What happened?^[1]

The administration announced last week that it’s nominating Dr. Overton to be the next FDA Commissioner. She’s about 10 years out of med. school and completed her residency under the tutelage of Marty Makary, the former FDA Commissioner. Since 2019, she’s worked for Trump or Trump-aligned organizations, most recently as deputy assistant to the president for domestic policy.

Notably absent is experience running a large public health organization, or any large organization for that matter. Josh Sharfstein, a former FDA leader, cautioned that heading FDA is among the toughest health jobs in government, requiring commissioners to juggle political, business, and advocacy pressures. Other candidates for the commissioner spot have those skills – one leads DoD’s health force, another runs a million-patient cancer/hematology group, and the third is FDA’s well-regarded acting commissioner. Overton shares Trump/Kennedy’s vaccine distrust and anti-abortion stance, but that’s not the same as heading a complex public health organization.

So, why did Trump nominate Overton? There are a couple of good theories. First, better candidates passed on the job. CNN reports that Congressman Brad Wenstrup and acting Commissioner Kyle Diamantas were Trump’s preferred choices and both declined. Overton was available and willing to step up.

That’s no surprise given Overton’s “gold standard” loyalty to MAGA/MAHA. She stood behind Trump and Kennedy while they culled the childhood vaccine schedule, and Trump called Overton a “ROCKSTAR in my Administration.” Cementing Overton’s bona fides is an anti-trans/anti-abortion screed that she wrote with ex-Louisiana Governor Bobby Jindal. In an administration known for loyalty tests, Overton happily drinks the Kool-Aid, making her a passable choice when better-qualified FDA-chief candidates demur.

What does this mean for device firms?

It’s tempting to answer, “very little.” As I’ve written, CDRH Director Michelle Tarver is a capable and experienced leader who insulates her team from some of the tumult hampering FDA. This protection likely will continue regardless of the Commissioner pick. So, the daily experience of CDRH staff and stakeholders may not meaningfully change.

Plus, it’s unclear if Overton will get the Commissioner job. She needs Senate HELP Committee approval and her support for vaccine cuts alienated committee members, including committee chair Bill Cassidy. Cassidy lost his reelection bid after Trump lined up against him, so he may not be inclined to ignore his concerns about Trump’s candidate. Cassidy claims that, as a doctor, he’s a strong advocate for established vaccine policy. That didn’t stop him from endorsing RFK Jr., but maybe he’ll grow some integrity (not the word I’m actually thinking of).

Regardless, device firms should weigh in on Overton’s candidacy. An unfit Commissioner disrupts the device ecosystem. Industry saw this with former Commissioner Makary and he mentored Overton. As FDA Commissioner, Overton risks her predecessor’s mistakes: decisions serving administration priorities that disregard career staff and stakeholder input.

That means bad policy, bad morale, and marketplace uncertainty. It disserves device firms when they – and patients and providers – question why products come to market. That skepticism followed the Commissioner’s National Priority Voucher (CNPV) program, a “favored-nation” drug plan promising expedited reviews. The CNPV claims to choose participants based on health priorities and medical needs, but initial picks benefited firms advancing administration political goals. The CNPV doesn’t cover devices yet, but it will and initiatives like the breakthrough device program and the TEMPO pilot are vulnerable to political pressure and stakeholder distrust. A Commissioner acting on FDA’s public-health mission helps defuse those risks; a Commissioner beholden to political whims assures them.

Also at risk is US marketplace dominance. CDRH is the world leader in device development and introduction, and other countries want to eat our lunch. China isn’t shy about this and it’s built technology and infrastructure behind that goal. Likewise, Europe’s new device regulations promise a more transparent, predictable, and sustainable marketplace. Michelle Tarver isn’t the right person to keep device firms looking to the US as they design and sell products. Success requires a savvy Commissioner coordinating with agency heads and administration leads to preserve the US’s first-in-the-world stance.

What should firms do?

Get in the game. Device firms must engage administration and Congress leaders about the next FDA Commissioner. Unfortunately, engagement appears to be moving in the wrong direction. A leading trade association quickly endorsed Overton’s nomination, predicting that her experience offers “deep understanding of the critical role medical technology plays in lowering cost and improving patient outcomes.”

Time for a reality check. Overton finished med school about 10 years ago and later worked for Trump and a Trump-aligned think tank. Those are good credentials, but do they qualify Overton to run a multi-billion dollar agency with thousands of employees tasked with developing policy for drugs, devices, biologics, tobacco, veterinary medicine, and foods, while assuring that these products are safe and effective for all Americans? The question answers itself.

Reflexive endorsement isn’t helpful. Device firms must at least ask whether the administration’s nominee is qualified for the massive responsibility of FDA leadership. And the answer can’t be that the nominee passes an ideological litmus test. If the administration doesn’t credibly back its nominee, then device firms should push for an alternate with the requisite skills and experience. Device firms and other stakeholders invest too much – money, expertise, experience, energy, and other treasure – to accept questionable candidates to lead the FDA.

This isn’t a call to arms. Going to war with this administration is bad strategy, but neither must device firms roll over and showing their bellies. FDA and its stakeholders benefit from core requirements for the FDA Commissioner. At minimum, that means an experienced, knowledgeable professional who can oversee the varied matters before the agency. And it means assuring product safety and efficacy while backing innovation, regulatory predictability, and marketplace stability. MAGA/MAHA bona fides are beside the point.

Certainly, industry can convey these points behind the scenes. And again, the messenger should be trade associations and similar representatives, not individual firms.

Working with the administration doesn’t mean working without a voice. The administration wants a thriving device sector and device firms need a Commissioner who can deliver that result. Speaking up for competence and experience isn’t disloyal; it’s the kind of engagement that keeps the device ecosystem stable, that keeps CDRH working effectively, and that keeps the US the best place in the world to bring devices to market.

^[1] I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?^[1]

FDA announced earlier this month that it was moving forward with OII redesign plans. I reviewed those plans, part of FDA’s “Simple Reform” program, in an earlier column.

I was sure my concerns would end Simple Reform (I have a healthy ego), but FDA proceeded anyway. In a July 2026 Federal Register notice, FDA announced agency-wide changes including the OII reorg. A blog from OII chief Elizabeth Miller and acting FDA Chief of Staff Lowell Zeta details the reorg.

Notably, the reorg replaces a prior OII improvement plan. That plan, which thoroughly engaged FDA leaders, targeted redundant enforcement work and aligned OII with FDA product centers to bolster investigators’ product expertise. It was also developed by the prior administration, making it non grata.

What does this mean for medical product firms?

Very little. OII redesign follows a long line of projects to improve how FDA oversees medical products. These projects haven’t meaningfully changed FDA’s facility inspection model.

Even the redesign’s Inspectional Affairs Council (IAC) (a cross-agency inspection governance body) is back to the future. The IAC promises “coordinate[d] strategic decision-making and enhance[d] alignment between OII and the Centers . . . .” Notably absent is how the IAC makes decisions, whether its role is advisory or directive, and whether OII must follow IAC decisions. Such questions are central to Center-OII coordination, but the OII improvement plan bypasses them – likely because the questions are perennial and they remain unresolved (although the canceled OII plan got closer to answers).

Also notable is the OII redesign plan’s disregard for expert input. Past reform efforts leveraged professionals (including career staff) who understood field operations, field-Center coordination, core practices, and improvement opportunities. OII redesign sidesteps these critical topics, adopting the shoot-from-the-hip style that derailed former Commissioner Makary.

In some respects, the effect of OII redesign is beside the point. What’s missing is a cogent rationale. Nowhere does Liz Miller, Lowell Zeta, or anyone else explain why their plan should replace detailed OII changes that were already underway. There’s no evidence of gaps in the prior work or explanation of how OII redesign will do better. And there’s no evidence that the OII redesign will produce better results than the canceled improvement plan.

That’s not to say that all the promised changes are bad. No question, FDA must innovate to expand beyond the inspect-and-cite model. But this administration has doubled down on excising – not refining – the last administration’s work. That leaves device firms facing business as usual. Inspections will continue, perhaps with more remote audits, paperwork submissions, and one-day visits. But innovation will remain on the periphery and the push towards generalization risks leaving investigators not knowing enough to oversee evolving medical products. And the risk of generalization is real: OII Director Miller assures blog readers that OII redesign “is not a move to generalists” but in the same breath promises “standardized training . . . and cross-functional development opportunities to ensure investigators can operate . . . across different product areas . . .” (emphasis added).

What should firms do?

Hold FDA accountable. Miller says that OII redesign will deliver “a more aligned, risk-informed, and transparent [inspectional] approach . . .” and promises everything from IAC-driven modernization to “a more deliberate and human-centered workforce strategy.” That last claim is especially eye raising after last year’s staff cuts – who knew that excising 20% of the workforce was human centered? Regardless, industry should rightly expect FDA to show how redesign delivers its promises, for example by improving OII reach, capability, and impact. If redesign can’t deliver results, then FDA leaders must show what they’re doing to course correct and when redesign will pay off.

Industry should also ask FDA to explain why it terminated OII-improvements underway when the administration changed. What justifies saying, “We’re scrapping the old plan and we’re spending time and money (lots of both) on our own plan.” How will OII redesign deliver better results than prior improvement work and where are the metrics showing this?

These points (no surprise) should come from trade associations and other stakeholder groups. I’ve written extensively about when stakeholders should act through proxies, so there’s no need to rehash those points here.

Of course, it’s not incumbent on industry to act. It can sit back and watch what happens: likely a weakened OII that’s less able to handle the technological and global demands underpinning terminated improvement work. But make no mistake – a weakened OII is bad for business. Industry pays millions of dollars a year to assure that FDA is well resourced and capable. When FDA initiatives make the agency worse than it was before, industry is justified in holding the agency accountable.

^[1] I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

In May, CDRH finalized its guidance, Content of Human Factors [HF] Information in Medical Device Marketing Submissions (HF Guidance), replacing the 2022 draft. HF has long been a murky corner in premarket review. CDRH has requested HF data in premarket submissions for years, but its messaging about which devices require HF testing and validation data was sometimes unclear.

The HF Guidance tries to close those gaps. It keeps the same basic architecture as the draft guidance: a risk-based flowchart sorting submissions into three categories, each requiring different documentation. The guidance also adds Decision Point D, instructing sponsors to weigh their devices’ interface use history, the strength of existing risk controls, and interface complexity.

Adding to these changes, CDRH updated its eSTAR templates (for electronic premarket submissions) to require sponsors to select an HF Submission Category when applicable. Selection is mandatory, so device firms must determine which HF category applies to their submissions. Depending on the result, device firms must also include corresponding HF documentation (including validation data when needed).

In a July public meeting, CDRH walked stakeholders through the HF Guidance, clarifying the guidance’s scope and highlighting the HF information that should accompany submissions (that’s increasingly validation data). CDRH said it doesn’t expect submissions before August 1 to satisfy HF Guidance criteria, although it will review HF content that submissions include. That grace period is over and HF Guidance requirements now apply.

What does this mean for device firms?

