Ion Feldman

Claude Code Agent Teams will introduce agent orchestration to the masses. This is a good thing. Agent orchestration adds another gear entirely and devs are going to be floored by how fast Clause Code will shred their work. They’re also going to learn a few more things. It’s expensive. The ride is a joy. But it’s expensive. They’ll have to start learning to judge when to use it and when not. Claude Code will help here but expect early adopters to be calling out for more budget. It needs ready work. Teams are going to need a lot of work if they are going to use the new top gear. We are now past the point where execution is the bottleneck. Business decisions, design and infrastructure are now pace setting. It’s lonely in an F1 car. The car can go at speeds never previously imagined. But it only has one seat. Coordination with other devs becomes a real problem. Putting a dozen F1 cars, with rookie drivers, on the same track creates traffic jams. Weak teams will blame the cars and drive slower. Strong ones will build more racetracks and live at top speed. It’s complicated. Orchestration feels like an entirely new skilll above and beyond single threaded Claude Code. Expect a new learning curve here. A lot of devs aren’t ready for it. With orchestration now emerging, we are into the final stage of Yegge’s framework; Agent Fleets are here. That means we need a new playbook to guide us. Excited to read it.

28

2 Comments

The CMS ACCESS Model FHIR Implementation Guide is Up. My co-founder Hadi Javeed sat down with me last night and we pulled up the specs. Four API operations, seven subscription event types, twelve FHIR profiles. More impressive than that- he created an interactive sandbox for builders!! We mapped each operation against the clinical tracks, walked through the payment model line by line, and sketched out how we would integrate this into our software infrastructure. Then we wrote it up on The Techy Surgeon! I still believe the most compelling component to this model is the partnership architecture. An MSK practice collecting $180 per care episode while forgoing downstream surgical revenue of FFS billing is thin. Impractical as a standalone play. But an MSK practice or PT group that functions as the specialty care node in a broader value-based network -- receiving referrals from ACOs, coordinating with rehab partners for triage and PROM collection, with co-management payments flowing to the upstream entities that feed the pipeline and expanding their tech enabled capabilities -- that's a different model entirely. The ACCESS Model doesn't reward isolation. It rewards coordination. And the participants who build the network around them will be the ones who make the economics work. We're in the MSK and Behavioral Health tracks. The MSK improvement-only design with early success reporting at Day 185 maps directly to what we've built at RevelAi Health -- PROM trajectory monitoring, clinically informed AI powered patient engagement, distributed care coordination, partner data exchange. And the BH track opens an adjacent market where the same partnership-dependent architecture applies and the need for coordination infrastructure is even more acute. And yes, without AI -- it becomes impossible for this program to work for the ACCESS participant. Will this program actually enhance access? That remains to be seen, but we're determined to try. The full analysis and sandbox on The Techy Surgeon (link in comments). If you're an ACO, PT group, or specialty practice exploring ACCESS and want to talk partnership structures: hello@revelaihealth.com or visit us at the MSK ACCESS website (link in comments) #ACCESSModel #CMS #ValueBasedCare #FHIR #MSK #BehavioralHealth #DigitalHealth #CareCoordination

40

7 Comments

Monday Musing: The dust hasn’t settled yet on the HHS reorganization, and we don’t yet know what things will look like six months or a year from now. But one scenario we need to prepare for is this: a leaner HHS that takes a lighter touch, with fewer new rules, less hands-on leadership, and limited capacity to convene the industry as it has in the past. If that’s where we’re headed, then it raises an important question. Who will keep us aligned and moving forward as an industry? Multi-stakeholder initiatives (MSIs) may be the answer. In fact, they may need to take on a bigger role than ever. Groups like HL7 Da Vinci Project, HL7 FHIR at Scale Taskforce (FAST), Gravity Project as well as non-accelerators like The Sequoia Project, Massachusetts Health Data Consortium (MHDC) and others have long brought together diverse voices across public and private sectors to tackle the toughest issues in health IT. They’ve helped us create scalable solutions, reduce fragmentation, and break down silos. In many cases, federal agencies have been active participants, helping shape priorities and drive consensus. Whether that level of federal involvement continues is still unclear. What is clear is that we cannot afford to lose momentum. If MSIs are going to step up and help fill any gaps, we may need to rethink how they operate, how inclusive they are, and how easy it is for new voices to get involved. The tents probably need to be bigger. A great example of the expanding role of MSIs is this week’s Da Vinci event, “Strategy to Execution: Better Prior Authorization by Integrating Operations, FHIR and CMS-0057.” It’s designed to help organizations jumpstart efforts to automate prior auth and meet CMS-0057 requirements—and it wouldn’t be happening without members who stepped up to sponsor and lead. The FAST Security Event in January is another strong example. MSIs are already driving progress, and they need more collaborators to keep the momentum going. https://hubs.li/Q03hZXyB0 For organizations that haven’t participated before, this may be the right time to reconsider. You don’t have to become a member to have an impact. There are public workgroup calls, testing opportunities at Connectathons, and open comment periods where fresh perspectives are welcome and needed. At POCP, we’ve always believed that MSIs are essential for tackling the big, messy challenges no single organization can solve on its own. That’s why we serve as the program management backbone for several leading MSIs, including Da Vinci, Gravity, FAST & MHDC. These aren’t just contracts for us. They’re passion projects. MSIs give the industry a collaborative space to sort through what’s next, especially when the policy outlook is murky. We’re proud to help keep that momentum going. Don't wait for the dust to settle. Stay focused & stay engaged.

