The Problem

The Michigan Medicine Physical Medicine and Rehabilitation Team experiences re-admission rates that are higher than the national average. Readmitting a patient who has been treated and discharged results in increased costs and takes a psychological toll on the patient. 

To help reduce their readmission rates, the Rehab Team was searching for an “off the shelf” solution to predict potential patient readmissions. However, they were unable to find anything that met their specific needs. After hearing about the success MCIRCC has had working on similar projects, they decided to consult with the MCIRCC Data Science team (pictured above). Through this collaboration, the teams were able to define requirements for a predictive analytic that will alert clinicians to patient deterioration in the rehab unit prior to any symptoms appearing. 

The Solution

The MCIRCC Data Science Unit modified its existing PICTURE (Predicting Intensive Care Transfers and other Unforeseen Events) analytic to predict deterioration in rehab patients. These predictions are based on patient vital signs (ex: temperature, respiratory rate, etc.) as well as labs.  

Following initial discussions with the Rehab Team administrators, the MCIRCC Data Science Team developed an application customized to fit the needs of the Rehab team. The application features MiChart integration and single sign on, allowing seamless access for clinicians. Clinicians are alerted in real-time when deterioration is detected in a patient and are supplied with the key indicators that alerted the analytic to the pending decline.


With support from Health Information Technology & Services (HITS), PICTURE entered the Rehab Unit on September 4. An initial three-month pilot will allow the entire Rehab team to work with the technology before further studies measure the effectiveness of the application. There will be continuous conversations between all teams involved to continue improving the product.

MCIRCC is currently exploring other areas where PICTURE can help to improve care. Current areas of focus include Cardiac Care and Pediatrics. If a predictive analytic application could help improve patient care in your unit, please contact MCIRCC– we’d be happy to find a way to work with you!