Huirong Hu, Ph.D. Candidate, Department of Biostatistics and Bioinformatics, University of Louisville

"Statistical methods for assessing treatment effects on ordinal outcomes using observational data"

Average treatment effect (ATE) is used to measure the outcome difference if all patients would have been treated under one treatment versus all patients would have been treated under another treatment. Many statistical methods have been developed to estimate ATE when the outcome is continuous or binary. In many cases, ordinal outcome is often used. We may lose information if we consider ordinal outcomes as continuous variable. In this project, we propose model the ordinal outcome directly using the marginal structure model (MSM), and we estimate the average treatment effect by the superiority score of the outcome under one treatment group versus another. To adjust the confounding factors between treatment and outcome in an observational study, we apply the inverse probability of treatment weighting (IPTW) to obtain a weighted sample, where the covariates become balanced among different treatment groups. We then assess ATE based on the weighted sample. Extensive simulation studies are carried out to examine the performance of the proposed method. We also applied the proposed method to access the ATE on patients’ recovery from alcohol use disorders using the Kentucky Medicaid data 2012-2019.

Stay connected TwitterFacebook LinkedIn YouTubeInstagram