New System Improves the Way Heart Failure Is Detected and Treated in Patients


At Intermountain, dictated reports and other data from electronic medical records are used by three different applications every 24 hours to determine a patient’s risk for heart failure. This data helps doctors determine each patient’s 30-day readmission and 30-day mortality risks. Patients with more severe cases are evaluated by a multidisciplinary team to receive more intensive treatment.
The previous method relied on manual data collection and took up to 40 minutes per patient to calculate. Thanks to this new automated system, Intermountain clinicians now have access to the Intermountain Risk Score and can review a patient’s status in just 10 minutes, allowing clinicians to intervene much more quickly.
The American Medical Informatics Association recently published Intermountain's study about the pilot program conducted at McKay-Dee Hospital in Ogden in 2014. The study showed a significant 67 percent decrease in 30-day mortality – and a significant increase in patient discharges to home health when using the automated heart failure applications. With the addition of natural language processing, accuracy of the applications rose from nearly 83 percent to more than 97 percent. 
“Clinicians could gather all the information manually by themselves, but the automatic detection allows them to spend more time treating and less time collecting,” says Jose Benuzillo, data analyst at the Intermountain Healthcare Cardiovascular Clinical Program. "We’re doing all that work for them so that they can concentrate their time and effort treating patients."
The application is currently installed in all 22 Intermountain hospitals and is receiving further evaluation at Intermountain Medical Center in Murray, Utah Valley Regional Medical Center in Provo, and Dixie Regional Medical Center in St. George. Algorithm-stratified care is expected to be used for heart failure patients at all Intermountain facilities by the end of 2016.
“When patient care doesn’t go as planned, it’s usually because someone didn’t get all the information they needed when they needed it,” said Scott Evans, medical informatics director at Intermountain Healthcare. “Clinical decision support applications work because they provide pertinent information when needed.”
Contributors to the study included an integrated team of cardiovascular and medical informatics professionals from Intermountain Healthcare. In addition to Evans and Benuzillo, the team included Benjamin D. Horne, James F. Lloyd, Alejandra Bradshaw, Deborah Budge, Kismet D. Rasmusson, Colleen Roberts, Jason Buckway, Norma Geer, Teresa Garrett and Donald L. Lappé.