Out of the FHIR Podcast

Gene Vestel

Talking about FHIR evestel.substack.com

  1. Nurses need AI too, and how it needs to be deployed at scale to ease administrative burden.

    Jun 12

    Nurses need AI too, and how it needs to be deployed at scale to ease administrative burden.

    Gene Vestel sits down with Michelle Skinner, Chief Clinical Executive at TeleTracking, to unpack the operational side of healthcare execution. Michelle is a nurse by background with an MBA who spent decades running emergency departments and trauma centers before moving into health-tech leadership. In this episode, she breaks down how TeleTracking a rare, 35-year-old owner-operated pillar in a sea of PE-backed digital health firms is using computational twin technology to radically optimize hospital operations without breaking clinical workflows. Listen now on YouTube, Spotify, and Apple Podcasts. We discuss: * The Reality of Hospital Patient Flow: Why emergency department boarding is a symptom of systemic operational gridlock, not an ER failure. * Computational Twins in Action: How simulating real-time capacity scenario planning can drop a hospital’s length of stay by over a full day. * The Nursing Cognitive Load Crisis: Why AI strategies must pivot from administrative data logging to keeping nurses at the bedside. * The Business vs. Care Matrix: How having clinical leadership embedded directly within engineering teams alters how code is written. * The Imperative of Rural Healthcare Access: Why urban-centric health models collapse when applied to regional communities. My 3 Biggest Takeaways from This Conversation 1. Hospital crowding is a patient flow problem, not a capacity problem When patients are held in emergency department hallways for days, the default reaction is often to blame ER throughput or demand more physical beds. The tactical reality is that ER boarding is a lagging symptom of poor downstream operational orchestration. When a hospital cannot cleanly coordinate transitions from the post-anesthesia care unit (PACU) to intensive care or general medical floors, the entire pipeline backs up. TeleTracking’s deployment of computational twin software builds a predictive digital replica of a facility’s entire capacity landscape, running scenario trade-offs 48 hours in advance. The result isn’t just arbitrary data tracking; it’s a systematic blueprint that has driven over a 50% reduction in ED holds while simultaneously allowing hospitals to scale up overall volume. 2. If technology doesn’t actively reduce a nurse’s cognitive load, it’s a failure While ambient listening models have made incredible strides in reducing “pajama time” and burnout metrics for physicians, the wider health-tech ecosystem has largely ignored the operational burden placed on nursing staff. Nurses have been turned into administrative traffic controllers spending critical clinical hours manually tracking down bed availability, coordinating discharge paperwork, or calling radiology to check on exam slots. We must evaluate new technology platforms through a singular, hyper-focused product lens: Does this give clinical hours back to the patient, or does it add friction to the system?. If it doesn’t systematically strip administrative steps out of the clinical loop, it shouldn’t be built. 3. Engineering teams need immediate clinical guardrails A distinct trap for tech-first companies entering healthcare is treating healthcare metrics as abstract, unfeeling datasets or lines of code. True product maturity occurs when engineering squads have an operational bridge to the clinical frontline. Having nurses embedded directly into development processes creates a permanent shift in engineering empathy. When developers understand that a minor database lag or a clunky workflow pattern directly delays a bed placement for a critical trauma patient, the quality of execution spikes. We must build software with a human-in-the-loop mentality, ensuring code serves the explicit, real-world workflow realities of active caregivers. Where to find Michelle Skinner & TeleTracking: * LinkedIn: Michelle Skinner * Website: TeleTracking If you found this operational breakdown valuable, consider subscribing to Out of the FHIR for weekly technical product leadership deep dives. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    31 min
  2. Navigating the Shift to Bulk Data and AI

