Journal of ISAKOS’ Post

It will be interesting to observe the legal and ethical implications of this tech if there was a mis-diagnosis. *But* if the overall outcomes of an AI driven medical expert is significantly better than humans, would it be poor judgment to not use the tech due to fear. Some considerations ⭕️ Guardrails against misuse ⭕️ Missing factor of human empathy ⭕️ Who is responsible if mis-diagnosis leads to death ⭕️ Diagnosis and Treatment options based on ability to pay or social factors #aidoctor #medicalcompliance #ai #medtech

After 60 years of escalating medical costs (compared to baseline inflation) we have the technology to dramatically drop in the cost of primary care. Dr. GPT should be allowed to handle the basics. Here's how it could work initially: 1. Dr. GPT verbally take the patient's history. 2. Anything flagged as high risk routes to an actual physician. 3. Allow Dr. GPT to request routine blood work. 4. Any highly abnormal blood work routes to an actual physician. 5. Allow Dr. GPT to prescribe routine things like blood pressure medicine, statins, etc for routine, chronic conditions. #OpenAI, #Oracle, or #epic really ought to be running a clinical trial on this. The clinical endpoint should be something like the number of adverse outcomes by the GPT as compared to the human physician. Success would be the GPT having either the same or less adverse outcomes. Our biggest obstacle, at this point, is not technical, but regulatory.

HKUST pioneers 'medical ChatGPT' - RTHK: Researchers from the Hong Kong University of Science and Technology on Thursday said they have developed a new AI medical model that answers ...

Today the most important article in chatgpt and "Assessing the utility of artificial intelligence throughout the triage outpatients: a prospective randomized controlled clinical study" - Utilizes ChatGPT, a natural language processing tool powered by artificial intelligence technology, for outpatient triage evaluation. - Conducted a study with 30 common outpatient questions, rated by five experienced doctors, showing high consistency with manual triage. - Prospective cohort study with 45 patients revealed a high level of agreement between manual triage and ChatGPT triage. - ChatGPT's responses were professional, comprehensive, and humanized, potentially improving patient treatment time, diagnosis, and alleviating medical staff shortage. #AI #Healthcare #Triage #ArtificialIntelligence #PatientCare

As we celebrate the strides in language models like ChatGPT, a pressing question arises: Are we overlooking technologies that could directly save lives, such as improved medical imaging tools? While language AI is impressive, enhancing medical imaging could significantly impact real-life health outcomes. Do you think data privacy issues in gathering medical imaging data may contribute to this perceived gap, as language models often rely on publicly accessible data? #LLMs #MedicalImaging #HealthcarePriorities #DataPrivacy #InnovationBalance

👉🏼 Evaluating the Proficiency of ChatGPT in Emergency Medicine: A Comparison Across AI Models 🤓 Chiu-Liang Liu 👇🏻 https://lnkd.in/eBCd-9hi 🔍 Focus on data insights: - 📊 GPT-4o outperformed other models, answering 138 out of 200 questions correctly, indicating its superiority. - 🔍 Custom GPTs performed better than GPT-4 but not as well as GPT-4o, showcasing potential for tailored applications. - 📈 Statistical analysis

Interesting results from the field of medical AI research. A recent study including 50 physicians asked: Can #ChatGPT diagnose patients better than doctors? And what happens when doctors use ChatGPT for help? The results are interesting: 🟢 ChatGPT alone scored 90% in diagnostic accuracy. 🔴 Doctors using ChatGPT only scored 76%, barely higher than doctors without it (74%). Why the gap? Two possible explanations stand out: 🔸 Overconfidence: Doctors often stuck to their original diagnosis, even when ChatGPT suggested a better one. 🔸 Underuse: Doctors treat ChatGPT like a "smarter Google" rather than leveraging its full capabilities and engaging with its reasoning. The potential is clear: AI could become a powerful “doctor extender,” offering second opinions and solving complex diagnostic problems. For this to work, medical professionals need training in how to fully utilize these tools. To all the doctors out there: it might be time to learn how to use AI effectively. Full study here: https://lnkd.in/gA6_h9GP

Two days left to register for our next #ChatGPT focused webinar! 👨💻 Analysing accuracy, balancing bias: Can ChatGPT be trusted to ease the #CareDocumentation burden? Register here to join us on 23rd May, 15:30 BST! 👉 https://lnkd.in/erjUtvDq This webinar will cover: 1. Racial disparities in automated speech recognition (Allison Koenecke, Cornell Department of Information Science) 2. Adapted large language models can outperform medical experts in clinical text summarization (Dave Van Veen, Stanford Center for Artificial Intelligence) 3. Discussion: What does this research tell us about using AI in long-term care? #GOLTC #DataScience #Webinar

Are medical studies being written with ChatGPT? We all know ChatGPT overuses the word "delve". Look below at how often the word 'delve' is used in papers on PubMed (2023 was the first full year of ChatGPT). ↓ Check out https://AlphaSignal.ai to get a daily summary of breakthrough models, repos, and papers in AI. Read by 180,000+ engineers and researchers.

8 Awesome ETHICAL uses of Chat gpt in medical research that you will never know alone 😏! I have tested their efficacy in this video. 1-Generating meta-analysis ideas 2-Generating ideas from datasets like DHS 3-Meta-analysis idea validity 4-Primary studies idea validity 5-Generating search term of your meta-analysis 6-Journals’ recommendation 7-Calculating the sample size of your paper 8-Conferences finding https://lnkd.in/dWgGVXkD #research #paper #publication #chatgpt #AI #acceptance #ideavalidity #systematicreview #metaanalysis #dhs #idea #samplesize #journal #conference #publichealth #editor