Using analytics to improve outcomes at the point of care: 3 lessons

As the population continues to age and many chronic conditions and co-morbidities increase, predictive analytics well become crucial to intervening earlier at the point of care and improving patient outcomes.

Nav Ranajee, the global marketing leader of healthcare analytics at IBM , Steven Stepp, director of business informatics and engineering at Harrisburg, Pa.-based Pinnacle Health System, and Matt Sabo, PhD, the healthcare data science practice lead at Waypoint Consulting, discussed the matter recently at Becker's Hospital Review's CIO/HIT + Revenue Cycle Summit in Chicago.

"You can't drive population health and quality improvement strategies without analytics," said Mr. Ranajee. "Analytics help providers get to know their patients, including their patients' medical histories, genetics and other factors."

Analytics also help providers make predictions about the direction of a patient's health. As a result, many providers have begun building their own predictive models, including PinnacleHealth.

According to Mr. Stepp, PinnacleHealth is currently building an infrastructure to support the complexities of a tremendously transitioning industry, called a "Closed Loop Awareness System." In this system, PinnacleHealth alters its behavior in response to patient patterns to make the system more successful at pursuing its goals.

PinnacleHealth is using two tools to support and integrate its data and analytics: Epic and a new proprietary system called Pulse

"We just feel that with so much development going on, Epic is just not going to do everything that we want to at the pace that we want to do it," said Mr. Stepp.

According to Mr. Stepp and Dr. Sabo, before a hospital begins integrating predictive analytics across the board, it should have a system in which it can create and operate the predictive data. The system should follow three steps:

1. Identify an area where operations need to improve

2. Use data to explain variation in the outcomes in that area; and

3. Use that information to create a standardized response to reduce the variations and evaluate the efficacy of the response.

PinnacleHealth has used this model to work on clinical and operational goals, such as extending the time between acute chronic obstructive pulmonary disease events and improving staff-to-patient ratios, across its hospital cost centers and emergency departments.

"Pinnacle has made a real commitment to predictive and prescriptive analytics. They've invested in the hardware, software, staff and training to develop a proficiency in this type of analysis," said Dr. Sabo. "By doing so, they have put themselves in the position to prove that this model of analyzing variations and creating a standard response is practical."

© Copyright ASC COMMUNICATIONS 2021. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.