Project Summary

One of PCORI’s goals is to improve the methods that researchers use for patient-centered outcomes research. PCORI funds methods projects like this one to better understand and advance the use of research methods that improve the strength and quality of comparative effectiveness research.

This research project is in progress. PCORI will post the research findings on this page within 90 days after the results are final.

What is the project about?

Researchers can use data in electronic health records, or EHRs, to compare two treatments to each other. Doing so can help show which treatment works better for patients. But comparing more than two treatments can be hard. Patients who get different treatments may differ from each other. Current methods for analyzing data may not fully account for differences among patients when many treatments are available to choose from, especially if information on patient traits is missing from EHRs.

In this study, the research team is developing new methods that let researchers make accurate comparisons across many treatments. The new methods also assess how conclusions will vary when information is missing from EHRs.

How can this project help improve research methods?

Results may help researchers when considering methods to compare multiple treatment options. 

What is the research team doing?

In this project, the research team is using machine learning techniques to compare multiple treatments. In machine learning, computers use data to learn how to perform tasks with little or no human input. The team is creating data to see how the new machine learning methods work in different settings. The team is also comparing the new methods to current methods. In addition, the team is testing how well the new methods account for traits that aren’t included in EHRs. Finally, the team is applying the new methods to existing data comparing treatment options for lung cancer. 

Research methods at a glance

Design Elements Description

The aims of this study are to

  • Develop new statistical methods for estimating the effects of multiple treatments with a binary outcome
  • Examine the operating characteristics of the new methods under a variety of contextually motivated settings and compare the new methods’ performance against existing methods
  • Develop a Bayesian framework for evaluating the sensitivity of the new methods to unmeasured variables
  • Apply the new methods to Surveillance, Epidemiology, and End Results-Medicare (SEER-Medicare) data to compare the effect of multiple surgical options (robotic-assisted surgery, open resection, video-assisted thoracic surgery) for treating lung cancer on perioperative mortality and adverse events
Approach Bayesian machine learning techniques involving Bayesian Additive Regression Trees (BART); simulation analyses

Project Information

Liangyuan Hu, PhD
Rutgers, The State University of New Jersey
Bayesian Additive Regression Trees for Causal Inference with Multiple Treatments and a Binary Outcome

Key Dates

August 2018
May 2023

Study Registration Information


Award Type
Intervention Strategy Intervention Strategies PCORI funds comparative clinical effectiveness research (CER) studies that compare two or more options or approaches to health care, or that compare different ways of delivering or receiving care. View Glossary
State State The state where the project originates, or where the primary institution or organization is located. View Glossary
Last updated: January 20, 2023