Precision Health is collaborating with the Michigan Medicine Central Transitions of Care (TOC) team, and the Wiens Lab to evaluate 2 new AI tools that are available to help identify patients at high risk for readmission. Testing a model in the clinical setting is a complex process requiring dedicated resources to help navigate the challenging pathway by making connections with subject matter experts, clinical champions and governance committees that will elevate the likelihood of successful evaluation.

Quality Improvement Pilot

The team will launch a 3-month quality improvement pilot to test the new tools in the local setting at UMHS. Specifically, the goal is to test 2 new machine learning algorithms, HARPP and version 2 Epic Readmission model. The HARPP model, built by the Wiens Lab for Michigan Medicine is a hospital-developed (custom) machine learning algorithm utilizing electronic health record (EHR) data combined with data used to compute the LACE score to predict risk of readmission and the likelihood of benefiting from a post discharge intervention. Using data from this pilot, the study team will compare HARPP to LACE and compare version 2 Epic Readmission model to HARPP and LACE.

More to Come

Whichever model turns out to have the best performance will be used operationally at Michigan Medicine to support the Central TOC service by enhancing the transition of care from hospital to home and promoting optimal patient outcomes. Learn more about the evaluation in upcoming communications.

The Team, The Team, The Team

This effort has been an interdisciplinary collaboration, as most Precision Health projects are! Project champions, investigators and contributors include the following folks: