Billing and Reimbursement Prime Targets for AI in Healthcare | IDWeek 2023


The chair of Harvard Medical School’s bioinformatics department says the deterministic, discrete data of billing and reimbursement means that artificial intelligence’s first big impact in healthcare is likely to be in “the business of healthcare.”

When artificial intelligence (AI) starts making a major difference in healthcare, it won’t be by hastening diagnoses or analyzing complicated datasets, says Isaac Kohane, M.D., Ph.D.

Instead, it will follow money and be used to manage billing and reimbursement, predicted the chair of the Department of Biomedical Informatics at Harvard Medical School and the featured speaker today at the opening plenary session today of the IDWeek 2023 meeting in Boston.

Isaac Kohane, M.D., Ph.D., said the future of artificial intelligence in healthcare is "muddy" but predicted that it would first make a major difference in the billing and reimbursement.

Isaac Kohane, M.D., Ph.D., said the future of artificial intelligence in healthcare is "muddy" but predicted that it would first make a major difference in the billing and reimbursement.

“All those healthcare professionals who are working for insurance companies to vet or not vet authorizations — they’re gone. And in the basements of hospitals all those upcoding teams, they’ll be gone soon as well because (AI) can do (their) job much, much better,” Kohane said.

“So the bottom line, for take-home message one, the first big impact (of AI) will be in the business of healthcare,” he added.

Kohane is the co-author of the 2023 book “The AI Revolution in Medicine: GPT-4 and Beyond” and the editor-in-chief of a new journal, NEJM AI, being published by the group that publishes the New England Journal of Medicine.

Kohane, who cracked several jokes, walked a fine line between talking up an AI-mediated future and issuing warning about shortcomings.

Even so, he strongly urged the audience of infectious disease clinicians, researchers and pharmacists to start exploring AI or risk falling behind.

“You’ve got to start using it otherwise you’re going to be really dumb compared to your patients and your students.”

He cautioned that the future of AI is “very muddy, it’s very unclear. And anybody who understands it well is, putting one over.” Kohane mentioned the absence of mechanisms of evaluating the current crop of AI models and clinicians ceding too much control to AI. He noted after order entry systems were put in place years ago, doctors starting using the default choice most of the time.

“How do we actually make sure that they — us — clinicians stay awake and not just let (AI) autonomously do things that it may not do well just because it can do so many things?”

In the same cautionary vein, Kohane also pointed to a program developed by Epic, the electronic health record company, that was billed as being able to predict the likelihood of COVID-19 patients needing intensive care unit care. But the program was inaccurate at other hospitals, and Kohane said Epic had to eventually stop using the program where it was implemented.

“Here’s the message — medicine changes all the time and our systems have to be updated for it.” said Kohane, adding that “we should not let EHR vendors implement things in our hospitals without peer review.”

But Kohane hit plenty of optimistic notes, sharing bright-spot anecdotes about AI helping with difficult diagnoses of rare diseases and reading echocardiograms. He envisioned AI-equipped nurse practitioners and physician assistants as a way to bridge the yawning gap between the dwindling supply of primary care physicians in the U.S. and the need for their services.

“The question becomes, can the combination of those professionals — who don’t happen to have medical school debt — with AI, can we bring them (to being) close or maybe even better than the average doctor?” Kohane asked.

Related Content
© 2024 MJH Life Sciences

All rights reserved.