Having worked at companies that built software for health insurers, I have seen the "evil" you describe. From "hey, can we mine the claims database for suspected/confirmed familial relationships and look at possible diagnoses to assign risk profiles?" No, you can't. "Why not? It's in the database." Because it's federally illegal. "Oh. So you won't expose that data?" We won't.

In this case, two things:

The system decides on the initial denial at most insurers. And when a claims adjuster reviews, the system is presumed to be accurate, and the adjuster has to provide reasoning to overturn the system's denial (this is before the denial has been returned to the provider). It's not "assume the provider was correct", but "we've decided to deny it, give the system reasons why we shouldn't". And that person reviewing it is often an LPN (no shade thrown at LPNs, but they shouldn't be overriding physician decisions, doubly so given an absent history).

How this has affected me personally: I had, for most of my life, a severely deviated septum. I spent most of my life mouth breathing because I could barely pull enough air through my nostrils to make breathing that way not an active effort. I finally went to an ENT who confirmed, sure enough, an approximately ninety per cent deviation. "Great, so lets schedule surgery". ENT: "Slow down. First I have to prescribe you these two nasal sprays so that when you come back in four weeks and report no change, because to both our disappointment, the sprays didn't realign and open up the cartilage in your nose, then I can submit the pre-auth to your insurer and they won't immediately reject it." What a fucking joke.

> "we detected your doctor is wrong"

It's not even that your doctor is wrong, it's "our nurses/expert systems disagree with your doctor so we're not paying".