Abstract:
|
Large-scale observational databases are increasingly used for comparative effectiveness research on therapeutic interventions. Exposure modeling techniques (e.g., propensity score methods) are popular for addressing confounding bias in observational studies. Theoretical work and extensive simulation studies have explored the properties and statistical performance of propensity score methods for estimating average treatment effects of point (or fixed) exposures. However, attempts to "replicate" the findings of randomized clinical trials using propensity scores in observational datasets have produced mixed results. In this talk, we present the rationale, methods, and preliminary results of a large-scale attempt to "replicate" the findings of randomized clinical trials (RCTs) using observational datasets (e.g., claims data combined with electronic health records). We discuss how this work overcomes some of the limitations of previous attempts to compare the results of RCTs and observational studies, and highlight challenges in using routinely collected data to assess the effects of interventions.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.