UPMC Spins Out AI Company to Abstract Clinical Notes
The University of Pittsburgh Medical Center (UPMC) has spun out a company that uses artificial intelligence (AI) and natural language processing (NLP), with the goal to identify precise patient populations with chronic diseases and cancer to ensure they receive the most beneficial treatments.
The company, Realyze Intelligence, has a clinical intelligence platform that officials say “reads” both the detailed clinical notes and structured data from patients’ electronic medical records (EMR). This deeper knowledge can be used by clinicians to identify precisely defined cohorts of patients who may be at higher risk of poor health outcomes, and those individuals can then be prioritized to receive appropriate care in a timely fashion, they say.
Providers use Realyze’s web-based tools to quickly assess the data that already are present in patients’ notes. The platform can be deployed within a provider’s workflow, aiming to make the solution effective for a variety of patient conditions, including various cancers, chronic kidney disease (CKD), diabetes and inflammatory bowel disease, officials explain.
“Patients aren't defined by their primary diagnosis. They are not just dealing with that one condition, but also many other factors that make them complex, and it currently is a manual, time-consuming effort to extract and use the relevant information from the EMR to ensure effective care,” said Gilan El Saadawi, M.D., Ph.D., Realyze chief medical officer. “Realyze streamlines this process and quickly decodes the patient’s ‘story’ so they get the best care.”
Officials note that the U.S. healthcare industry spends $8.7 billion per year employing people primarily to read clinical notes to abstract data for various clinical workflows. Aaron Brauser, Realyze president and CEO, points out that providers need a complete understanding of their patients and all of their comorbidities, and that this solution helps them find specific patients and intervene at the correct time with the correct treatment. “This can improve a patient’s overall health while hospitals benefit from avoidance of unplanned events and reduction of abstraction costs,” he says.
El Saadawi and Brauser founded Realyze in conjunction with UPMC Enterprises, the innovation and commercialization arm of UPMC.
UPMC leaders offer one example of an analysis of more than 100,000 CKD patients at UPMC, in which Realyze was able to enhance the work done by UPMC’s clinical analytics team by deriving new insights from clinical notes that were not available in the structured data. “With this information, UPMC will improve its ability to implement more precise patient segmentation, which, in turn, will allow clinicians to better deploy resources for more appropriate care,” officials assert.
Oscar C. Marroquin, M.D., UPMC’s chief healthcare data and analytics officer, adds, “As UPMC continues to pursue data-driven, high-quality health care, we always are looking for ways to get more information that can enhance the analytic-driven insights our clinicians need. Realyze’s ability to assess massive amounts of unstructured data has added new dimension and depth to the analyses we do, making them even more useful for our clinicians as they deliver life-changing medicine to the communities we serve.”