The real shift isn’t the HF framework – it’s the mandatory terms. Under the draft guidance, device firms might apply judgment, for example to select HF categories if they were confused about whether submissions required validation data. With eSTAR now mandating HF Submission Category selection, there’s no judgment calls. Every submission must correctly disclose its HF category and reviewers see every disclosure.

That will catch some companies off guard, especially those applying existing risk controls when Decision Point D requires heightened justification. As one observer notes, the HF Guidance greatly expands the assessment examples, covering scenarios like pediatric use, augmented-reality interfaces, and devices with use-related problems. The expansion gives sponsors clearer templates to pick the right HF Submission Category.

There’s also a submission-cycle cost. HF validation studies take time to design, populate, and run. A sponsor that discovers mid-review that its HF decisions don’t hold up, or that a reviewer doubts its analysis, is looking at a deficiency letter and months of delay. The HF Guidance helps sponsors avoid these risks before filing submissions.

What should device firms do?

Device firms won’t learn of HF deficiencies in postmarket inspections; this is a premarket challenge. Firms must include HF data as they plan submissions; bolting on HF inputs as filing nears is bad strategy.

Firms can start by running Decision Point D against their review pipelines. That includes documenting interface use history and assuring adequate risk controls for devices with design changes. Firms that are unclear about which HF Submission Category applies, or whether CDRH expects validation data, can use Q-subs to get answers. The HF Guidance points sponsors there and Q-Subs are a cheaper way to get answers than deficiency letters.

Firms choosing not to submit validation data must be ready to produce their underlying HF study results. CDRH reviewers want these results (sometimes even if the HF Guidance doesn’t strictly require it), so firms should have the information ready regardless of their submission decisions.

Finally, firms should double-check submissions that were built around the 2022 HF draft guidance. Confirming that HF categories, their rationales, and corresponding eSTAR fields satisfy the final guidance isn’t minor cleanup; it’s a compliance gap to close before the next submission goes out the door.

CDRH has turned previous HF gray areas into hard checkpoints. There was room for judgment when the HF Guidance was in draft. Now device firms must show their work – with HF categories, rationales, and validation data. The final guidance helps firms add these inputs into submission planning. Firms that instead treat HF data as a late-stage add-on risk bad regulatory results.

^[1] Note to readers: What the FDA will not publish on August 14 and will return August 22.

^[2] I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?^[1]

CDRH hasn’t launched conventional (non-digital) device programs to match digital ones. The disparity isn’t surprising given FDA staff cuts under the Trump administration. CDRH’s workforce dropped more than 20% last year and analysts report lengthening device review times. That’s noteworthy because CDRH emphasizes that it met MDUFA V goals, a claim in tension with analysts’ figures.

But here’s a caveat: there’s no evidence showing conventional device reviews lagging digital reviews. What’s asymmetric is CDRH attention to, and program building for, digital devices. “Attention” here means where and how CDRH deploys bandwidth – developing programs, standing up pilots, drafting guidance, and briefing the press, for example. So far, CDRH has launched digital initiatives while device review queues move uniformly. So far.

What does this mean for device firms?

The question is whether CDRH digital initiatives match device use and patient need, and the numbers aren’t encouraging. CDRH received about 22,000 device submissions in 2025 and the AI/ML device category produced around 300 510(k) clearances. That’s about 1.4 percent of the submission volume. As illustrative, the 2025 monthly clearance rate for orthopedic devices – a ubiquitous conventional-device type – was 1.5 to 2.5 times the monthly AI/ML device pace.

Patient need tells much the same story. Doctors perform about 1.5 million hip and knee replacements in the U.S. annually, along with roughly 965,000 angioplasties and 900,000 cardiac surgeries. Adding to this is more than 2 million coronary and peripheral stent implants each year. These procedures rely on conventional hardware devices.

Digital health’s reach looks different. No question, the impact is enormous – one large radiology practice, for example, reported running AI across more than 20 million exams annually. But as FDA and patient-procedure numbers show, patients face myriad illnesses, most of which require conventional devices.

None of this means that conventional devices are being shortchanged on reviews. It means that if CDRH’s limited bandwidth gets used for programs in areas representing a single-digit share of submissions, stakeholders rightly question whether enough resources remain for conventional device oversight and innovation.

What should device firms do?

First, track specific review figures instead of extrapolating from CDRH aggregates and press coverage. CDRH data includes review volumes and timelines for orthopedic, surgical, and other conventional device. Checking these numbers against digital device timelines and staffing reveals whether disparities exist.

Second, use the MDUFA VI public comment period (which closes next week) to make the innovation-versus-volume argument. Unlike prior user-fee agreements, MDUFA VI doesn’t hold FDA to specific hiring targets, opting instead for “shared intent to hire . . . additional review staff.” If digital-device hiring remains a priority (formally or in practice) while conventional-device coverage reverts to vague “shared intent” goals, that’s the moment to put patient and submission volume data on the record.

Third, don’t wait passively for CDRH support. Submission quality matters when CDRH resources are scarce. Firms competing for resources must file submissions that are well organized, well supported, and complete. CDRH would be hard-pressed to explain why it’s bypassing ready-to-go submissions in favor of digital initiatives.

Last, pay attention to CDRH’s guidance agenda. The 2026 guidance list is shorter than in past years and it carries holdover priorities like diagnostics validation and patient preference information. Even when guidance development affects conventional and digital devices alike, the impact may be greater for devices without a Center of Excellence advocating for them.

None of this is reason to panic. It’s reason to watch where CDRH’s energy and resources go. Visibility of and attention to conventional and digital devices is uneven and will remain so while CDRH resources are depleted. Firms that track product reviews, that attend to MDUFA VI hiring specifics, and that make clean submissions will be least exposed if scarce resources produce review gaps.

^[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?^[1]

Last December, GAO reported on FDA’s 2020-2024 device recall numbers. The results aren’t great. FDA missed its 3-month recall termination target 74% of the time and about 60% of the surveyed recalls remained open as of the report’s publication. GAO’s explanation – FDA short staffing pushing required work into the can’t-get-to-it bucket –is familiar. GAO recommends that HHS build a recall strategy for problems including delayed closures. Advising HHS to help FDA is ironic given its steady work to derail the agency.

Late closures aren’t a new recall problem. GAO checked FDA recall handling in 2011 and noted a 17-month timeline for Class I recalls (the highest-risk category), and a 2018-2022 review shows a longer timeline. This is a structural, resourcing, and prioritization challenge that’s outlasted multiple commissioners and that’s rediscovered whenever high-profile recalls occur.

What does this mean for device firms?

Senators Durbin and Blumenthal requested the GAO report, using it to push their Medical Device Recall Improvement Act, which mandates electronic recall notifications and bolsters FDA recall staffing. Here’s a safe prediction: that act never sees the light of day. That’s not cynicism, it’s pattern recognition. The act’s been introduced and reintroduced since 2023 with no floor vote , and it sits in a Congress with no appetite to expand FDA’s authority or budget. Plus, FDA lacks full data on adequate recall staffing, which doesn’t incline Congress to send more money.

Device firms face a tough reality: recall closures will keep lagging recall completion. FDA admit that it deprioritizes this work in favor of higher-risk tasks. That’s a rational choice for an under-resourced agency, but it leaves a process gap for device firms to fill.

What should device firms do?

The most important shift is in communications, not operations. Device firms need to clarify – to patients, healthcare providers, and others – that FDA recall closure isn’t the measure that matters because many recalls stay open even after firms finish remediation.

Firms that don’t disclose FDA delays risk others telling their stories. Competitors can misrepresent open recalls as showing product defects after corrections are complete. Or patients don’t see that “open” recalls may concern long-fixed devices. And it’s unrealistic to look to FDA for proper messaging. An agency that struggles with on-time audit checks won’t timely report when manufacturers finish recalls. So, device firms must add closure communication to their recall playbooks. That includes announcing when corrections are done, demonstrating recall effectiveness, and explaining that FDA closure delays don’t mean ongoing product problems.

In other columns, I’ve recommended that device firms collectively engage FDA. That makes sense for matters uniformly affecting firms, with similar impact and results. Recalls are different. These aren’t policy asks; recalls concern unique devices and firm-specific events. Firms need to step outside the collective and speak for themselves.

Some more concrete steps: firms must track corrective action, not FDA recall termination, as their progress measure, and employees must use this measure for stakeholder communications (to avoid messaging that recalls are open after fixes are done). Firms should prepare plain-language notices – for sales staff, call centers, and patient/provider communications – that explain product problems, fixes, when the fixes were verified, and why FDA records don’t yet reflect this work. Last, firms should check how competitors and analysts describe their recalls. If open recalls get miscast as continuing safety issues, firms must counter with timely, factual accounts.

None of this fixes FDA’s recall-closure backlog and nothing suggests that the backlog will shrink. But device firms don’t need FDA to fix recall handling to protect their own reputations. They need to speak for themselves so that FDA silence and competitor misdirection don’t speak for them.

^[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?^[1]

A recent study found that 2026 average PMA review times jumped 49% from 2025. Likewise, de novo review times rose by about 8% and 510(k) review times grew by about 5%. To be clear, these are calendar days (i.e., submission-to-decision times), not FDA days used to measure MDUFA commitments (which include review-clock stops).

The results won’t surprise anyone paying attention to FDA’s headcount. Last year saw 20% workforce reductions and former CDRH staff describe a center with large workloads and little support. The Trump administration likes to frame this as doing more with less. That’s a tag line for press releases; the data shows that having less – fewer reviewers, thinner institutional memory, and paltry administrative support – means doing less.

The twist is that approval volume is up. This year through June, CDRH approved 23 original PMAs, ten more than at the same point last year, and analysts project that 2026 PMA approvals will outpace 2025. CDRH is approving more devices but taking longer to do it, and that’s a problem for capital runway and demand forecasting, especially for smaller firms.

Contrast that with CDRH’s 2025 annual report. The center acknowledged “resource challenges” but said it was on track to meet key review timeline targets. That framing hasn’t aged well and it echoes Makary’s “trains on time” promise.

What does this mean for device firms?

The draft MDUFA VI commitment letter largely copies MDUFA V’s review goals. The PMA shared outcome goal stays at 285 calendar days and the 510(k) goal is on a glide path from 128 days in 2028 to 112 days by 2032.

These aren’t new, ambitious goals and holding the line isn’t reassuring when review times are moving in the wrong direction. As troubling, MDUFA VI doesn’t set specific hiring targets, replacing them with staffing reports and an agreement to hire “sufficient” reviewers. Doing better requires CDRH to staff up and vague hiring language doesn’t shorten review times.

Plus, there’s the priorities that CDRH can’t get to. For example, industry relies on FDCA section 513(g) to classify devices and set regulatory requirements, but 2025 staff cuts depleted this program. A CDRH manager recently warned that the center is running out of bandwidth, with reviewers stretched thin and too few staff for non-review activities.

CDRH laboring to meet its review clock leaves little time for initiatives outside that clock – things like compliance innovation, digital health policy, and effective hiring. This work depends on staff time and resources that CDRH can’t spare and MDUFA VI funding is “largely the same” as MDUFA V, so there’s no resource infusion on the horizon.

What should device firms do?

