or purchase a subscription now.
Please note: institutional and Research4Life access to the MJA is now provided through Wiley Online Library.
- 1. Attia JR, Jones MP, Hure A. Deconfounding confounding part 1: traditional explanations. Med J Aust 2017; 206: 244–245. https://www.mja.com.au/journal/2017/206/6/deconfounding-confounding-part-1-traditional-explanations
- 2. Von Elm E, Egger M. The scandal of poor epidemiological research. BMJ 2004; 329: 868.
- 3. Manson JE et al. The Women's Health Initiative hormone therapy trials: update and overview of health outcomes during the intervention and post‐stopping phases. JAMA 2013; 310: 1353–1368.
- 4. Petitti D. Commentary: hormone replacement therapy and coronary heart disease: four lessons. Int J Epidemiol 2004; 33: 461–463.
- 5. Lee PH. Is a cutoff of 10% appropriate for the change‐in-estimate criterion of confounder identification? J Epidemiol 2014; 24: 161–167.
- 6. Vanderweele TJ. Principles of confounder selection. Eur J Epidemiol 2019; 34: 211–219.
- 7. Attia JR, Oldmeadow C, Holliday EG, Jones MP. Deconfounding confounding part 2: using directed acyclic graphs (DAGs). Med J Aust 2017; 206: 480–483. https://www.mja.com.au/journal/2017/206/11/deconfounding-confounding-part-2-using-directed-acyclic-graphs-dags
- 8. Jones MP, Walker MM, Attia JR. Understanding statistical principles in correlation, causation and moderation in human disease. Med J Aust 2017; 207: 104–107. https://www.mja.com.au/journal/2017/207/3/understanding-statistical-principles-correlation-causation-and-moderation-human
- 9. Webb P, Bain C, Page A. Essential epidemiology 3rd ed. Cambridge: Cambridge University Press, 2017.
- 10. Banks E, Joshy G, Korda RJ, et al. Tobacco smoking and risk of 36 cardiovascular disease sybtypes: fatal and non‐fatal outcomes in a large prospective Australian study. BMC Med 2019; 17: 128.
- 11. Johansen KL, Chertow GM, Ng AV, et al. Physical activity levels in patients on haemodialysis and health sedentary controls. Kidney Int 2000; 57: 2564–2570.
- 12. Karaca‐Mandic P, Norton EC, Dowd B. Interaction terms in nonlinear models. Health Serv Res 2012; 47: 255–274.
Series editors
No relevant disclosures.