Introducing MiCohort Screener, an AI-assisted tool that helps researchers determine eligibility for studies
The tool streamlines the process of reviewing medical records so that teams can focus more on recruitment.
ANN ARBOR – Research teams at the University of Michigan can now spend less time analyzing patient charts and more time advancing recruitment for their studies and clinical trials with MiCohort Screener, a new AI-assisted tool developed by AI & Digital Health Innovation (AI&DHI).
Today, research coordinators must manually review hundreds of medical records to identify eligible participants—a process that can take weeks or longer. MiCohort Screener streamlines this work by automatically screening patient charts against a study or trial’s inclusion and exclusion criteria.
How it Works
“MiCohort Screener is designed to support, not replace, research team review. By helping organize and summarize chart-based eligibility information, this tool allows teams to focus on confirming eligibility, resolving uncertain cases and preparing for participant recruitment.”
Megan Haymart, MD
Associate Director of Data Solutions, AI&DHI; Nancy Wigginton Endocrinology Research Professor of Thyroid Cancer and Professor of Internal Medicine, Michigan Medicine
To use MiCohort Screener, researchers first build a prospective cohort using DataDirect or any structured query tool. The team then defines their study’s screening criteria within MiCohort by creating clinical questions with detailed eligibility requirements or by adapting pre-made prompts that are available in the tool’s community library.
After the criteria are established, study teams may upload a list of up to 1,000 patients. MiCohort Screener will review each chart and return patient-level eligibility determinations as well as model reasoning and citations to source documentation. The results can then be exported to Excel for review by investigators and sponsors.
“MiCohort Screener is designed to support, not replace, research team review,” said Dr. Megan Haymart, Associate Director of Data Solutions at AI&DHI, Nancy Wigginton Endocrinology Research Professor of Thyroid Cancer and Professor of Internal Medicine at Michigan Medicine. “By helping organize and summarize chart-based eligibility information, this tool allows teams to focus on confirming eligibility, resolving uncertain cases and preparing for participant recruitment.”
MiCohort Screener uses U-M GPT, the University of Michigan’s institutionally hosted, HIPAA-approved large language model. The tool draws on the Michigan Medicine Research Data Warehouse, which means it does not require new data flows or third-party dependencies.
MiCohort Screener is available to all members of the U-M research community including faculty, staff and students. Teams must have an IRBMED protocol approved for use of PHI data before being granted access to the tool.
Investigators who are interested in using MiCohort Screener are encouraged to fill out the access form on AI&DHI’s website.
About AI & Digital Health Innovation
AI & Digital Health Innovation (formerly Precision Health at U-M) is dedicated to empowering researchers at the University Michigan to change the future of digital healthcare. They work with multi-disciplinary teams of health providers, basic scientists, engineers, and administrators to tackle the most difficult research problems and help rapidly bring ideas to the bedside. For more information visit aidhi.umich.edu.