Watch the premarket clock closely. Device firms must track their submissions against CDRH’s actual performance, not stated targets. That includes highlighting review delays. Firms should also include alternate review-time models in regulatory strategy, investor communications, and capital planning. These models are essential in case CDRH can’t meet MDUFA VI commitments. A multi-prong response would follow, including stakeholder engagement, possible renegotiation, and prioritizing core business operations.

If it’s not premarket, don’t wait for CDRH to fix it. Improving compliance oversight, developing digital health policy, and promoting RWE are critically important, but CDRH has sparse resources for this work. So outside stakeholders must take the laboring oar, proposing operating structures and regulatory frameworks, developing test plans and deployment models, and bringing the results to CDRH. That output fosters CDRH-stakeholder collaboration and reverses CDRH’s standard model (where it proposes and stakeholders respond), helping assure that unfunded work gets done. And with apologies for beating the same drum, industry should engage through trade associations or similar collectives. This is another case where industry operates best by combining perspectives and acting through a single representative.

And:

• Lean hard on the Pre-Sub program and TAP. Both programs support early CDRH-stakeholder alignment, before the review clock starts. Fixing problems on the front end is cheaper than managing delays on the back end.

• Attend next month’s MDUFA VI public meeting and file comments on the draft commitment letter. Meeting remarks and written comments should make the same point: beyond funding, MDUFA success requires enforceable staffing commitments, not aspirational language. Industry should also press CDRH to publish current review-time results alongside commitment letter targets, so Congress and stakeholders can track performance.

• Don’t mistake commissioner messages for CDRH operational data. Kyle Diamantes is a good acting Commissioner and he’s right about FDA’s mission and public health pillars, and Marty Makary’s “trains on time” remark is wrong but innocuous. Still, these comments don’t matter in judging device user-fee commitments. What matters is CDRH operational data: hard numbers showing review rates and averages, total days to decision, and business-planning inputs.

CDRH may be hitting “review day” targets, but full-review timelines are getting longer. This reality affects device firms as they develop products and plan for premarket review, marketing, and resource allocation. The data also shows that what gets paid for gets done, leaving few resources for unfunded critical initiatives. Resource scarcity requires industry to take the lead, developing initiatives and regulatory frameworks and then working with CDRH to refine and deploy them.

^[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

FDA needs a new Commissioner and most of the chatter about candidates centers on drugs and biologics. That leaves out device firms, which have just as big a dog in the fight. Some FDA watchers highlight CDRH stability, predicting that whoever gets the Commissioner job won’t change much for devices. But that forecast is misplaced. A Commissioner who understands device oversight can be an ally to industry and CDRH. One who doesn’t could default to neutrality, treating devices as somebody else’s problem. And one who misses device regulatory nuance could be a negative – second-guessing or slow-walking initiatives that industry and CDRH favor. Which of these scenarios industry gets depends on who’s nominated.

What happened?[1]

When Marty Makary left FDA, Kyle Diamantas took over as acting Commissioner. Weeks later, there’s still no permanent Commissioner nominee, but there’s three finalists: White House aide Heidi Overton, oncologist Jeffrey Vacirca, and Pentagon health official Stephen Ferrara.

Those candidates aren’t device savvy. An observer described Overton as untested on FDA program innovation – not a known champion, but not on the record opposing reform. Vacirca comes from the clinical care world, leading a prestigious New York oncology group and serving as past president of the Community Oncology Alliance. Ferrara has health-system management chops. He’s a senior Pentagon executive responsible for health affairs, following a 25-year Navy career that includes a stint as the CIA’s chief medical officer. That’s an operational, not a regulator résumé, but some early reporting has him as the frontrunner. That makes the gaps in Ferrara’s record more consequential, not less.

The device overlay is real. Under Makary, FDA revised the CDSS and General Wellness guidances, narrowing what FDA considers a regulated device, and FDA and CMS created a coverage fast track for breakthrough devices. None of this went through formal rulemaking and there was little or no public comment; it’s policy by fiat that the next Commissioner could endorse, ignore, or tear up.

What does this mean for device firms?

Diamantas shows what happens when the Commissioner has no device background. He’s kept a low profile and his stakeholder meetings – including BIO, PhRMA, and right-to-try advocates – skew toward drugs. CDRH Director Tarver and center career staff are maintaining device momentum through initiatives like the breakthrough device program, TAP (faster device development), and RAPID (faster device payments).

That approach is trickier for politically contested programs like RAPID. It was launched with a press announcement instead of a binding rule, meaning it needs a champion with political capital to defend it. Career staff don’t have that capital; they can administer programs but not always protect them.

With Overton as Commissioner, device deregulation likely rides on the signals she gets from RFK Jr. and the White House. Vacirca knows community drug access, but he doesn’t know and may not prioritize device issues. Ferrara has federal health-management experience but, like the other candidates, no device policy expertise.

These gaps matter for initiatives like RAPID, with thin political armor. RAPID is popular within CDRH and with industry, but popularity isn’t durability. RAPID lacks solid backing like agency rulemaking or guidance, and popularity might not protect it from attack. In that case, RAPID needs a sponsor with political capital. Without it, RAPID and similar initiatives are vulnerable.

What should device firms do?

Don’t sit this one out. Industry should identify which Commissioner candidates it supports and which ones it doesn’t and make both known to Congress and the administration before a nominee is picked. The right messenger is trade associations, not individual companies. Unified messages carry more weight with the White House and the Senate HELP committee, which holds the confirmation hearing, votes the nominee out of committee (or doesn’t), and builds a record that tracks nominee positions and commitments.

Beyond that:

• Get device questions into the confirmation record. Senators routinely submit confirmation questions for the record. Device associations should be feeding HELP Committee members questions to build a trail of positions and commitments that can be invoked later.

• Look beyond the Commissioner pick. CDRH initiatives run through Tarver and center staff who outlast Commissioners. Engaging CDRH leaders and staff counts no matter who sits atop the org. chart. Effective engagement means picking the guidances, policies, and programs that industry supports and working with the Center to drive them. CDRH is in charge, so the question is what can industry do to further initiatives that it cares about. In areas of novelty and uncertainty, the Center has a record of working with industry toward good results, and that doesn’t depend on the Commissioner pick.

• Oppose podium policy. Industry has a window, while Diamantas is in charge, to push CDRH to formalize initiatives like RAPID and the CDSS and General Wellness guidances. If rulemaking is too cumbersome, then the initiatives require at least public notice and comment. Formal measures insulate the initiatives from whoever takes the permanent Commissioner job. The alternative is podium policy and that’s bad strategy.

• Expect a long wait. The Senate calendar is overcrowded with pending nominations and this is an election year, meaning less time to vet and vote on nominees. Even after the administration announces its Commissioner pick, Senate confirmation will be slow. Diamantis can remain acting Commissioner through early December and device firms should build that trajectory into their regulatory planning.

The Commissioner search will dominate headlines as a drug and biologics story. Device firms on the sidelines cede control to stakeholders who aren’t considering CDRH. Innovation quickened when Makary spent political capital on it, but a Commissioner without device experience might reject that approach. Industry action now, with clear messaging to the White House and the Senate HELP committee, hedges against that result.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?

Makary’s announcement was podium policy and it’s notable that FDA issued updated CDSS and wellness device guidances without first seeking stakeholder input. The guidances substantively differ from earlier versions, making stakeholder input important and appropriate.

On May 12, Makary resigned and Kyle Diamantas took over as acting FDA commissioner. Diamantas has worked through established channels instead of the press announcements that Makary favored. Diamantas doesn’t want the permanent commissioner job, which means new nomination(s) and a new commissioner.

Meanwhile, FDA’s Digital Health Center of Excellence (DHCoE) has cycled through four directors since January 2025. In February 2026, AI-industry veteran Rick Abramson took over. Abramson has been advising the commissioner’s office since last summer and he helped shape FDA’s digital deregulatory push.

Where do things stand? The status quo remains (in part). The FDA/CMS TEMPO pilot, which gives some digital devices fast payment and market access, is on track. Likewise, the CDSS and wellness device guidances are unchanged since Makary announced them. But a recent public hearing on the Commissioner’s National Priority Voucher (CNPV) (giving accelerated drug review) tells a different story. Stakeholders there pushed FDA to add transparency and predictability, for example through formal guidance or rulemaking.

What does this mean for device firms?

The risk is that Makary’s public announcements won’t hold. He swapped speed for slower and more durable agency process. That process gives programs life beyond any one commissioner. Programs with documented rationales, prior notice, stakeholder input, and formal communication, are hard for successor commissioners to unwind. Programs created with press events – built on a commissioner’s directive and without stakeholder input – can be discarded with the same ad hoc process used to create them.

Plus, as the DHCoE shows, leadership churn creates uncertainty. Rick Abramson’s selection as the DHCoE chief is promising, but he faces zero job protection, an acting commissioner who doesn’t want the top job, and a question mark where the permanent commissioner should be. There’s no guarantee that the new commissioner will share Abramson’s deregulatory, industry-friendly stance. This is an administration where policy blows with the wind (much of it coming from the President). As the recent CNPV hearing shows, stakeholders sensing this fragility are seeking formal process instead of hoping that informality holds.

What should device firms do?

Device firms must choose their priorities. An FDA hands-off approach gets products to market faster but may not include policy predictability and durability. As troubling, fast tracking review can slow coverage. CMS and private payers want the kind of durable, documented markers that notice-and-comment actions provide. Firms should decide, feature by feature, whether speed to market or coverage certainty matters more, because the regulatory posture serving one doesn’t always serve the other.

Next, firms should push FDA to formalize the process features that they favor. That means engaging now, before a new commissioner arrives with views and goals that may not match industry priorities. The CNPV hearing is a template, with stakeholders seeking process clarity and durability. Device firms can follow suit, choosing process over podium policy. No question, this is a trade off with industry swapping speed for stability. But the switch helps ensure that initiatives like the TEMPO pilot and the CDSS and the wellness guidances outlast the commissioner who announced them.

Formal agency action takes time. Meanwhile, device firms should use the FDA guidances, pilot programs, and other initiatives now in place. That includes relying on the CDSS and general wellness guidances and joining the TEMPO pilot if worthwhile. But these initiatives are vulnerable to change. Firms protect themselves by documenting their reliance on agency directives and looking to Abramson and the eventual permanent commissioner as leading indicators for regulatory strategy. Firms that base strategy on podium policy are making a mistake.

Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?[1]

FDA and industry agreed in principle on MDUFA VI, the next round of device user fees, which runs through 2032. The deal covers industry priorities including FDA hiring and staffing transparency.

FDA is drafting a formal commitment letter to send to HHS before Congressional briefings, Federal Register publication, and a public-comment period. The letter is due to Congress by mid-January next year, giving lawmakers time to act before MDUFA V expires in September 2027.

What does this mean for device firms?

MDUFA VI has key objectives, including FDA hiring commitments that industry should watch.

MDUFA VI’s hiring language is vaguer than MDUFA V’s. Under MDUFA V, FDA and industry agreed to yearly hiring targets, with FDA promising at least 144 new hires in 2023 and 42 new hires in 2024. Missed hiring goals triggered fee reductions. MDUFA VI jettisons these guardrails. FDA proposes instead “language similar to the MDUFA IV Commitment Letter” (which lacks MDUFA V’s specific hiring agreements).

