Han-Shen (Han) Yuan

California is the top Home Health state The chart below shows the total Medicare amounts for home health agencies by state. The Centers for Medicare & Medicaid Services quietly updated the utilization statistics data for Post-Acute Care providers, including IRF, LTC, SNF, Hospice, and Home Health. And yes, they are now also available in mimilabs to be explored! Aside from the excitement of new datasets, the chart has some interesting story too. Texas and New York have been the top markets, but something changed dramatically since 2020. California, previously 3rd in the home health market size, finally became 1st in the nation and has been broadening the gap ever since. As can be seen, all other states show downward trends in home health spending. Not just TX and NY - you'll notice the lines are slightly heading down for other states too, making California's surge even more remarkable. So, what happened to CA? It may have been the perfect storm, though we can't pinpoint exactly what drove this shift. California had the nation's longest lockdown—March 2020 to January 2021—while home health was deemed "essential" and could keep operating. So 40 million Californians are stuck at home, hospitals are feeling unsafe, but home health agencies could still visit? Who knows, but one can imagine. The data also shows telehealth expansion (permanent reg changes in 2021), Medicare Advantage plans adding in-home benefits (from 223 plans in 2020 to over 1,000 in 2023), California's aging demographics, and nearly $1 billion in state investment for care worker training. But without more granular data, we're mostly connecting dots. My question would be: how does this dramatically different utilization pattern affect ACO benchmarks and performance? California ACOs might be seeing inflated home health spending built into their spending. It would be fascinating to dig deeper into how state-specific pandemic policies have created these lasting structural changes in post-acute care utilization patterns.

15

2 Comments

UCSF Chair of Medicine Robert Wachter (author of "The Digital Doctor" and "A Giant Leap") on Doctronic in his Substack: "Personally, I'd have no problem using this system to refill my Lipitor." He also flags that we may be the first company to buy malpractice insurance for autonomous AI in medicine. (We are.) I'm genuinely humbled by the mention. Bob has shaped how a generation of physicians thinks about technology in medicine. Everyone should read his take on why this week marked "the moment AI moves from advising on care to actually delivering it": https://lnkd.in/eqpFgC2Y

88

5 Comments

For the past year, we’ve been working closely with root-cause focused clinicians on some of the most complex cases in practice. Similar patients, comprehensive labs for both, deep expertise. And yet, one responds to treatment while the other develops unexpected side effects. Even with everything we know, there’s still so much trial and error. That’s why we built Diadia Health’s AI, designed differently from typical “AI health copilots” (which are GPT-4 + PubMed wrappers). Diadia is built on causal inference, using three distinct reasoning modes: • Direct mechanisms (A → B) • Chain inference (A → B → C, therefore A → C) • Pathway convergence (if a genetic variant affects pathway P, and a drug also affects pathway P → therapeutic match) In other words, Diadia finds the why behind the why using the depth of our genetics. Now, we’re opening a limited clinical pilot to explore a key question: "When clinicians are given systems-level insight into how genetics, biomarkers, and symptoms interact, can we reduce uncertainty and help complex cases stabilize faster?" And how much more can we accomplish with the use of genetics? In this pilot, clinicians select one difficult or non-responsive case, a patient they’ve already tried treating using their usual approach. They then apply Diadia’s AI-generated genetic and systems-level insights to adjust or refine the strategy and observe what changes. All patients remain under normal clinical oversight. We’ll measure what actually matters in real practice: • How many biomarkers move into optimal range • How quickly key systems stabilize • Whether patients feel meaningfully better What’s in it for you? For two months, you can explore how genetics and biomarkers influence patient outcomes, potentially solve existing difficult cases, and save time - all at NO COST and with NO STRINGS ATTACHED. The goal isn’t to change how clinicians practice medicine, it is to surface patterns humans can’t reliably track alone and create a new understanding of why patients stop responding as expected. We’re onboarding a small group of functional, integrative, hormone, and longevity clinicians who regularly work with complex or non-responsive patients. If you’re interested in helping shape the next generation of clinical decision intelligence, visit https://lnkd.in/efRd3YRt for more details and how to get on board.

42

3 Comments

🎉 Portfolio Highlight: mdhub 🎉 MdHub is re-imagining daily workflows for behavioral health practices with a clinical AI assistant that handles the paperwork so clinicians can focus on patients. Imagine cutting hours of paperwork down to seconds. The platform records sessions, then auto-generates structured notes, treatment plans, patient summaries, and other documentation in seconds freeing up roughly two hours per provider every day and paying for itself in less than half a day of saved time. But that's not all. Their 24/7 AI admissions coordinator “Sarah” triages every inbound call, qualifies new leads, schedules intakes, manages reschedules, and even processes medication-refill requests, lifting new-patient volume by as much as 30 percent for partner clinics. Talk about efficiency! Founded by Dominik Middelmann and Efrén A. Lamolda, MdHub (YC S24) is on a mission to make high-quality mental-healthcare more accessible and affordable through intelligent automation. Super excited about the potential impact here!