    Jun 1

    Navigating the Shift to Bulk Data and AI

    Ron Urwongse, co-founder of Defacto Health, and I sit down to break down the rapid shifts hitting CMS regulations, the transition from standard FHIR APIs to national bulk data datasets, and how agentic AI workflows are compressing engineering timelines from months to an afternoon. Listen now on YouTube, Spotify, and Apple Podcasts. We discuss: * The Bulk Data Pivot: Why CMS is expanding beyond endpoint APIs into massive bulk NDJSON files for Medicare Advantage plans. * The National Provider Directory Ecosystem: A technical audit of the new data release, where it shines, and where the logical models are still failing. * AI as an Engineering Accelerator: How teams are using agentic workflows (like Claude Code) to build production-ready validation engines overnight. * Smart Scheduling Links: The inevitable roadmap toward universal, consumer-centric open appointment booking. * The CMS Feedback Loop: Why the newly established CMS Health Tech Ecosystem Slack channel is radically altering how regulations are refined in real time. My 3 Biggest Takeaways 1. Compliance cycles have compressed from six months to a single weekend In the legacy enterprise playbook, updating a platform to conform with newly dropped technical implementation guides took a quarter or more of roadmap planning. Today, that layout is dead. Ron noted that when CMS dropped updated technical guidance on a Friday afternoon, multiple forward-thinking payers had already fully conformed by Monday morning. The differentiator isn’t engineering headcount; it’s the shift toward AI accelerators. If your senior architects aren’t actively feeding CMS Implementation Guides into tools like Claude Code to interpret, write, and deploy schemas, you are building an operational bottleneck. 2. We are transitioning from simple Master Data Management to Federated Graphs The industry has long clamored for CMS to run a centralized database as a mastered system of record. Instead, the tactical reality looks much more like a federated graph across hundreds of independent nodes. CMS isn’t attempting top-down data cleansing; they are supplying the network scaffolding to link provider organizations, practitioners, endpoints, and digital footprints. Payers must now prioritize internal accuracy auditing because upcoming mandates like the Real Health Providers Act will require plans to publicly score and publish the validity of their directory data. 3. Open scheduling is the ultimate bottleneck for value-based care Up to 75% of open care gaps remain unfilled simply because of the high friction involved in patient engagement such as transcribing an identical medical history onto a 40-page clipboard during an intake cycle. Universalizing lightweight specifications like Smart Scheduling Links originally built to aggregate vaccine availability during COVID will allow insurance directories to natively embed real-time booking slots. The monetization model still needs guardrails to protect providers from high platform fees and patient acquisition gaming, but opening up EHR scheduling data to the wider ecosystem is an absolute necessity to drive actual consumerism in healthcare. Deep Dive: Auditing the National Provider Directory The launch of the National Provider Directory marked a major milestone for healthcare data liquidity, but looking under the hood reveals clear technical hurdles that the developer community is currently solving. [ Practitioner ] │ Is associated with ▼ [ Provider Organization ] ── Publishes ──► [ Bulk NDJSON Dataset ] │ │ Resolves endpoint to │ Contains ▼ ▼ [ Patient-Centric Endpoint ] ◄── Audited by ── [ AINPI.dev Engine ] The Architectural Gaps in the NPD Release To test the real-world utility of the new data, Gene imported the entire publicly available directory into an open-source tool built over a weekend to evaluate and audit conformance: AINPI.dev. The audit highlighted several distinct areas where the logical models require iteration: * Endpoint Association Confusion: There remains an ongoing architectural debate within CMS working groups regarding where endpoints should sit logically. Attaching a FHIR connection endpoint directly to an individual practitioner creates massive, unmanageable data duplication. The correct semantic approach maps endpoints strictly to the Provider Organization, which then establishes relationships down to the underlying practitioners. * The Specialty Taxonomy Mess: There is still no clean, unified consensus on processing specialty codes. Payers are left navigating multiple conflicting sources of truth published across PECOS, NPPES, and specialized CMS charts, leading to distinct fragmentation in search results. * Missing Endpoints: The front door to patient-directed data access relies on clean endpoint visibility. Currently, a vast percentage of active provider organizations feature zero mapped digital endpoints, making true interoperability a fragmented experience depending entirely on where a patient lives. Where to find Ron Urwongse & Defacto Health: * LinkedIn: Ron Urwongse * Website: De facto Health Referenced in the show: * The Open-Source Audit Tool: AINPI.dev * The Agentic Healthcare Assistant Concept: HealthClaw.io * The Technical Repository Framework: Smart Health Connect * Gene’s AI Builder Cohort: FHIRIQ Workshop If you found this breakdown valuable, consider subscribing to Out of the FHIR for weekly deep dives into technical health-tech leadership. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    39 min
  3. State of Prior Authorization with Mark Fleming (Availity)