That leaves FDA and industry with “shared intent to hire . . . additional review staff.” MDUFA VI also lacks an independent assessment to validate CDRH hiring metrics. But variable and sometimes opaque metrics make this assessment essential. Without independent oversight, it’s hard to confirm that FDA hiring translates into real, functional capacity.

Delays and salary disparities weaken FDA hiring. FDA hiring is notoriously slow. The federal hiring target timeline is 80 days but filling scientific and technical roles can take hundreds of days more. CDRH needs expert scientists, engineers, and regulatory professionals to handle areas like AI/ML, digital health, and cybersecurity. GAO found that FDA struggles to recruit and retain staff because it competes with industry, which pays more. Congress gave FDA more salary authority under the 21st Century Cures Act, but disparities remain.

Political uncertainty makes things worse. DOGE cut swaths through FDA in February 2025, including CDRH new hires with technology expertise. Staff reductions two months later emptied FDA offices responsible for employee recruiting, onboarding, and training, gutting the infrastructure that CDRH needs for hiring. These losses left remaining staff with heavy workloads and a culture of “fear and anxiety,” driving more departures.

HHS brought back some fired workers, but that model won’t help CDRH. The administration spent early 2025 defaming and firing federal workers. What sane person would return to a toxic environment for a job they get pushed out of?

And outside candidates aren’t a better option. Even were CDRH to target private sector or academic applicants, who’d choose government work knowing that political whims could get them fired? This month, the administration threatened to reclassify thousands of federal employees as at-will workers stripped of civil service protections. That’s not a compelling pitch for job candidates with other choices.

Attrition makes things worse. Even in ordinary times (which these aren’t), FDA loses staff to retirement, better offers, and relocation. Now, attrition is accelerating due to political uncertainty and deteriorating morale. When CDRH must put time and energy into hiring just to keep staff levels flat, gains become an unlikely goal.

With vague targets, no independent oversight, and a damaged recruiting infrastructure, delivering on hiring commitments will be tough. Will CDRH hire more staff? Yes. Will it fully meet hiring promises? Not likely.

What should device firms do?

Track CDRH’s hiring progress. General hiring pledges aren’t enough. During user-fee negotiations, industry sought office-level staff reports and CDRH demurred. Industry should keep pushing for hiring granularity. The metrics that matter show actual recruiting progress, staff in specific program areas, and attrition and retention rates.

Comment on, and recruit others to comment on, the user-fee agreement. The MDUFA VI commitment letter will publish later this year, opening a public comment period. Device firms should comment on how hiring commitments are defined, measured, and enforced. The MDUFA V model – with numeric targets, fee-adjustment triggers, and independent assessment – is the right benchmark.

A recurring question is whether firms should act alone or through industry associations. The answer depends on whether engagement topics are firm specific or common concerns. For common concerns, firms and even industry associations shouldn’t act alone. Public-interest groups, patient advocates, and healthcare providers all need effective CDRH staffing. Industry should tell these groups about the staffing challenges facing CDRH: slow hiring, damaging administration actions, and tepid MDUFA VI commitments. These groups can then join industry in pushing for transparent, robust, and verifiable hiring measures.

Share staffing strategies with FDA. Industry has navigated competitive talent markets and developed approaches – university partnerships, internship, and targeted compensation – that CDRH can adopt. Firms should share their effective hiring strategies with CDRH. Offering practical ideas creates goodwill, helps CDRH hit staffing targets, and serves industry’s need for an adequately resourced regulator.

MDUFA VI produced a fast deal, but speed and execution are different. Making MDUFA VI work requires hiring, training, and retaining qualified staff. Vague hiring language, no independent oversight, gutted recruiting teams, and political leaders hostile to staff aren’t a recipe for success. Device firms should welcome what MDUFA VI gets right, help fix what it gets wrong, and expect an under-staffed CDRH.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?[1]

For decades, device regulation was largely a dialogue between device firms and FDA. Patients were the implicit beneficiaries but rarely participated. FDA’s patient preference information (PPI) guidance helps change that.

PPI refers to qualitative and quantitative patient preference data. The data shows device attributes, risks, and outcomes, i.e., the trade-offs patients will accept when choosing a device. FDA issued its first PPI guidance in 2016. That version was limited, focusing primarily on PMAs, HDEs, and De Novos, and preferencing premarket topics.

On March 30, FDA issued updated final guidance, with two changes standing out. First is scope expansion. PPI now informs FDA decision-making widely, extending even to IDEs and breakthrough devices. Second, PPI now influences postmarket enforcement: the guidance clarifies that FDA administrative and enforcement actions should consider PPI.

The 2026 guidance also explains how to conduct PPI studies. This includes preference elicitation and an attribute-development process that sponsors should coordinate with FDA as a keystone to study utility.

But the guidance doesn’t create new legal obligations. PPI studies remain voluntary, although their expanded role and visibility change the strategic calculus for firms seeking market authorization.

What does this mean for device firms?

Voluntary doesn’t mean unimportant. The 2016 guidance was voluntary and FDA folded PPI into device decisions in the years that followed. The 2026 guidance formalizes this trend. Firms can now use PPI to strengthen premarket submissions, develop labeling, and include patient interest in regulatory strategies.

The enforcement angle matters. The 2016 guidance was quiet on postmarket enforcement; the 2026 version isn’t. FDA notes that PPI may be relevant to administrative and enforcement actions, including whether patients want continued access to noncompliant devices. This doesn’t mean using PPI to dodge enforcement; it means that, for high-impact and high-demand devices, PPI may affect FDA’s oversight approach.

High-stakes decisions are where PPI earns its keep. PPI is most valuable when benefit-risk calls are close – such as devices with risks and benefits and devices for patients with few alternatives. In these cases, credible PPI can inform FDA about the risks that patients will accept.

Study design sets a higher bar. The 2026 guidance’s emphasis on attribute development is important. Flawed PPI studies – which omit key risks, misalign attributes with clinical endpoints, or misinterpret preference data – carry little weight in device reviews. Pre-study FDA coordination is crucial; firms that skip this step may produce expensive data that doesn’t move the needle.

What should device firms do?

Confirm PPI relevance. The first question is whether PPI matters. Devices with patient interfaces, serious-condition indications, and benefit-risk tradeoffs are the primary targets. Commodity and low-risk devices, where patient experience adds little to safety and effectiveness profiles, are poor PPI candidates. Firms should review their product portfolios to flag devices that need PPI assessment.

Use the Q-Sub program. The 2026 guidance highlights that attribute development – picking the benefits and risks to include in PPI studies – should be coordinated in advance with FDA. The Q-Sub program allows device firms and FDA staff to align on study design before firms invest in data collection. PPI can help shape clinical trial design, endpoint selection, and benefit-risk documentation. But PPI studies planned late in device development – which may patch submissions without informing them – will underperform.

Engage the Medical Device Innovation Consortium (MDIC). MDIC partners with FDA on PPI development and promotion. MDIC’s Science of Patient Input (SPI) program helps integrate patient preference into clinical studies, device design and development, and other points across the device lifecycle. Device firms should support the SPI program, which allows them to engage FDA as it promotes PPI.

Document patient input in business activities. Many device firms already engage patients, getting data from advisory boards, clinical trials, online channels, and other sources. Firms that document this input can narrow subsequent PPI submissions – studies with PPI from business sources need less additional data.

Patient voices have been part of FDA’s benefit-risk framework for a long time – in theory. The revised PPI guidance highlights patient input, applying rigorous methodology to an expanded scope. Device firms can use these changes as a competitive advantage. Well-developed PPI will speed premarket review, with benefits extending to postmarket events. Firms that don’t meet requirements can show FDA that patients still need their devices. That’s not the point of the revised guidance, but its effect matters.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?[1]

FDA launched the CNPV program last June as a tool to shorten drug and biologics review times – targeting one to two months instead of ten months for standard reviews. Participating products must advance one of five national priorities, such as representing innovative therapies, targeting unmet medical needs, or improving drug affordability. The program doesn’t replace other FDA expedited review programs.

The program gained traction fast. FDA granted seven CNPV approvals before Makary’s exit, ranging from gene therapies to affordable generics. But momentum isn’t clarity. Critics, including Congress, cited the program’s opaque selection criteria, the lack of statutory authorization, and political influence replacing scientific decisions.

FDA leaders echoed these concerns. Former CDER director George Tidmarsh reportedly questioned the program’s legality, and his successor Richard Pazdur raised concerns about the program’s transparency and legitimacy. FDA held a hearing yesterday to get feedback on the program’s eligibility criteria, selection process, sponsor responsibilities, and review procedures. Written comments are due by June 29. The program gained traction fast. FDA granted seven CNPV approvals before Makary’s exit, ranging from gene therapies to affordable generics. But momentum isn’t clarity. Critics, including Congress, cited the program’s opaque selection criteria, the lack of statutory authorization, and political influence replacing scientific decisions.

What does this mean for device firms?

The CNPV is a drugs and biologics program, so device firms might be tempted to treat it as someone else’s problem. That’s a mistake.

The CNPV has staying power. A program that outlasts the commissioner who created it is sticky. HHS’s decision to keep the CNPV is no accident. The program furthers administration priorities: cutting “red tape,” faster access to treatments, and sidelining notice-and-comment rulemaking. Ideologically aligned programs tend to grow, not disappear.

Device firms can address CNPV problems. That’s true even for firms that didn’t attend the June 4 hearing. Written comments give stakeholders who missed the hearing time to highlight concerns like CNPV priority categories that are so broad that almost any product qualifies. Firms should push for objective, measurable thresholds for these categories, including minimum evidence standards and inclusion/exclusion criteria. Without guardrails, eligibility is a judgment call, and judgment calls produce confusion and lawsuits.

Plus, the CNPV gives selection authority to a review council with no public charter and no external members. An independent structure – modeled on FDA’s advisory committee system or NIH’S study section model (both of which include outside participants) – insulates selections from political pressure, creates auditable records, and fosters the scientific and regulatory credibility needed to face judicial review.

A mature CNPV will eventually include devices. If a priority voucher program is durable, transparent, and successful, then political and industry pressure to extend it to devices will build. Device firms have the same regulatory needs as drug and biologics companies So, the question isn’t whether a device-equivalent program will happen; it’s how device firms can help shape the program before it goes live.

What should device firms do?

Submit comments. Device firms shouldn’t treat the June 4 hearing as a drugs and biologics event. Written comments on the CNPV will affect its structure and reach, including the criteria and governance measures that apply to devices. Comments are due June 29, and device firms should be on the record individually and through trade associations.

Build a device CNPV framework. In addition to comments, device firms shouldn’t wait for an FDA proposal before designing a device CNPV. The proposed design should reflect device efforts already in place like the Breakthrough Device Program and the Total Product Life Cycle Advisory Program (TAP), which promote early and focused FDA-industry engagement. A device CNPV built on these programs is faster and more defensible – scientifically, legally, and politically – than a coopted drug-review model.

Advocate for transparent, merit-based criteria. The biggest CNPV criticism is that its selection criteria are vague and politically driven. Device firms should push for objective, public standards – criteria applied to device submissions regardless of political relationships and administration priorities. Establishing these principles now for drugs and biologics creates a template for devices.

