Austin Perlmutter, MD

Three announcements in one week by OpenAI about health and healthcare. All three would be major on their own, taken together. They are a tsunami of new information. Here are the three announcements. 1) early January (Beckers article on Jan 5): 2025 ChatGPT usage by consumers: https://lnkd.in/e_xnd53z 2) January 7: “Introducing ChatGPT Health; A dedicated experience in ChatGPT designed for health and wellness.” https://lnkd.in/e_HxpjT4 3) January 8: “Introducing OpenAI for Healthcare; Secure AI products to help healthcare organizations scale high-quality care, reduce admin work for teams, and power custom clinical solutions—while protecting health data.” https://lnkd.in/ebY5hXX6 So let’s stay out of the details. Three major announcements in a week probably took some planning! OpenAI’s health team is early but serious — they recruited seasoned digital health and product leaders in 2025 (Gross and Alexander), are hiring dedicated health AI researchers, and are building product teams to support OpenAI for Healthcare and ChatGPT Health offerings. The agenda is a mix of go-to-market strategy, product execution, and research quality/safety. What do we make of a foundation model company verticalizing this explicitly? Well if you are in healthcare then all you are seeing this week is the water going out and so, maybe, you are just admiring all the pretty seashells on the beach. Or, maybe, you are running for high land. The same playbook is visible across the stack: Anthropic is verticalizing through enterprise compliance and cybersecurity, Google is explicitly fusing Gemini with domain data in bio/geo/finance, and Meta’s open LLaMA models are being privately fine-tuned into legal, finance, and biotech tools. Microsoft is doing something similar with CoPilot for large “late” adopters like the public sector. Geopolitical segmentation has already occurred and is visible in Microsoft ’s global diffussion numbers (and its globalist positioning) Industry segmentation will be read as aggressive but it may actually be defensive. There is no moat around undifferentiated foundation models and people are already pointing out that “90% national models” may be three times as good as last year’s US model and in a couple of years…. What if it is defensive and OpenAI is spending money to become a vertical product company while it still has VC dollars. AOL and TimeWarner anyone? Honestly, I don’t know …. and the sea shells are so pretty! [More on the healthcare specific implications in another post; but for now I am taking my time to read the three announcements separately as well as collectively]

30

9 Comments

Software for your DNA? How bioregulators are redefining cellular resilience in 2026. 💻 If you could support one specific organ system for longevity: heart, brain, or gut: which would you prioritize right now? Bioregulators act like precise signaling molecules, helping your body repair and regenerate at a cellular level. It’s the next frontier in personalized health. Investing in your cells today means more vitality tomorrow. 🌱 #Bioregulators #CellularRegeneration #AgingWell #FutureOfHealth #Vitality

4

Children's Hospital Colorado is ahead of its time in using whole-genome sequencing to treat its smallest patients. Read about it in my latest article, "Children's Hospital Colorado Integrates Genomic Insight Into Day-to-Day Care." Many thanks to Austin Larson, MD, Caitlin Jenney, and Rachael Fowler for their help and to Robin Hocevar for the assignment! #ChildrensHospitalColorado #genomesequencing #ASAMonitor

4

What patient satisfaction questions do you find most helpful in assessing the patient experience? I have been giving this a lot of thought lately. Value is relative, and patient-reported outcome measures provide limited information. Even if someone makes a clinically important improvement, that doesn't mean their treatment was a success. MCID can be reached with walking programs or just waiting for injuries to recover on their own. One patient may struggle to afford a $75 copay finds a way because they believe the service is worth it. Another may be wealthy, not bat an eye at spending $75, but still find it to be a waste of money, given their experiences with PT. I think objective measures are important, but I also think we need to understand what patients believe about their care plans and the value they perceive. I gravitate towards these four measures: 1. Outcome satisfaction (how satisfied they are with the amount they believe they recovered/improved) 2. Duration of care satisfaction (how satisfied they are with how long the plan of care took) 3. Cost of care satisfaction 4. Experience satisfaction (the overall experience of the plan of care) What about you? Do you have standard questions or areas of focus, or do you stick with the single 5-star Google rating or NPS?

7

3 Comments