Of course, the real problem is running medicine like a business, but technology can only serve the megamachine...
Having been exposed to this only a little, I can imagine what kind of garbage the medical staff has to endure to make their reports and time keeping.
It's insane that a truck driver _must_ have 8 hours of sleep before work (they should, not arguing against it), but a surgeon can work with a couple of naps for days.
Also, this is one place where AI can potentially provide both a more efficient system and greatly improve job satisfaction for physicians. Automatically generating and updating medical records based off ambient recordings of patient interactions fits right into LLMs wheelhouse.
> Also, this is one place where AI can potentially provide both a more efficient system and greatly improve job satisfaction for physicians. Automatically generating and updating medical records based off ambient recordings of patient interactions fits right into LLMs wheelhouse.
No, LLM's are not reliable enough for that.
If the technology sucks and is in fact a regression, the solution is not to hope yet more half-backed technology will fix the problem. The solution is to dump the technology until someone can actually do it right.
If paper charts are better, go back to paper charts...and maybe hire a digitization/data entry team to realize whatever benefits to electronic records there are.
It requires physicians to read and verify the LLM output before signing it. And not just rubber stamping.
There is a precedent for this with speech recognition, which a lot of doctors use to dictate their documentation. Speech recognition isn't completely reliable either, so the doctor needs to be diligent to review before signing. And not just assume it's all correct.
Then what's the point? That's not going to provide a "more efficient system and greatly improve job satisfaction for physicians."
AI would only improve the situation if it can actually be trusted. You're actually making things far worse if you replace a "production" task with a "continuous monitoring for errors. Humans are terrible at the latter.
> There is a precedent for this with speech recognition, which a lot of doctors use to dictate their documentation. Speech recognition isn't completely reliable either, so the doctor needs to be diligent to review before signing. And not just assume it's all correct.
The one transcript like that I've seen had obvious speech recognition errors, when the notes got to my primary care doctor.