Device firms have largely watched the CNPV from the sidelines. Now it’s time to get in the game. Firms that engage – shaping the program’s structure and insisting on transparent, defined criteria – will be best positioned when the CNPV incorporates devices.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What Happened?

Marty Makary resigned as FDA Commissioner earlier this month after weeks of reported clashes with White House and HHS leaders. President Trump announced that Kyle Diamantas, then the Deputy Commissioner for Foods, would take over. Within days, FDA underwent a broader leadership overhaul – CDER and CBER both received new acting leaders and FDA stalwart Lowell Zeta was elevated to acting Chief of Staff.

Diamantas isn’t a doctor or a scientist. He spent his early FDA career on food regulation, but lately he’s focused on third-rail topics like abortion-pill studies and lost scientific integrity. Comments from industry on Diamantas have been good, with an observer calling him a “drama-free, stabilizing caretaker.” Congress has been more tepid: Senator Bill Cassidy, who chairs the HELP Committee, cautioned that Diamantas lacks the expertise to oversee FDA’s complex device and pharmaceutical functions.

What Does This Mean for Device Firms?

Diamantas will be around for a while. The FDA Commissioner is a Senate-confirmed position, and the administration has a crowded confirmation queue with bigger tickets than FDA’s chief. The Surgeon General slot – vacant after multiple failed nominations – remains unfilled, and the CDC director position is likewise fraught. These slots will take up Senate floor time and administration political capital. Factor in the compressed 2026 legislative calendar and approaching midterm elections, and there’s little prospect of a permanent FDA commissioner before 2027. Diamantas isn’t a quick placeholder; he’s the acting commissioner into next year.

Deimantas’s device experience is thin. Deimantas’s career has focused on foods. Before his promotion, Diamantas helped run FDA’s Human Foods Program, leading Senator Cassidy to question if he has the expertise to oversee the device program. None of this means that Diamantas will harm device regulation. It means that device firms can’t expect an acting commissioner with a feel for their priorities or the stakes in user fee negotiations and other high-profile device efforts.

Lowell Zeta helps (a bit). Zeta’s move to acting chief of staff is meaningful but not dispositive. His former portfolio encompassed “all FDA-related product[s]” and he’s spoken and published broadly on FDA topics. But he focused previously on food and drugs, not devices. Zeta’s presence in FDA’s inner circle likely won’t adjust those priorities.

CDRH leadership changes the calculus. The most stabilizing factor is Michelle Tarver. She’s the only senior FDA leader to survive from the Biden administration, showing that she’s politically adept. Adding to this is Tarver’s institutional and technical expertise, giving CDRH operational continuity amid commissioner-level chaos. Tarver can – and given her track record will – run CDRH capably. But there are limits to what a center director can do without an engaged commissioner. Budget negotiations, user fee reauthorization, and policy battles require support at the commissioner level. Tarver provides a floor of competent device regulation, but she’s not a substitute for an informed and engaged commissioner.

What Should Device Firms Do?

Watchful waiting is a legitimate strategy. This FDA leadership environment rewards restraint. Diamantas hasn’t revealed a device-specific agenda, and the administration’s policy bandwidth is consumed by battles over abortion drugs and MAHA priorities. Device firms that keep low profiles steer clear of the crossfire. That means tracking matters relevant to the device sector – like MDUFA VI, CDRH staffing, and AI policy – and engaging when not acting would produce worse results.

Invest in CDRH leadership. Good CDRH leaders fosters regulatory continuity, so device firms should enhance their leadership engagement. Effective engagement includes core leaders such as OHT chiefs and the managers they direct. Core staff was hard hit by 2025 cuts and device firms can help CDRH identify and plug gaps. Engagement opportunities run from Q-Subs and PCCP meetings to check-in calls, all of which allow discussions about how to grow CDRH leadership capability. Device firms know how to identify and develop leaders, and they should share those practices.

Protect MDUFA VI. FDA and industry crafted the MDUFA VI agreement in about half the time such negotiations normally take. The risk now is that administration mismanagement derails the agreement as it moves through Congress. Device firms, working through trade associations, should monitor congressional review and engage if administration friction threatens the agreement.

Build the education brief. A commissioner who lacks deep device expertise isn’t an adversary; he’s a blank slate. Device firms and their trade associations should develop concise, credible materials explaining the stakes of key device policy issues. The goal isn’t lobbying; it’s inoculation – ensuring that when Diamantas and his staff encounter device issues, they’re hearing industry’s framing instead of someone else’s.

Watch the commissioner’s calendar. Diamantas will eventually take meetings with device-sector stakeholders. Getting on his calendar early – before positions harden and before other regulated industries crowd it out – is crucial. Firms with a strong Washington presence will outperform those relying on less experienced, dispersed resources.

The key question facing device firms is whether and when Diamantas needs to be educated about devices. Tarver’s capable CDRH leadership means that most device questions can be answered at the center level. But when larger questions arise – like user fees, CDRH staffing and budget, and AI regulation – the Commissioner must be in the room. Industry can’t then afford an FDA leader who encounters device issues for the first time.

I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?[1]

Last June, FDA launched Elsa, a generative AI tool designed to assist reviewers, investigators, and other agency staff. Ex-FDA Commissioner Makary touted Elsa as holding “tremendous promise in accelerating the review time for new therapies.” FDA says it’s using Elsa to summarize adverse event reports, speed up clinical reviews, and identify inspection targets.

Inside CDRH, a companion tool called CDRH-GPT has been deployed with limited success. Per FDA, CDRH-GPT and Elsa operate in secure, closed environments. The reality is a bit messier. Reports emerged that CDRH-GPT was struggling with basic tasks like document uploads, submitting queries, and connecting to FDA systems. Meanwhile, Elsa sometimes hallucinates, referencing studies that don’t exist. And in February 2026, a politically-forced directive required FDA to migrate Elsa from Anthropic to Google, raising questions about data handling, system validation, and whether FDA’s administrative record can withstand scrutiny if different AI models reviewed different parts of the same submission.

All this is unfolding against a backdrop that should concern device firms. DOGE-driven layoffs hit CDRH hard, with AI and digital health staff among those terminated first. Many were re-hired, but April 2025 RIFs took another bite. The result is an agency that’s accelerating its use of AI after jettisoning the experts needed to validate it. The irony isn’t subtle.

What does this mean for device firms?

Industry conversations about Elsa often ask if it speeds reviews, but there are other important topics:

Transparency in AI-assisted decisions. FDA hasn’t publicly documented what Elsa does – or doesn’t do – when reviewing device submissions. That means sponsors don’t know when Elsa flags submissions, how Elsa’s output affects review decisions, or when regulatory correspondence has AI fingerprints. As an observer notes, sponsors face “limited visibility into how much AI influenced [regulatory] decisions, raising concerns about opaque or unvalidated reasoning.” FDA asserts that human reviewers retain final decision authority and they probably do. But the assertion side-steps how much human decisions are informed by upstream machine analysis.

Legal exposure and the administrative record. The Administrative Procedure Act (APA) requires agencies to explain and support their decisions with a reviewable record. When Elsa affects a regulatory outcome, experts caution that the administrative record must include both AI-generated artifacts and human reviewers’ rationales. FDA hasn’t said whether that’s happening, allowing device sponsors contesting adverse decisions to argue that the administrative record is insufficient.

The Anthropic-to-Google migration adds complexity: for submissions reviewed under different AI models, the resulting administrative records may contain inconsistent analyses. That creates APA vulnerability, possibly requiring FDA and sponsors to reconstruct what happened.

The Pre-Sub dynamic. Pre-Subs promote FDA feedback, reducing reviewer confusion and clarifying data requirements before costly device studies. That value model depends on human continuity – reviewers who remember conversations, guide evidence generation, and understand device functions, operations, and classification. When AI summarizes meetings, flags patterns, and generates questions, the FDA-sponsor exchange changes. Firms need to consider what they put in writing and how they document Pre-Sub interactions, especially when those interactions may be recorded by a tool with opaque processing.

Predictability. Device firms make multi-year investments based on their read of FDA expectations. FDA uses methods like guidance documents, precedent citation, and public forums to convey those expectations. AI tools pose new risks: an unvalidated, undocumented layer that can generate inconsistent outputs across similar submissions or cite imagined studies. That’s a predictability problem that challenges strategic business planning.

What should device firms do?

FDA use of AI is a net positive, allowing faster, more efficient, and more effective reviews. But the technology is new and the risks are real, and that calls for proactive industry action.

Document everything. Device firms should treat key FDA interactions as administrative record materials – timestamped, indexed, and preserved. Firms receiving FDA input that’s misaligned with their submissions should ask FDA to explain (in writing) the discrepancies. Experts recommend retaining copies of submissions, datasets, and communications with FDA, and exercising care in disclosing confidential information beyond what’s required.

Ask FDA directly. Sponsors can ask FDA whether it used AI to review their submissions and, if yes, which tools. FDA isn’t obliged to respond, but a written request creates a record. For adverse decisions that seem thinly reasoned or inconsistent with prior communications, AI use is a reason to seek dispute resolution or supervisory review.

Revisit submission architecture. Submission structure matters. Submissions that are disorganized, that bury key data, or that require reviewers to synthesize inputs may fare poorly when the first pass is algorithmic. Device firms must consider how their submissions read to a tool using pattern-matching and extraction, not a human conducting holistic review.

Engage collectively – and at the right place. Individual firms pressing FDA about AI may get limited traction. Collective action (e.g., through trade associations) amplifies concerns and signals that they’re industry wide. And picking the right FDA counterparts is key. The place to start is FDA’s Digital Health Center of Excellence (DHCoE), which identifies and helps resolve digital heath challenges (such as how CDRH uses AI). If engagement requires additional subject matter experts, the DHCoE can help recruit them.

Well-designed AI tools that reduce reviewer burdens, add data inputs, and enhance review consistency are genuinely valuable. The problem isn’t the concept – it’s execution challenged by staff depletion, political turbulence, and opacity about how these AI tools work. Device firms can’t wait for FDA to resolve these challenges; helping FDA fix them now will promote AI’s positive impact.

[1] I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

Makary announced the pilot program last week, a year after FDA’s push to expand unannounced foreign inspections. That push yielded little; don’t expect one-day site assessments to do better.

What happened?[1]

Most domestic inspections are unannounced, but foreign inspections require advance notice. FDA promised to expand unannounced foreign inspections, with Makary claiming last year that the domestic/foreign inspection double standard “ends today.” Last week, FDA rolled out a pilot program for one-day assessments – shorter, focused site screenings to “complement standard FDA inspections.”

The link between FDA’s inspection and audit announcements isn’t coincidental. Foreign inspections are logistically complex, making pre-announcement almost always necessary. Before investigators arrive, investigators must secure visas, arrange international travel, engage translators, and confirm that key personnel are available. Foreign inspection windows are short (about a week) and investigators who arrive to find employees or records unavailable waste weeks of planning and thousands of dollars. Logistics aren’t bureaucratic inconveniences; they’re structural prerequisites and they make unannounced foreign inspections difficult. Staffing shortages compound this challenge, restricting FDA’s use of unannounced inspections.

