Posted on 11/18/2024 4:21:06 AM PST by dennisw
A small study found ChatGPT outdid human physicians when assessing medical case histories, even when those doctors were using a chatbot.
Dr. Adam Rodman, an expert in internal medicine at Beth Israel Deaconess Medical Center in Boston, confidently expected that chatbots built to use artificial intelligence would help doctors diagnose illnesses.
He was wrong.
Instead, in a study Dr. Rodman helped design, doctors who were given ChatGPT-4 along with conventional resources did only slightly better than doctors who did not have access to the bot. And, to the researchers’ surprise, ChatGPT alone outperformed the doctors.
“I was shocked,” Dr. Rodman said.
The chatbot, from the company OpenAI, scored an average of 90 percent when diagnosing a medical condition from a case report and explaining its reasoning. Doctors randomly assigned to use the chatbot got an average score of 76 percent. Those randomly assigned not to use it had an average score of 74 percent.
The study showed more than just the chatbot’s superior performance.
It unveiled doctors’ sometimes unwavering belief in a diagnosis they made, even when a chatbot potentially suggests a better one.
And the study illustrated that while doctors are being exposed to the tools of artificial intelligence for their work, few know how to exploit the abilities of chatbots. As a result, they failed to take advantage of A.I. systems’ ability to solve complex diagnostic problems and offer explanations for their diagnoses.
A.I. systems should be “doctor extenders,” Dr. Rodman said, offering valuable second opinions on diagnoses.
But it looks as if there is a way to go before that potential is realized.
(Excerpt) Read more at nytimes.com ...
More medical research money...public and private (private meaning things like The American Cancer Society)...flows into Boston than into any other city/region on earth (I've seen the numbers).
“””I would trust AI before I’d trust about 80% of doctors. Feed in all the vitals - blood work etc and hands down I’d go with the computer.”””
I would agree. Most doctors do not take the time to look at prior blood work results when they are looking at the current lab results. I have found that I need to do that myself. I maintain an excel spreadsheet where I have recorded my lab results over the past 20 years.
I forgot. BU must be in the Boston medical complex too. I did not know that Boston was number one for Medical of all kinds. Research. Treating patients and so on. I’ll bet loads of rich South Americans and other fly in for crucial medical care. This helps with our trade deficits.
I would trust AI before I’d trust about 80% of doctors.
~~~
My problem with AI is that it’s a threat to your privacy.
If you thought google was tracking people, I can’t imagine how well AI is planning to keep people’s information.
If I am not mistaken, you can’t even use ChatGPT without being logged in.
Back in the day, in the 1990s, I worked on a computer research project with a hospital to add interactive computers into the mix for doctors treating leukemic children.
One aspect was to provide videos showing how parents had to handle the treatments they were responsible for immediately after being shocked with the news of what they were about to deal with. Very hard to concentrate on instructions after hearing do this wrong and your child dies. Have a nice day. This when you’re still mentally frozen from the news, so videos were a good idea. Had a camera crew at the hospital and my audio guy slid down the wall watching a treatment.
Another aspect encouraged fathers who would be more reluctant to ask questions of other human beings. It was meant to help them deal with their feelings in the middle of the night with no one watching. An emotionless support group. Get fathers who COULD deal with their feelings talk about them in videos to the ones who weren’t dealing as well.
The third part was to answer medical questions. After an answer, almost all questions were followed up with Ask Your Doctor.
This was close in time when another part of our research was starting to work with robotic surgery. The company was just then trying to move our research into more practical, money-making work. Not a bad approach, but it was traumatic after having always lived in a pure world of ideas.
No...the three complexes are in different parts of the city. Boston Medical Center,BU’s main teaching hospital,is located (IIRC) not far from “Chinatown”. And Tufts hospitals are scattered throughout the city,Tufts Medical Center and Saint Elizabeth’s Medical Center being their two most important teaching hospitals. IIRC Tufts Medical Center is located in,or near,the “Financial District”...some distance from Harvard’s complex.
Dr. Rodman may be surprised but I am not. Too many bad experiences have conditioned me to expect less of doctors. Like surgeons, so many have become “sometimes wrong and never in doubt”. Dr. Rodman’s study supports this statement. Even when presented with a justified contrary diagnosis they reject it out-of-hand. I believe they are taught they are infallible in order to give them the confidence necessary for so many critical decisions. I also believe they have to be come numb to bad decisions to keep on making enough good ones. Medicine has become like a forced march.
You may be interested in a book called “How Doctors Think”.
I’m going through brain cancer and have used grok, x version of ai to make understandable mri scans and other medical reports. Just paste in report and it will spit it back with deeper explanations.
The bright side is, DEI taking over health care 'education' isn't as likely to kill you now.
(Gina Kolata wrote a book "Influenza", well worth reading.)
Stopping teaching DEI, and social justice to docs would be a start. Stop forcing them to believe that female+naughty bits = male, or that unborn babies in the womb are clumps of cells, and that mRNA injected into human cells only does one thing,(namely :make vaccines).
Teach them how to discriminate between real data vs. “massaged” data. Get rid of their ridiculous paperwork overload, and documentation.
Finally, get rid of the bad docs.
P.S. did these AI tests used general doctors vs AI in general types of cases? Or AI vs a specialist?
The whole test could be skewed...example: AI vs general doc testing with a case of obscure endocrine disorder...
Wouldn’t it be better to have AI vs endocrinologist?
Don’t forget...you will not be able to change AI’S conclusions. E G.....if one of your symptoms is a “rare” or low incidence symptom, it may be downplayed when the diagnosis is made. The algorithm will disregard the pertinence of your symptoms.
If AI is misprogrammed you might as well be dead. Cause none of the medical “professional” sheeple will dare to go against the “robot”...
I see AI as a “prompt” to remind docs of obscure diagnostic or therapeutic considerations.
Dumb docs may use AI incorrectly...when general euthanasia is well accepted...expect it presented as an option for your AI generated reatment plan.
That’s the beauty of AI - they have all the information on the rare symptoms and can call for tests to rule out the less than 1% stuff.
Doctor’s often say, ‘common things are common’ which is why our medical school can graduate ‘doctors’ without the necessary skills or brain power to do the job correctly. Common things are common. Most of their calls will be correct - where it will fail will be the one off diagnosis of the person who has the unusual.
Program computers right and everyone will be better off with AI than with a live doctor in the bottom 25% of his graduating class. We’ve had DEI crap in medical schools for some time now... too long to feel safe with many ‘doctors’...
Does AI treat symptoms instead of causes better than doctors do?
Or do they actually cure disease despite making less money for the industry?
The question is: can we take the good from AI without the bad?
If not I'd rather have some unqualified idiot who got into medical school based on white guilt - - as a doctor. Eventually that person would find the right diagnosis or pass the case on to someone who's qualified. The only drawback there would be if a form of socialized medicine existed that stuck with a doctor 'assigned' to you.
I agree.
Did anyone ever feed it all the stuff Biden or Harris has said and asked for a diagnosis?
Thanks. Seems Harvard is the one and only, and overlord of the Boston medical complex.
Now do covid.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.