18

We talk a lot about FHIR APIs. But if your automation strategy relies only on APIs, you have a 40% failure rate. Real-world access means handling the "long tail" of payers who lack APIs, whose APIs are down, or who require a specific fax form. This is where our "Agentic" approach wins. We prioritize the API (the fast lane). But when that isn't available, our system seamlessly falls back to AI-driven faxing or even AI-assisted phone agents. We don't just "connect"; we complete the task. That is the difference between "Interoperability" and "Results." #Interoperability #HealthTech #PriorAuthorization #AI #AgenticAI

2

2 Comments

One question when buying clinical software: Where does accountability live? It breaks when context is split, ownership is unclear, or escalations rely on “someone noticed.” Look for structured handoffs, clear triggers, and audit trails not nice-to-have, a risk requirement. DM your letter for the checklist.

5

Thrilled to announce this big step forward in surgical inventory automation — seamless integration is now live between ReadySet Surgical and Ascendco Health. Built to streamline today and scale tomorrow.

17

1 Comment

When I look at the CMS roadmap for 2026 and 2027, it’s clear: healthcare industry has reached a point of no return. While much of the current discussion focuses on the "how" of FHIR API implementation, the real strategic challenge lies in the "why." Why are clinical decisions still being stalled by data silos in an era of instant connectivity? The mandate for a 72-hour turnaround on urgent Prior Authorizations is more than a regulatory hurdle. It is a real call to eliminate the friction that costs time, money, and, ultimately, patient outcomes. In our experience , meeting these deadlines shouldn't be treated as a painful compliance exercise. It is a unique opportunity to finally re-engineer the fragmented workflows that have slowed the industry down for decades. 2026 is here! The time to align technical infrastructure with clinical reality is now. How is your leadership team preparing for the shift from "storage" to "action"? #HealthTech #CMS #Interoperability #DigitalHealth #CEOInsights #ClinicalWorkflow #HealthIT #ITIRRA #AIDA

47

20 Comments

🚀 Launching Centered Care’s Partner-Facing API   Interoperability has been core to how we built Centered Care from day one.   Our Care Concierge model depends on unified data. CareFlows depends on secure integrations across EHRs, eMARs, ambient sensors, nurse call, and other operator point solutions.   Until now, most of that integration work has been outbound — us building to others.   Today, we’re opening that door in the other direction with our Partner-Facing API.   From a technical standpoint, this release focuses on:   🔹 Standards-based architecture designed for secure healthcare environments 🔹 Structured, versioned endpoints with clear documentation 🔹 Tenant-aware authorization to support multi-community deployments 🔹 Event-driven workflows to enable real-time data exchange 🔹 Built-in auditability aligned with regulated care environments   This isn’t about exposing data for the sake of it. It’s about enabling controlled, secure extension of the CareFlows platform — so partners can build meaningful workflows on top of the same infrastructure that powers our AI and care coordination tools.   We’ve always integrated into the ecosystem. Now we’re enabling the ecosystem to integrate with us.   Interested in building together? 👉 https://lnkd.in/gmZ77h-e   #SeniorLiving #HealthTech #APIs #Interoperability #HealthcareInnovation Holly Hrabik Rocky Samuel James Lydiard Joelle Poe Steve Weissblum Forrest Walker

21

1 Comment

3 reasons why we build software and service as one team: We don’t build tech *for* operations at Tandem. We build tech *with* operations. Here’s why that matters; 1) Fix friction at the root. If a vet has to click 5 times to do something that should take one, we fix it. Not just because it’s “inefficient”. But because it creates cognitive load, slows care delivery, and compounds under pressure. Less friction = faster throughput, fewer errors, and lower burnout. 2) Close the loop between builders and users. The person who writes the feature talks directly to the person who uses it. That means issues surface faster, fixes are more accurate, and nobody ships something that “technically works” but operationally fails. It’s not just “faster iteration” - it’s tighter alignment with reality. 3) Scale the system, not the sprawl. New clinics, new roles, new workflows. We don’t layer on SaaS duct tape. We evolve a system designed to flex with operational growth. That means fewer integration points, tighter control, and a consistent experience from core to edge. We don’t treat tech like a layer. We treat it like a nervous system. Because when you’re delivering care in the real world, you can’t afford delay, dissonance, or dropped signals. That’s why we built Tandem the way we did: One system. One team. Zero silos. Because when service and software evolve together, the whole organism gets smarter, faster, and stronger over time. ------------- Hi, I’m Doug. I’m the Chief Technology Officer at Tandem. We’re engineering the backbone for modern pet care. Follow me for posts on: * How we build tech with ops, not for them * How we build tech tools that scale cleanly * Scaling hybrid companies without SaaS bloat

