Ruth, here. The Food and Drug Administration is in turmoil as drug, device, and clinical trial technology races ahead.

In the last week, the Trump administration has fired Food and Drug Administration commissioner Marty Makary, head of artificial intelligence, Jeremy Walsh, and the controversial lead of the Center for Drug Evaluation and Research, Tracy Beth Hoag. Others may be on the way out, too. 

At the same time, the FDA has seen major upheaval in the last 18 months. At the end of last year, during a Prescription Drug User Fee Act reauthorization meeting, the agency shared that it lost a fifth of its reviewers. 

“There’s been so many departures, and that includes senior career scientists, and it’s not the sort of thing that’s easy to rebuild,” Lowell Schiller, who led FDA’s policy office under the first Trump administration and now consults on regulatory policy, told me. 

Still, he thinks that amid all the upheaval is an opportunity to rebuild.  

“My sense is that the folks that are working on this have heard the criticisms and what needs to change and it’s a matter of execution,” he added.

It won’t be easy though, he said, noting that newer reviewers tend to be skittish about approving products and need training to be able to make tough calls. 

Going forward he thinks the agency needs to be especially focused on addressing two issues: 

— Staying on top of emerging technology: He said we’re at an inflection point where the technologies exist to radically transform clinical trials. Policy needs to move forward in order to capitalize on those innovations, such that companies are not forced to use outdated technologies and protocols in order to comply with FDA standards. 

For example, “Do we still need to randomize people into a placebo group just because that’s what we would have done 50 years ago?” he said.

— Responding to China’s biotech boom: He said that the FDA needs to make changes to its Investigational New Drug policies that would make it easier for biotech companies, especially smaller ones, to conduct early phase trials in the U.S. This has spurred more companies to run early phase trials in China. Schiller said it’s hard for the FDA to validate the quality of labs overseas as well as the data generated there. 

Last month, Makary laid out legislative proposals that would address this problem. But Schiller says that FDA, without Congress, could take steps to incentivize biotech companies to run their early phase trials in the U.S.    

But as Schiller said, whether the FDA rises above the last 18 months of distress depends tremendously on its ability to execute. Now onto the news of the week with Annalisa Merelli. 

The Ebola outbreak affecting the DRC and Uganda is already the third largest on record, only a week after it was detected. By then, it had been raging for weeks without any coordinated response. Insiders say the Trump administration's cuts to international aid and the demise of USAID have weakened the response, Daniel Payne reports in STAT. One can see how: USAID sent $1.2 billion to the DRC in 2024; that number went down to zero as of 2026.

Twenty-nine large health organizations, including Cleveland Clinic, Oracle, and Epic as well as large insurers like UnitedHealthcare, Cigna, and Aetna, have joined a CMS initiative aimed at reducing prior authorization friction. “Fifty percent of prior auth questions are sent back by fax. It’s very inefficient, often is delayed, is lost, confusing, doesn’t get correlated right, it doesn’t get exchanged back and forth, you take scans of fax pages—it’s just dumb. We should just ax the fax,” said CMS administrator Mehmet Oz, presenting the program.  

Here is the full list of organizations that have joined so far. Whether the initiative will be effective remains to be seen; so far, physicians are skeptical.

Internal documents from CMS obtained by news outlet NOTUS showed that 20% of the people who purchased their health care coverage through the HealthCare.gov marketplace were disenrolled for failing to pay their premiums. This tracks with widespread concerns that the rise in price of healthcare policies would make them unaffordable for a large portion of patients. The administration is seeking to blame the  drop in enrollment on weeding out fraud, though internal sources cited by NOTUS say it’s unlikely that would be the reason. 

A decade-long process to rename the condition formerly known as polycystic ovary syndrome (PCOS) — whose patients crucially do not have cysts on their ovaries — led to a new label: polyendocrine metabolic ovarian syndrome, or PMOS. The new name stresses the importance of the metabolic and endocrine elements of the syndrome, which patient advocates had long been stressing. The change doesn’t just update the understanding of a complex, common, and consequential condition, it affects the many businesses and organizations that offer support and services for patients, and that may be forced to costly rebranding.

In this week's episode of Lifers, Chrissy sits down with Heather Fernandez (CEO of Solv and former Trulia executive) to dig into the question that has stumped healthcare for decades: why don't you know what your doctor visit will cost? They also discuss:

Our very own Christina Farr speaks with Garner Health CEO Nick Reber and Business Group on Health CEO Ellen Kelsay about what AI is doing for employer-sponsored health care and the benefits field generally.

CF: Lumeris is building AI-enabled infrastructure for primary care through Ask Tom, which the company announced is already being deployed in a major health system. Why do you believe primary care is the foundation for the future of healthcare delivery?

