US Health IT Chief Keane Sees AI Rivals Cracking EHR Oligopoly
Julie Yoo, Colin Rom, Thomas Keane
National Coordinator Tom Keane predicts artificial intelligence will yield a direct challenger to entrenched acute care record vendors within two years as federal regulators dismantle certification hurdles and penalize data hoarding.
AI-assisted summary of The Next Era of Health Data: AI, Open APIs, and the End of the EHR Oligopoly — original source. Check the original for full context.
Source published:
Key ideas
- National Coordinator for Health IT Dr. Tom Keane predicts a viable new competitor will emerge to disrupt the entrenched acute-care EHR oligopoly within just one to two years, driven by generative AI and vibe coding. (31:25)
- ONC is moving to eliminate its certified AI transparency 'model card' mandate after public filings revealed near-zero market adoption, with Oracle reporting roughly 2 queries per month and Epic around 50. (20:00)
- TEFCA has scaled to approximately 1.7 billion cumulative transactions and is processing roughly 200 million transactions per month as a national public utility for health data exchange. (10:00)
- ONC is partnering with the HHS Office of Inspector General, FTC, and DOJ to actively police whether incumbent data restrictions reflect legitimate security concerns or competitive moats masquerading as compliance. (25:00)
Detailed notes
## Background Published on RSS and the web on October 29, 2026, by Julie Yoo, Colin Rom, and National Coordinator for Health IT Dr. Tom Keane, this episode explores healthcare's shift from basic EHR digitization to data liquidity, patient access, and open competition. According to the show description, the conversation covers how data moves via APIs and TEFCA, how wearables and AI integrate into clinical workflows, and how privacy concerns can turn into competitive moats, culminating in Dr. Keane's prediction of a new EHR competitor within one to two years.
## TL;DR Dr. Tom Keane details the evolution of federal health IT policy, explaining how early HITECH incentives unintentionally fueled a closed hospital EHR oligopoly and how subsequent 21st Century Cures Act mandates aim to return data stewardship to patients. He highlights ONC's current shift toward acting as an open platform builder through standardized patient-facing APIs, the nationwide scaling of TEFCA, and the streamlining of rules like HTI-6, which scraps ineffective AI model card mandates to reduce compliance burdens for new entrants. Keane also details the specific diagnostic tests federal agencies use to police information blocking disguised as security concerns. Ultimately, he predicts that the combination of open APIs, vibe coding, and generative AI will give rise to a viable challenger against incumbent EHR giants within one to two years.
## Atmosphere & style The conversation strikes an engaging, collegial tone that seamlessly bridges deep federal policy wonkiness with startup-world technical optimism. Julie Yoo and Colin Rom bring seasoned venture insider perspectives, trading lighthearted banter with Dr. Keane about technology lore at UIUC alongside Marc Andreessen and the absurdities of health system HIPAA compliance. The narrative feels lucid, candid, and pragmatic, moving briskly from historical statutory reflections to concrete enforcement benchmarks, leaving listeners with an energized sense that healthcare's long-stagnant infrastructure is on the cusp of genuine disruption.
## Bulletin - `31:25` National Coordinator for Health IT Dr. Tom Keane predicts a viable new competitor will emerge to disrupt the entrenched acute-care EHR oligopoly within just one to two years, driven by generative AI and vibe coding. - `20:00` ONC is moving to eliminate its certified AI transparency 'model card' mandate after public filings revealed near-zero market adoption, with Oracle reporting roughly 2 queries per month and Epic around 50. - `10:00` TEFCA has scaled to approximately 1.7 billion cumulative transactions and is processing roughly 200 million transactions per month as a national public utility for health data exchange. - `25:00` ONC is partnering with the HHS Office of Inspector General, FTC, and DOJ to actively police whether incumbent data restrictions reflect legitimate security concerns or competitive moats masquerading as compliance.
## Key takeaways - `00:00` Healthcare has completed its digitization phase and entered an interoperability phase where data must follow the patient rather than stay siloed in hospital enterprise software. - `01:46` Dr. Tom Keane brings a multi-disciplinary perspective to ONC, having worked five years in engineering, fifteen years in clinical medicine, and crossed paths with Marc Andreessen at UIUC during the Mosaic launch. - `04:47` EHR adoption stalled between the 2004 Bush executive order and 2009 because, unlike digital PACS in radiology, general EHRs lacked organic financial and operational return on investment. - `04:47` The 2009 HITECH Act drove EHR adoption from 15% to over 90% via $30 billion in incentives, but accelerated market consolidation into an oligopoly tailored around hospital billing and coding rather than clinical usability. - `04:47` The 21st Century Cures Act established the principle of patient data sovereignty, prohibiting information blocking in healthcare in contrast to private data hoarding in search or real estate. - `07:22` The modernization of the banking sector via the SWIFT network in the late 1970s serves as an external precedent for federally guided interoperable infrastructure. - `10:00` ONC's core regulatory foundation rests on three pillars: standardized patient-facing APIs under the 2020 Cures Act rule, TEFCA exchange architecture, and codified information blocking enforcement. - `15:00` Healthcare's structural 'original sin'—where payers and providers change frequently while the patient remains the sole lifelong constant—demands autonomous software agents to manage longitudinal care. - `15:00` Upcoming HTI-6 rules will incorporate non-traditional clinical data such as consumer wearables and biometric readings into standardized observations for AI agents and clinicians. - `20:00` ONC is pivoting its certification strategy away from prescriptive feature mandates and toward interoperability, preventing burdensome compliance programs from blocking nimble startups. - `20:00` Mandated AI model cards failed due to rigid schemas, regulatory arbitrage where AI operated outside certified EHR modules, and preference for bilateral commercial NDAs. - `25:00` Federal regulators assess whether data frictions are anticompetitive by evaluating affiliate favoritism, fee transparency, bounded integration timelines, and parity across EHR vendors. - `25:00` Roughly 90% of reported integration frictions are non-malicious interoperability problems that ONC resolves collaboratively without resorting to inter-agency enforcement. - `30:00` ONC's six-month agenda centers on finalizing HTI-5 to lower market barriers, expanding TEFCA to 100% reliability, and coordinating enforcement with the OIG, FTC, and DOJ. - `31:25` The convergence of open APIs, reduced certification hurdles, and rapid AI development will yield a direct competitor to incumbent hospital EHRs within one to two years.
## Detailed walkthrough ### `00:00` Introduction: Moving from Digitization to Interoperability
Healthcare spent the last fifteen years and tens of billions of dollars migrating from physical paper charts to electronic health records. Now, the battleground has completely shifted: the objective is no longer merely digitizing records, but ensuring that clinical information moves fluidly to patients, third-party applications, and emerging artificial intelligence agents. As Dr. Tom Keane, the National Coordinator for Health IT, and the hosts outline at the opening of the discussion, constraints on how data is shared create profound friction around trust, scale, and security. In the absence of enforceable, standardized ground rules, organizations naturally default to hoarding information rather than leveraging its full potential. The ultimate regulatory and technological vision is to turn the patient into the primary steward of their own medical records, dismantling historical gatekeeping.
The stakes of this transition are heightened by an acute care market dominated by a deeply entrenched hospital software oligopoly. A central question facing the industry is whether modern developer tools, patient-facing application programming interfaces (APIs), the Trusted Exchange Framework and Common Agreement (TEFCA), and generative AI can crack that stronghold. When asked by Julie Yoo how soon the market might witness a viable contender emerge to dismantle the incumbent electronic health record (EHR) oligopoly, Dr. Keane offers an aggressive timeline: within one to two years. Reaching that point requires modernizing the Office of the National Coordinator for Health Information Technology (ONC) regulations, lowering compliance overhead for tech startups, and critically evaluating privacy and security arguments that incumbents often deploy as anticompetitive moats.
- **Regulatory Scope:** ONC is tasked with the direct federal oversight of health information technology standards and nationwide health information exchange. - **Core Policy Objective:** Mandating open, standardized, patient-facing APIs so clinical data follows the patient rather than staying locked inside hospital enterprise systems. - **The Competitive Horizon:** A forecast that generative AI and open data infrastructure will produce direct, viable competitors to dominant acute care EHR incumbents within one to two years.
---
`01:46` Dr. Tom Keane’s Path: Engineering, Medicine, and Marc Andreessen's Mosaic Code
Welcomed by Julie Yoo and Colin Rom, Dr. Keane reflects on an unconventional path that spans three distinct disciplines: engineering, clinical medicine, and federal policy. Julie introduces him as a rare "unicorn" across these domains, prompting Keane to jokingly shoot back: “And thank you for the implication that having done many things, I'm very old.” After completing an undergraduate engineering degree, Keane spent five years working in engineering before deciding to apply to medical school. He practiced clinical medicine for approximately fifteen years, experiencing firsthand the mounting operational headaches, clunky software interfaces, and structural administrative burdens that physicians navigate daily. Motivated by an ambition to reduce those frictions and advocate directly for physicians and patients, he joined the government, taking on roles within ONC before rising to lead the agency.