27

2 Comments

For many years at Canvas Medical we resisted the “headless EMR” label ⛔ It implied a back-end data architecture and developer tools (correct, we do have those) without a front-end interface (WRONG — Canvas is also the certified EMR front end for our customers). Turns out we fought the good fight and won it at a perfect time. The market has woken up to the need for a programmable UI, not a headless EMR. Ours is built with our server-side SDK, accessible to end users to customize and extend with AI-assisted coding. No developer needed. ✅ It’s only because we stuck to our guns, refusing to sell a product without our certified EMR front-end, that we're in this position to win. And now, I can’t see a world where the EMR GUI as we know it survives. The destination is an entirely new paradigm for interaction between the clinician, the patient, their data, and the steps to provide and document care. We will use natural language (voice, written text, inline commands) to prompt, instruct, and monitor agents taking action on your behalf. Examples: 🧠 Behavioral health — a psychiatrist wraps up a telehealth visit and agents schedule the GAD-7 follow-ups at 2 and 6 weeks, submit the prior auth for CBT, and e-prescribe buspirone 10mg twice daily. 💊 Chronic care management — clinician dictates “Refill the metformin, schedule a follow-up in 3 months, and flag if their A1C is overdue” and three workflows are executed. ⚖️ Weight management — a care team lead says “show me every patient outside of their target weight” and an agent evaluates smart scale data, trends results against patient-reported and physician-documented targets, builds a custom visual dashboard on the fly, and automates follow-up tasks for the patient. There will be a tremendous shift in expectations and capabilities in healthcare UX over the next 12-18 months, and our team is ready for it.

68

7 Comments

🚀 We absolutely love Apache Camel! At 314e Corporation, Apache Camel is the powerhouse behind our SaaS-based healthcare integration engine 'ZSegment'. It quietly runs 10K–20K messages per second, connecting EHRs, labs, payers and more in real-time. 🧠 Why Camel? Because it just works. It gives us: ✅ 300+ built-in components (REST, HL7, JDBC, Kafka, SFTP, and more) ✅ Rich support for EIP patterns like splitter, aggregator, wire tap, content-based routing ✅ Clean DSLs in XML, YAML for readable and maintainable routes ✅ Seamless data transformation using XSLT, JSONata, custom beans ✅ First-class error handling, retries, circuit breakers, DLQ ✅ Production-ready observability via Prometheus, OpenTelemetry, and MDC logging 🎨 We’ve taken it further by integrating Camel with Apache Kaoto, We allow end users to visually design and deploy integration routes without writing a single line of code. That’s the kind of flexibility and extensibility Camel offers. 🤖 And now, we’re super excited to see AI and LLM enhancements making their way into the Camel ecosystem. The idea of intelligent, self-optimizing integrations is no longer futuristic. it's becoming real. 🔥 Our dream feature? Native step-level message queuing, enabling pause/resume, reprocessing, and full decoupling between processors. That would truly enable next-level resilience and observability. 👏 Huge shoutout to Claus Ibsen and the entire Apache Camel community. You’ve built something incredibly powerful, modern, and elegant and it’s time more teams discovered the magic of Camel. ❤️🐪 #ApacheCamel #SaaS #HealthcareIT #IntegrationEngine #ZSegment #Kaoto #FHIR #HL7 #Kafka #EventDriven #LLM #OpenSource #ClausIbsen #WeLoveCamel

45

1 Comment

Everyone is talking about AI scribes.  Fewer people are talking about system design.  The real questions I’m discussing with health systems right now: • How does AI documentation integrate into your EHR without adding review fatigue?  • Where does coding ownership live, and how is quality governed?  • What does a responsible human-in-the-loop model actually look like at scale?  • How do you measure ROI beyond time saved? AI in healthcare isn’t just about speed. It’s about architecture. If documentation and coding aren’t aligned from the start, you risk creating a faster front end and a more complicated back end. At HIMSS26, we’re going deeper than demos. We’re talking about:  ✔️ Designing AI with accountability built in  ✔️ Mid-revenue cycle integration that reduces denials, not just keystrokes  ✔️ Building physician trust through transparent accuracy models  ✔️ Creating products that scale across specialties and governance structures This space is evolving quickly. Strategy matters more than hype.  If you’re evaluating solutions, or rethinking an early deployment, I’d welcome the conversation! 📍 Booth #747  📅 March 9–12 | Venetian Expo Center #HIMSS26 #HealthIT #AIinHealthcare #EngineeringLeadership #ClinicalAI #RevenueCycle

