Richard A Hansen

Confounding control in a nonexperimental study of STAR*D data: logistic regression balanced covariates better than boosted CART

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.