```mermaid flowchart LR accTitle: Dr. Tom Keane Career Trajectory accDescr: Flowchart tracing Dr. Tom Keane's career from engineering to medicine, government policy, and heading ONC. A["Engineering Graduate<br/>(5 Years in Industry)"] --> B["Clinical Medicine<br/>(15 Years in Practice)"] B --> C["Government Service<br/>(ONC Policy & Leadership)"] C --> D["National Coordinator<br/>for Health IT"] ```
The conversation takes a nostalgic detour into technology history when Julie notes that Dr. Keane attended the University of Illinois Urbana-Champaign at the exact same time as Netscape co-founder Marc Andreessen. Keane confirms that their tenures directly overlapped, with Andreessen graduating in 1993 and Keane in 1994. In March of 1993, Andreessen and Eric Bina publicly released the Mosaic web browser, transforming UIUC into ground zero for the modern commercial Internet. As an engineering student, Keane recalls walking over to inspect the physical boards where top students' names were posted to see who this suddenly famous creator was, and digging directly into the raw source code. Keane recalls looking through the original Mosaic codebase and noting that every single developer comment left by Andreessen was signed under the handle “Marca.” While he cannot recall if they ever crossed paths in the campus hallways, Keane remembers the electric atmosphere of the department, prompting Julie to remark that “there was something in the water.”
- **Career Path:** Transitioned from five years in engineering to fifteen years in clinical practice before entering public service at ONC. - **University of Illinois Overlap:** Attended UIUC alongside Marc Andreessen during the development and launch of the Mosaic web browser in March 1993. - **Historical Milestone:** Mosaic is remembered as the first commercially viable graphical web browser that ignited consumer adoption of the World Wide Web.
---
`04:47` The History of ONC: From Bush's 2004 Order to the Cures Act
To demystify ONC’s mandate, Dr. Keane traces its institutional trajectory over the past two decades. He notes with amusement that until a recent agency reorganization, his formal title was Assistant Secretary for Technology Policy / National Coordinator for Health IT, which gave him the single longest title in the entire federal government. ONC was originally established by executive order in 2004 under President George W. Bush with a 10-year goal to have every hospital in America operating on electronic health records. However, virtually zero meaningful adoption occurred between 2004 and 2010. Hospitals had readily adopted Picture Archiving and Communication Systems (PACS) for digital diagnostic imaging—starting with the first PACS system at the University of Maryland in 1989 and reaching near-universal hospital penetration by 2000—because digital radiology provided clear internal economic returns, natural clinical utility, and native interoperability. General EHRs lacked those commercial and operational tailwinds.
The stagnation ended abruptly in 2009 when President Barack Obama signed the Health Information Technology for Economic and Clinical Health (HITECH) Act, attaching $30 billion in direct financial adoption incentives alongside future regulatory penalties. Putting that scale in perspective, Keane observes: “And to give you an idea, back then, the total budget of the NIH was $30 billion. So, the incentives provided for EHR adoption were equal to the incentives for the United States.” The massive cash infusion worked: hospital and provider EHR adoption skyrocketed from roughly 15% in 2010 to over 90% by 2015. ONC served as the structural architect, managing program design and establishing a functional software certification floor to prevent providers from purchasing low-quality software just to capture incentive checks.
