This article is based on a presentation by Ariel Chandler, Senior Healthcare Technology Consultant at Protiviti at IBA’s Spring 2023 Conference on Healthcare Analytics.
According to Ariel Chandler, the healthcare is an industry is in flux.
Aside from the effects of the COVID-19 pandemic, one of the main drivers of change in the industry is the transition from a fee-for-service model to a value-based model, Chandler said. She explained this transition and the current state of provider analytics in her presentation “Using Provider Analytics to Improve Patient Health Outcomes” during the Spring 2023 Institute of Business Analytics Conference on Healthcare Analytics.
Provider analytics are analytics related to any healthcare provider, Chandler said, including physicians, nurses, and physical therapists. In her role at Protiviti, Chandler works to systematically improve patient outcomes through provider analytics.
The Current State of Provider Analytics and Patient Outcomes
Historically, healthcare systems and providers have been paid for each service performed, following the “fee-for-service” model. Now, members of the industry are starting to think about healthcare from a value-based perspective, Chandler said, and focusing on keeping the population healthy. Because of the history of the fee-for-service model, though, Chandler noted that there is minimal outcome measurement for value-based models.
“Right now, a lot of analytics that have been used are very volume-based,” Chandler said. “How many patients are using a day? What number of procedures per provider? Providers are thinking about how to manage their daily tasks instead of trying to help the patient get the best outcome possible.”
Another challenge the healthcare industry is facing is data quality and data siloing barriers, Chandler said. The data currently used in a lot of processes is from claims or electronic health records, which were created for other purposes but are now being reused for operational optimization purposes.
“There’s a huge opportunity to improve patient outcomes, and in a way that’s financially viable,” Chandler said. “Right now, 20-25% of American healthcare spending is wasteful because of failures in delivery of care and care coordination, and low value treatment. All of these things can be addressed at the provider level.”
How to Successfully Engage Provider Analytics
Chandler offered three key areas to consider to successfully use provider analytics: recognizing system context, designing analytics that enable action, and standardizing data.
To improve outcomes, you have to understand and improve the system, Chandler said. This involves understanding the multiple, interconnected factors that account for variances in outcomes including individual clinicians, collaboration between clinicians, and the treatment a patient receives. You have to account for these differences in your model to get an actionable result from it, Chandler said.
Next, Chandler suggested asking key questions to enable action on the client’s part and make sure they are getting the information they need from the data they receive.
“What problem are they trying to solve?” Chandler said. “Who is the end user and what is their goal? What are they trying to achieve — financial stability for their organization or improved clinical outcomes? Will medical care impact outcomes?”
Chandler’s final key consideration for success was making sure that healthcare providers can be compared to one another and communicating the limitations of data. Providers have very diverse roles in our healthcare system, Chandler said, because they treat different patient populations.
“There are physicians, physical therapists, nutritionists, and more,” Chandler said. “They’re going to see different patients, so we don’t want to all just compare raw numbers to each other. Risk adjusting is a critical first step for any analytic solidity and in addition to that risk adjusting, it’s really important to bring in any benchmarking, similar peers, or proportional measures to help drive the next step.”
Provider Analytics in Action
One use case of provider analytics that Chandler provided was from the dissertation research she completed as a Ph.D. student at Northwestern University. The research focused on measuring teamwork to improve quality outcomes, and she specifically studied teamwork among clinical members of a patient care team in an inpatient neurological unit.
To measure the positive patient outcomes among high teamwork connections versus low teamwork connections, Chandler and her team put all of the providers into a network and weighted the network based on their teamwork score. The research showed that when the high-performing group worked together, they had a 60% rate of positive patient outcomes, versus a 9% rate of positive patient outcomes for the low-performing group.
Next, the team hypothesized that during certain treatments, and certain steps in treatment, teamwork was much more critical to patient outcomes than in other areas. They found that there were certain critical clinical steps the care team took that led to improved patient outcomes, including a higher rate of documented communication.
“We did Natural Language Processing on CT scan notes to determine if the connection between measured and actual teamwork was really there, and we could find it in the clinical documentation,” Chandler said. “We were able to take this data mining technique of looking at teamwork and identify just a couple of steps in treatment where quality improvement can strengthen those steps to make sure that communication at these key steps occurred.”
Leave a Reply