The Policy Stack

The CMS Innovation Center has run payment experiments for 15 years. Most paid providers for activity: enrolling patients, filing reports, hitting process measures. At the end of 2025, CMS launched a model that changes the unit of payment itself. It’s called CMS ACCESS, and it pays for outcomes. Did blood pressure come down? Did HbA1C come down? Did depression scores improve? The first cohort went live weeks ago.

Underneath that idea is a much harder one. To pay for outcomes, you have to measure outcomes at the patient level, across thousands of organizations, in something close to real time. As host Lisa Bari puts it, this is a data problem before it’s a policy problem. Her guest is Dr. Will Gordon, who was the first Chief Informatics Officer at the CMS Innovation Center and a senior advisor to the CMS Administrator on technology and interoperability, and who helped design several of the models now in the field. He’s a practicing physician and clinical informaticist, now a senior advisor at Manatt Health.

Together they trace how CMS ACCESS came to be: why activity-based billing and technology-driven care fundamentally don’t fit together, why an AI agent managing blood pressure isn’t limited the way a physician’s day is, and why paying for outcomes has always been the obvious idea that turned out to be brutally hard to operationalize. They get into the mechanics (the four clinical tracks, the roughly 50% outcomes withhold, the twelve-month episode), the data infrastructure that finally makes it feasible (Model Space, bulk FHIR APIs, pre-adjudicated claims), the interoperability gaps CMS still can’t close on its own, and whether commercial and Medicaid payers will adopt the same chassis. A clear-eyed look at what it actually takes to turn good policy into something real.

Medicaid covers more than 70 million people, and last year it went through the largest change to its financing in its history. The 2025 reconciliation law cut an estimated $911 billion in federal Medicaid spending over ten years, added work requirements, shortened eligibility checks to every six months, and limited how states can raise money through provider taxes. For most of last year those were words on a page. Now the deadlines are real, and many of the biggest changes begin January 1, 2027.

Lisa Bari’s guest is Verlon Johnson, EVP and Chief Government and Corporate Affairs Officer at Acentra Health. Verlon also chairs MACPAC, the commission that advises Congress on Medicaid, and spent decades at CMS. She has seen the program from the federal side, the state side, and the technology side. Together they get past the politics to the harder question of implementation: what it actually takes for states to redefine business rules, retrain eligibility workers, stand up appeals processes, and modernize eligibility systems before the clock runs out. Verlon’s throughline is one every operator will recognize: good policy deserves good implementation, and data alone doesn’t improve health, better decisions do. It’s a clear-eyed look at whether the program can hold, and what it will take to make sure the people who depend on it don’t fall through the cracks.

79% of hospitals are running AI tools built by their EHR vendor. Not because those tools are necessarily the best, but because the structural advantages of staying in suite make it the easier choice every single time.

In this episode of The Policy Stack, taken from Innovaccer's Context 2026 Healthcare Autonomy conference, Lisa Bari, Head of Policy and Partnerships at Innovaccer, and Julia Adler-Milstein, PhD, Professor at UCSF, joined Ruth Reader of Second Opinion for one of the most substantive policy conversations of the conference.

The data from Julia's recent JAMA perspective piece makes it clear: health systems are not always choosing the best AI tools available. They are choosing the most accessible ones. Until the information blocking rules that are already in law get enforced, until API performance is publicly reported, and until patients are empowered to direct their own data, that dynamic is not going to change on its own.

#ThePolicyStack #HealthcareAutonomy #HealthPolicy

Why does almost 80% of America’s healthcare spend go to treating chronic disease while prevention barely receives a fraction of the $5 trillion health budget?

On this episode of The Policy Stack, Lisa Bari tries to get to the root of the problem with Dr. Anand Parekh, Chief Health Policy Officer at the University of Michigan School of Public Health and author of Prevention First: Policymaking for a Healthier America.

A veteran of two decades at HHS under George Bush and Barack Obama, Parekh explains what's actually being cut when the CDC loses 53% of its budget, why the wins from prevention are politically invisible, what the Supreme Court's ruling on ACA preventive services actually means for your next cancer screening, and why the ratio of social services to healthcare spending may be the single most predictive number in population health.

Brian Janssen reflects on running Medicare operations at WPS Health Solutions, processing 130+ million claims annually through an infrastructure that many healthcare professionals don't know exists. He shares why MACs are "the largest backbone of processing in the country that the healthcare world doesn't really know about," how regional variation matters when Flint, Michigan's environmental crisis affects local coverage decisions differently than in rural Kansas, and the challenge of innovating with AI while maintaining FISMA High security in physical data centers in systems that can't afford a single day of payment disruption.

Liza Health’s Duncan Reece reflects on decades in value-based care, from Medicare Health Support's failure to today's explosion of CMMI models launching simultaneously. He shares why MSSP's $10 billion in savings gets unfairly criticized against Medicare's $1.2 trillion spend, how Iora built its own EHR from scratch when nothing else worked for accountable care, and why the real unlock isn't payment models: it's separating necessary accountability from unnecessary complexity while technology costs plummet and AI adoption in healthcare outpaces every other industry.

Kristen Valdes from b.well reflects on building CMS's voluntary Health Technology Ecosystem framework after discovering that, despite a decade of interoperability policy, live demos showed only 1 out of 30+ providers could return data automatically. She shares lessons on community organizing when 500+ companies need coordination, but the government can't lead, why the real barriers are business and liability issues rather than technology, and how to prevent legacy players from relitigating settled policy when the goal is implementation, not debate.

Most healthcare decisions happen far from the headlines. The Policy Stack, hosted by Lisa Bari, reveals the people and processes shaping how American healthcare really works. From state leaders to industry innovators, each episode examines the laws, funding, and political realities behind the policies that shape care.

In this live episode from Innovaccer’s Mini-Xccelerate conference in DC, Blair Childs and Dave Johnson examine the sweeping policy and market shifts transforming U.S. healthcare. They discuss the move from government-designed systems toward market accountability, how “liberating data” is key to real value-based care, and why structural incentives, not funding, determine the pace of reform. With their deep experience in health policy and strategy, they offer a grounded look at where healthcare is heading, and what it will take to finally realign incentives for better outcomes.

Lisa Bari speaks with Timi Leslie about implementing California's groundbreaking data exchange framework, where healthcare organizations must share data across sectors for the first time. Timi shares lessons from her career in health IT policy, about why technical interoperability is just the beginning, and how to build trust between organizations that don’t usually work together: from hospitals to housing agencies.