```mermaid flowchart TD accTitle: Federal Health IT Policy Evolution accDescr: Diagram showing the timeline of ONC policies from the 2004 Bush Executive Order through the HITECH Act, the Cures Act, to the modern platform builder era. subgraph S1["Digitization Phase (2004–2015)"] A["2004: Bush Executive Order<br/>(10-Year EHR Goal; Market Stalls)"] --> B["2009: HITECH Act<br/>($30B Incentives; 15% to 90% Adoption)"] end subgraph S2["Unintended Frictions (2015–2016)"] B --> C["CFO-Driven Purchasing<br/>(Billing/Coding Priority; Poor UX)"] B --> D["EHR Oligopoly<br/>(4-5 Vendors Lock Up 95% Market)"] end subgraph S3["Interoperability & Platform Phase (2016–Present)"] C & D --> E["2016: 21st Century Cures Act<br/>(Reduce Burden; Mandate Data Sharing)"] E --> F["Modern Mandate: ONC as Platform Builder<br/>(APIs, TEFCA, Patient Data Sovereignty)"] end ```
By 2015 and 2016, two severe structural market failures emerged from that rapid rollout. Because systems were procured by hospital chief executive officers and chief financial officers rather than bedside clinicians, software vendors tailored their products almost exclusively to revenue cycle management, billing, and coding rather than clinical ergonomics or patient utility. Concurrently, the market consolidated into an oligopoly where four to five dominant EHR vendors controlled 90% to 95% of the sector, locking proprietary customer data inside closed silos. In response, Congress passed the 21st Century Cures Act in 2016 with near-unanimous bipartisan support. The Cures Act established two clear legal mandates: drastically reduce administrative documentation burdens on clinicians, and outlaw information blocking by requiring data sharing. Dr. Keane highlights how unique this standard is across the broader economy:
> “When a company like Google acquires data from prior searches, they're under no obligation to share that. When a real estate company acquires data about what units are renting for and what units are selling for, they are not under any obligation to share that freely. They can sell that information and it can be leveraged by their customers for competitive advantage. The same is not true in healthcare. Information belongs to the patient and has to move for the benefit of the patient at the direction of the patient.”
Under Dr. Keane's tenure, ONC's focus has evolved from certifying closed EHR modules into acting as an open platform builder. The mandate is to establish transparent regulatory rails so clinical information flows freely, empowering patients and enabling third-party developers to lower systemic costs, improve clinical outcomes, and build thousands of novel healthcare applications.
- **2004 Mandate:** Executive order under President George W. Bush set a 10-year timeline for universal hospital EHR adoption, but stalled due to misaligned financial incentives. - **The PACS Contrast:** The first Picture Archiving and Communication System deployed at the University of Maryland in 1989; by 2000, market adoption was near 100% without federal funding because diagnostic imaging offered inherent financial and clinical returns. - **The HITECH Juggernaut (2009):** Injected $30 billion—matching the total annual budget of the National Institutes of Health at the time—propelling electronic records adoption from 15% to 90% between 2010 and 2015. - **Market Failures:** Accelerated purchasing created an entrenched oligopoly of 4 to 5 vendors covering 90% to 95% of the acute care hospital landscape, heavily biased toward enterprise billing and coding. - **The 21st Century Cures Act (2016):** Bipartisan legislation legally prohibiting data blocking, directing the reduction of physician administrative burden, and codifying patient data ownership.
---
`07:22` A Precedent for Mandated Interoperability: Financial Networks
Julie Yoo reflects on launching her first health tech startup right as the HITECH Act was rolling out nationwide, recalling how surreal it felt to see an entire, historically sluggish industry flip almost overnight from paper files to digital records. She asks whether ONC or federal policymakers looked to any historical precedents outside of healthcare when structuring a program that explicitly tied billions of taxpayer dollars to the adoption of a specific technological architecture. While such top-down federal technology mandates are exceptionally rare in the private sector, Dr. Keane points to the modernization of the financial sector as a historical analog. In the late 1970s, government authorities actively pushed for the adoption and maturation of the Society for Worldwide Interbank Financial Telecommunication (SWIFT) network. The goal was to ensure disparate banking institutions could securely exchange transaction data across standardized networks, establishing an interoperable backbone for the global economy.
- **Historical Precedents:** Direct federal subsidies tied to specific private software implementations are essentially unprecedented in American commercial industry. - **Financial Analog:** Dr. Keane cites the late-1970s push for the SWIFT banking network as a comparable instance where public policy helped drive private financial institutions onto a standardized, interoperable communications infrastructure.
`10:00` The Core Mandates of ONC: Patient APIs, TEFCA, and Information Blocking
The discussion begins by examining the government's role in standardizing infrastructure at national scale, comparing health IT to sectors like air traffic control. The unique aspect of the healthcare transition was the deployment of federal financial incentives under President Barack Obama, Jonathan Gruber, and administration officials, driven by the belief that digitizing healthcare would lower costs and improve system management. However, as medical records moved onto digital rails, intense lobbying emerged across stakeholders who believed data belongs to the patient and must move freely rather than being monopolized by a handful of well-intentioned electronic health record (EHR) vendors.