David Carmouche: Lumeris isn't a new company. We've been operating for two decades in and around primary care, helping large medical groups de-risk their shift to value-based care. In that work, we saw firsthand how high-quality primary care lowers costs by a third and that people with primary care relationships actually live longer. It's the only specialty that spans your entire life longitudinally. So we see primary care as the bedrock of a high-functioning healthcare system — and right now, it's a distressed, underleveraged asset.

Nobody has been giving primary care physicians the technology to make their jobs easier. Instead, we've made them harder with more documentation, more quality metrics. That's part of why there's such a lack of satisfaction in the field, and why access has gotten so constrained that patients self-refer to specialists for things primary care should handle. A health system recently told me that a third of their cardiology appointments next year will be blood pressure management and checkups. That's a mismatch.

AI changes that equation. Take heart failure: we know it can be effectively managed with four drugs that need titration, and that doing so drops hospitalization rates by over 35 - 50%. But less than 10% of patients with heart failure actually get treated that way. With AI monitoring patients, tracking compliance, and surfacing clinical decision support to the primary care physician, that changes. AI is the great equalizer for cognitive specialties. What we'd pay a cardiologist or rheumatologist for their expertise, AI can bring that to the primary care visit. Everything short of procedural and super-specialist work is fair game for disruption.

CF: You've sounded the alarm about the shortage of primary care physicians. How do we make it more attractive for the next generation coming out of medical school?

David Carmouche: The work has to change. One study found that if you actually tried to do all the chronic disease management, acute care, and administrative work a primary care doctor is expected to do, it would take 27 hours a day. We've created a job that's virtually impossible. You start every day knowing you won't be able to do what your patients need… that's a brutal way to practice medicine.

AI can compress that 27 hours into something manageable. Imagine a physician whose patients are largely monitored by a connected AI solution, and their human time is spent on the 5 to 10% who truly need them. That’s the complex cases requiring reasoning, judgment, and real decision-making. The others are being watched continuously; you can adjust care without requiring an in-person visit. That's closer to what internal medicine was supposed to be.

The economics follow. If I can take responsible care of 5,000 patients using connected technology, see the 10 to 12 people a day who actually need me — virtually or in person — and get paid for the full panel, the math works completely differently. Primary care physicians should be compensated at the level of specialists today. Then demand for the field comes back. Right now, the job we've created isn't attractive to most people coming out of training. I've had the chance to talk with people at HHS about changing payment models to unlock this. That's the lever.

CF: The $50 billion going into rural health is an impressive number. But there's also $1 trillion in cuts around the corner. Can we integrate technology into rural healthcare while keeping health systems sustainable?

David Carmouche: We're not going to have more primary care physicians in rural communities anytime soon. That's just the reality. So the question is whether we can build end-to-end solutions that reach patients who may be many miles from a clinician and connect them to world-class care. That means integrating community-based resources to address social needs, which is why we recently announced partnerships with companies like Instacart and Unite Us, as well as clinical needs. 

The business case has to be loss avoidance. When a rural patient shows up needing to be admitted for heart failure or sepsis, the health system doesn't make money on that. If tech-enabled primary care prevents those admissions, that's the ROI. And if we can demonstrate that the model works (improved community health, downstream savings), then the people who purchase healthcare, state Medicaid agencies, managed care organizations, and health plans will have reason to sustain it.

We're currently in conversation with about 30 states interested in this approach. The $50 billion rural health initiative is an opportunity to test whether new models can be built that are actually sustainable.

CF: At Second Opinion, we've been thinking a lot about how AI changes the fundamental reimbursement structure for healthcare. What are the right ways to pay for AI as it becomes more autonomous?

David Carmouche: Fee-for-service doesn't incentivize health — it never has. But you can layer on top of it. We did something effective at Blue Cross of Louisiana: a monthly care management fee for providers managing qualifying chronic disease patients, tied to actual outcomes pulled from the EHR. Do well, and the payment increases. If you do poorly, it decreases. It moved outcomes without blowing up the existing FFS structure.

That's a template. You start paying for outcomes alongside FFS, let physicians deploy AI, and modernize remote monitoring payment to account for AI-driven monitoring — not just human time.

Ultimately, primary care needs a subscription or direct primary care model: a per-member-per-month base that pays for 24/7 always-on AI connected to an accountable human clinician, plus outcomes bonuses for population health. The office visit should become the exception, not the default. But right now, if there's no payment for the work you're doing outside of an E&M code, there's no business model.

The PMPM doesn't need to be prohibitive. We're talking $50 a month. Deployed at scale across a payer or health system, which funds the investment in AI infrastructure. And when you add outcomes bonuses, you're not just rewarding a physician is taking care of a panel, you're rewarding them for doing it well. That's the model that can actually unleash what primary care is capable of.