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

I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

inflecthealth.medium.com

Post reply on HN
Print view View original

261–270 of 368 posts

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#261

Post by ChuckNorris89 »

RandomLensman wrote: post #259
ChuckNorris89 wrote: post #254

Earlier quoted context omitted.

> The goal is not to see more patients Not when you're a doctor in the EU dealing with the shortages of socialized care where you have a limited amount of time to spend diagnosing a patient because there's dozens waiting in the queue. As a doctor you'd like to give everyone better care but you can't because you have too many people to see and too little time. Doctors here are busy AF and sometimes misdiagnose you bec…

The EU has a very heterogenous healthcare landscape

That's why I narrowed it down to the socialized part, not the private part as that has next to no shortages.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#262

Post by HelloMcFly »

SanderNL wrote: post #251
lumb63 wrote: post #219

The tail end of this article, where the author talks about how many more patients he could see in his life if he had AI assistance, made me realize that part of healthcare cannot be solved by AI. The goal is not to see more patients; the goal is to help more patients get better. For a lot of patients who are frustrated and have not had their problems validated and have been to many doctors and seen no results or poor…

"Seeing the patient" is an important step towards "patient getting better". > have not had their problems validated and have been to many doctors and seen no results or poor results, Also, being validated or not is not going to change with AI as you yourself state that seeing actual physical doctors is not helping in the cases you allude to. I agree that having access to professionals that care about their profession…

> Who thought it was a good idea to have a fixed amount of medical students?

A major limiter on the amount of trained medical doctors is not to do with someone "deciding" to have an artificially scarce number of medical students. Two important things come immediately to mind:

1) Medical schools rely on clerkships and preceptors to deliver actual clinical experience critical for learning. These cannot be created out of thin air and often rely on access to urban centers. There is already a shortage of clerkships and preceptors. https://www.aamc.org/news-insights/so-many-medical-students-... (not my original source but the nearest to hand one)

2) The real chokepoint for creating practicing clinicians is the number of residency spots. This year there were 43k graduating MD and DO students, but only about 40k residency spots. This doesn't count international medical school graduates. https://www.medicaleconomics.com/view/match-day-2023-a-remin.... Creating a new residency program is significantly more challenging than graduating more medical students.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#263

Post by analog31 »

funcDropShadow wrote: post #253
detrites wrote: post #240

Earlier quoted context omitted.

One thing the article misses is doctors are not all equal. Some are freaking terrible. Others may be competent but susceptible to bad decisions when overworked. Others are simply great. How does ChatGPT fare against the whole group? For one, it can't become "overworked", so at least better in that regard. It may well be better than a terrible doctor too. Given malpractice is consistently a leading cause of death, it…

That's a horrendous claim. > Given malpractice is consistently a leading cause of death Can you back it up?

The above post might have been thinking of medical error:

https://www.hopkinsmedicine.org/news/media/releases/study_su...

"Malpractice" might be harder to measure since it's a legal term.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#264

Post by aix1 »

jspdown wrote: post #232

Is it legal in the US to send patient data to a third party service? In the context of a scientific study, with explicit patient agreement things are different of course. But I haven't seen any of that in the article.

Doing this sort of thing would typically require approval from an ethics committee (called IRB = Institutional Review Board).

From my experience of going through IRB reviews, I would guess that an IRB review for what's described in the blog post would be focussed on the privacy of subjects whose data is to be entered into a non-HIPAA-compliant third-party system. My understanding is that privacy requirements can typically be met either by de-identifying the data to a certain standard, or obtaining patients' consent.

The following doc is about a different type of thing (case reports in medical journals) but gives a good idea of the required standard of de-id:

https://hipaa.yale.edu/sites/default/files/files/Case%20Repo...

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#265

Post by RandomLensman »

ChuckNorris89 wrote: post #261
RandomLensman wrote: post #259

Earlier quoted context omitted.

The EU has a very heterogenous healthcare landscape

That's why I narrowed it down to the socialized part, not the private part as that has next to no shortages.

Even the socialized part is very different in terms of load/structure/waiting times across the EU. Compare Sweden with Germany, for example.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#266

Post by shagymoe »

ChuckNorris89 wrote: post #254
lumb63 wrote: post #219

The tail end of this article, where the author talks about how many more patients he could see in his life if he had AI assistance, made me realize that part of healthcare cannot be solved by AI. The goal is not to see more patients; the goal is to help more patients get better. For a lot of patients who are frustrated and have not had their problems validated and have been to many doctors and seen no results or poor…

> The goal is not to see more patients Not when you're a doctor in the EU dealing with the shortages of socialized care where you have a limited amount of time to spend diagnosing a patient because there's dozens waiting in the queue. As a doctor you'd like to give everyone better care but you can't because you have too many people to see and too little time. Doctors here are busy AF and sometimes misdiagnose you bec…

Do you live in the EU and have experience with the healthcare system there?

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#267

Post by thelittleone »

lumb63 wrote: post #219

The tail end of this article, where the author talks about how many more patients he could see in his life if he had AI assistance, made me realize that part of healthcare cannot be solved by AI. The goal is not to see more patients; the goal is to help more patients get better. For a lot of patients who are frustrated and have not had their problems validated and have been to many doctors and seen no results or poor…

During my wife's second pregnancy she was admitted to hospital at 6 months due to some intense abdominal pain and inability to hold down food.

The put her on PNT, pumped her stomach daily and run a bunch of tests. I logged every test. Weeks passed she was slipping away and the specialist had no real game plan. So I got another Doctor. A busy professor who travelled a lot and was clearly passionate about his field.

Yet still dozens of different tests and scans and they couldn't find the cause. Meanwhile she was getting worse.

One of the scholar articles I found matched a test result where blood from upper body had a different profile to that in urine (which matched her result) which pointed to a condition called placenta parcreta. The Dr agreed it was possible and as my wife was slipping away we did an emergency c section. Sure enough the cause was placenta parcreta and the Dr showed a picture of a small tear in uterus that he said was about to rupture anytime. The operation saved her life and perhaps my son too.

I'm no Doctor but good at research. Doctors are busy AF. They don't have time to research cases like I did. This is a big gap and while it may not always identify cause and save a life, its damn important. And I can't see why AI could not find what I did.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#268

Post by aix1 »

amai wrote: post #241

This „study“ is missing a control group. He should have given his data also to some humans and see how they would do compared to ChatGPT.

There's also an implicit assumption that his ground-truth diagnoses are 100% correct.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#269

Post by giraffe_lady »

funcDropShadow wrote: post #253
detrites wrote: post #240

Earlier quoted context omitted.

One thing the article misses is doctors are not all equal. Some are freaking terrible. Others may be competent but susceptible to bad decisions when overworked. Others are simply great. How does ChatGPT fare against the whole group? For one, it can't become "overworked", so at least better in that regard. It may well be better than a terrible doctor too. Given malpractice is consistently a leading cause of death, it…

That's a horrendous claim. > Given malpractice is consistently a leading cause of death Can you back it up?

They probably meant medical error which kills like 200k people/y in the US. Malpractice is I think more like a legal/administrative judgement that indicates culpability that may or may not exist in a specific error.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#270

Post by n8henrie »

ChuckNorris89 wrote: post #254
lumb63 wrote: post #219

The tail end of this article, where the author talks about how many more patients he could see in his life if he had AI assistance, made me realize that part of healthcare cannot be solved by AI. The goal is not to see more patients; the goal is to help more patients get better. For a lot of patients who are frustrated and have not had their problems validated and have been to many doctors and seen no results or poor…

> The goal is not to see more patients Not when you're a doctor in the EU dealing with the shortages of socialized care where you have a limited amount of time to spend diagnosing a patient because there's dozens waiting in the queue. As a doctor you'd like to give everyone better care but you can't because you have too many people to see and too little time. Doctors here are busy AF and sometimes misdiagnose you bec…

Or a doctor in a US emergency department, where EMTALA essentially socializes access to emergency care without creating any provisions for funding said care.
Post reply on HN