    May 28

    State of Prior Authorization with Mark Fleming (Availity)

    Mark Fleming is Senior Director of Prior Authorization, Interoperability, and Portal Solutions at Availity, a leading healthcare clearinghouse and data network. With over 25 years of experience in healthcare IT and revenue cycle management starting back when Epic had only 500 employees Mark is one of the industry’s foremost experts on modernizing the administrative friction between payers and providers. Listen on YouTube, Spotify, and Apple Podcasts. We discuss: * Why a staggering two-thirds of prior authorizations are still stuck on manual faxes, phone calls, and isolated web portals. * The massive structural shift behind the CMS-0057 mandate and how standardized FHIR APIs will force standard authorization timelines from weeks down to a strict 72-hour window. * Moving from isolated transactions to real-time clinical transparency—letting providers query exact documentation and medical policy rules directly inside their EHR at the point of care. * How digitizing clinical data allows modern AI platforms to parse requirements instantly, letting patients schedule sensitive procedures within days rather than waiting for weeks. * The daunting scaling bottleneck of point-to-point connections, why the average health system routinely deals with 40 to 80 distinct payers each month, and why the industry must look toward centralized networks over customized developer builds. My biggest takeaways from this conversation: * The Stagnant State of Healthcare Administrative Friction: Despite immense technological progress in other areas of our daily lives, healthcare transactions remain stubbornly legacy. Currently, only about a third of prior authorization transactions utilize automated electronic X12 standards; the remaining two-thirds are split evenly between manual payer portals and decades-old faxes and phone calls. * The Clinical Shift of CMS-0057: The incoming federal FHIR API standards mandate a massive operational pivot. Historically, providers gathered documentation and “threw it over the fence,” resulting in back-and-forth rejections because of highly specific medical policies. By introducing Coverage Requirements Discovery (CRD) and Documentation Templates and Rules (DTR) directly into the point-of-care workflow, providers will instantly know exactly what clinical information is required before a submission occurs. * Real-Time Automated Care Approvals: Integrating real-time bi-directional FHIR streams with clinical decision platforms paves the way for immediate automated processing. By utilizing modern AI architectures to evaluate digital clinical datasets against explicit payer criteria, current production implementations (like Availity’s authAI tool) are already approving up to 78% of initial submissions within 60 seconds. This eliminates the safety buffer where providers schedule slots weeks out just to wait for a manual determination. * The Network Scalability Challenge: Point-to-point custom integrations simply do not scale for provider ecosystems. Because an average mid-sized health system must route documentation to 40 distinct payers every single month and larger ones route to up to 80 building out separate point-to-point lines of communication is logistically unfeasible. Centralized networks must step in to act as translation and trust clearinghouses to standardize operations between varying EHR versions and complex payer architectures. * The Cost-Burden Equivalence: Transitioning away from legacy administrative manual procedures can remove immense financial waste from the healthcare system. Current metrics show that a manual submission for a prior authorization costs an average of $9.00 per submission, whereas an fully electronic transaction utilizing standardized networks drops that cost to just $0.25. Where to find Mark Fleming: * LinkedIn: Mark Fleming on LinkedIn * Website: Availity Official Portal Referenced in the show: * CMS-0057 (Interoperability and Prior Authorization Final Rule): CMS Official Summary * CMS-0062 (Proposed Rule Expanding FHIR to Medications): Federal Register Rule Details * HL7 Da Vinci Project & Burden Reduction Group: Da Vinci Framework Overview * Trebuchet Project: Trebuchet Connectivity Infrastructure Initiative * HealthClaw: Open Source Fire Data Quality Assessment Layer * Epic Systems: Epic Corporate Page * Athenahealth & Humana Joint Case Study: Reference Implementation Learnings * Medical Group Management Association (MGMA) Survey: Prior Authorization Burden Metric Report * TEFCA (Trusted Exchange Framework and Common Agreement): HealthIT.gov TEFCA Details * FAST (FHIR At Scale Taskforce) Security Initiative: ONC FAST Security and Identity Working Group This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    44 min

