Our study of AMIE at @BIDMC_Medicine is out! You can read about what we did in posts from my co-authors (including the Google post below).
But I wanted to talk about some background for this study, and what I think are most interesting findings. 🧵⬇️
https://t.co/H8nQBdF18R
Thanks to @EricTopol for highlighting our latest study, which looks at some of the failure points in the clinical reasoning performance of reasoning models.
What does this mean for the future of diagnostic decision support? And can LLMs actually medically reason?
A 🧵⬇️
Huge update to our preprint today on the superhuman performance of reasoning models in medical diagnosis!
TL;DR – they don't just surpass humans in meaningful benchmarks, but in actual medical care from unstructured clinical data:
A 🧵⬇️:
https://t.co/7fYxNfebiF
@KevinH_PhD I am so sad to read this. I have followed you online and learned so much from reading your posts and papers. You have moved the field forward!
There is a lot of buzz about our new paper in Nature Medicine on the effects of LLMs (GPT-4) on physician management reasoning! I had TONS of fun working on this -- but what it MEANS requires some unpacking.
A 🧵⬇️
https://t.co/yLZJw1U5IE
Preprint out today that tests o1-preview's medical reasoning experiments against a baseline of 100s of clinicians.
In this case the title says it all:
Superhuman performance of a large language model on the reasoning tasks of a physician
Link: https://t.co/QB0mawHJNE
A 🧵⬇️
Our first RCT on using an LLM on diagnostic reasoning is out!
And the results are 🔥🌶️... adding ChatGPT did NOT improve diagnostic accuracy or reasoning, and the AI alone outperformed ALL the humans.
What does this mean? A 🧵⬇️
https://t.co/MgkhrrkKmT
What's a typical day in a GP practice in Denmark 🇩🇰 where 98%+ of people have access to primary care?
Here's some of what I learned spending the day at a practice in Aarhus last month 🧵
(and yes, the space really is as beautiful as it looks in these photos!)
🚨#FLOW: The most anticipated trial of the year now published in @NEJM with #ERA24
Semaglutide 1.0mg once weekly disease modifying with kidney, CV, & mortality benefits in people w T2D+CKD, cementing itself as a new pillar of care.
🧵 on key thoughts 👇🏾
https://t.co/5fbSxPSAMM
@zach_london Great 🧵Are there any general principles to adress patients with pre-existing peripheral neuropathies who have an indication to begin one of these anti-cancer drugs? Are some types of neuropathies more susceptible to worsen with exposure?
#idboardreview unimmunized child w/fever, cough, nasal inflammation followed 2 days later by diffuse nonpruritic maculopapular rash on face then chest then arms/legs. Exam: red conjuctiva, cbc, chem normal, chest clear, cxr b/l interstitial infiltrates #idtwitter#idxposts#meded
@MichaelAlbertMD@soonergise “Cardiovascular disease was defined as previous myocardial infarction, previous stroke, or symptomatic peripheral arterial disease.” I wonder how prior auth will limit access for those with sub clinical ASCVD (positive CAC…CCTA, etc) or elevated clinical risk scores
@MichaelAlbertMD I’ve learned a lot from following you and @BevTchangMD …do you have any favorite publications or online grand rounds that cover some of the practical issues re; titration or adjustment and optimizing longitudinal success?