Eric Doherty (@EDohertyGlobal) header
Eric Doherty (@EDohertyGlobal) profile picture

Global Telehealth | AI Data |Medical Devices and Pharmaceuticals | Patient Access & Quality | Global Health | Global Commercialization |

Atlanta, GA
Joined March 2021
Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)Photo by Eric Doherty (@EDohertyGlobal)
Turning advances in AI into real outcomes is the right framing. In healthcare the outcome test is simple: less paperwork for clinicians and faster answers for patients. We're watching how platform announcements like these translate into virtual-first care at QC Healthcare. @satyanadella teapawt.net/satyanadella/status/21…
Our ambition with everything we are announcing at Ignite today is to parlay technological advances into real outcomes and new points of leverage for people, organizations, and society at large. Just a few of the things we're announcing:
4
HLTH USA 2026 is underway through today, Nov 18, at The Venetian in Las Vegas, bringing healthcare innovators, payers and providers together on AI, digital health and care delivery. For virtual care teams, events like this are where partnerships and policy conversations take shape. Worth following alongside the work of @AmericanTelemed and @HIMSS. Event info: hlth.com/2026event At QC Healthcare, My Pediatric Doctor (mypediatricdoctor.com, [email protected]) and My Adult Doctor (myadultdoctor.com, [email protected]) bring telehealth to families. About us: myadultdoctor.com/about-qc-h… cc @ONC_HealthIT #HLTH2026 #DigitalHealth #Telehealth #QCHealthcare
13
AI scribes went from novelty to a crowded market fast. The real test for any of them is whether clinicians get time back and patients feel more heard in the visit, not just a cleaner note. That matters in virtual care too, where the conversation is the whole visit. teapawt.net/statnews/status/212313…
2
Agent safety beyond the prompt is the right conversation. In care delivery, an AI agent touching scheduling, messages or records needs guardrails and a clear human owner long before it earns patient trust. teapawt.net/nvidia/status/21230683…
The AI Podcast is back. @huggingface's @Thom_Wolf and @ServiceNow's Amanda Grady join us to talk open models, agent safety, and helping more people build with AI. 0:00 AI safety beyond the prompt 1:50 Open and proprietary models for cyber defense 3:28 Giving agents their own identity and permissions 5:07 AI Passport and NVIDIA Sentry 6:03 What better security unlocks 8:32 Open source and security 10:58 Bringing safety into the physical world 13:10 Making fine-tuning more accessible 14:09 What’s next—and why they’re optimistic
9
Models improving the chips they run on is a striking loop. Healthcare AI sits on the same compute stack, so efficiency gains like this eventually decide whether AI-assisted care stays affordable enough to reach every patient, not just the best-funded systems. teapawt.net/OpenAI/status/21231003…
Working with @AMD, our models identified up to 14% PPA improvement on flagship GPU IP in less than five months—building on designs AMD’s engineers had already spent years optimizing. The same approach helped improve Jalapeño, our first custom inference chip, with gains verified on real silicon. New evaluations also show our models getting better at certain chip design tasks. We’re sharing these results and how we’re working with @Synopsys to bring our model capabilities and their trusted EDA design and validation tools to more chip designers. openai.com/index/advancing-c…
1
AI moving from answering questions to actually handing off everyday tasks is a big shift. Healthcare is full of that kind of life admin, from scheduling to follow-ups, and patients feel every bit of it. Consumer tools like this will shape expectations for how easy care access should be. @sundarpichai teapawt.net/sundarpichai/status/21…
Knowing things → doing things → finding ways to save you time and money We’re expanding Tasks in Gemini free of charge to hundreds of millions of people in six countries. Powered by Gemini 4 Hydrogen, Tasks lets you hand off everyday life admin directly: - Forward emails to [email protected] to RSVP and add it to your family calendar - Have Gemini call local businesses to compare price quotes and check inventory (with full transcripts + approval before dialing) - Check out securely via Google Pay (uses the Agent Payments Protocol - AP2) - Seamlessly sign in via Google Password Manager We wanted to make this helpful + secure, we’re also partnering directly with major websites and security providers so it works well for site owners and agents.
2
Frontier models keep getting cheaper per task, and that's what makes AI practical for everyday care operations, not just demos. For virtual-first care, the bar is still the same: does it save clinicians time and help patients get answered faster? That's how we judge every new model at QC Healthcare. teapawt.net/OpenAI/status/21227501…
Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. openai.com/index/introducing…
2
Integration into the workflow is the real gate for health IT. If a tool doesn't fit how clinicians and patients already work, it rarely gets past the pitch. That's the lens we bring to virtual-first care at QC Healthcare. Sharp read from @chrissyfarr. teapawt.net/chrissyfarr/status/212…
Health systems I'm chatting with are saying the challenge for them is they'd probably go with a B minus solution from Epic even over an A plus from a vendor. There's basically no point in hearing the pitch if the solution isn't integrated into the Epic workflow.
5
Agents are only as good as what they can retrieve. In healthcare the useful version connects scattered information across a patient's history without losing context, and that's a retrieval problem as much as a model problem. Interesting direction from Microsoft Research. teapawt.net/MSFTResearch/status/21…
GraphRAG Zero solves a fundamental challenge that has limited AI agents' ability to reach their full potential. It’s a technology for holistic discovery that reflects years of Microsoft research, transforming graph-based retrieval from an experimental technique into a deployable system. Optimized for real-world applications and agents, it enables organizations to explore connections across vast amounts of their own data, quickly and at low cost. It’s available now in Foundry IQ and @github’s Repo Mind Light, supporting applications from scientific discovery to repository understanding. msft.it/6017xE5xL
2
Patient access to their own records started as a policy idea and is now table stakes for digital health. The next step is making that access usable, so patients can actually act on what they see between visits. Good timeline from @Berci. teapawt.net/Berci/status/212297803…
Look at this timeline of digital health. Decades ago, the need for patients to be able to access their own medical records appeared. Now, the US Department of Health and Human Services has announced changes to strengthen individuals’ rights to access their own protected health information (PHI), including electronic PHI through personal health apps; improve information sharing for care coordination and case management; and reduce regulatory and administrative burdens on covered entities. Great steps (not bold enough, but still great) forward, but it has taken decades of e-patients standing up for their rights. I hope we do a better job when AI takes over every repetitive task in healthcare. The announcement: hhs.gov/press-room/hhs-issue…
1
Patients are increasingly handed consumer health tests, so the question of what's trustworthy matters. Trust in digital health gets built when the evidence is clear and clinicians can help patients interpret what they're seeing. Useful reminder from @EricTopol. teapawt.net/EricTopol/status/21230…
Direct-to-consumer gut microbiome tests are still not ready to be trusted Feature @nature nature.com/articles/d41586-0…
9
Discharge summaries are exactly the kind of clinician busywork AI should take on, but this is the right question: how do AI-drafted summaries hold up against the clinician-written standard? At QC Healthcare, we think about this for virtual-first care, where clear handoffs between visits matter. Worth a read from @EricTopol. teapawt.net/EricTopol/status/21227…
When AI-generated hospital discharge summaries meet the clinician-authored current standard of care jamanetwork.com/journals/jam…
6
Pharmacogenomics only helps if the result is in front of the prescriber at the moment of the decision. For virtual-first care, where visits are quick and remote, that kind of decision support has to be built into the workflow rather than left to a separate lookup. teapawt.net/NIH/status/21230830396…
New findings from the NIH-funded All of Us Research Program highlight the potential value of using genetic information to guide medication prescribing. Read more:
CMS released a Clinical Super Intelligence roadmap on Nov 13 ('Better Answers, Faster Care, and Lower Costs'), coordinating @CMSgov, @HHSGov and @US_FDA efforts to speed clinical AI adoption. Four priorities, as reported: health data access, AI evaluation systems, regulatory modernization, and payment models tied to patient outcomes. CMS Health Tech Ecosystem commitments reportedly grew from 60+ to 800+ companies, and the Medicare App Library is live. It does not announce new reimbursement rates or guaranteed coverage for any AI tool. Source: distilinfo.com/2026/11/16/cl… CMS roadmap: cms.gov/initiatives/health-t… At QC Healthcare, My Pediatric Doctor (mypediatricdoctor.com, [email protected]) and My Adult Doctor (myadultdoctor.com, [email protected]) keep care accessible. About us: myadultdoctor.com/about-qc-h… cc @HIMSS @ONC_HealthIT @AmericanTelemed #HealthcareAI #HealthIT #DigitalHealth #QCHealthcare
