Parkland Health and Hospital System

Using an Automated Model to Identify Heart Failure Patients at Risk for 30-Day Readmission or Death Using EMR Data

Heart Failure (HF) is the leading cause of readmission among patients age 65 and older. While numerous studies have shown that interventions that include discharge planning, provider coordination and patient counseling can be effective in preventing readmission, targeting all HF patients for such an intervention can be very costly. However, by using an electronic medical record (EMR) to identify HF patients upon their initial admission, patients at risk for readmission could be targeted for an intervention at the first point of contact, thus potentially resulting in a decline in readmission rates for HF. Targeting only patients at risk for readmission could also reduce costs of an intervention that may otherwise be cost prohibitive.

Researchers at Parkland Health and Hospital System hypothesized that:

  • An automated model derived from clinical and nonclinical factors, including social, behavioral and utilization factors, could accurately stratify the risk for readmission among HF patients; and, 
  • The clinical factors that ordinarily predict mortality may be insufficient to predict readmission by themselves.
To test these hypotheses, they used real-time EMR admission data for patients admitted to Parkland Memorial Hospital between January 1, 2007 to August 31, 2008 with a principal discharge diagnosis of HF to construct and validate a model of 30-day readmission risk and 30-day mortality.

By using real-time data in the electronic model, researchers found it was possible to predict mortality and readmission from HF within 24 hours of the patients’ admission. Using the model, clinicians would be able to make more informed decisions about the course of care for HF patients. For example, patients at risk for readmission may be targeted for additional intervention services aimed at preventing readmission, whereas patients found to be at low risk may be discharged early. The ability to target patients at risk can help clinicians and institutions use limited resources on those who need it the most.

For more information on Parkland’s automated electronic model, or for reprints of the Medical Care article which describes the model, please contact:
Ruben Amarasingham, MD, MBA
Center for Clinical Innovation
Parkland Health and Hospital System
5123 Harry Hines Blvd.
Dallas, TX 75235
[email protected]  

"Most Wired" Innovations

  • Other Innovations

  • Denver Health

    Electronic Dashboards for Quality, Safety, and Operations: Enchancing Feedback and Transparency

Upcoming Events

From: 
Email:  
To: 
Email:  
Subject: 
Message: