A boy saw 17 doctors over 3 years for pain. ChatGPT found the right diagnosis

https://www.today.com/health/mom-chatgpt-diagnosis-pain-rcna101843

by birriel • 3 years ago

109 147 3 years ago

today.com

elromulous• 3 years ago

https://archive.ph/TuI0Y

joe_guy• 3 years ago

It was Tethered Cord Syndrome.

https://en.m.wikipedia.org/wiki/Tethered_spinal_cord_syndrom...

rowanG0775• 3 years ago

Let's see how the AI haters here are going to spin this one...

lcnPylGDnU4H9OF2• 3 years ago

It is a worthwhile practice to open one’s mind to disagreeable views to find where their disagreement lies. If you were to spin this against ChatGPT, would you really have nothing to say?

rowanG0775• 3 years ago

Yes it's very important to have an open mind. I'm very curious how the AI haters will spin chatGPT diagnosing a kid while 17 doctors could not in anything but a win for AI.

rafaelero6• 3 years ago

What explains people's knee jerk reaction to something as awesome as this? It's weird reading comments here.

rowanG0775• 3 years ago

Because everytime a positive AI article is posted people come out of the woodworks trying to paint the current state of AI in the most negative light possible.

fungiblecog• 3 years ago

One anecdote is not data. You need controlled trials that include both correct and incorrect diagnoses not just a headline

corethree25• 3 years ago

Wrong. One anecdote is one data point. If one person witnesses a murder that is a data point. The court doesn't just dismiss the data point because they need a sample size of 1000 witnesses for any statement to be meaningful. This article here says something meaningful about AI and it also says something damning about doctors.

Also there have been studies: https://www.sciencedirect.com/science/article/abs/pii/S01960...

smlavine• 3 years ago

What an irresponsible article.

“It’s a super high-powered medical search engine,” Beam says.

NO IT IS NOT. THAT IS NOT WHAT IT IS AT ALL. How dare they publish this for a general audience.

I'm sure Today will also publish all the stories of mothers who kill their kids because they use ChatGPT like those infamous lawyers did a few months ago.

Shameful.

mjr003• 3 years ago

Not sure what's ChatGPT-specific here; patients discovering their own diseases has happened a lot since the internet has gone mainstream. Google has been the most prolific doctor of the 21st century. The only comparison given is

But both seem to work better than the average symptom checker or Google as a diagnostic tool. “It’s a super high-powered medical search engine,” Beam says.

Which is, arguably, more of an indictment of how bad Google's search capabilities are in 2023 than anything else.

achow• 3 years ago

Google is like going to an individual doctor which this mother was doing already.

ChatGPT synthesized its result from multiple sources and communicated that - no source existed which understood the connections between various symptoms. This is exactly the thing that the mother was quoted at one point - every doctor was looking at things from their specialization and not holistically.

mjr003• 3 years ago

ChatGPT synthesized its result from multiple sources and communicated that - no source existed which understood the connections between various symptoms.

Is this true? Or is the source that had all of the symptoms buried under WebMD/MayoClinic SEO spam?

warent2• 3 years ago

Bizarre take.

  Not sure what's ChatGPT-specific here

They literally used ChatGPT, specifically.

You may as well hear the story of William Kamkwamba building a wind turbine for his village and say "Not sure what's African-village specific here. People have been building wind turbines all over the world with various resources."

corethree25• 3 years ago

It's not bizarre at all. Tons of people are in denial of the capabilities of chatgpt. The current strategy is to position it as nothing more than an advanced search engine.

Clearly chatGPT has flaws. Clearly it's more than a search engine. And clearly there's a large contingent of people who don't want to think it's anything more than a search engine. Probably was Bizarre when the hype first started but now these people in denial are a dime a dozen, as cliche as the people they want to bring down.

mjr003• 3 years ago

A more apt comparison would be seeing one of the many "we rewrote our Python app in Go and improved performance by 100x" articles, and pointing out that most of the improvement comes from the rewrite and not the language.

Similarly, the real moral of the article is that doctors miss things and doing independent research, whether that's ChatGPT, Google, Bing, Yandex, or reading research papers yourself is sometimes necessary.

nojs• 3 years ago

Assuming they meant “couldn’t you already do this with Google” I think the big difference is the context length. You can describe way more symptoms, and include a lot more data, than you ever could in a single search query. Even if Google worked properly, in its current form it’s just not capable of parsing pages of info and drawing conclusions beyond keywords.

downWidOutaFite• 3 years ago

Health info on google was captured by the likes of webmd and healthline 20 years ago and they've just been farming their seo position ever since, absolutely no innovation, just regurgitating medical literature 101.

But it really feels like Google's entire ecosystem is collapsing rapidly as seo bots cannibalize it. Ad-supported content is a local optima that LLMs will thankfully destroy.

ceejayoz4• 3 years ago

... patients discovering their own diseases has happened a lot since the internet has gone mainstream...

With, of course, a corresponding increse in patients thinking they've discovered their own diseases.

For every "ChatGPT diagnosed my rare illness" story there's a "the Internet helped me find a bunch of people who are certain I've got Morgellons and chronic lyme".

JKCalhoun• 3 years ago

Whatever happened to "expert systems" from the 80's, 90's? They were supposedly going to be able to solve this exact problem — medical diagnosis.

Perhaps this was the camp of AI that thought you could just build a decision tree (top down) ... that appears to have failed?

Instead the bottom-up (neural networked) AI seems to be winning out.

I wonder though if a ChatGPT would have been possible in 1985 with no rich internet to feed it data. I wonder if it would have been possible with even 2000 internet?

jeffwiederkehr• 3 years ago

I don't think anything close to the compute needed to train chatgpt or any of the other LLM's would be possible in 1985 or 2000

OJFord• 3 years ago

Decision tree is basically how doctors work, at least how they're trained to work. I think for a machine to be significantly better, for it to be impressive, it stands to reason it would work differently.

koolba2• 3 years ago

Not really. A programmatic decision tree will actually try out all the steps rather than glancing over the ones it thinks are not relevant. It also would not forget to test branches.

If all a doctor is doing is following a flow chart, you can be damn sure that a machine programmed with the same flow chart would do it more effectively.

kayodelycaon3• 3 years ago

I’ve written several 1980s-style expert systems.

In practice they are basically simplified programming languages. If this, then that. The application of rules can lead to surprising emergent behavior. You might be able to find a pattern in it but you’re more likely to get wrong results and need to update the rule set.

There is also the problem of inputs. You end up needing to be an “expert” on the system to get good results.

tyingq• 3 years ago

The ones I've worked with weren't solely just if->then->else trees. There were also things like weightings, probabilities, risks, etc.

arcticbull• 3 years ago

Instead the bottom-up (neural networked) AI seems to be winning out.

What do you mean by ... winning out? It seems to have made one novel diagnosis, which is great for the individual in question. But how many times has it diagnosed something completely absurd? What's the ratio of signal to noise here?

If it's anything like trying to use it to solve difficult engineering problems it must have hallucinated turbocancer hundreds of times over.

famouswaffles• 3 years ago

What do you mean by ... winning out?

Symbolic AI has been less performant than NN AI for decades now and by a large margin. Language, speech, images etc.

It seems to have made one novel diagnosis

Ok so let's get this out the way. GPT-4 is good at diagnosis

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10425828/

Six patients 65 years or older (2 women and 4 men) were included in the analysis. The accuracy of the primary diagnoses made by GPT-4, clinicians, and Isabel DDx Companion was 4 of 6 patients (66.7%), 2 of 6 patients (33.3%), and 0 patients, respectively. If including differential diagnoses, the accuracy was 5 of 6 (83.3%) for GPT-4, 3 of 6 (50.0%) for clinicians, and 2 of 6 (33.3%) for Isabel DDx Companion

janalsncm• 3 years ago

When used in conjunction with an actual expert, ChatGPT’s statistical search can be very powerful. It’s essentially performing a statistical search over the space of all diseases it knows about and returning (at a first order approximation) the most reasonable diagnosis. What a doctor should do is to perform tests to rule the suggestions out.

The point isn’t to replace doctors but to empower them. They still need to come up with tests to rule out conditions.

We also have an algorithm for the tests: information gain. If you think of diagnosis like a game of Guess Who, we need to come up with tests which can narrow down the space of possible conditions. (It’s a bit more complicated of course since tests have a margin of error too.)

potatoman22• 3 years ago

The human body is incredibly complex so most expert-derived models are less accurate than purely statistical methods. There are ways to marry the two, like Bayesian networks, but they take much longer to develop and aren't typically more accurate.

The thing is, a model doesn't have to be accurate to be useful. Expert-derived systems tend to operate closer to how doctors understand the problem, making them more interpretable.

GuB-42• 3 years ago

Expert systems are not dead.

We use them at work for the diagnosis of mechanical parts rather than human bodies. We don't call them expert systems but they are exactly that. There is a database of potential failures, repairs and verifications, and it will suggest a troubleshooting procedure taking cost into account. It means that it may try unlikely causes first if they are easy to check. It also has simple learning capabilities.

The problem with expert systems is that it needs a curated and complete database. It is possible to do with mechanical parts as they are well specified and failure modes are well documented. The top-down approach comes naturally because that's how components in a mechanical system are designed. By comparison, human bodies are a complex mess.

As for ChatGPT, it is a pure product of technical advances. Neural networks are an old idea, they knew about the theory in the 60s, and we had implementations in the 80s. The problem is that with little data and limited computing power, they were little more than toys. They have been made practical in the last few years only because of the "rich internet" on one side and lots of powerful GPUs on the other. And the way I understand it, recent advances in deep learning technology is not really fundamental research, more like trying a bunch of stuff and see what sticks, again made possible by lots of data and computing power.

kayodelycaon3• 3 years ago

As much as I want to bash ChatGPT, this is actually a good use for it.

It’s been possible to google this stuff for a long time but you need to know how to use a search engine. And the results have slowly been getting worse over time.

ChatGPT adds another dimension to search. It’s not an oracle of truth but it can point you to things you would have thought of.

keyle3• 3 years ago

LLM's are amazing. Truly amazing stuff. Let's not screw this up.

mjhay• 3 years ago

Articles breathlessly getting excited about anecdotal cases are a good way of doing just that.

(It's good this article is on HN, but I don't like this reporting)

keyle3• 3 years ago

I agree, it reads like the doctors didn't do their job properly.

Wohlf• 3 years ago

Says more about doctors really, even dealing with something as banal as back pain it took years and finally visiting a physical therapist to get the issue properly identified and resolved.

vatican_banker• 3 years ago

According to the article, it is difficult to diagnose babies/kids/toddlers because "[...] they can’t speak”.

The article also mentions that the kid didn't have the symptoms or physical manifestations of spina bifida, complicating the diagnosis.

This case reads as a genuinely difficult case to diagnose.

rafaelero6• 3 years ago

This is amazing and we are only getting started.

andrewstuart• 3 years ago

Somehow we need to get to the point that patients expect doctors to use AI to assist their diagnosis.

I think at the moment using AI would lead to the perception that the doctor is incompetent.

But I would always rather a doctor who consulted AI and then overlaid their personal experience and expertise to rule in and out what the AI suggests.

The idea that any/every doctor knows everything, or even could know everything is so false and wrong and out of date.

There is probably a business opportunity in this somehow, for "AI first doctors" - their target customers being people who want the best human advice PLUS the best AI advice.

corethree25• 3 years ago

The goal should not be to expect doctors to use AI to assist.

The goal is to get AI to be superior to doctors such that AI replaces doctors.

AI should augment your skill such that the doctor is no longer needed. The problem with the US medical system is that they hold your life hostage then rip you off. Make no mistake people attribute this to complexity and other bullshit but the money ends up in the pockets of administrators and doctors.

You want to fix the system? Attack the root. Replace doctors.

radus• 3 years ago

  sed 's/doctor/programmer/g'

KerrAvon2• 3 years ago

It’ll have to be better than today’s spicy autocomplete technology to do that, though. ChatGPT will never be in the driver’s seat; it can’t be trusted.

flir2• 3 years ago

Why would you expect "we have a better solution but we'll charge less for it" to be the outcome in a free market? That's just leaving money on the table.

topspin3• 3 years ago

It's going to be interesting to see this play out with the AMA et al. The raging cynic in me thinks this is going to get very serious, very quickly. There are a lot of lifestyles at risk. On the other hand it will soon be possible to supply an endless parade of mothers offering politically unimpeachable testimony and demanding freedom to employ these tools.

I know who wins that, but there will be many pounds of flesh rent on the field in the meantime.

colordrops• 3 years ago

Yeah I feel the same. As soon as someone comes out with a medical AI service that gets traction, the AMA is going to go nuclear and get legislation to ban medical diagnosis AI.

incongruity• 3 years ago

I’d be amazed if one or more of the large EMR vendors wasn’t already developing some sort of diagnostic co-pilot. They along with insurers have unique access to critical data for such applications and both have motivations… having worked for a major insurer, however, I don’t believe they have the technical competence or willingness to see longer-term value from true improvements in diagnostics vs. short-run thinking. (But I really hope I’m wrong there).

d_burfoot• 3 years ago

Somehow we need to get to the point that patients expect doctors

Yes, and I think malpractice suits can work well here. If a doctor misdiagnoses a patient, leading to an injury, s/he needs to be able to document that an AI system was consulted and the AI-generated diagnosis was checked or considered. Maybe there's a good reason to discard the AI's suggestion, but you better at least have asked the AI, or you're going to slapped with a big malpractice penalty.

readthenotes1• 3 years ago

The irony is that the lawyers profiting are doing so because they have not been replaced by AI

drewcoo4• 3 years ago

There should, I would hope, be more to social considerations than litigation.

JumpCrisscross2• 3 years ago

at the moment using AI would lead to the perception that the doctor is incompetent

There are numerous examples, in popular culture, of geniuses working with AI assistants. This shouldn't be a difficult bridge to sell. Particularly if the patient can witness the whole conversation, possibly even feel involved in it.

drewcoo4• 3 years ago

Honestly, I think that's the opposite direction for human consideration.

I think patients would like doctors or nurses or nurse practitioners or whatever their insurance pays for . . . to pay attention to them and actually listen and care. Maybe AIs "do it better" in that case.

If computer systems care more about us than other human beings, the problem is not the computer systems.

davidguetta• 3 years ago

I think at the moment using AI would lead to the perception that the doctor is incompetent.

Then it's the patients that are stupid. i've seen my GP check some very specific vaccines / test procedures on the main gov body website and I was like "hey at least he's not making things up".

General medicine is not tailored for specific cases so AI is more than filling a need. It may be doing the job in the future.

kenjackson• 3 years ago

I had a similar experience where ICU doctors and nurses missed the diagnosis, but I put the info into ChatGPT and it found the right diagnosis. I just wish I had tried it 12 hours earlier -- it would have saved a lot of suffering.

theptip• 3 years ago

This is an ideal usecase for ChatGPT; synthesize a huge body of knowledge, and come up with some suggestions in an environment where a wrong answer doesn’t really hurt (since it’s either an answer she heard already, or a doctor shoots it down).

I think we can safely bet that doctors will be using these systems to generate suggestions (human in the loop) very soon, but making the diagnosis reliable enough that you’d take the human out of the loop could be extremely challenging.

behnamoh• 3 years ago

How about the other cases that ChatGPT misdiagnosed? We don't see them on media because, you know...

corethree25• 3 years ago

We don't see it because chatgpt is garbage in terms of reliability.

This article is on the media because Doctors claim they're good and effective. But when a unreliable product beats doctors it turns their claim on it's head.

This article isn't saying anything about chatGPT we don't already know. It's saying something damning about doctors.

rafaelero6• 3 years ago

According to this study, they are as reliable as doctors (for ED's context)[1]. Also, Google has developed Med-PaLM 2, which they also say matches human doctors. So yeah, they may be unreliable. But they are as unreliable as humans.

[1]https://www.sciencedirect.com/science/article/pii/S019606442...

iopq• 3 years ago

She went to 17 doctors that misdiagnosed the cause, so my question is: how about the doctors that can't find the right cause of the pain? You seem to be giving them a free pass

rvz9• 3 years ago

Exactly. We know that the AI bros won't tell us that, since it spoils their narrative.

For each guess in its diagnosis that is correct, there are around 1,000 incorrect suggestions that it makes, hence why it is untrustworthy in its diagnosis and for every answer it gives it lacks transparent reasoning, making it close to dangerous to rely on for medical advice.

JumpCrisscross2• 3 years ago

How about the other cases that ChatGPT misdiagnosed?

She didn't consult ChatGPT and then start feeding her kid Tide pods. She took the diagnosis to a doctor. This is fine. This is productive. One could argue it wastes doctors' time, but in this case, it clearly didn't; our traditional random-search system did.

renewiltord• 3 years ago

A friend got the radiologist report for a musculoskeletal issue. Pasted into early GPT-4 and got a list of things to do, prognosis etc. I ran it by my dad (orthopaedic surgeon) and my friend asked his doc what to do.

Matched almost exactly. My dad said it was very impressive. Great software.

ecommerceguy• 3 years ago

I wonder how https://www.drgupta.ai/ would have performed...

matt3210• 3 years ago

ChatGPT 4 won’t answer medical questions for me.

highstep• 3 years ago

Ask it to do a role play as a doctor, or maybe just rephrase your question so it's less of a request for advice, but instead a discussion of scientific data/facts. It should be easy to get around thus guardrail.

hn_throwaway_99• 3 years ago

I've used ChatGPT for medical diagnoses support, and I thought it was fantastic.

I was most impressed that I could simply cut and paste lab results from a PDF, with its wonky formatting, and it still interpreted everything correctly.

manicennui• 3 years ago

And how much bad medical information has it given?

hervature• 3 years ago

She scheduled an appointment with a new neurosurgeon and told her she suspected Alex had tethered cord syndrome. The doctor looked at his MRI images and knew exactly what was wrong with Alex

Doesn't this kind of summarize the article? She went to 17 doctors who did not know about the illness. ChatGPT pointed her to a competent individual who immediately diagnosed it. My understanding is that ChatGPT's diagnosis is wrong because tethered cord syndrome typically occurs after a surgery to correct a spinal bifida which is what the neurosurgeon diagnosed.

jononomo• 3 years ago

One comment in the story struck me -- doctors have become so specialized now that no single doctor is able to see the big picture anymore.

RcouF1uZ4gsC• 3 years ago

“I went line by line of everything that was in his (MRI notes) and plugged it into ChatGPT,” she says. “I put the note in there about ... how he wouldn’t sit crisscross applesauce. To me, that was a huge trigger (that) a structural thing could be wrong.”
She eventually found tethered cord syndrome and joined a Facebook group for families of children with it. Their stories sounded like Alex's. She scheduled an appointment with a new neurosurgeon and told her she suspected Alex had tethered cord syndrome. The doctor looked at his MRI images and knew exactly what was wrong with Alex.

Notice how it gets very vague about the timing and specificity. It seems not that ChatGPT diagnosed tethered cord as tethered cord but that it happened to come up in one of the many chat sessions that led to the mother doing a deeper investigation of that.

RScholar• 3 years ago

Maybe I'm just getting old, but my reaction to this article had nothing to do with LLMs but was instead that this is an indictment of the state of medicine in America. At each diagnostic milestone it seemed obvious that the conclusion didn't adequately explain the totality of the boy's condition and the search needed to continue for the real culprit. Instead it seems like each one of these diagnosticians felt they'd found the "real problem" and pushed ahead with treatment as such.

Now I'm not saying that back in the day all doctors were Gregory House or holders of forgotten knowledge, but I have to think that the lower pressure to make medicine behave like a business and deeper ties one would form with their patients would've benefitted this family. For this to have gone on for three years without getting escalated to the attention of one of the real-life Dr. Houses out there is appalling. I feel like ChatGPT got the answer right because it allowed her to input all of the minor observations from the charts; perhaps had the doctors done likewise they'd have connected the dots as well.

When I was growing up and in my early 20s you went to the doctor and they got you talking about your life, teasing out the observations you had on your own life and addressed things from there. Then about 15 years ago everything changed and doctor visits became this twisted form of speed dating where you never see the same person twice and any question or answer that takes longer than ten seconds to express is taboo. I hate it, and I can't believe that we keep moving along like it's acceptable.

corethree25• 3 years ago

It stems from low supply of doctors. There aren't enough doctors because if there were more doctors they'd make less money.

salawat2• 3 years ago

More like hospitals refuse to self-fund residency slots. It's pretty bad when the number of doctors entering practice is fundamentally gated by residency positions funded by Congress.

moribvndvs2• 3 years ago

I’m more inclined to go with the fact that medical school is extremely expensive and the process is long, laborious, and difficult but sure, ok.

anon-3988• 3 years ago

Not sure how increasing the supply of doctors would help in this case? If doctors begin to earn way less money than they do now then that would create an incentive to find ways to make money, whereby increasing the amount of deceptive practices of pushing certain medicines/medical procedures to prop up sales and not cure diseases.

I personally think money have no business (pun intended) in medicine. I would go as far as say that if a trillionaire parent have a sick child that have a disease that requires 1 trillion dollar to cure, the procedure should not go through and that is immoral. Which is to say that at a governmental/societal level, yes, money should absolutely be considered and allocations should be discussed. But when it comes to each individuals, money shouldn't be considered at all.

wyldfire• 3 years ago

Instead it seems like each one of these diagnosticians felt they'd found the "real problem" and pushed ahead with treatment as such.

More likely that they heard hoofbeats and were content to use the effectiveness of their treatment plan as another diagnostic signal.

When I was growing up and in my early 20s you went to the doctor and they got you talking about your life, teasing out the observations you had on your own life and addressed things from there

The history and physical are a critical part of every exam. But much less so at an urgent care where they intentionally disavow long term health to focus explicitly on acute issues.

drewcoo4• 3 years ago

Go back further when there were family doctors who knew generations of their patients' families and that's my memory as a kid. When I was in grade school, Doc Carpenter mentioned treating family members I'd never met because they were 3 generations older!

It was shocking to me that I heard about family mythos from my kindly old doctor. Luckily it was followed up with a lollipop to soothe me.

Ekaros• 3 years ago

Actually how things usually worked out in House is likely realistic. I'm not talking about the simple cases, but main cases in the show. Which took multiple false starts and wrong treatments to get to final conclusion. In non-critical cases patient is either dead or it will take months or years.

ilyacooks• 3 years ago

Unfortunately it’s the same in other healthcare systems - in public UK healthcare you have 10 mins appointments with doctors from history to diagnosis to treatment plan

Xyem• 3 years ago

Not just the US. Here in the UK, I recently went to a follow-up appointment regarding a non-healing ulcer and the NHS doctor took the dressing off and said "It is healed" despite the fact that:

1) the ulcer is deep enough that you can see the muscle below the skin fat layer (she said it no longer needed dressing!)

2) it has looked that way with nearly no change at all for nearly 3 months[1] (expected heal time is roughly 7 weeks)

3) there is a large area around the wound that is still red and inflamed (ulcers are not even regarded as healing until the inflammation is gone)

I've had to go for private care and I am currently scheduled in for minor surgery to see if a foreign object caused the ulcer and is still present, preventing it from healing. Probably no surprise to you, dear reader, that I am pursuing a case of medical negligence against the doctor.

[1] When I pointed this out to her, she said she "didn't have a crystal ball, so [she doesn't] know what will happen in the next week or month". Literally nonsense due to being a non-sequitur. I was informing about the past, not asking about the future.

kumarvvr• 3 years ago

I wonder how many people used ChatGPT to find out their disease, got absurd results for their symptoms, and just shrugged it.

topspin3• 3 years ago

Does your dichotomy matter? When the system that curates the supposedly existing sources is impenetrable they're as good as fiction.

If it does matter then what might you propose? Hint: to be viable it must involve some professional class interest making bank.

sargenem• 3 years ago

It will be a shame if AI doesn’t become more mainstream in healthcare. AI and AR have the potential to disrupt the medical field.

I am sure we are not getting rid of actual doctors anytime soon, but having AI as a “co-pilot” would be immensely beneficial.

Gunnerhead• 3 years ago

Anybody have a good starting prompt to prime ChatGPT to answer medical questions or diagnoses?

Engineering-MD• 3 years ago

Without knowing the MRI report it’s hard to interpret this. If the radiologist had missed the spinal bifida then all the other doctors would rely on this report. Then it’s all a single point of failure: a radiologists report. May be other things, but I suspect the report is the root cause of the issues here.

Fire-Dragon-DoL• 3 years ago

This is so terrifying and true.

A boy saw 17 doctors over 3 years for pain. ChatGPT found the right diagnosis

Related Stories