When asked about the most underrated accomplishments of the Office of the National Coordinator for Health Information Technology (ONC) that empower today's builders, Dr. Tom Keane jokes that none of his agency's accomplishments are underrated, before outlining three foundational pillars:
```mermaid flowchart TD accTitle: ONC Interoperability and Governance Pillars accDescr: Three pillars enacted by ONC to free healthcare data from EHR vendor lock-in.
subgraph Pillars["ONC Policy Foundations"] A["2020 Cures Act Final Rule<br>Standardized Patient-Facing APIs"] B["TEFCA & QHIN Technical Framework<br>Nationwide Exchange Infrastructure"] C["Information Blocking Rules<br>Legal Deterrents & Codified Exceptions"] end
subgraph Impact["Ecosystem Impact"] D["Secure Patient-Permissioned<br>Third-Party App Access"] E["1.7B Transactions Handled<br>(~200M per Month)"] F["Elimination of Anti-Competitive<br>Data Hoarding"] end
A --> D B --> E C --> F ```
First, under the 2020 Cures Act final rule, ONC mandated that every certified EHR expose a standardized, patient-facing application programming interface (API). This established strict security postures and standardized protocols, giving third-party applications the right to retrieve patient data with verified patient identity and permission. Second, ONC operationalized the Trusted Exchange Framework and Common Agreement (TEFCA), which was originally mandated in the 21st Century Cures Act of 2016. Keane highlights that healthcare operates under extreme constraints regarding trust, liability, and the risk of patient harm—illustrating this with the absurd reality that major health systems routinely call ONC for clarification on whether sending a simple bereavement card to a deceased patient's family constitutes a HIPAA violation. To solve the friction of disparate legal agreements, privacy safeguards, and "minimum necessary" payment rules, TEFCA establishes a unified national baseline:
> “By standing up TEFCA, we've tried to address this under a common agreement. We've tried to say here's one agreement, one set of protocols, one stand, you know, QHIN technical framework, we call it, to allow for the exchange of information.”
Third, Keane cites the codification and federal enforcement of information blocking rules. By laying out clear exceptions where withholding data is justified for privacy and security—while making anti-competitive data hoarding illegal across health systems and technology vendors—ONC established clear ground rules for market conduct.
- TEFCA currently operates at a volume of approximately 1.7 billion cumulative transactions, scaling at roughly 200 million transactions per month. - ONC's goal for TEFCA is to establish a 100% reliable, robust national public utility for data liquidity. - Enforcement of information blocking operates across inter-agency partnerships within the federal government to ensure strict market compliance.
---
`15:00` Expanding Beyond EHRs: Wearables, HTI-6, and AI Agents
Julie Yoo reflects on how unusual healthcare is compared to other industries, noting that it benefits from federal regulation actively mandating modern technological primitives like APIs that empower software builders. However, she points out a looming regulatory boundary: existing rules target traditional EHRs, acute care hospitals, and outpatient clinics. Today's explosive growth in health data happens outside traditional clinical and billing workflows—through continuous consumer wearables, patient chat platforms, and direct-to-consumer biomarkers. Yoo warns that founders risk stalling if they fear future rules might penalize their current data practices, and asks how ONC intends to address this frontier.
Keane responds by confirming that ONC has formally accounted for this migration in its upcoming HTI-6 (Health Data, Technology, and Interoperability) rule, published in the federal Unified Agenda. He ties this evolution directly to resolving what Yoo has previously characterized as healthcare's "original sin"—the structural reality that healthcare providers and payers who invest in keeping a patient healthy rarely capture the financial return, because patients constantly switch employers, insurers, and hospital networks.
Keane argues that the only lifelong constant is the patient themselves, supported by autonomous software and clinicians:
> “Our vision is that the patient will be the ultimate steward of their information and that in a really well-functioning market, you'll have an agent, whether it's a piece of software or some piece of software in collaboration with a doctor that will longitudinally be responsible for that patient and be able to benefit from that patient being taken care of.”
To make this vision real, ONC is focusing on standardizing device data into structured, clinically reliable observations that flow bidirectionally between traditional EHRs and external consumer applications. Keane highlights the present reliability challenge: wearing two different smartwatches on opposite wrists often yields conflicting biometric readings. Addressing this requires robust data standards so that traditional clinical systems trust external signals, while artificial intelligence agents—and synthetic intelligence (SI)—can access rich clinical histories to safely execute actions on behalf of patients.