  4. May 5

    The AI Paradox: Why LLMs in healthcare actually require more structured data, not less | Ewout Kramer & Ward Weistra (Firely)

    Ewout Kramer is the “head nerd” and founder of Firely, and one of the original architects of the FHIR (Fast Healthcare Interoperability Resources) standard. Ward Weistra leads data modeling tools at Firely and curates the content for FHIR DevDays. Together, they have spent over a decade transitioning healthcare from messy legacy standards to a modern, developer-friendly ecosystem. Listen on YouTube, Spotify, and Apple Podcasts We discuss: * The Origins of DevDays: How a kitchen-table meetup 15 years ago turned into the canonical global event for health tech developers. * AI as a FHIR Catalyst: Why AI doesn’t replace the need for structured data—it actually makes the “Step Zero” of standardization more critical. * The Human Side of Interoperability: Why building trust between competitors is more important than the JSON schemas themselves. * The EHDS and Global Regulation: How the European Health Data Space and U.S. Cures Act are forcing a “bottom-up” shift in software engineering. * Community & Culture: From the “Nerd Awards” to student tracks and even forming a “FHIR band.” My biggest takeaways from this conversation: 1. Standardization is only “Step Zero” A common mistake in health tech is assuming that once data is standardized into FHIR, the job is done. Ewout argues that standardizing data is merely the baseline. The real work and the focus of this year’s DevDays is extracting meaning. This involves moving from static data to national-scale workflows, clinical decision support (CQL), and figuring out how data travels with a patient across institutions without losing context. 2. The AI Paradox: More AI requires more structure, not less There is a contrarian view that LLMs are now so good at reading unstructured text that we no longer need to invest in the “hard manual work” of FHIR mapping. Ward and Ewout share the results of their global “State of FHIR” survey, which suggests the opposite. Government leaders and engineers agree that AI actually increases the demand for FHIR. To prevent hallucinations and ensure clinical safety, AI agents need the “guardrails” of a structured schema to reason over data reliably. 3. Interoperability is an “Inter-human” problem Technology rarely solves the hardest problems in healthcare; communication does. Many data mapping issues stem from “decades of legacy data” where the original developers are gone, and no one knows how a specific field is used in a specific hospital. Solving this requires what Ward calls the “trust layer” getting competitors in the same room to agree on implementation guides so that the software actually talks to each other in the real world. 4. Regulation provides the “Bottom-Up” power The EHDS (European Health Data Space) is set to mandate that by 2030, every piece of health software in the EU must implement the same interfaces. While this is a top-down mandate, it empowers the “single developer” within a large organization to convince their management to do the right thing. It shifts FHIR from a “nice-to-have” innovation project to a legal requirement for market entry. 5. The “Tiny Core” of the FHIR Community Similar to how great products have a “tiny core” (like the Notion block or the GitHub PR), the FHIR community relies on a core group of “head nerds” who have grown from junior devs to national thought leaders over the last 15 years. Events like DevDays maintain this culture through informal “nerd-outs” like automating pet turtle enclosures or building “Back to the Future” garage doors ensuring the community remains focused on building, not just policy-making. Where to find Ewout and Ward: * Ewout Kramer: LinkedIn * Ward Weistra: LinkedIn * Firely: https://fire.ly Referenced: * FHIR DevDays (June 15-18, Minneapolis): https://devdays.com https://www.hl7.org/ * The EHDS (European Health Data Space): Official Overview * Vivian Lee’s “The Long Fix”: Book Link * Kill the Clipboard Initiative: https://killtheclipboard.com * Simplifier.net: FHIR Registry Important Disclamer: DevDays is organized by both HL7 International and Firely This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    48 min

  5. Apr 17

    Why TRUST is the most important element in enabling Healthcare Interoperability

    Ever wondered why security is key for FHIR interoperability? Here’s what you need to know! Tag someone who needs to understand this! In the healthcare world, we face an ongoing challenge: scaling security effectively. Fast security initiatives aim to solve this age-old problem by automating processes that once required human intervention. With dynamic client registration, organizations can connect without lengthy manual steps, ensuring quicker and more secure data exchanges. But it’s not just about speed trust is essential. The FAST Security IG establishes a foundation for technical trust, allowing seamless interactions across the FHIR ecosystem. Imagine a world where you can access healthcare data without the friction of individual Portal Logins, or organization data sharing without CSV files or endless emails! SMART on FHIR solved the auth flow in 2015. It did not solve trust at network scale. Plaid proved network trust in finance. Visa proved it at the point of sale. Sabre and Amadeus proved it in travel. All three work because there is a cryptographically signed identity you present once, and the whole network honors it. UDAP is that identity layer for healthcare. SMART handles the authorization. FAST Security IG is the rulebook that composes them into something that can actually carry TEFCA, CMS-Aligned Networks, Patient Right of Access, and the full B2B / B2C / B2B2C surface without requiring a bilateral deal for every new endpoint. Let’s keep this conversation going what are your thoughts on the future of healthcare data exchange? Listen to the full episode for deeper insights! This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    51 min

  6. Mar 25

    Episode 24 - HIMSS 2026 Roundup

    I just got back from HIMSS 2026, and my ears are still ringing mostly from the sound of vendors pivoting to “AI-First” strategies overnight. In Episode 24 of Out of the FHIR, we’re skipping the press releases. I’m breaking down the three biggest lies being told on the conference floor and highlighting the few innovators who are actually using FHIR to move the needle. The “Spicy” Summary: * Stop calling it AI if your data is a mess. Clean FHIR data is the tax you haven’t paid yet. * Interoperability isn’t a tech problem anymore. It’s a “who-gets-paid” problem. * The patient is still an afterthought. Until we stop building silos, “patient-centricity” is just a marketing slogan. Listen to the conversations from the show floor! While at HIMSS I had a chance to look at the National Health Digital Architecture discussion paper released by the National Academy of Medicine. Below are my thoughts on it. The National Academy of Medicine’s (NAM) perspective on a National Health Digital and Data Architecture serves as a direct counterpoint to what is currently unfolding at HIMSS 2026. While NAM proposes a unified, foundational “operating system” for healthcare data, the reality on the conference floor is a fragmented landscape of competing interests. Here is a review of the NAM strategy and why it is not gaining traction in the current HIMSS environment. The Suggested Strategy: A Unified National Architecture The NAM strategy calls for a fundamental shift in how health data is managed: * A Shared Public-Private Utility: Treating data infrastructure as a “common good” (like roads or the power grid) rather than a competitive advantage. * Standardized Semantic Foundation: Moving beyond just “sharing files” to a universal understanding of data meanings (FHIR-native, semantically aligned). * A Single “Architecture of Trust”: A unified framework for identity, consent, and security that removes the need for thousand-point-to-point connections. * Separation of Data from Applications: Storing data in a vendor-neutral layer so that clinical apps can be swapped in and out without data loss. Why This Isn’t Happening at HIMSS 2026 1. The “FHIR-Washing” of Legacy Silos Strategy: Standardized semantic alignment. Reality: While every booth at HIMSS 2026 prominently displays “FHIR,” most vendors are still “FHIR-washing.” They use FHIR as a thin translation layer on top of their proprietary, legacy databases. This maintains the vendor lock-in that the NAM strategy aims to destroy. True digital transformation requires data to be FHIR-native from the start, but vendors have no financial incentive to rebuild their core architectures. 2. The AI Distraction (Infrastructure vs. Shiny Objects) Strategy: Laying the foundation first. Reality: HIMSS 2026 is dominated by Generative AI. Companies are racing to sell the “shiny” end-product (AI clinical scribes, predictive models) while ignoring the “plumbing” (data architecture). AI built on the current fragmented architecture results in “Black Box Silos” AI that only works within one specific EHR or payer system, which is the exact opposite of NAM’s vision for a national learning health system. 3. Competitive Moats vs. Common Utilities Strategy: Infrastructure as a public-private utility. Reality: In the HIMSS exhibit hall, data is still viewed as an asset to be guarded, not a utility to be shared. The business models of many HIMSS “Anchor Tenants” (large EHR and data aggregators) rely on being the gatekeepers of patient data. There is significant resistance to a national architecture that would commoditize the data layer and force companies to compete solely on the quality of their software applications. 4. The “Information Blocking” Dance Strategy: Seamless interoperability by design. Reality: Despite CMS and ONC mandates, “information blocking theater” is in full effect. Companies are technically compliant but operationally obstructive charging high API fees or creating complex workflows that discourage third-party integration. The NAM’s “Architecture of Trust” is impossible to build when the underlying participants are still incentivized to keep data within their own ecosystems. This is why the NAM emphasizes that surface-level technical solutions alone won’t work. You have to resolve the underlying incentive misalignments through policy levers, shared commitment frameworks, and cross-sector collaboration. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    38 min
  7. Episode 23 - Max Nussbaumer

