How Eric Poon Eliminates Cracks in Patient Care at Duke Health

Eric Poon on Duke University Health System’s boundary-pushing, patient-centered tech adoption

Photo by Erin Hull/Duke University Health System

Eric Poon readily dismisses any notion that technological innovation and adoption at Duke University Health System is specifically under his purview. The chief health information officer refers to the forward-thinking ideals at his organization as a “constellation of collaboration.”

“I think there are a lot of people in our organization who really have that mind-set,” Poon says. “All of us share the common experience of being clinicians. We know that things fall through the cracks. We know the way we’ve done things may not be the way we should continue to do them.” It’s a mind-set that led Duke Health to achieve a Stage 7 Maturity Score from HIMSS Analytics Adoption Model for Analytics Maturity in 2018, becoming the first health system in the US to do so. The Stage 7 status signifies an organization’s ability to leverage predictive, personalized, and prescriptive analytics in its care environments, and while it may be a mark of distinction, Poon insists it’s merely a byproduct of organizational rigor at Duke.

“I think in my role, part of my job is to make sure that we cultivate great ideas from all corners of the organization,” Poon says. “I find it really helpful to bring the discussion back to the patients.” That philosophy underpins almost every decision his team makes, Poon says. Progress isn’t worthwhile if it doesn’t better serve the patient in some capacity.

That line of thinking also often helps settle divergent approaches to the same problem. “People can have differing opinions about what’s the best approach, vendor, or way to take care of patients,” Poon says. “But once you ask, ‘What if the patient is one of your loved ones?’ it really helps people step back and think about how they would design a system that best meets the needs of patients.”

Eric Poon, Duke University Photo by Erin Hull/Duke University Health System

In November 2018, Duke introduced Sepsis Watch, an AI algorithm that will help identify patients in the early stages of sepsis based on predictive indicators including vital signs, lab test results, and medical histories. “All of us as clinicians have had experiences where we know the patient isn’t quite looking right, but with so many things happening, it’s hard to pick out those faint signals from the noise,” Poon says. “In retrospect, we always wondered if we could have intervened earlier during the early stages of sepsis and improved the outcome.”

Poon says he and his team were proud to be part of the design process and are now awaiting results from the system’s first evaluation. “We’re not afraid to put something in, but we want to evaluate rigorously whether it makes an impact in patient care,” Poon says. “If it doesn’t, then we turn it off. We want to innovate but make sure we are doing it smartly.” All indicators are pointing to a success, however.

Another area where Poon says Duke Health is looking to push medicine forward comes in the design of learning health units: an evolving concept that is focused on redesigning the fundamentals of patient care. “We want to create real-life, living, breathing learning places within the clinical care unit, both in the hospital setting and the ambulatory setting where we can test out new ways of taking care of patients,” Poon says. “It would be a great way for us to apply ourselves to define new care paradigms and to help advance the state of science.”

Poon says developing the learning health units involves bringing together clinicians, data science professionals, tech leaders, and implementation experts to think creatively about what they can test out and share results of with the rest of the world.

Although some of the approaches may be new, the problems they’re trying to remedy often remain the same. “We can’t continue to spend more time and more resources that we don’t have as a society on healthcare,” Poon says. “At the same time there are so many different options right now. What combination of drugs, machines, social support, and outpatient education would best help patients achieve their best health outcomes?” Duke Health believes the health learning units may be instrumental in helping find new answers.

Poon says his current focus involves working to loop in more of the health system’s clinicians into understanding the potential for technology to better their patients. After spearheading a consultation service for the clinical faculty and trainees at the Duke University School of Medicine to leverage EHR and analytics technologies to advance their research, Poon says the demand has further highlighted the need for this service. “We got the sense that while the technology is mature and we have a lot of adoption, not everybody is taking advantage of those opportunities,” Poon says.

And while organization-wide adoption may be a work in progress, Poon says that Duke Health’s fundamental ideals are helping fill in those cracks that are so common to clinicians, Poon says. “This is an environment where we want folks to innovate and break some boundaries.”