As the healthcare industry moves from a fee-for-service model of patient care to one that is based on value and population health, use of predictive analytics is playing a larger role in care delivery. Aurora’s use of a predictive analytics tool has positioned us well for this new environment.
Aurora began exploring the use of such a tool in 2012. We recognized that our patient population had a high incidence of heart failure (HF) and chronic obstructive pulmonary disease (COPD) compared to patients at other organizations of our size, and we wanted to develop a care model that used proactive care management to reduce avoidable hospital admissions.
As part of a pilot program with our HF patients, we used the tool to identify the likelihood that a given patient would be admitted to the hospital within the next six months. Patients who fell in the 80th percentile or higher, per the tool, received proactive support from an RN care coordinator working as part of the patient’s primary care team.
After one year, we recorded a 60 percent decrease in HF-related admissions among the 126 patients in the pilot. In addition, those patients had a 20 percent decrease in all-cause admissions compared to the previous year.
The same model was implemented for COPD patients, with 240 patients monitored by the RN care coordinators. The overall admission rate for those patients also decreased by 20 percent.
In addition to the decreased admissions, subsequent data analysis showed that many of the patients being monitored had shifted into lower-risk percentiles.
A New Care Model
When we began the project, we knew that the standard, reactive model of care was not going to keep high-risk patients from being admitted, and that simply adding an RN care coordinator to the existing model was insufficient. We developed standards of care for these high-risk patients, including best practices for physicians and training for all members of the care team.
Our RN care coordinators help formulate a plan of care for patients in collaboration with the patient’s primary care physician, and they monitor the patient through regular contact. The coordinators emphasize the importance of medication adherence, educate patients about the early warning signs that their health status is deteriorating, and monitor patients’ symptoms between physician visits.
Our care coordinators also address potential barriers to care, such as access to transportation, the ability to afford medications, and the need for assistance with household tasks. While such topics go beyond the scope of a typical patient visit to a doctor’s office, they have a huge impact on healthcare outcomes.
The program has succeeded in its goal of empowering patients to take care of their own health in a proactive manner. We have expanded the program throughout the Aurora system and are now monitoring 2,500 high-risk patients with HF, COPD, or diabetes.
We also are in the process of adding patients’ risk scores to electronic health records, making it easier for clinicians to recognize patients who may benefit from additional support.
In the big picture, predictive analytics are a great tool, but they are just that—a tool. They identify opportunities, but the real work is in developing a care model that is consistent and replicable and that engages all members of the care team.
Success in caring for these high-risk patients does not come from the data points provided by an analytic tool, but rather in what you do with them.