While a new AI platform at Cleveland-based MetroHealth System was “not cheap,” the expected benefits include documentation burden relief and better patient outcomes, a leader told Becker’s.
The public safety-net health system recently adopted an AI system from Pieces Technologies that uses generative AI to summarize patient journeys for providers and predicts discharges and helps with discharge planning.
“We like the idea that a strong enterprisewide platform can create a rising tide that lifts all boats,” said Yasir Tarabichi, MD, chief health AI officer of MetroHealth. “So we’re moving toward solutions that can solve multiple pain points across the system.”
The company also has “strong guardrails” in place to monitor its AI for accuracy and bias, Dr. Tarabichi said. In addition, the health system is codeveloping solutions with the vendor, such as electronic phenotyping, and plans to use the platform to research social drivers of health.
“You can cut down on some of the operational inefficiencies,” Dr. Tarabichi said. “Physicians can spend more time with patients. Physicians can be better informed about patients under their care.”
The return on investment comes from increased efficiency on the inpatient side, or “moving patients faster through the machine that is inpatient care,” as Dr. Tarabichi put it. The AI might can also identify extra reimbursement opportunities during a patient’s stay. He expects the platform to “at the very least” cover its cost but also have “increased revenue potential.”
The software-as-a-service nature of the platform — rather than being an implementation-heavy tech solution — also allows MetroHealth to quickly change course if it doesn’t perform.
The health system’s previous AI successes include predictive tools that forecast sepsis risk and patient no-shows, as well as AI for prior authorizations and to read diagnostic studies. MetroHealth is also piloting ambient AI for clinical documentation.
“A big part of that work is showing how we can implement the solution in a way that really is fair and equitable and supports our patient population,” Dr. Tarabichi said. “When we deploy these solutions, we want to make sure we are helping our patients, not hurting them, so improving disparities, for instance, rather than widening them.”
He takes it as a responsibility to educate fellow senior leaders on the technology, weeding out AI solutions that won’t live up to their promises. “Everybody says AI and all of a sudden people assume magic is going to happen,” he said. “For those of us who live on the implementation side, we know that’s far from true. It takes a lot of energy to get something to the point where we know it’s going to work.”
