Eric Murray – Fairview Health Services

As soon as a patient exhibiting COVID-19 symptoms enters the emergency department at one of Health Fairview’s hospitals, clinicians are on high alert—a crucial first step being to order a chest x-ray.

Those captured images immediately go through a specially developed algorithm that scans for patterns associated with the virus.

Within seconds, the health care team can determine whether the patient has the coronavirus, or may be afflicted by something else. In either case, an action plan follows.

Eric Murray, director of IT at Fairview Health Services, points to this as a prime example of how Fairview and its academic partner, the University of Minnesota, use data, machine learning and predictive modeling to tackle real-time health care needs. Particularly with COVID-19, those needs have been accelerated, he adds, and that’s served to push advancements in care enabled by data and technology.

“It is rewarding to see our work used in a real-world application,” Murray says, “and one that so quickly came to life.”

Data: The future of healthcare

A nonprofit academic health system and Minnesota’s largest hospital system by volume of patients, Fairview Health consists of 12 hospitals and medical centers and more than 150 primary care and specialty clinics across the midwestern state.

Since joining in January 2020, Murray has capitalized on Fairview’s unique relationship with the University of Minnesota to use data and analytics to improve quality of care, enhance the patient and physician experience, and reduce health care costs.

Take, for instance, the chest x-ray algorithm. The system and the university collaborated to write and test the model, which can not only help diagnose, but serves as a baseline to determine whether a patient with the virus is getting better or worse. And while currently specific to COVID-19, the plan is to use similar algorithms as a guide to improve care for other viruses and conditions, Murray explains.

He describes the model as “a perfect storm between the skillset and the technology already there, and the real-world need for an answer.”

In addition, Fairview is using artificial intelligence to optimize operations: For example, a patient flow command system allows floor staff and clinicians to visualize and monitor their environment in real time and receive “nudges” (alerts) when something requires their attention.

“What makes this really exciting is that the nudges are predictive,” Murray explains. “Using the command center, we can optimize the flow of patients across our hospitals, allowing us to act as one system and adjusting patient flow proactively—including guiding a care team to enable the expedited discharge of patients that are ready to go home.”

Under one roof

Passionate about predictive analytics and machine learning, Murray’s goal is to modernize Fairview’s data warehouse to provide easier access to reliable data. The ideal scenario, he says, is a continuously evolving health system, where models are implemented, monitored and enhanced to better care for specific disease states and the overall patient population.

All of this, in turn, must be understandable for clinicians using the system. Data must present not just as numbers, Murray emphasizes, but more visually digestible information that inspires action and leads to improved outcomes.

“These days, many physicians are in information overload,” he says. “Providing clear, concise information, and at the right time, is going to be the key to success. In the past, a doctor may have received a report saying ‘This is what happened with your recent patient population.’ But the current implementation is more focused on ‘This is what is happening right now, and this is what you can do with this information to change the course for your patients.’”

A beneficial partnership

Equally important initiatives include the ongoing buildout of a master data management strategy and a data archival service. In a matter of days, Murray and his team created the infrastructure for a new data repository that pulls together all patient data from across multiple systems, providing what he calls a “longitudinal view” of the patient.

“It’s all about creating an environment where algorithms can be tested and modified in a secure location,” Murray explains. “We can then use the infrastructure to train and validate our models and link patient data across systems.”

That, in turn, will allow the organization to drive down operational costs by decommissioning unneeded apps while still providing easy access to the data within those apps. Patient information from different systems will be united to create “a total view of care, which is absolutely critical,” Murray says.

What makes these initiatives possible? The “strong, beneficial partnership” between the system and the university, Murray says.

It’s a symbiotic relationship: Fairview has the data, the access to patients and the great skillsets of physicians—while the university has academics versed in machine learning and AI. The latter’s research with real-life and real-time hospital data helps them learn and hone their skills, and, in turn, their discoveries can be plugged back into Fairview’s systems to impact real health care change.

“We need each other to be able to do all this,” Murray says.

Technology for the greater good

A graduate of the University of Wisconsin at Madison, Murray spent the first 15 years of his career at Boston Scientific. He started as an intern, then ascended the ranks from that of developer to, finally, the director of information technology for digital health and data services reporting to the chief information officer

Fairview presented the opportunity to move from the product side to the provider side of healthcare.

“They were really looking for someone to come in and drive transformation,” he says. He adds that the pressure to reduce health care costs has never been greater, and that many organizations are looking for IT professionals to take the lead. “A big part of my job comes down to the question, ‘How do we reduce the cost of patient care by modernizing and advancing technology, while simultaneously driving down the cost of IT?’”

Still, the human element was what most attracted him to his role, he says. As a point of fact, he directly reports to chief analytics and care innovation officer Dr. Genevieve Melton-Meaux, an active surgeon at Fairview.

“It was a unique opportunity to work for someone who has the perspective of helping patients every day,” he says.

An experience that he now shares.

“The work that I’ve done has validated why I came over to Fairview,” he says. “I put a lot of time, energy, and passion into my job, and I have the opportunity to have my work help a broad range of patients and the community.”