FDA’s made limited progress on unannounced foreign inspections. The foundational barriers – visa mandates, travel planning, personnel coordination – are more difficult than ever. As ProPublica reported last November, 65% of the FDA staff responsible for coordinating travel and budgets has been fired or left FDA. Now investigators must add booking flights, obtaining diplomatic passports, and fronting expenses to their long list of duties. Current and former FDA staffers call the unannounced foreign inspection program “comical” given these constraints.

The administration won’t fix these problems. The GAO documented a 36% decline in FDA inspections between 2019 and 2023 (largely due to Covid). And that’s before the administration cut FDA staff. Bridging inspection gaps requires resources and expertise, including support staff to manage complex travel and investigators with the experience and expertise to do the work. With FDA gutted, those resources are gone.

Ex-Commissioner Makary floated hiring 1,000 then 3,000 new employees to backfill lost positions. That’s a pipe dream. Federal hiring is detailed, deliberate, and time-consuming. Filling a single position can take months. How FDA fills 3,000 positions efficiently is a mystery – and that assumes 3,000 qualified candidates would join an agency that eviscerated its workforce at the direction of an unelected billionaire.

One-day assessments try to show progress on inspections without fixing underlying inspection problems. And there are significant questions about what these assessments even are – legally and procedurally. They’re not true inspections. FDA calls them “assessments,” which raises regulatory and administrative law questions that the press release conspicuously ignores. Questions like:

• Must FDA make formal findings and must firms respond?

• Will findings include observations and Form FDA 483s that firms must address?

• Are assessment findings evidence for import alerts, warning letters, and enforcement actions?

What the press release does include is Makary’s prediction that assessments will provide feedback for “lower-risk establishments.” Why FDA would dedicate limited resources to lower-risk assessments, and why it would highlight that strategy, is a head scratcher.

Here’s the most fundamental question: can FDA meaningfully assess device sites in a single day? These aren’t dirty food warehouses. Device production is complex, requiring careful and informed review. The QMSR demands risk-based thinking across interconnected quality processes, which FDA reviews for systemic effectiveness. Charting a firm’s risk-management approach and deciding if it’s compliant takes time – time to review documents, interview personnel, and trace decisions. If one day is insufficient for this evaluation – and it is – then assessments will default to checking what can be checked in a day: simpler, lower-risk processes. That’s theater, not meaningful investigation.

What does this mean for MedTech firms?

Despite the public announcement, the press release, and the media coverage, one-day audits won’t meaningfully change device firms’ inspection experience. FDA has so far completed about 45 assessments across all regulated products, a number that doesn’t signal urgency or scale. Most of these assessments confirm compliance and that’s no surprise given the one-day timespan

Limited impact is more certain after Tuesday’s report that Commissioner Makary got fired (maybe he’ll come back as a reviewer to help FDA hit hiring targets). With Makary’s exit, the assessment program is liable to wither. New leadership in an agency with depleted resources will have little appetite for a predecessor’s uncertain pilot program.

There’s a further irony: Makary’s complaints about foreign inspections include claims that advance notice allows bad actors to conceal violations. One-day assessments may inadvertently cause this. Firms being assessed – especially when assessments go badly – now know two things: they’re on FDA’s radar and a full inspection is coming. Those firms can clean up their acts (or appear to do so) before FDA returns. That’s precisely what Makary claims to oppose.

Last, assessments’ unclear regulatory status and effect undercut their impact. If assessments can’t independently support enforcement, then their primary use is for triage – identifying sites for follow-up inspections. For a well-run organization, that’s not a threat; for firms with compliance troubles, it’s advance warning.

What should MedTech firms do?

Nothing beyond what they’re doing now.

One-day assessments aren’t likely to happen in meaningful numbers or produce significant findings. Even if they produce findings, they’re unlikely to produce adverse results (like warning letters or enforcement action), especially given open questions about assessments’ legal and regulatory status.

Still, there are constructive steps for firms to take:

Treat every day as inspection day. That’s the baseline for any compliant business. QMSR requirements don’t change based on inspection frequency or format. Firms must show compliance regardless of whether investigators show up for a day or a week.

But considering the assessment format, firms must also judge which aspects of their quality systems can be reviewed in a day. Those are the likely targets for one-day assessments. And as always, risk is the overriding consideration in QMSR inspections. Preparation for, and communication during, these assessments must clarify how risk permeates quality assurance.

Treat assessments as early-warning systems. If FDA shows up for a one-day assessment, take it seriously. Any findings – formal or informal – signal areas for scrutiny during a subsequent inspection. Address findings, show effective corrective and preventive action, and document closure. For pending actions, provide timelines that include stage gates, closure dates, and effectiveness checks.

Monitor regulatory events. FDA leadership is in flux and the incoming commissioner will bring new priorities. If one-day assessments survive this transition, note whether FDA resolves open procedural and administrative questions. The answers will determine whether assessments are a meaningful tool or an artifact of the Makary era.

Engage FDA through trade association. If assessments continue and questions remain, industry should press FDA for clarity. This is another instance where sending unified messages is more effective than separate firms making the same points. Trade associations, convening members and developing consensus views, are good forums to create and convey unified messages. And as always, the best strategy is to propose solutions for problems raised.

Don’t be distracted. One-day assessments are unlikely to affect device firms. What needs attention is the depletion of FDA investigators and the staff supporting them. Capacity for meaningful device inspections – the kind that find real problems and require real responses – is low. That has implications for compliance investments, inspection readiness, and risk-based quality planning. These concerns dwarf whatever the pilot program produces.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

Enforcement discretion is a rational choice for many wellness products. Skipping premarket review is faster and cheaper, letting firms focus on product development and commercialization instead of regulatory submissions. FDA designed the general wellness carveout because these products are low risk and the agency has limited bandwidth to review them. Taking the regulatory off-ramp makes sense (for firms where it makes sense).

But enforcement discretion isn’t free. There’s an opportunity cost for less regulatory oversight. So, before choosing enforcement discretion, firms should weigh what they’re giving up against what FDA review offers.

What happened?[1]

FDA’s January 6 guidance supersedes the 2019 version of its General Wellness policy, granting enforcement discretion to sensor-based products that estimate physiologic parameters. The guidance sets specific conditions: to qualify, products must be for wellness use only; must not substitute for an FDA-cleared device; and mustn’t prompt clinical action. Products meeting these conditions can go to market without FDA review.

The guidance is part of an FDA deregulatory push that includes a revised Clinical Decision Support Software Guidance, also published January 6. Commissioner Makary announced the changes at the Consumer Electronics Show, framing them as deregulatory steps to promote AI and ease digital health regulatory burdens. The message from FDA is clear: if you’re low-risk and wellness only, go build your product. That’s an important message, but “you don’t have to” isn’t the same as “you shouldn’t.”

What does this mean for MedTech firms?

Enforcement-discretion removes regulatory burdens, but not the strategic calculus for firms deciding whether to pursue premarket review. That calculus has four components.

Market differentiation. The wellness device market is crowded, with fitness trackers, sleep monitors, smart rings, and hundreds of other products competing for attention. FDA review signals credibility in this environment. Health systems, employers, and consumers increasingly note products that have been through scientific review, a distinction shared by 510(k)-cleared and De Novo-authorized wellness devices.

Coverage. This is the big one. CMS doesn’t cover wellness products that aren’t FDA-authorized, and neither do most private insurers. Enforcement discretion may put wellness products on the market, but those products don’t satisfy coverage requirements.

The coverage gap is widening. FDA and CMS recently announced the RAPID program, which offers Medicare payment for certain FDA-authorized breakthrough devices. Products that get to market quickly leave just as quickly when there’s no payment path. RAPID ties FDA review to Medicare coverage, with no similar payment prospects for products getting enforcement discretion.

Liability. Enforcement discretion may protect firms from FDA enforcement, but there’s no protection against private lawsuits. Federal preemption protects FDA-approved devices from suits, but there’s no preemption for devices on the market through enforcement discretion. And even without preemption, as one analyst notes, a device’s compliance with FDA regulations is relevant in product-liability suits. A wellness product with no regulatory record gives plaintiffs’ counsel a compelling narrative: “Company X made health-related claims with no independent validation.” Rebutting that account is hard without the evidence that premarket review generates.

Future ambitions. Firms may not plan to stay in the wellness lane forever. Commercial logic can point toward clinical indications – a sleep tracker that detects sleep apnea; a fitness wearable that flags cardiac arrhythmias; or an exercise monitor that helps manage chronic disease. Such pivots require FDA authorization and that’s faster and cheaper for firms with regulatory files, FDA relationships, and quality capability. Firms that skip FDA review aren’t saving regulatory costs; they’re deferring them.

International markets are another factor. The EU’s MDR framework, Canada’s Medical Devices Regulations, and many Asian markets require regulatory submissions that enforcement discretion doesn’t supply. Firms with 510(k)s, De Novos, and the like can better navigate those markets than firms starting from scratch.

What should MedTech firms do?

The first step is to stop treating enforcement discretion as the default answer and start treating it as one option among several – to be selected deliberately and based on commercial strategy, not regulatory minimalism. If a strategy involves Medicare or private insurance coverage, enforcement discretion is likely a dead end. The question isn’t whether FDA product review is required – it’s whether the product’s revenue model works without coverage.

Firms should also assess their litigation exposure, factoring in the data that the product generates, its wellness claims, and what plaintiffs’ counsel could make of no premarket review. Products without review are a better litigation target than products with agency oversight.

Firms should think next about their 3-5 year trajectory. If clinical claims, health system partnerships, or international expansion are in strategic plans, investing early in FDA review is cheaper than bolting it on later. Premarket review produces a product baseline, a regulatory record, and the predicate infrastructure that future submissions will need.

Last, firms choosing enforcement discretion should still maintain and document effective quality management systems. FDA offers enforcement discretion when deciding where to focus its resources – that’s not a promise of permanent inaction. Firms that operate without quality controls, performance validation, or complaint handling are exposed if enforcement priorities shift or if safety issues arise.

Enforcement discretion is faster and cheaper than premarket review, so it’s the right call for many firms. But it trades lower regulatory costs for exposure in areas like payment coverage and liability. Plus, enforcement discretion can limit market differentiation, product evolution, and OUS expansion. Smart firms will weigh these factors – the positives and the negatives – before choosing enforcement discretion.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

Readers will be shocked to learn that I’m sometimes critical of FDA leadership. But I admit when FDA gets it right and last week’s FDA-CMS announcement of the Regulatory Alignment for Predictable and Immediate Device (RAPID) pathway is one of those moments. RAPID allows FDA and CMS to speed Medicare coverage for some Class II and all Class III breakthrough devices. That’s a win.

What happened?^[1]

FDA’s Breakthrough Device Program promotes devices that address unmet needs for life-threatening and debilitating conditions. The program offers real benefits and, with more than 1,200 breakthrough designations, it’s popular.

But payment for breakthrough devices has been the missing piece to supercharge the program. A Stanford/Duke survey reports that nationwide Medicare device coverage can take nearly 5 years, and there are longstanding complaints about the timeline for breakthrough coverage.