`20:00` HTI-6, Vibe Coding, and Why ONC Is Scrapping AI Model Cards
Dr. Tom Keane outlines the upcoming regulatory priorities at ONC, noting that the agency is aiming to release the HTI-6 (Health Data, Technology, and Interoperability) proposed rule in the fall. The primary objective is establishing provenance, auditability, and execution traces for autonomous AI actions in clinical settings to support post-market surveillance against AI model drift. Addressing Julie Yoo’s observation that the very definition of an electronic health record (EHR) is fundamentally changing as care and data move beyond hospital walls, Dr. Keane explains that ONC is pivoting: instead of mandating basic in-system clinical and billing features, the agency is substantially dialing back its certification program to focus almost exclusively on interoperability and lowering barriers to entry.
```mermaid flowchart TD accTitle: Failures of Mandated AI Model Cards Under Health IT Certification accDescr: Diagram illustrating the five core reasons why ONC proposed eliminating the certified model card requirement.
subgraph CertificationIssues["Structural Flaws of Model Card Mandates"] A["Static Requirement<br/>(Rigid 31 source attributes)"] --> F["Policy Flaws"] B["Narrow Scope<br/>(Applied only to certified EHRs;<br/>apps simply migrated outside)"] --> F C["Near-Zero Adoption<br/>(Oracle: ~2 queries/month;<br/>Epic: ~50 queries/month)"] --> F D["False Assurance<br/>(No audit backing or verification;<br/>empty 'Good Housekeeping' seal)"] --> F E["Bilateral Alternatives<br/>(Commercial buyers prefer<br/>proprietary NDAs for secret sauce)"] --> F end
F --> G["HTI-6 Deregulation:<br/>Drop Model Card Requirement &<br/>Slash Costs for New EHR Entrants"] ```
Dr. Keane illustrates this deregulatory push by discussing the decision to eliminate the AI transparency "model card" requirement that took effect on January 1, 2025. Practical market feedback demonstrated that the mandate was fundamentally not fit for purpose: - **Commercial reality and IP protections:** Enterprise buyers and physicians do not rely on generic public model cards; if a model's capabilities or data are proprietary, vendors prefer disclosing their "secret sauce" directly under non-disclosure agreements (NDAs). - **Inflexible data schemas:** ONC mandated 31 static source attributes, which industry participants immediately criticized as either the wrong attributes or vastly too complex compared to a practical five-attribute standard. - **Regulatory arbitrage:** Because certification requirements only govern certified health IT modules, vendors merely unbundled their software, running standalone AI tools outside the EHR architecture to evade compliance entirely. - **Dismal market utilization:** Certified EHR vendors reported virtually no engagement. Public comment filings revealed Oracle received only approximately 2 queries per month, while Epic reported roughly 50 queries per month across their entire client footprints. - **Lack of authentic verification:** From a physician's perspective, Dr. Keane stresses that the cards functioned as an unvetted facade: “It had the purpose of serving as a good housekeeping seal of approval that didn't actually have any audit or backing or assurance behind it, so that's why we proposed removing it.”
Reflecting on the pace of modern software development, Dr. Keane notes that burdensome federal certification cannot be allowed to freeze out nimble founders. As he puts it: “In the days of vibe coding, you can create a new electronic health record or a company that is a system of record or a company that performs many of the functions that a system of record would, you can create that quickly, probably in a year or two with vibe coding. And we did not want to be standing in the way with a costly, burdensome certification program.”
---
`25:00` Identifying "Competitive Moats Masquerading as Privacy and Security Concerns"
Julie Yoo brings up the inherent friction between incumbent health IT vendors and new tech entrants. While incumbents routinely justify data restrictions on the grounds of patient privacy, cybersecurity, and risk aversion, entrants view these obstacles as deliberate anti-competitive walls. Yoo observes that while it is easily 100 times easier for a digital health founder to integrate with an EHR today than it was 15 years ago, substantial friction persists. Dr. Keane agrees that the balance is an evolving target: advances in identity, authentication, and authorization elevate baseline expectations, while external risks like quantum computing and sophisticated AI create genuine security vulnerabilities.
```mermaid flowchart LR accTitle: Diagnostic Tests for Distinguishing Legitimate Security from Information Blocking accDescr: Flowchart showing criteria used by ONC, OIG, FTC, and DOJ to determine whether vendor data frictions are anticompetitive.
