Will AI help patients appeal insurance denials? – Healthcare Economist

It looks like the answer is ‘yes’. KFF reports that Holden Karau, a software engineer from the San Francisco Bay Area, created an AI tool to help appeal insurance denials. She calls it: Fight Health Insurance, 

How does it work?

You’ve gotta make a scan of the denial letter from your insurance company, and run it through “optical character recognition” — turn it from an image of text into actual text. Oh, and zap personally-identifiable information — like your name and address — from the document. So none of that gets captured by any machines.
You can also write up a narrative with any details — that’s optional, but seems like a good idea. And you can upload your documents from your insurance company that describe your benefits. That also seems smart. 
You feed it everything, and it gives you back a draft of an appeal letter — actually, more than one, so you can pick and choose, and make edits. 

How did she get the AI to actually write a good letter? The key was training the model on a large database of health insurance appeals.

Holden needed to train her model on a bunch of factual data for health insurance appeals and denials. And she found it. Thanks to the California Department of Insurance. If your insurance denies a claim, and it’s regulated by the state of California, you can request an independent medical review from the Department of Insurance. 
Which decides whether your procedure was medically necessary. Every decision gets published online. Describing the facts. The diagnosis. The procedure. And the reasoning behind the decision.

If health insurance companies start using AIs to review appeals letters, we may soon devolve into a battle of AIs for insurance coverage. The future is coming quicker than we may think!