The first Trump administration tried to fix this, ordering CMS to streamline payment for innovative technologies. CMS responded with the Medicare Coverage of Innovative Technology (MCIT) rule, which gave breakthrough devices four years of coverage. The Biden administration paused MCIT and CMS later rescinded it, promising alternative coverage that arrived in August 2024 – the Transitional Coverage for Emerging Technologies (TCET) pathway. Instead of automatic coverage upon FDA authorization, TCET conditioned coverage on actual evidence generation.

TCET’s biggest flaw was its narrow scope. It accepted only five candidates per year and that cap caused criticism. By mid-2025, a coalition of patient groups and life science organizations petitioned CMS for a better approach, stating that TCET was “incapable of providing robust and exhaustive coverage of breakthrough devices and diagnostics.”

It was just a matter of time before the second Trump administration acted. Nobody who hangs a picture of an auto pen for Biden in the Presidential Walk of Fame was going to let the MCIT rescission stand.

Last week, CMS and FDA jointly announced RAPID, which uses evidence from FDA reviews to support Medicare coverage decisions. CMS paused TCET as it focuses on RAPID implementation.

RAPID is available for some Class II breakthrough devices in FDA’s Total Product Life Cycle Advisory Program (TAP) and all Class III breakthrough devices. RAPID devices must be part of a clinical study that enrolls Medicare beneficiaries and assesses health outcomes agreed upon by FDA and CMS. RAPID could enable Medicare coverage within two months after market authorization.

RAPID also has constraints. Breakthrough designation isn’t enough for class II devices; they must also be TAP participants. That’s not true for class III breakthrough devices, but clinical trial requirements apply to all participants. Even so, RAPID does better than preceding programs (like TCET) in extending Medicare coverage to breakthrough devices.

What does this mean for device firms?

Some firms embrace breakthrough selection for reasons besides regulatory benefits. Selection offers visibility within FDA and in clinical and investor communities. Breakthrough devices address serious unmet needs, promoting credibility with physicians, payers, and strategic partners. This status is valuable in a crowded marketplace and RAPID adds a predictable path to faster Medicare coverage.

RAPID participants will co-design device clinical studies with FDA and CMS, helping assure that study endpoints support FDA market authorization and Medicare payment. That’s a meaningful commercial head start. Historically, even after marketing authorization, devices face coverage uncertainty, clouding investor expectations and launch timelines.

RAPID also has clinical trial implications. Studies enrolling Medicare beneficiaries – a RAPID requirement – must now reflect the realities of patient populations, including age profiles, comorbidities, and care settings. Medicare beneficiary data has been missing from breakthrough coverage decisions and firms whose clinical trials generate this data will be well-positioned for RAPID approval.

For the roughly 40 qualifying devices (with about 20 more candidates) RAPID is a strategic opportunity. Leveraging that opportunity requires agency engagement, even before the planned Federal Register (FR) notice announcing RAPID. The FR notice opens a 60-day comment period, allowing industry to clarify that RAPID’s success requires meaningful timelines, accountability, and effective program management.

What should device firms do?

First, firms should give credit where it’s due. That’s no small thing. With FDA Commissioner Makary facing broad criticism – visible industry support for RAPID matters. Positive press is currency in Washington and Makary could use some.

RAPID is a good initiative and it deserves backing. It promotes device innovation while preserving FDA and CMS jurisdiction. By aligning the agencies’ timelines and evidence requirements, RAPID reduces redundancy and unnecessary delays. As a proof-of-concept showing that interagency programs strengthen regulation, RAPID fosters similar programs.

Second, firms should give feedback. The RAPID FR notice will solicit input and stakeholders need answers about program eligibility, TAP mandates, and clinical study requirements. Now is the time to ask questions, raise concerns, and propose solutions. Firms with eligible devices – or firms planning them – should submit detailed, data-driven comments and coordinate with other stakeholders to present unified views.

Last, savvy firms will use RAPID as a commercial differentiator. Investors, hospital systems, and health care providers respond to payment risk. Expedited coverage is a commercial differentiator and RAPID participants will have a meaningful story to tell in fundraising and partnership discussions.

Breakthrough device coverage has left wrecks along the highway, including good plans (like MCIT and TCET) with disappointing execution. Device firms must help assure that RAPID is different. Success requires specificity: defined timelines, clear accountability, and transparent outcomes. Device firms should work alongside FDA and CMS to develop and implement these measures.

[1] I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

Trump administration workforce reductions hit CDRH hard. In February 2025, DOGE-directed firings targeted probationary employees, including experts needed to oversee new technologies. More firings followed last April, removing administrative, communications, regulatory, and management staff. From September 2024 to January 2026, CDRH lost about 20% of its workforce.

The damage was concentrated in areas where CDRH needs depth, such as AI and digital health. CDRH’s science and engineering office, which reviews complex technologies, lost about one-quarter of its staff. The April firings made things worse, gutting the offices tasked with recruiting, onboarding, and training staff.

Into this vacuum, Commissioner Makary announced that FDA will hire more than 1000 new employees. Part of those hires must go to CDRH, commensurate with its losses and technical needs.

Hiring is central to MDUFA VI negotiations. Industry agreed to pay for CDRH hiring under MDUFA V, but the Trump administration fired many of those hires and industry now demands accountability. FDA has acknowledged the “shared intent to hire . . . additional review staff” without disclosing the details or accountability measures behind that pledge.

Regardless, headcount promises are just a start. Effective hiring requires a functioning administrative team to recruit, screen, select, and onboard staff. The April 2025 firings eliminated offices responsible for those functions. HHS says it will fill gaps by consolidating administrative support. We’ll see.

As HHS Sec. Kennedy often shows, this administration neglects staff recruitment and retention. Morale at FDA is at a historic low, with a culture of “fear and anxiety” accelerating attrition. Job protections that made federal service attractive are weakened or gone. As one analyst observed, FDA hiring can take 6-12 months and fired staff will think twice about returning to an agency that treats them as disposable. Worse, the separation of science and politics– part of FDA’s DNA – has been replaced by subordinating science to political priorities.

New CDRH staff also face onboarding challenges that hiring numbers don’t measure. Integrating staff into an organization as complex and technical as CDRH takes years. Mentoring, training, and knowledge transfer require the same experienced managers and experts that the administration pushed out. Those who remain are stretched thin, creating more burnout and departures.

What does this mean for device firms?

The first risk is industry paying for results that CDRH can’t deliver. If CDRH doesn’t hire adequately, or hires without effective staff integration, user fees won’t build its capacity and capability. MDUFA V authorized more than $420 million for 2026 alone and MDUFA VI will add to this. The gap between fees paid and services rendered is visible: Q-Sub meetings are being converted to written exchanges, review feedback is inconsistent, and experienced contacts are harder to reach.

Submission risk is also rising. CDRH is meeting review timelines (for now), drawing on work begun when staffing was higher. The lagging indicators are okay; the leading indicators aren’t. Firms making submissions in 2026 – reviewed by a smaller and less experienced team with less oversight – face risk and, with lost CDRH expertise, novel and complex devices face higher risks.

CDRH capacity during MDUFA VI (which runs through 2032) is uncertain, clouding industry’s device development and regulatory plans. If CDRH doesn’t meet hiring commitments – or meets them in headcount but not with qualified, trained, and integrated staff – hiring efforts won’t work. For example, Q-Subs have been critical to de-risk submissions, but reduced CDRH capacity now limits Q-Subs, threatening submission quality and timelines.

A depleted CDRH isn’t just a review-timing problem. It hurts feedback quality and consistency, the pace and scope of guidance, and responses to enforcement and compliance concerns. These results serve no one – not industry, not patients, and not the administration that caused the problem.

What should device firms do?

Hold CDRH accountable for real staffing, not just numbers. The MDUFA VI commitment letter must include hiring promises beyond advertised positions and new employees. CDRH must deliver effective staff training, retention, and integration. Tracking and accountability are key and industry’s proposal for quarterly office-level staffing reports makes sense. CDRH agreed only to center-level reporting; if that’s insufficient, industry should again push for more granularity.

Demand administrative infrastructure. Hiring announcements are meaningless if CDRH lacks the support to recruit, screen, onboard, and train new staff. The offices performing those functions were gutted, so CDRH must now explain, transparently, how it’s rebuilding this infrastructure. Industry must require milestones to gauge whether this rebuilding is working.

Stress-test regulatory timelines. It’s a mistake to use past benchmarks to estimate CDRH timelines. Device firms must prepare for slower or shallower feedback and consider whether submitting sooner – before capacity degrades further – reduces risk. AI and digital health devices face an additional concern: lost expertise means that novel submissions may be reviewed by staff without the skills and training needed to evaluate them.

Protect Q-Subs. Q-subs remain among CDRH’s best engagement tools and device firms should use them proactively, even if they’re weakened and even if regulatory strategy seems clear. When CDRH converts Q-Subs to written exchanges, firms must deliver high-quality submissions. If CDRH response quality drops, firms should push for better information, including by effectively engaging CDRH supervisors. Firms must also preserve submission records and FDA correspondence. Institutional memory at CDRH is degraded and firms may need to educate new reviewers on device administrative and regulatory history.

Document gaps between fees paid and services received. If CDRH doesn’t meet MDUFA VI commitments – in hiring, training, or reviews – device firms should tell Congress what happened. MDUFA is a Congressional program and Congressional oversight spurs FDA accountability. Firms should track and document poor FDA performance and report to Congress, especially members on the House Energy and Commerce and the Senate HELP Committees, which oversee FDA.

The window to shape MDUFA VI is closing fast. February agreements are being translated into a commitment letter with hiring promises, reporting requirements, and accountability measures. The time to act is now – through trade associations, through congressional outreach, and through CDRH engagement. CDRH’s hiring challenges are solvable, but only if industry holds it accountable for real solutions, not just reassuring numbers.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?[1]

The Home as a Health Care Hub’s key theme is that the home should be an integral part of healthcare – not just a place to recover, but an active site of care, prevention, and wellness. As CDRH noted when launching the initiative, the goal is to shift care models from systems to people, addressing chronic shortages, rising costs, and traditional healthcare’s limits in reaching rural and lower-income communities.

The initiative has traction and it’s growing. CDRH unveiled a virtual reality prototype to visualize how products fit into home environments, including affordable housing. The Home Innovation Challenge will select up to nine devices from different manufacturers and participants will receive early FDA engagement, feedback on device design and testing, and technology review at FDA research facilities. Clinicians support this momentum: 81 percent of surveyed clinicians want to expand home-care investments and 70 percent say these programs enhance access to care. This support helps make the Home Innovation Challenge a strategic regulatory advantage.

Two forces ensure that home healthcare isn’t a passing fad.

The first force is CDRH Director Michelle Tarver. She built her FDA career around patient empowerment and patient-centered device regulation. Before leading CDRH, Tarver directed Patient Science and Engagement and helped lead CDRH’s Office of Strategic Partnerships. Home as a Health Care Hub was among Tarver’s signature projects before she became CDRH Director, and she describes the animating vision succinctly: achieving the smallest footprint with the biggest benefit for home-based care. Her philosophy is embedded in CDRH’s home-use device plans.

