Background: Randomized clinical trials (RCTs) are the gold standard for identifying optimal treatment options. However, RCTs usually compare two treatments on a small homogenous group of patients. Thus, observational data are the main tool to identify the optimal treatment among several when applied to a large cohort.
Objective: To test and provide guidance for utilization of statistical methods aimed at estimation of optimal treatment among a group of three or more treatments.
Methods: We will use extensive simulation analyses that are based on plausible realistic settings to examine current as well as newly proposed procedures for estimating treatment effect with multiple treatments. These simulations will allow us to identify the optimal method for investigators to use taking into account their own questions and data sets.
Outcomes: This study will provide researchers with guidelines for choosing the most appropriate statistical method when comparing more than three treatments. We will also develop new methods for estimation, when currently used methods are inadequate, and will provide software implementing the methods for use by other researchers.
Anticipated Impact: To substantially extend the current standards for estimating causal effects to situations where comparison between multiple treatments is required. This addresses a significant methodological gap in the literature and is essential for patient-centered outcomes research.
Lopez M, Gutman R. Estimation of causal effects with multiple treatments: a review and new ideas. (January 2017).
Document Type: working paper.