51

1 Comment

Delivering AI-assisted ICD10 coding, AlphaCoding is built on AI models and guided by Centers for Medicare & Medicaid Services (CMS) standards. #meaningfulai #healthcaretech #augmentedintelligence #revenuecycle

15

1 Comment

🌍 Proof of Concept: Standardizing Oncology Care with FHIR. I'm excited to present Oncology Implementation Guide;a proof of concept developed to demonstrate how HL7® FHIR® standards can streamline oncology care. This guide provides a structured framework for capturing, managing, and exchanging oncology data—spanning prevention, diagnosis, treatment, and follow-up—enabling interoperability across healthcare systems. Explore the guide here: https://lnkd.in/ggMugMUj #Kenya #OncologyCare #FHIR #DigitalHealth #HealthTech #Interoperability #CancerCare #ProofOfConcept

11

5 Comments

Flexpa just published our response to CMS and ASTP/ONC's request for information on the state of healthcare interop Drawing from our experience operating the nation's largest payer API network, we provide concrete recommendations for improving Blue Button 2.0, addressing USCDI implementation gaps, standardizing digital identity, and building sustainable network infrastructure Big thanks to Angela Liu for driving this internally and to Brendan Keeler for draft reviews https://lnkd.in/g6J-QAVJ

21

To me this exemplifies a combination of shifts that are happening in AI. You hear this from Satya, Meredith and others. If your proprietary data and business process and acumen are your moats why would you give this away to systems that are training on it? (Don't come at me that enterprise accounts solve this bc the tech world is riddled with say/do mismatches thru history and it's a risk paradigm equation not an accusation). So if you're well funded (see payers) building your own systems/agents/model/solutions on your own data/process become a logical choice. It then began the question set of what else you can internalize, maybe even the models themselves aka open source etc .. Couple this with the friction to build is perpetually lowering. See AI beyond the vibe code slop. Likely there will be an orchestration of everything including LLMs that are used private vs public based on the sensitivity of the data and the risk profile of having that data outside your moat. The classic "and vs or" paradigm. To quote Trading Places... "Why can't we have both?" Factor the emotional/relevance/ego/opportunity matrix of CIO relevancy into this and I think it's just the beginning. The next thought exercise is what are the broad downstream implications across the system vs just the vendors which is zeroed in on below. My 3.1415 cents ;) Happy Friday.

21

5 Comments

Surprisingly, EMR pricing is still wildly misunderstood. 🤯💰 And a common belief goes like this: EC2 is the real cost. 💻 EMR is just a thin, cheap management layer on top. 🪶 That’s… not how the math works. ➗❌ If you run 10 × r8g.4xlarge nodes on EMR, here’s what happens: • On-Demand → ~80% of your bill is EC2, ~20% is EMR • Reserved (50% EC2 discount) → EMR becomes ~33% of total • Spot (70% EC2 discount) → EMR becomes ~45% of total Notice the pattern? With on-demand instances, EMR markup looks ~20%. Reasonable. 👍 But once teams move to Spot (which many do), the markup overhead grows close to 50%, with no real extra value. 💸📈 Why? 🤔 Spot can be up to 90% cheaper than on-demand. It drops EC2 line item fast. 📉 But EMR pricing doesn’t. 🚫 On-demand EC2 is the cost base. 🧮 So as you optimize compute with Spot, “managed Spark” quietly becomes the largest tax on your cloud bill. 🧾💵 I spend 10s of hours on calls with teams running large Spark installs. 🔥This is exactly why most customers I talk to want to at least explore running Spark themselves on Kubernetes instead of EMR. 🚢 Not because they love undifferentiated ops work — but because they want control over the economics. ⚙️📊 #spark #aws #cloudcost #kubernetes

459

16 Comments

Great to see Cibolo Health continue its mission of supporting independent rural hospitals by introducing MyoVista Insights as part of their technology ecosystem. Their work helping hospitals stay independent while strengthening access to modern, practical technology makes a real difference, and it’s exciting to see our ECG Management System become part of that broader effort to support sustainable rural care!

11

From 2010 through 2015, medical coding changed significantly. Code structures expanded, character length increased, and documentation requirements became more detailed. Many clinicians felt ICD-9 supported care delivery reasonably well. After the transition, healthcare costs increased, yet outcomes did not clearly improve. The added complexity affected everyone: - physicians - billing teams/ RCM firms - EHR systems - insurers - patients Documentation grew heavier as systems grew more detailed and complex. The question many clinicians still ask is simple. Question: If you could design the next ICD version, what would you simplify first?

3