    Feb 4

    Episode 23 - Max Nussbaumer

    In this episode, I sat down with Max Nussbaumer to talk about why he founded Max Health, why he’s bringing his work from Europe into the U.S. market, and what he believes is about to change in healthcare faster than most people realize. Max has deep roots in the FHIR ecosystem from academia at the Technical University of Munich, to Firely, to years of consulting across Europe. Now he’s building something bigger: a company focused not just on FHIR infrastructure, but on how that infrastructure enables real, visible impact for patients through AI, Smart on FHIR, and even wearable integrations. But the real theme of this conversation wasn’t “FHIR adoption.” It was this: Healthcare is about to be reshaped not by standards alone, but by what AI can now do with standardized data. FHIR Is No Longer Niche Max pointed out something many of us have felt over the past year: FHIR used to be a niche standard discussed at connectathons and DevDays.Now it’s the plumbing behind consumer AI health integrations from tools like ChatGPT and Claude. For the first time, health data is entering general-purpose reasoning systems that people use every day. That changes the game. Not because FHIR is new, but because AI can finally use FHIR data at scale. “Software Is Inflationary” One of Max’s best lines: What used to take a year to build can now be built in days. This is showing up at hackathons, startups, and even among 18-year-old builders experimenting with AI, spatial awareness, and real-time health event detection. The barrier to building healthcare software has collapsed if you understand the underlying problems. The differentiator is no longer coding skill.It’s domain understanding and knowing what problem is worth solving. The Real Problem: Incentives, Not Technology Max made a powerful comparison: In aerospace, countries cooperate on standards to prevent planes from colliding mid-air.In healthcare, we still don’t have true global data standards because the financial incentives don’t demand it. Healthcare costs continue to rise (approaching a quarter of U.S. GDP), yet most digital tools help people navigate the system, not reduce the need to use it. The real opportunity for AI + FHIR is: * Preventing unnecessary ER visits * Supporting preventative care * Giving patients usable access to their data to power their self care * Enabling decision support before care becomes expensive Where AI Changes Healthcare First According to Max, the most immediate impact won’t be in replacing doctors. It will be in: * Extracting structured FHIR data from unstructured inputs * Real-time scene and event understanding (AI + spatial awareness) * Reducing overload in emergency systems * Supporting remote diagnostics and distributed care models * Turning patient data into actionable, evidence-based guidance We’re moving from: “How do we move data between systems?”to“How do we use this data to keep people out of the system?” The Future of FHIR: Less Open, More Precise We also discussed the direction of FHIR R6 and the move toward tighter, more stable core resources while allowing innovation to happen in defined spaces. Max’s view: this won’t reduce innovation it will accelerate it. Builders will have more certainty that what they implement won’t shift under them, while AI makes data mapping across standards increasingly trivial. A Global View Having worked across Europe and now the U.S., Max sees strengths and weaknesses in both systems. What’s clear is that: Both systems need better use of data, and patients need more control and transparency. That’s a central theme of Max Health and why he remains deeply committed to open source and global standardization efforts. Where to Find Max * Newsletter: maxhealth.tech * LinkedIn: Max Nussbaumer * Conferences, connectathons, and anywhere FHIR builders gather The Big Takeaway This episode wasn’t about FHIR history. It was about this moment. AI has arrived at the exact time when healthcare finally has a usable data standard. The combination means we can stop talking about interoperability as a goal and start using it as a foundation to solve real problems. And the people who understand both the data and the problems are about to build very quickly. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    28 min
  8. Episode 22 - Chris Hutchins

