Health IT Analytics has revealed an artificial intelligence algorithm capable of extracting electronic health record data and suggesting the best diagnostic methods, leading to improved diagnosis and treatment.
Although AI performs very well when trained in years of human data in specific areas, the technology has not been able to manage the vast number of diagnostic tests and disorders of modern clinical practice, a team from USC Viterbi at the University of Southern California worked On developing an artificial intelligence algorithm that can learn and think like a doctor, but with countless experience.
“The algorithm thinks about what to do after each stage of medical work,” said Gerald Loeb, a professor of biomedical engineering and neuroscience at the University of Southern California. “The difference is that it leverages all the experiments in the collective healthcare records.”
The algorithm developed by the university’s researchers reflects this process, as the new AI model seeks those tests that are most likely to identify the correct disease or condition, regardless of how ambiguous it is. The algorithm can also take into account the costs and delays associated with a set of diagnostic tests.
The team noted that the new algorithm could help service providers make better diagnostic and testing decisions by suggesting many good options, including some options that doctors might not think otherwise. The algorithm could also help service providers create records. Medically accurate and more easily completed, rather than searching for icons or working through traditional menus.