Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ontario family doctor says new AI notetaking saved her job

globalnews.ca

Post reply on HN
Print view View original

31–40 of 274 posts

Re: Ontario family doctor says new AI notetaking saved her job

#31

Post by switch007 »

PlunderBunny wrote: post #22
nvy wrote: post #2

The root cause of this is that the billing system (which varies by Province) by which doctors actually get paid is a fee-for-service model. That means the system incentivizes doctors to see patients by only allowing them to get paid for actually seeing the patient: Give someone a vaccine? $15. Family doctor routine checkup? $34. The system is set up to reward the kind of doctor who sees 6 patients an hour and is on t…

My limited experience with dental practice payors (e.g. the NHS in England) is that it's an ogre's choice: Pay for over-treatment, or pay for neglect.

Damn that is accurate

Re: Ontario family doctor says new AI notetaking saved her job

#32

Post by elitan »

monocultured wrote: post #30

Just the other day there were news about how many mistakes the AI-assisted journaling caused – wrong names, wrong diagnoses, typos (buksmärta -> kuksmärta which is hilarious but serious). Some of it is surely teething problems, but unless there is a robust check upon implementation it might just add another layer of inefficient new public management make-work to the system. https://sverigesradio.se/artikel/ai-journal…

I wouldn't be surprised if Region Blekinge were using something much worse and much more expensive than Whisper for their transcription.

I've been transcribing A LOT of SR (Swedish Radio) shows as part of https://nyheter.sh/, and Whisper (self-hosted) has been very accurate.

Re: Ontario family doctor says new AI notetaking saved her job

#33

Post by cjk2 »

thih9 wrote: post #20
cjk2 wrote: post #14

Earlier quoted context omitted.

No one will be double checking. On at least three occasions now I’ve had to correct monumental fuck ups in mine and my ex wife’s medical records. One nearly lead to a transfusion of an incompatible blood type. I suspect this will lead to a decline in accountability with there being another party to blame rather than the medical professional. The LLM did it, not me.

Alluding to FAFO, a computer cannot find out, so a computer shouldn’t fuck around. I’m hoping for a lot of legal precedent showing that an AI cannot be blamed, especially in a medical context.

I would hope that would be the case but a conservative safety culture is unfortunately built on piles of dead people.

Re: Ontario family doctor says new AI notetaking saved her job

#34

Post by tomfreemax »

I think these tools are the solution to the shortage of doctors, and will greatly improve performance by taking more parameters into account. Not sure though if a simple LLM can do that. (Shortage of doctors is a thing at least in DE and CH)

Doctors study like 10 years to follow a decision tree based on patient background and symptoms.

Especially in countries where education is harder to get by this can be useful, I imagine.

Then we transition from doctor to med-tech, like in the sci-fi movies.

Re: Ontario family doctor says new AI notetaking saved her job

#36

Post by wernsey »

r0ckarong wrote: post #5

I just hope someone is double checking whatever notes are taken. My experience with LLM tells me they're fantastically bad at remembering and even worse at producing factual stuff that can be used without context.

Agreed, but now the doctor has to do her job and the AI's job.

Cory Doctorow wrote about it a while back. I think it was this article "Humans are not perfectly vigilant" [0]. It explains how technology is supposed to help humans be better at their jobs, but instead we're heading in a direction where AIs are doing the work but humans have to stand beside them to double check them.

[0] https://pluralistic.net/2024/04/01/human-in-the-loop/

Re: Ontario family doctor says new AI notetaking saved her job

#37

Post by mial »

Scoundreller wrote: post #6

> The Ford government has been so impressed with the technology that it announced a pilot program to allow 150 family physicians to use AI Scribe as part of their practices. The health minister said the early signs were promising but stressed the government would proceed carefully. What I find fascinating is that doctors are independent contractors but the government appears to be buying their tools to make them more…

When the state subsidizes healthcare, it is in its interest to lower the costs/improve productivity.

Re: Ontario family doctor says new AI notetaking saved her job

#38

Post by popinman322 »

elitan wrote: post #32
monocultured wrote: post #30

Just the other day there were news about how many mistakes the AI-assisted journaling caused – wrong names, wrong diagnoses, typos (buksmärta -> kuksmärta which is hilarious but serious). Some of it is surely teething problems, but unless there is a robust check upon implementation it might just add another layer of inefficient new public management make-work to the system. https://sverigesradio.se/artikel/ai-journal…

I wouldn't be surprised if Region Blekinge were using something much worse and much more expensive than Whisper for their transcription. I've been transcribing A LOT of SR (Swedish Radio) shows as part of https://nyheter.sh/ , and Whisper (self-hosted) has been very accurate.

Tangent here: really? I've found base Whisper has concerning error rates for non-US English accents; I imagine the same is true for other languages with a large regional mode to the source dataset.

Whisper + an LLM can recover some of the gaps by filling in contextually plausible bits, but then it's not a transcript and may contain hallucinations.

There are alternatives that share Whisper internal states with an LLM to improve ASR, as well as approaches that sample N-best hypotheses from Whisper and fine-tune an LLM to distill the hypotheses into a single output. Haven't looked too much into these yet given how expensive each component is to run independently.

Re: Ontario family doctor says new AI notetaking saved her job

#39

Post by ThePowerOfFuet »

monocultured wrote: post #30

Just the other day there were news about how many mistakes the AI-assisted journaling caused – wrong names, wrong diagnoses, typos (buksmärta -> kuksmärta which is hilarious but serious). Some of it is surely teething problems, but unless there is a robust check upon implementation it might just add another layer of inefficient new public management make-work to the system. https://sverigesradio.se/artikel/ai-journal…

> typos (buksmärta -> kuksmärta which is hilarious but serious)

To save people looking it up, that one-char difference changes "abdominal pain" into "cock pain".

Wow.

Re: Ontario family doctor says new AI notetaking saved her job

#40

Post by freilanzer »

notachatbot1234 wrote: post #24
h_tbob wrote: post #7

I talked to a therapist who hated doing notes. So I proposed a solution: Use Microsoft's "seeing AI" to describe the children playing during "play therapy". Then have a camera take a picture every few minutes, and then have chatgpt come up with the story based on the pictures and transcript of the audio, transcribed with openai whisper. Business in a nutshell!

Ah, the freedom of having no consideration for privacy or ethics. No one should record children and feed the imagery to a ominous third-party.

A worrying trend indeed.
Post reply on HN