August 7, 2014

Hospitals Use Predictive Analytics to Allocate Resources

George Leopold
Health-Care-System.jpg

Data analytics continues to make inroads in the efficient allocation of expensive healthcare resources.

Among the latest examples is the use of predictive analytics at a Texas hospital that has helped reduce its 30-day readmission rate for heart failure patients by nearly half. Meanwhile, other efforts around the nation are designed to reduce initial hospitalization and readmission.

Observers note that analytics is gradually moving from the administrative to the clinical side of healthcare delivery. At either end, big data tools are being deployed to help reduce the soaring costs of U.S. health care, advocates say.

A clinical effort at the 213-bed Harris Methodist Hospital outside of Dallas is using predictive analytics to scan medical records to help clinicians determine what type of care will improve medical outcomes. An algorithm looks at multiple data points like blood pressure and blood glucose and “tells us which patient is at higher risk for heart failure,” explained Dr. Susan Land, the hospital’s chief medical officer.

The algorithm generates a “30-day risk score” for heart failure that allows physicians to target cardiac patients most in need of intensive follow-up care.

As healthcare costs rise, administrators and clinicians are trying to leverage analytics to maximize scare resources. “We have to pick and choose which [patients] we manage more intensely,” Land stressed.

The big data tool has been used to analyze the medical records of more than 14,000 patients admitted to the Texas hospital since 2012. The predictive analysis software was developed at the Parkland Center for Clinical Innovation. Since its adoption, 267 patients at the Texas hospital have been flagged as high risk while another 139 were classified as moderate risk heart patients.

The result was a reduction in the hospital’s 30-day readmission rate from 23 percent to 12 percent.

Increasing use of predictive analytics has also improved natural-language processing technology that enables software tools to read and physician’s notes and other electronic health records.

Other healthcare networks around the country are also using predictive analytics to rein in costs and target resources for patients most in need. The University of Pittsburgh Medical Center is spending $105 million on a data analytics initiative while Ohio State University’s Wexner Medical Center is using data algorithms to help triage cardiology and cancer patients.

Examples abound of clinical use of predictive analytics. For example, the 14-hospital Aurora Health Care system based in Milwaukee is using a predictive analytics tool developed by Humedica. Aurora is a member of the AMGA Collaborative, a data-sharing collective organized by the American Medical Group Association.

Two predictive analytics pilots focused not only on readmissions, but also on treating high-risk heart and chronic obstructive pulmonary disease patients at outpatient clinics.

Related items:

Healthcare Organizations Faced Daunting Data Challenges

Fixing Healthcare with Big Data