13
Useful read from the floor at HLTH. "The shift left and who's going to serve" is the question I keep coming back to: as care moves earlier and closer to home, the winners will be whoever makes access simple and reliable for families. Thanks @chrissyfarr for the roundup. teapawt.net/chrissyfarr/status/212…
What everyone is talking about so far at HLTH for those who aren’t here: - The smart healthcare VCs are waiting for the AI bubble to pop - Oura/ Counsel and the deal mechanics - The shift left and who’s going to serve - Who’s really making margin & why (it’s not health systems) - What “human in the loop” even means to different parties & more
1
25
Good health journalism helps patients and clinicians make sense of a fast-moving field, and that matters more as AI reshapes care. Congratulations to Helen Branswell on this recognition from @TheLancet. teapawt.net/EricTopol/status/21224…
6
Enterprise AI keeps moving from chat to agents, and the same shift is coming for care delivery. The measure that matters for any platform shown at Ignite: does it take work off clinicians and make access easier for patients? That's where we focus at QC Healthcare. teapawt.net/satyanadella/status/21…
Looking forward to Ignite, kicking off today at 10am PST! Hope you’ll join us live. teapawt.net/i/events/2116180615661…
5
The AI-native playbook starts with people, not tools, and healthcare is no different. Clinicians adopt what visibly gives them time back and keeps patients at the center. Rollout order matters as much as the technology. teapawt.net/mardehaym/status/21224…
I've watched a lot of PE-backed companies try to go AI-native this year. The ones that get there do it in roughly this order: The people. 1. Give everyone a paid AI seat, with a budget per team so the bill doesn't surprise your CFO. 2. Find the 5 people already using it every day. They teach everyone else. 3. Ask every team lead to write down the 10 most boring tasks their team does each week. 4. Sit with your best people and write down how they do the job. If it only lives in their head, an agent can't do it. The data. 5. List every system you run. After a few acquisitions it's usually a zoo, with Excel holding it together. 6. Pick one source of truth for revenue, customers and headcount. Agents can't fix numbers that disagree. 7. Give agents read access first. Write access gets earned, same as a new hire. The workflows. 8. Put a dollar number next to each boring task. Hours saved only count if somebody spends them on something else. 9. Pick one workflow, with one owner and one number you want to move. 10. Put every fixed rule in plain code. The model only handles the judgment calls. 11. Build a golden set: 100 past cases your best people already judged. 12. Run it in shadow mode next to your team for a few weeks and read every disagreement. 13. The agent prepares, a person approves. Anything touching money or customers waits for a human. 14. Ship it in weeks. If it takes a quarter, the scope is wrong. The guardrails. 15. Every model call goes through one gateway, with logs and a budget. 16. Keep a kill switch, and test it. 17. Re-run the golden set every time the model changes. The engineering team. 18. Small teams ship. Two or three engineers with agents and one tech lead reviewing everything. 19. Every pull request goes through the same checks, whoever or whatever wrote it. 20. Measure AI spend against what shipped to customers. The part nobody wants to hear 21. Never vibe-code your ERP or your billing. 22. Finish one workflow properly before you start the next five. Skip the boring half of this list and you end up with 40 pilots and nothing in production.
4
Longevity research keeps pointing to the gut as a quiet driver of healthy aging. The practical question for primary care is how fast findings like this can become earlier, simpler ways to keep people well, not just treat them once they're sick. teapawt.net/EricTopol/status/21223…
Exceptional longevity in centenarians and mice linked to the gut microbiome's augmented ability to modify (deconjugate) bile acids @Cell_Metabolism cell.com/cell-metabolism/ful…
8
Multi-step research problems are exactly where AI earns its keep, and the same pattern applies in care delivery: the value shows up when it handles the tedious steps and people keep the judgment calls. Worth watching how these tools hold up with real-world rigor. teapawt.net/GoogleDeepMind/status/…
What happens when you put Gemini 4 Carbon to work in the lab? 🔬 Researchers are using our latest model in Google @Antigravity to tackle multi-step problems across physics, math, and the life sciences.
3