An Efficient Distributed Learning Framework for Integrating Evidence in Clinical Research Networks
Study Dates
Last Modified
Tags
Publisher
Abstract
Study to develop a framework of distributed algorithms that efficiently synthesize evidence in large clinical data research networks for studying impacts of risk factors for rare adverse events.
Affiliation(s)
Funder(s)
Provenance
Description
Rare events such as gastrointestinal bleeding, clinical sites independently often contain insufficient number of cases leading to challenges in characterizing the effect sizes of risk factors for adverse events and conducting accurate risk prediction. Individual patient-level information at each site is often protected by privacy regularities and rules, and direct data integration across multiple clinical sites is infeasible or requires a large amount of operational effort.
Literature presents gaps in how to fully utilize all electronic health records (EHR) data from different sites to characterize the impacts of risk factors on adverse events following certain medications and how to borrow information using all EHR data across different sites to better predict the incidence of adverse events in patients.
In this study, the research team is to develop a framework of distributed algorithms that efficiently synthesize evidence in large clinical data research networks, for studying impacts of risk factors for rare adverse events. The framework of algorithms will be tailored and applied specifically for clinical data research networks (CDRNs), with a focus on enabling more efficient privacy-preserving sharing of data with the PCORI-funded pediatric learning healthy system, PEDSnet.
Development Code
Vocabulary
Clinical Subjects Headings
Creative Commons license
