The New York Times ran a piece this week on hospital AI versus insurer AI.
Most of the attention will go to coding. The part I keep coming back to is a phrase from Abridge founder Shiv Rao: "bots fighting bots."
Emory Healthcare's CEO voiced the same fear from the provider side: automated systems on each side of a claim, talking to each other with no end point.
One detail deserves more attention. Insurers report a surge of appeals showing signs of language model authorship. Providers likely gave up on too many denials in the past. Now appeal volume is rising faster than either side's ability to review it.
When both sides automate, the question stops being whether the machine can produce the work. It becomes who stands behind it.
An appeal is a claim that a patient needed care. A denial is a claim that they didn't. Each should have a person accountable for it. Payers say they never deny care without oversight from a medical professional. That is authorization. Whether anyone verified what the machine produced is a separate question, and it applies to both sides.
Will payers start weighting appeals by who stands behind them?
Automation lowers the cost of each round. It does not lower the cost of being wrong.
Paywalled Article: https://lnkd.in/e2G4Wrvx
Xillium keeps accountable human review inside automated healthcare workflows. What that looks like in practice is on our Solutions page.
