Many times, that risk isn’t identified until well after the patient has been discharged, which is too late, he says. However, “if we can identify them electronically early in their hospital course, we can direct our care transition resources more effectively and reduce the readmission rate,” says Amarasingham. The top 20% of risk generates 60% of the readmissions, he adds.
The model uses both clinical and physiological factors as well as social and environmental factors to generate a risk score. It’s able to run in real time—that is, while the patient is still in the hospital—through its ability to extract data from Parkland’s comprehensive electronic medical record. That information is used to compile the risk score. A separate software system extracts that data within 24 hours of admission and identifies the top 20% to 40% who are most at risk of readmission and sends an alert to a dedicated team of pharmacists, nutritionists, nurse practitioners, and home visit nurses who do what Amarasingham calls “very aggressive discharge planning and care transition.”
They help transition such patients to an outpatient environment. After discharge, a trained home-visit nurse visits the patient within two weeks and makes sure the patient attends an outpatient primary care visit within 30 days. After implementing the tool early in 2010, Parkland’s preliminary results through last fall show a nearly 40% reduction in readmissions for Medicare patients and 25% among all patients.
“As an industry we have to start looking at value and not just volume of services,” says Ron Anderson, Parkland’s CEO. “Five percent of the people cost 49% of the money, and the top 1% use 30% of the resources. A lot of those people have these chronic conditions. Before we haven’t had very good ways to predict their resource needs, particularly on the day of admission, as Ruben’s tool allows.”
Based on the success of the heart-failure tool, Amarasingham is hard at work developing one for pneumonia and heart attack. He also has a $400,000 grant to try to develop a similar tool that works regardless of disease state.
“What we want to do is find a way to change the cost curve without affecting quality adversely,” Anderson says. “There’s an opportunity here to have real reform that’s not about payment mechanisms but delivery models.”