35

9 Comments

At HIMSS, I met Saamer Siddiqi M.D., CEO of LIVEMED Telehealth™ - and we didn’t talk tech. We talked about the truth. Here’s one: Telehealth adoption exploded. But meaningful outcomes? Not so much.  👉 The digital divide is widening. 👉 Maternal mortality is still a crisis in rural communities. 👉 Providers are overwhelmed by disconnected workflows. 👉 Reimbursement models remain a puzzle for most practices. Why? Because too many platforms are built around features, not outcomes. ✅ What we need: Seamless integration with EHRs and patient data. Engagement models for chronic and maternal care, not just urgent visits. Smart triaging + follow-up systems are baked into the platform. Human-centered UX that earns trust across diverse populations. I believe telehealth can be the bridge only if we start solving real operational pain, not checking tech boxes. If you're a healthcare leader tackling this, I’d love to hear your perspective—or brainstorm a better way forward. Justin Schreiber Michael Shope Ido Schoenberg Fay Rotenberg Abhi Pathak #HealthcareLeadership #TelehealthTransformation #MaternalHealthCrisis #HIMSS2025 #VirtualCareThatWorks

39

5 Comments

Agentic AI startup VoiceCare AI has launched with $3.8 million in funding and a pilot with the Mayo Clinic to automate healthcare back-office work. VoiceCare AI has developed an enterprise-grade platform to automate routine phone conversations for healthcare companies, tackling tasks like benefit verification, prior authorization, and claims management. Its voice AI agent, "Joy," can handle long, complex, and nuanced conversations, including navigating IVR systems and holding for up to 2.5 hours, to ease the administrative burdens for healthcare staff. The funding comes from investors including Caduceus Capital Partners, LLC, and Bread and Butter Ventures. The capital will support team expansion, sales and marketing, and continued investment in enhancing platform accuracy, security, and compliance. Congratulations to founder and CEO Parag Jhaveri and the VoiceCare AI team. Fierce Healthcare: https://lnkd.in/dNeBZHNV #HealthTech #AI #AgenticAI #VoiceAI #Automation #Funding #VentureCapital #MayoClinic

14

2 Comments

Client Success Story: MedSure Health — Securing Patient Diagnostics with Honeypotz Confidential Computing TEE Client Overview MedSure Health, a fast-growing telemedicine startup, provides AI-powered diagnostic services to patients across North America. Their platform enables users to upload medical images and receive instant analysis from AI models, followed by virtual consultations with certified physicians. The Problem As MedSure scaled, they faced a critical challenge: protecting sensitive patient data during AI processing. Although data was encrypted at rest and in transit, it remained vulnerable during computation — a major compliance and trust issue, especially under HIPAA and GDPR regulations. MedSure’s CTO, Elena Ramirez, explained: “Our AI models needed access to raw patient data to deliver accurate diagnostics. But that meant exposing sensitive information during processing — a risk we couldn’t afford.” The Solution: Honeypotz Confidential Computing TEE MedSure partnered with Honeypotz Inc to integrate their Confidential Computing TEE platform into the diagnostic pipeline. Honeypotz’s solution offered: • Trusted Execution Environments (TEEs) that isolate and protect data during processing. • Remote attestation to verify the integrity of the computing environment. • Seamless integration with MedSure’s existing cloud infrastructure. Implementation Highlights 1. Secure AI Inference Medical images and patient histories are processed within Honeypotz’s TEEs, ensuring that data remains encrypted and inaccessible to unauthorized parties — even during computation. 2. Compliance Automation Honeypotz provided real-time attestation logs and audit trails, simplifying MedSure’s compliance reporting and reducing audit preparation time by 30%. 3. Patient Trust & Market Expansion With enhanced data security, MedSure launched in European markets, confident in GDPR compliance. Patient onboarding increased by 25%, driven by trust in the platform’s privacy guarantees. Results • Zero data exposure incidents since deployment. • 90% reduction in internal data access risk. • Faster regulatory approvals for new markets. • Improved brand reputation and patient retention. Client Testimonial “Honeypotz didn’t just give us a security solution — they gave us a competitive edge. Our patients feel safer, and our team sleeps better at night.” — Elena Ramirez, CTO, MedSure Health

9

1 Comment

OpenAI is formally entering healthcare with ChatGPT Health. OpenAI says more than 230 million people ask health-related questions in ChatGPT every week, and now those interactions are being structured into a dedicated ChatGPT Health experience. The focus isn’t diagnosis or treatment. It’s preparation: helping users understand test results, track trends from wearables, and walk into medical appointments with better questions. The most important feature isn’t the model, it’s the trust layer: -Separate health space -Compartmentalized memory -Added encryption Health conversations not used to train foundation models → This will be valuable if privacy protections hold at scale. #OpenAI #ChatGPTHealth #AIinHealthcare #DigitalHealth #HealthTech

75

94 Comments

Early sign of real-world AI application: new research shows AI has boosted labor productivity by 1.3% since ChatGPT's release. Vertical AI apps will target underused industries to broaden adoption. We are seeing it play out in Healthcare with K Health

25

1 Comment

KLAS Research recently evaluated Suki's ambient clinical AI and published a detailed report. KLAS evaluated real-world deployments of Suki’s ambient documentation across three large Epic health systems: FMOL Health, McLeod Health, and Rush University System for Health. Highlighted results: • ~27% reduction in documentation time, translating to ~3.6 hours saved per provider per month • ~$1,000+ in additional revenue per provider per month driven by improved E/M coding capture • ~$862 added revenue per user per month at another system from more complete documentation • 21% reduction in time spent on notes • 43% reduction in notes remaining open after seven days These numbers reflect gains across clinician productivity, chart closure, and revenue capture, measured using production EHR data rather than pilot metrics. This kind of validation signals where the market is heading. AI solutions are increasingly being judged on quantified time savings, financial impact and durability at scale. #AIinhealthcare #KLAS

25

1 Comment

Healthcare today is drowning in administrative burden—clinicians spending more time managing paperwork than treating patients. At Autonomize AI, we're challenging this unacceptable status quo. Our AI-driven agents are reshaping healthcare workflows by automating complex, tedious tasks. This isn't incremental improvement; it's transformative change, enabling healthcare professionals to reclaim their focus on patient care and clinical outcomes. Dive into this piece from TMC highlighting how Autonomize AI is pioneering smarter healthcare workflows: https://lnkd.in/gR_QM7fr #healthcareAI #workflows #AI

75

3 Comments