    Jan 21

    Episode 22 - Chris Hutchins

    Top 10 Takeaways: COVID, Data, and the Coming AI Reckoning in Healthcare 1) Healthcare didn’t lack data. It lacked urgency.The pandemic didn’t introduce new analytics capabilities. It changed the cost of being slow. When delay becomes lethal, organizations suddenly discover they can make decisions in hours instead of quarters. That tells you something uncomfortable: speed was optional until it wasn’t. 2) The winning move wasn’t better dashboards. It was deciding which questions mattered.Pre-COVID analytics chased curiosity. During COVID, analytics chased survival. The shift wasn’t technical sophistication it was ruthless prioritization. Moneyball lesson: when resources are constrained, focus beats breadth every time. 3) Interoperability works best when you shrink the problem space.Northwell didn’t unify 70+ EHRs. They built a currently admitted patient index a small, high-value dataset tied directly to decisions. That’s classic systems strategy: optimize the part of the system where leverage is highest. 4) Real-time analytics requires trust more than compute.Two daily huddles. Locked pipelines. Tight access controls. The goal wasn’t “more data.” It was shared belief in a small number of metrics. In complex systems, trust is the scarcest input. 5) AI turns data quality from a nuisance into a risk multiplier.Bad data used to waste time. Now it produces confident, well-phrased errors at scale. AI doesn’t clean your data it accelerates whatever state your data is already in. This changes the ROI math on governance overnight. 6) The most dangerous bias isn’t malicious. It’s missing context.Models assume you’ve provided enough information. Healthcare almost never does. Missing baselines, fragmented history, and unspoken nuance quietly distort outputs. This is the hidden error term no benchmark fully captures. 7) Consumer AI creates a parallel healthcare system with no referee.Patients are already using AI for triage, interpretation, and reassurance outside clinical workflows. There’s no visibility, no accountability, and no feedback loop when the model is wrong. That shadow system will shape outcomes whether clinicians like it or not. 8) Accountability in healthcare AI is misaligned and unstable.Clinicians and health systems bear liability once AI output enters care. Vendors largely don’t. Patients bear risk when they self-diagnose with consumer tools. That imbalance won’t survive contact with real harm. Regulation is coming but likely late and blunt. 9) AI exposes healthcare’s incentive structure, not just its data gaps.If AI reduces unnecessary visits, insurers benefit first. If it increases demand through anxiety, providers feel the strain. Like Moneyball, the advantage won’t come from better tools it will come from understanding who wins and loses under new rules. 10) The real competitive advantage isn’t smarter models. It’s judgment.AI can summarize, predict, and suggest. It can’t know what matters most right now. The organizations that win won’t be the ones with the fanciest AI. They’ll be the ones that combine clean data, tight feedback loops, and humans who know when not to trust the machine. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit evestel.substack.com/subscribe

    44 min
See All (30)

About

Talking about FHIR evestel.substack.com

Information

  • Creator
    Gene Vestel
  • Years Active
    2025 - 2026
  • Episodes
    30
  • Rating
    Clean
  • Copyright
    © Eugene Vestel
  • Show Website
    Out of the FHIR Podcast

You Might Also Like