A["Vendor Restricts Data Access"] --> B{"Diagnostic Tests"}
subgraph Tests["Evaluating Friction Legitimacy"] B --> C["Affiliate Favoritism<br/>(Easier access for internal partners?)"] B --> D["Commercial Terms<br/>(Are API fees & terms public & at market rates?)"] B --> E["Timeline Delivery<br/>(Do development milestones have concrete ends?)"] B --> F["Vendor Outlier Status<br/>(Works with 3 vendors, blocked by 2?)"] end
C & D & E & F --> G{"Adjudication"} G -->|"Legitimate risk"| H["Work with ONC on standards"] G -->|"Bad-faith barrier"| I["OIG / FTC / DOJ<br/>Enforcement Action"] ```
However, Dr. Keane makes clear that federal regulators actively evaluate whether these frictions are legitimate technical hurdles or bad-faith anticompetitive behavior: “There are certain ways that we can evaluate whether the frictions that somebody in the market are experiencing are legitimate privacy and security concerns or whether they are competitive moats masquerading as privacy and security concerns.” To police this boundary, ONC coordinates with the HHS Office of Inspector General (OIG), the Federal Trade Commission (FTC), and the Department of Justice (DOJ), evaluating friction using concrete benchmarks: - **Affiliate disparity:** Inspecting whether a health data holder grants its own corporate affiliates or preferred commercial partners faster onboarding, lower development hurdles, and easier data access than outside third parties. - **Transparency of commercial terms:** Checking whether API documentation, onboarding timelines, and access fees are published openly and set at fair, standard market rates. - **Bounded integration timelines:** Verifying that vendor integration timelines have explicit, realistic completion dates rather than open-ended, indefinite delays. - **Cross-vendor parity:** Checking if an application functions smoothly across multiple peer EHR vendors while being uniquely stalled by one or two outliers failing their information-blocking statutory obligations.
Dr. Keane directly solicits input from digital health founders and technology developers, pointing out that 90% of observed frictions stem from standard market friction or interoperability gaps that can be resolved via technical coordination. Emphasizing that “the C in my name is Coordinator—we are there to coordinate with the healthcare market,” he urges startups to report friction directly to ONC or via the official information blocking portal. When problems are flagged with specific potential solutions, ONC can often prompt incumbents to alter their behavior collaboratively; only when deliberate noncompliance persists does the agency transition to an aggressive enforcement posture alongside its regulatory partners.
`30:00` Agency Priorities: HTI-5, TEFCA Expansion, and Inter-Agency Enforcement
Dr. Tom Keane lays out the immediate operational and regulatory roadmap for the National Coordinator's office, stressing that active industry engagement is central to the agency's mission. He encourages health tech founders and organizations to work directly with his team and participate in initiatives like CMS’s health tech ecosystem, which currently convenes around 1,700 participants to shape both policy and shared technical solutions. Outlining goals for the immediate six months ahead, Keane highlights the finalization of the HTI-5 rule (Health Data, Technology, and Interoperability) to lower barriers to market entry, alongside ongoing refinements to federal information blocking regulations.
At the same time, the agency plans to ramp up investments in the Trusted Exchange Framework and Common Agreement (TEFCA) national infrastructure. The goal is to ensure TEFCA achieves 100 percent reliability, expands its active use cases, and solidifies its standing as an indispensable, ubiquitous rail for health data exchange. Crucially, Keane makes clear that facilitation will be paired with an assertive enforcement posture, coordinating across federal bodies to hold bad actors accountable under congressional mandates.
```mermaid flowchart TD accTitle: Health IT Interoperability and Enforcement Framework accDescr: Diagram illustrating the policy initiatives and joint enforcement agencies overseen by the National Coordinator.
subgraph Policy ["Immediate Policy & Infrastructure (Next 6 Months)"] A["Finalize HTI-5 Rule"] --> D["Lower Market Entry Barriers"] B["Refine Information Blocking Rules"] --> E["Eliminate Anti-Competitive Lock-In"] C["Scale TEFCA National Rails"] --> F["Expand Use Cases & Ensure 100% Reliability"] end
subgraph Enforcement ["Coordinated Enforcement Posture"] G["Congressional Mandate"] --> H["National Coordinator / ASTP"] G --> I["HHS Office of Inspector General (OIG)"] G --> J["Federal Trade Commission (FTC)"] G --> K["Department of Justice (DOJ)"] H & I & J & K --> L["Strong, Unified Market Signal"] end ```
- In describing the open-door ethos of his agency, Dr. Tom Keane emphasizes: “We will engage with everybody. It is the main purpose of our organization of the agency is to coordinate.” - Addressing compliance and market expectations, Keane asserts: “We believe that we've been given a charge by Congress with laws to make sure that these laws are complied with. And between us, the OIG, the FTC, and the DOJ, we believe that we will send a strong signal to the market in a way that everybody can be agreeable to.” - All major updates—finalizing HTI-5, clarifying information blocking boundaries, and fortifying TEFCA—are slated for rollout within the upcoming six months.
