Propensity Score Methods for Merging Observational and Experimental Data Sets

• Activity Number: 99 - Causal Inference with Non-Traditional Designs
• Type: Invited
• Date/Time: Monday, July 29, 2019 : 8:30 AM to 10:20 AM
• Sponsor: IMS
• Author(s): Evan Rosenman* and Art Owen and Michael Baiocchi and Hailey Banack
• Companies: Stanford University and Stanford University and Stanford University and University at Buffalo
• Keywords: causal inference; randomized controlled trials; observational studies; external validity

Abstract

This project considers how to augment a limited amount of data from a randomized controlled trial (RCT) with more plentiful data from an observational database (ODB), in order to estimate a causal effect. We work with strata defined by the propensity score in the ODB. RCT subjects are placed in strata based on the propensity they would have had, had they been in the ODB. Our first method spikes RCT data into their corresponding ODB strata. Our second method takes a data-driven convex combination of the ODB and RCT treatment effect estimates by stratum. Using the delta method and simulations, we show the spike-in method works best when RCT covariates are drawn from the same distribution as in the ODB. Our convex combination method is more robust than the spike-in to covariate-based inclusion criteria that bias the RCT data. We apply our methods to data from the Women's Health Initiative, a study of thousands of postmenopausal women which has both observational and experimental data on hormone therapy (HT). Using half of the RCT to define a gold standard, we find that a version of the spiked-in estimate yields stable estimates of the causal impact of HT on coronary heart disease.