The implications of propensity score variable selection strategies in pharmacoepidemiology: an empirical illustration
Article first published online: 10 MAR 2011
Copyright © 2011 John Wiley & Sons, Ltd.
Pharmacoepidemiology and Drug Safety
Volume 20, Issue 6, pages 551–559, June 2011
How to Cite
Patrick, A. R., Schneeweiss, S., Brookhart, M. A., Glynn, R. J., Rothman, K. J., Avorn, J. and Stürmer, T. (2011), The implications of propensity score variable selection strategies in pharmacoepidemiology: an empirical illustration. Pharmacoepidem. Drug Safe., 20: 551–559. doi: 10.1002/pds.2098
- Issue published online: 23 JUN 2011
- Article first published online: 10 MAR 2011
- Manuscript Accepted: 7 DEC 2010
- Manuscript Revised: 9 NOV 2010
- Manuscript Received: 2 JUN 2010
- confounding factors;
- epidemiologic methods;
- hydroxymethylglutaryl-CoA reductase inhibitors;
To examine the effect of variable selection strategies on the performance of propensity score (PS) methods in a study of statin initiation, mortality, and hip fracture assuming a true mortality reduction of < 15% and no effect on hip fracture.
We compared seniors initiating statins with seniors initiating glaucoma medications. Out of 202 covariates with a prevalence > 5%, PS variable selection strategies included none, a priori, factors predicting exposure, and factors predicting outcome. We estimated hazard ratios (HRs) for statin initiation on mortality and hip fracture from Cox models controlling for various PSs.
During 1 year follow-up, 2693 of 55 610 study subjects died and 496 suffered a hip fracture. The crude HR for statin initiators was 0.64 for mortality and 0.46 for hip fracture. Adjusting for the non-parsimonious PS yielded effect estimates of 0.83 (95%CI:0.75–0.93) and 0.72 (95%CI:0.56–0.93). Including in the PS only covariates associated with a greater than 20% increase or reduction in outcome rates yielded effect estimates of 0.84 (95%CI:0.75–0.94) and 0.76 (95%CI:0.61–0.95), which were closest to the effects predicted from randomized trials.
Due to the difficulty of pre-specifying all potential confounders of an exposure-outcome association, data-driven approaches to PS variable selection may be useful. Selecting covariates strongly associated with exposure but unrelated to outcome should be avoided, because this may increase bias. Selecting variables for PS based on their association with the outcome may help to reduce such bias. Copyright © 2011 John Wiley & Sons, Ltd.