The New CMS Health Tech Ecosystem

The New CMS Health Tech Ecosystem

You’ve probably heard by now that the US Centers for Medicare & Medicaid Services (CMS), at its July 30 “Make Health Tech Great Again” (I can’t help wondering what “again” refers to in this context) event, announced its proposed Digital Health Ecosystem” that it calls “a bold new vision built on collaboration, not just compliance”.

CMS is “calling on the healthcare industry – data networks, Electronic Health Record (EHR) systems, health app developers, providers, and innovators – to voluntarily align around a shared framework for data and access that empowers people, improves care, and accelerates progress”.

Importantly, CMS says it will be “launching key public infrastructure like the National Provider Directory”. The infrastructure will supposedly also include developing “Modern Identity for Medicare.gov” but, of course, it would be great if that was expanded to be a universal medical ID for patients, and perhaps it will in time (but, given the public's suspicion of government, I wouldn't bet on it).

According to the new CMS Administrator, Dr. Mehmet Oz, “The Administration’s efforts focus on two broad areas: promoting a CMS Interoperability Framework to easily and seamlessly share information between patients and providers, and increasing the availability of personalized tools so that patients have the information and resources they need to make better health decisions.”

According to CMS, to provide that, a major focus of the program is “Patient Facing Apps” apps “to unlock the full potential of modern digital health, CMS is supporting a voluntary, standards-based ecosystem of private-sector technology partners - including apps, EHRs, and providers - that integrate with CMS Aligned Networks. This ecosystem is designed to deliver immediate value to patients through real-world, high-impact tools that are secure, user-centered, and connected”.

Toward that end, CMS says it “is looking for early adopters to highlight and prove that modern technology is possible to deliver quickly.”  A web page provides the initial early adopters, categorized into these six groups (with the number of participants): Data Networks (21), Health Systems and Providers (11), Patient Facing Apps - Conversational AI Assistants (18), Patient Facing Apps – Diabetes and Obesity (8), Patient Facing Apps – Kill the Clipboard (12) and Payers (5).

The apps, among other things, must:

  • support data exchange after patient identity verification,
  • connect for communication with patients via medicare.gov,
  • offer a free trial if there is an app charge, and
  • participate in a future “app store” interface on Medicare.gov.

They must be for one of these use cases:

  • Diabetes and obesity management;
  • The use of conversational AI assistants to help patients check symptoms, navigate care options, and schedule appointments, among other tasks; and
  • Tools to “kill the clipboard” by replacing paper intake forms with seamless digital check-in methods; and more.

We will focus on the middle one - conversational AI Assistants

Here, CMS says the objective is the use of AI “to deliver personalized, context-aware guidance to patients by securely accessing and interpreting their medical history in real time”. 

To do that, the patients must have their medical history, and, elsewhere in the “kill the clipboard” section, the announcement says that EHRs “must provide patients the ability to retrieve a visit record in FHIR format” and that this “should not require portal credentials or additional account setup to accept or return data when the patient’s identity is verified using a CMS-approved service”.

In addition, apps must:

  • Enable patients to transmit information using FHIR, including their digital insurance card and health history (via QR code, Smart Health Card, or Smart Health Links). As shown in the graphic, SMART Health Cards are a vehicle for sharing information, such as a vaccine record, via a QR code. They are incorporated into the State of Louisiana’s digital driver’s license app.  SMART Health Links are a way of sharing FHIR based health information in software, best appreciated by reading the User Story posted by the SMART team.


Article content
The State of LA Digital Driver's License App Provides a SMART Health Card

  • Allow patients to retrieve a summary of their visit (e.g., notes, diagnoses, instructions) from the provider at the end of the encounter in FHIR format and then presented in a user-friendly form.

Once all of that is in place, the AI assistants should provide:

  • Personalized AI-driven support across the patient’s clinical record - including symptom checking, care planning, coordination, and chronic disease support.
  • Responses that must clearly indicate when the results are AI-generated and include appropriate disclaimers when not intended to replace clinical judgment.
  • Conversational AI tools that will clearly distinguish educational content from clinical guidance, assist patients directly when appropriate and guide them to care from a health professional when needed.

