Two-Stage Meta-Regression Framework for Precision Medicine Using Data from Clinical Data Research [Methods Study], United States, 2018-2023 (ICPSR 39739)
Version Date: Mar 23, 2026 View help for published
Principal Investigator(s): View help for Principal Investigator(s)
Yong Chen, University of Pennsylvania
https://doi.org/10.3886/ICPSR39739.v1
Version V1
Summary View help for Summary
Network meta-analysis, or NMA, is a statistical method that researchers use to combine results from many clinical trials done within a research network. A research network is a group of scientists and doctors from different places, like hospitals and research centers, who do studies together and share data. Researchers can use NMA to compare how well different treatments work for a specific health problem. But current NMA methods don't work well when comparing three or more treatments across many health outcomes.
In this study, the research team developed new NMA methods to compare three or more treatments that bring on labor to start the process of childbirth across many health outcomes using research network data.
Citation View help for Citation
Export Citation:
Funding View help for Funding
Subject Terms View help for Subject Terms
Geographic Coverage View help for Geographic Coverage
Distributor(s) View help for Distributor(s)
Study Purpose View help for Study Purpose
To develop new network meta-analysis (NMA) methods for comparing three or more treatments at the same time across multiple health outcomes using research network data
Study Design View help for Study Design
The research team developed new methods, called PAtient-centered treatment ranking via Large-scale Multivariate network meta-analysis (PALM), for comparing multiple treatments. PALM consists of three parts: (1) a framework that accounts for multiple outcomes and treatment comparisons, (2) an algorithm that calculates the model's parameter values, and (3) plots to visually compare treatments.
The research team tested the methods with existing data from completed clinical trials for labor induction treatments among pregnant people. First, the team used PALM to estimate the effects of eight treatment options on five labor induction outcomes. Using the visualization plots, the team ranked the treatments based on the likelihood that each treatment would result in each of the outcomes compared with a placebo. Then they compared the eight treatments to each other based on each treatment's combined risk for two outcomes: cesarean section and serious neonatal morbidity.
To test the likelihood of errors with the new methods, the research team also conducted simulation studies under equal and unequal variance assumptions between individual studies. Equal variance assumes trials results are similar, whereas unequal variance assumes trial results are different.
Patients and clinicians helped design the study.
Data Source View help for Data Source
NMA data set with data from 48,068 women in 280 randomized clinical trials on prostaglandin treatments in the Cochrane Pregnancy and Childbirth Database
Notes
The public-use data files in this collection are available for access by the general public. Access does not require affiliation with an ICPSR member institution.
ICPSR usually offers files in multiple formats for researchers to be able to access data and documentation in formats that work well within their needs. If you have questions about the accessibility of materials distributed by ICPSR or require further assistance, please visit ICPSR’s Accessibility Center.