Leveraging the integrated EHR for trial matching across a nationwide network.

Abstract

165 Background: The Guardian Research Network (GRN) is a nationwide consortium of integrated health systems, who share their electronic health records (EHR) to democratize clinical trial access through improvements in “process”. The GRN is a unique-in-class, free-to join, non-exclusive consortium leveraging the digital EHR---including labs, medications, demographics and non-discrete data (all text) data--- mining nightly for clinical trial candidates. Using a suite of NLP and AI tools, the GRN dramatically improves the efficiency of the clinical research staff, by electronically searching all records for the I/E criteria for trials. The GRN uses a central IRB, one contract and legal review, promising to revolutionize the trial accrual process and speed drug development. Methods: With a database of > 1.0M patients, the GRN reviews all active cancer records nightly from > 110 member hospitals to produce lists of trial candidates. Comprehensive electronic screens were filtered by manual reviews to rapidly find best candidates. Results: Our data collected over 10 mo suggest comprehensive electronic queries examining hundreds of thousands of records daily eliminate > 90% of ineligible patients in minutes. Manual review further refines eligible list. This is vastly different from current opportunistic screening approaches that examine only a tiny fraction of potential trial candidates (last week's new patients). Conclusions: The GRN has executed an all-inclusive approach to trial accrual, embedding a scalable database search technology within an integrated trial network. The novel approach seeks to exponentially expand operational capabilities of CTOs with limited staff, review all eligible patients, and solve for a large unmet need for democratizing trial access in the community. [Table: see text]