Since launching last March, Cleveland Clinic's Center for Artificial Intelligence has begun 20 AI-related research projects focused on improving disease prediction and diagnostics, according to a Feb. 27 Wall Street Journal report.
The AI center uses machine learning and other AI methods to examine more than 1.5 million patient records, pinpoint data patterns and improve the accuracy of medical predictions. The center's researchers are developing systems to predict patient readmissions and refine chemotherapy treatments, among other functions.
Developing machine learning algorithms for healthcare providers is challenging because the data is stored in EHRs, which are designed for billing and not necessarily for research or data collection, said Aziz Nazhar, MD, director of the AI center, according to the report.
"Sometimes you look at the text and as a human, you don't really recognize what it's actually trying to say," Dr. Nazha said. "That becomes a very difficult problem for the computer."
Cleveland Clinic aims to address these challenges by channeling expertise in both medicine and computer science. Medical professionals from several of the health system's departments, including pathology, radiology and IT, are involved with the AI center. The initiative is part of the Cleveland Clinic Enterprise Analytics division.
The AI center's machine learning projects are just beginning, so it is unclear when the algorithms will be ready to be widely deployed, Dr. Nazha said. He added that he's working to get more funding from Cleveland Clinic, as well as grants and philanthropic donors, to continue the research.