---
`31:25` A "Spicy Question": Predicting the Fall of the EHR Oligopoly
Julie Yoo closes the interview with a provocative question targeting the consolidated acute care electronic health record (EHR) market. Given federal efforts to promote open APIs, liquid data, and competition, she asks Keane directly: when will the market finally see a genuine new contender challenge the dominant hospital EHR oligopoly—in two years, five years, or ten years? Keane reflects on the broader history of computing, comparing the current artificial intelligence revolution to witnessing the birth of the personal computer and the advent of the internet.
Keane argues that the capabilities unlocked by AI are so profoundly transformative that market structures will inevitably shift. While existing systems of record recognize these shifts and are attempting to adapt to find their place, Keane delivers a surprisingly aggressive timeline: he expects a serious competitive contender to enter the field within just one to two years. Welcoming the forecast, Yoo and Colin Rom promise to bring him back onto the podcast at the two-year mark to assess how new EHR CEOs are performing on the playing field.
- Asking for a concrete forecast, Julie Yoo challenges: “When do you think we will see a contender enter the playing field to break that oligopoly? Is it another two years? Is it another five years? Is it another 10 years?” - Placing the current technological wave in perspective, Keane notes: “Of all the things I've gotten to see in my life develop from the personal computer I'm old enough to remember when those came online to the internet SI is truly transformative.” - Delivering his exact forecast, Keane declares: “So to be more concrete one to two years.” - Setting up the future reunion, Keane jokes: “Invite me back in two years and in addition to telling me I've gotten two years older you can and then we can see how some of the CEOs of the companies that are in the new playing field of the future EHR.”
## Notable quotes > `04:47` Information belongs to the patient and has to move for the benefit of the patient at the direction of the patient. — Dr. Tom Keane
> `25:00` There are certain ways that we can evaluate whether the frictions that somebody in the market are experiencing are legitimate privacy and security concerns or whether they are competitive moats masquerading as privacy and security concerns. — Dr. Tom Keane
## Mentioned Dr. Tom Keane, Julie Yoo, Colin Rom, Marc Andreessen, Eric Bina, George W. Bush, Barack Obama, Jonathan Gruber, ONC (Office of the National Coordinator for Health Information Technology), ASTP (Assistant Secretary for Technology Policy), TEFCA (Trusted Exchange Framework and Common Agreement), HITECH Act, 21st Century Cures Act, HTI-5, HTI-6, PACS (Picture Archiving and Communication System), Mosaic, University of Illinois Urbana-Champaign, University of Maryland, NIH (National Institutes of Health), CMS (Centers for Medicare & Medicaid Services), HHS Office of Inspector General (OIG), Federal Trade Commission (FTC), Department of Justice (DOJ), SWIFT, Epic, Oracle, Google, a16z
## Video description Healthcare spent the last 15 years digitizing medical records. Now the challenge is making that data useful to patients, AI agents, and new health technology companies. National Coordinator for Health IT Dr. Tom Keane says the next phase is about interoperability, patient access, and lowering barriers for new entrants. Tom joins a16z’s Julie Yoo and Colin Rom to discuss how healthcare data can move more easily through APIs and TEFCA, how wearables and AI-generated data fit in, and where privacy and security concerns can become competitive moats. And when Julie asks when AI might produce a serious new competitor to today’s dominant electronic health record companies, Tom predicts: within one to two years.
[Source](https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/d1df597a-30d8-4cae-bb4c-22a8ec496e6d/episodes/744c6b9f-a3a1-4ac3-89f6-80ea9336212c/audio/128/default.mp3?aid=rss_feed&awCollectionId=d1df597a-30d8-4cae-bb4c-22a8ec496e6d&awEpisodeId=744c6b9f-a3a1-4ac3-89f6-80ea9336212c&feed=BXDamaKF)