Here are the early adopters for Conversational AI Assistants:

Anthropic, Brado, Citizen Health, Curai Health, Ellipsis Health, Google, Hippocratic AI, K Health, Microsoft AI, NantHealth, Nightingale MD, Nuna, OpenAI, Oura, Polygon Health, Sharecare, UnitedHealth Group, and Zocdoc.

Interested companies can contact CMS at HealthTechRFI@cms.hhs.gov.  There is also a CMS Healthtech Ecosystem YouTube channel.

It will certainly be interesting to see how this pans out. It is an exciting vision, but digital tools that patients, and particularly older ones, can use effectively has proven over the years to be difficult, so app design will be a crucial factor. Perhaps AI can help overcome that challenge and it will be exciting to see how that evolves.

To think about other challenges, you might want to consider these questions posted on LinkedIn by Stat News Healthtech Reporter, Mario Aguilar.

  • What happens if the twenty-odd "CMS Aligned Networks" can't get on the same page about important topics like delegation? What if they miss the July 4th, 2026 deadline to make data available over FHIR?
  • What exactly did OpenAI and some of these other tech companies agree to do?
  • Where are the CMS incentives to play nice?
  • Why so few health systems?
  • This is all about the patients, but who represents them?

I would add, why so few payers?

On a somewhat related theme, I suggest Prior Authorization Reform Is Finally on the Table — Standards Must Be Too, a MedCity article by HL7 President Chuck Jaffe.

Paris Majot

🔗 I connect patients, clinicians and data | Director – Clients & Partners at Orion Health | EMR • HIE • AI

3mo

Thanks for the podcast! Great article. Plenty of parallels with Australia’s national HIE program HealthConnect. We’re also moving towards a FHIR enabled, patient facing ecosystem with its key goal is to deliver realtime data access. We’ve also got a national front door 1800Medicare. So we’ve got the foundations for a truly connected healthcare ecosystem. Now for the execution…

Mike McGuire

Looking for my Next Adventure in Healthcare. Interested in working with companies focused on the digital transformation of Healthcare. Want to help companies align all the gears to gain traction and visibility.

3mo

Mark, this is exciting but we’ve been here before. Granted, without the current state of technology but with the same level of enthusiasm. You’re right to cite missing players especially the payers who can be instrumental in really reimagining how we render services.

Ian Siemon

Retired financial planner

3mo

I love the clear language, even I can understand all of it.

Paris Majot

🔗 I connect patients, clinicians and data | Director – Clients & Partners at Orion Health | EMR • HIE • AI

3mo

Can we expect a podcast version? 😊

To view or add a comment, sign in

More articles by Mark Braunstein

  • The Spark That Lit Today’s AI

    This is first of a two-part article written to put today’s AI into context and explain the ‘spark’ that led to its…

  • Deep Thinking AI

    This post is an experiment. I've been increasingly impressed with Google NotebookLM's ability to summarize my articles,…

  • Discussing Genome Language Models

    Recently I posted Can AI Design New Life Forms? after I read Generative design of novel bacteriophages with genome…

  • AI versus Malaria

    Recently I was discussing with friends AI and DeepMind’s AlphaFold solving the long standing protein folding problem…

  • How Novel AI is Helping Treat Cancer

    Gary Marcus is professor emeritus at New York University, a company founder, book author, and publisher of Marcus on…

    1 Comment
  • AI: Your Health ‘Thought Partner’?

    In its August introduction of GPT-5, OpenAI said that it “is our best model yet for health-related questions…

    4 Comments
  • Can AI Design New Life Forms?

    Normally I focus on territory I’m familiar with - the applications of informatics and AI in healthcare delivery. Almost…

  • AI for Clinical Decision Support in Kenyan Primary Care

    We are all used to research and implementation stories about advanced health IT applications such as Clinical Decision…

    6 Comments
  • Interoperability Leaders Propose an Agentic AI Healthcare Process Automation Standard

    This is yet another, and perhaps my most important, article about the incredible pace of innovation in applying agentic…

    7 Comments
  • Can AI Explain Epic's Cosmos AI to Physicians?

    Comments about my recent Epic AI from Wisconsin article predictably and appropriately asked about obtaining physician…

    6 Comments

Others also viewed

Explore content categories