Purpose: Propensity scores (PSs), a powerful bias-reduction tool, can balance treatment groups on measured covariates in nonexperimental studies. We demonstrate the use of multiple PS estimation methods to optimize covariate balance. Methods: We used secondary data from 1292 adults with nonpsychotic major depressive disorder in the Sequenced Treatment Alternatives to Relieve Depression trial (2001-2004). After initial citalopram treatment failed, patient preference influenced assignment to medication augmentation (n = 565) or switch (n = 727). To reduce selection bias, we used boosted classification and regression trees (BCART) and logistic regression iteratively to identify two potentially optimal PSs. We assessed and compared covariate balance. Results: After iterative selection of interaction terms to minimize imbalance, logistic regression yielded better balance than BCART (average standardized absolute mean difference across 47 covariates: 0.03 vs. 0.08, matching; 0.02 vs. 0.05, weighting). Conclusions: Comparing multiple PS estimates is a pragmatic way to optimize balance. Logistic regression remains valuable for this purpose. Simulation studies are needed to compare PS models under varying conditions. Such studies should consider more flexible estimation methods, such as logistic models with automated selection of interactions or hybrid models using main effects logistic regression instead of a constant log-odds as the initial model for BCART.
|Original language||English (US)|
|Number of pages||6|
|Journal||Annals of epidemiology|
|State||Published - Apr 2013|
Bibliographical noteFunding Information:
Supported by a contract from the Agency for Healthcare Research and Quality to The University of North Carolina Developing Evidence to Inform Decisions about Effectiveness Center (contract no. HHSA290200500401). At the time of this work, Dr. Dusetzina received funding from a National Research Service Award Pre-Doctoral Traineeship from the AHRQ , sponsored by the Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill , Grant No. 5-T-32 HS000032-20 . Dr. Til Stürmer’s work on this project was funded in part by R01 AG023178 from the National Institute on Aging at the National Institutes of Health .
- Epidemiologic methods
- Estimation techniques
- Propensity score
- Statistical models
- Statistics as topic