The second force is political palatability. This administration has been aggressively hostile to diversity, equity, and inclusion. On Inauguration Day, Trump signed Executive Order 14151, terminating DEI programs across the federal government and directing agencies to dismantle DEI offices and DEI-related websites and terminate contracts promoting DEI. FDA DEI programs are now dead on arrival.

Home healthcare is different. It’s diversity-neutral in the most politically useful sense: everybody gets sick, everybody ages, and everybody benefits from cheaper care where they live. Crucially, rural communities – a reliable Trump stronghold – are among the populations most underserved by traditional healthcare. Trump campaigned on a promise to improve rural health and home healthcare delivers on that pledge.

What does this mean for device firms?

CDRH is more receptive to home-use medical products and that trend will continue.

For marketed devices, the question is whether products can transition to a home setting. The analysis is familiar – does home use mean that devices are better or easier for patients and does home use lower cost? If yes, then pursuing home use makes sense, especially with now-smoother CDRH regulatory pathways.

For firms developing new devices, home-use capability is a design factor to be built now, not bolted on later. Devices empowering care at home will fare better at CDRH than similar devices requiring clinical supervision. This even applies to AI devices. The home-use path is available if firms demonstrate that patients can use their devices safely at home and without professional oversight.

This analysis matches healthcare-delivery trends. Ambulatory surgery centers account for 72 percent of U.S. surgeries, offering patients lower costs compared to hospitals. The shift away from hospitals is reshaping how and where healthcare gets delivered, and home-use devices are a logical next step in that progression.

What should device firms do?

Device firms must design and innovate based on where healthcare is actually delivered. That means asking whether a device can be used safely and effectively at home and then supporting related CDRH initiatives. Firms have until September 30, 2026, to apply for the Home Innovation Challenge and participants will begin CDRH engagement about two months later. Like CDRH’s breakthrough device program, there’s no guarantee of device approval. But also like the breakthrough program, participants will have early and iterative CDRH engagement, including “sprint” discussions and frequent feedback on device design and testing. This extra attention requires A-game device development and submissions. Done right, it establishes credibility and regulatory insights that pay dividends for upcoming submission and ones down the road.

Device firms should also invest early in human factors and usability engineering for home environments. Home settings are fundamentally different from clinical ones: lighting is variable, users are lay caregivers, and distractions abound. The human factors guidance already requires device firms to demonstrate safety in intended-use environments, and that standard will apply more rigorously to home-use devices. Firms that invest in human factors testing early – including in diverse, real-world environments – will generate evidence that CDRH needs and reduce the risk of late-stage redesign.

Finally, firms should track and participate in the policy process. CDRH is soliciting stakeholder input as it builds out the Home as a Healthcare Hub framework. Input allows firms to influence the standards, guidance, and coverage policies that will govern home-use devices.

But device firms must also act on marketplace reality. Home use is not a panacea and some devices are designed for, and should remain in, traditional clinical settings (no one is putting MRIs in home garages). The question is not whether home use is universally appropriate, but whether it’s appropriate for a particular device and patient population.

Understanding when the home-use pathway is right – and how to navigate it – is key to device firm competitiveness. Firms that build this knowledge now, while CDRH is mapping the home-use terrain, will do better than firms that treat home use as a future concern. Healthcare is going home and the question is which firms will be there when it arrives.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.

What happened?[1]

The draft guidance was long-awaited. FDA missed a 2024 deadline and published the draft last year. The guidance covers AI device submissions, including device description, risk assessment, validation, performance monitoring, and cybersecurity. It also takes a Total Product Life Cycle approach, meaning that FDA oversight extends through the device’s market life.

The comment period generated substantial feedback, including questions about generative AI, foundation models, and large language models – technologies that the draft doesn’t squarely address. FDA must work through those comments before finalizing the draft, meaning that the final version could change materially.

Meanwhile, CDRH’s track record on B-list guidances is poor. In 2025, the center published only 35% of its A-List guidances and just one B-list guidance. Plus, CDRH’s commitment to finalize draft guidances 3 years after comments end puts the clock at April 2028. CDRH faces real constraints in meeting that deadline. The center lost staff with AI and machine learning expertise in the 2025 reductions-in-force, and the administration’s 10-for-1 deregulatory executive order has chilled guidance publication.

What does this mean for device firms?

The draft guidance reflects FDA’s current thinking and CDRH reviewers are using it. But draft guidances can change and there’s no guarantee that the final version will match what the draft requires. This creates risk in three areas.

The first is submissions. The draft requires AI-specific documentation, including model descriptions, bias analysis, and performance monitoring. Firms building submissions around these requirements face uncertainty. Comments on the draft are contentious; with industry pushing back on the draft’s treatment of generative AI and foundation models, which represent a growing share of AI device submissions and which the draft handles only obliquely. The draft also creates ambiguity by using terms – especially around validation – that differ from how AI-developers use those terms. Last, the draft was written before the QMSR brought cybersecurity into device quality management systems. So, the final AI guidance must integrate a QMS framework different from the one in place when the draft was published.

Second is Predetermined Change Control Plan (PCCP) risk. PCCPs are a powerful tool, supporting FDA pre-authorization for planned modifications to AI algorithms. That’s a big regulatory efficiency, but PCCP value depends on how carefully and comprehensively they’re scoped. PCCPs must specify, upfront, how planned modifications will be developed, validated, and deployed. Modifications outside the PCCP require a new submission before they happen.

As worrying, the final PCCP guidance was written against the AI draft guidance. The draft’s framing of what’s a significant versus non-significant device modification informs how firms scope their PCCPs. If the final AI guidance redraws those lines, firms could find that their PCCPs no longer cover planned changes. Bridging such a gap would be expensive, time-consuming, and disruptive.

Third is post-market risk. AI devices pose transparency, validation, and drift concerns. Without final guidance defining the performance monitoring that FDA requires, and what thresholds trigger new submissions, firms face uncertainty managing post-market obligations.

What should device firms do?

Use the draft guidance – but don’t treat it as final. The draft guidance is the best chart of FDA expectations and it should set the floor for AI device submissions. But firms should build their submissions anticipating that final requirements could be more demanding, particularly around generative AI, bias documentation, and real-world performance monitoring. There’s no value in over-optimizing for draft language that could change.

Use the Q-Sub program. Absent final guidance, a Q-Sub meeting is the best way to confirm what FDA requires for AI devices. Device firms should treat Q-subs as a required step, not an option, especially for novel AI applications and devices using foundation models. Firms must also remember that Q-subs are device specific, so the lessons learned in one meeting may not apply to other AI device submissions.

Seek PCCPs. The PCCP guidance is final and more stable than the draft AI guidance. A well-scoped PCCP with closely-defined device modifications protects against disruption from AI guidance finalization. It’s better to scope device modifications narrowly and expand them later than define them broadly based on draft guidance that may not survive.

Plan for the B-list reality. Firms should war-game two adverse scenarios: (1) FDA finalizes the AI guidance with material changes that require reworking submissions; or (2) FDA doesn’t finalize the AI guidance and firms must operate indefinitely with a draft. Neither scenario is a catastrophe, but both threaten firms that don’t plan submissions. Firms minimize risk by monitoring the CDRH guidance page and like sources, and then building strategies and review milestones into their AI plans.

The AI device market is moving faster than the regulators who oversee it. The resulting gap creates strategic opportunities, and risk. Firms that treat the draft AI guidance as a floor (not a ceiling), that act on direct FDA messages (not assumptions), and that build regulatory strategies to survive guidance changes will fare better than competitors who wait for final guidances that may never come. Executives who miss this lesson may find themselves explaining why their AI-review strategies didn’t survive FDA guidance.

What happened?[1]

In February, FDA launched the QMSR and retired the Quality System Inspection Technique (QSIT), replacing it with Compliance Program Manual CP 7382.850 (CPM).

The biggest CPM change is inspection of internal company records. Under the QSIT, routine inspections didn’t assess internal quality audits, management reviews, and supplier appraisals. But the QMSR authorizes investigators to check all three.

Two months in, what has industry actually seen? Publicly available data – like QMSR inspection observations and warning letters – is sparse. That’s not surprising: inspection results lag inspections, with worse delays for warning letter publication. QMSR enforcement likely will face similar delays. Still, observers correctly predicted that FDA’s April 1 QMSR Town Hall would address what investigators are finding.

What does this mean for device firms?

Some lessons are emerging:

Internal audit reviews are real. Removing the audit shield creates a strategic dilemma that pre-QMSR comments (mine included) underestimated. Under the QSR an effective internal audit found and documented regulatory problems, independent of FDA. Under the QMSR, that audit becomes a roadmap for FDA investigators who then ask whether CAPAs are open, timely, and effective. An audit linking a finding to a stalled or inadequate investigation now effectively provides FDA with evidence for a 483 observation. This isn’t a reason to avoid internal audits – it’s a reason to ensure that what gets documented gets fixed.

Risk management is a lifecycle obligation. Under the QMSR, risk-based thinking isn’t confined to device design and development. FDA expects industry to integrate risk across the entire product lifecycle. Investigators won’t simply ask whether a risk management file exists; the file must be a living document, updated as new complaint data, adverse events, and field experiences emerge. This is a cultural shift. FDA’s QMSR preamble expressly frames risk management as a behavior and an organizational attitude, not just a record type.

ISO 13485 certification is little shelter. FDA has plainly stated that ISO 13485 certificates will not exempt device firms from inspection and QMSR requirements. As an observer notes, FDA-specific requirements sit on top of the ISO standard – including tighter record controls, new labeling and packaging requirements, and device-identification requirements. ISO-certified firms that haven’t done a gap analysis against these provisions face unidentified exposures.

Investigator knowledge is variable. A frequent complaint starts with FDA training investigators on ISO 13485 ahead of QMSR implementation. But workforce reductions jettisoned many of those investigators. The result is that some inspections are being conducted by investigators still learning the QMSR framework. That doesn’t reduce a firm’s compliance obligations. But firms should be ready to explain their compliance rationales to FDA staff lacking the expertise to understand the new framework.

What should device firms do?

First, firms must treat internal audits as a two-part process. Audits must identify problems and produce effective and timely solutions (often CAPAs). During inspections, audit findings are an inspection roadmap. The question investigators will ask isn’t “did you find the problem?” but “did you fix it?” Audit findings with open, aging, or ineffective CAPAs are a priority remediation item.

Second, risk management applies across the device lifecycle – not just in design controls. Firms should map risk-based decisions to supplier qualification, production process controls, complaint handling, and post-market surveillance. A risk file that doesn’t update in response to post-market signals is a structural gap. If an investigator asks how a complaint influenced the risk file, the firm needs a real answer, not a reference to a pre-launch document.

Next, treat Remote Regulatory Assessment (RRA) readiness as a baseline. FDA may request records in advance of or in lieu of an in-person inspection, and RRAs can be used to identify gaps before investigators arrive. Firms that can’t produce organized QMS records on demand risk bad RRA outcomes.

The QMSR’s two-month anniversary is not a milestone that comes with a clean empirical verdict. Enforcement signals are still developing. But the QMSR shift is already shaping inspections for FDA and industry, and firms will fare best by treating the new framework as an operational reality.

[1] Note: I sometimes use artificial intelligence in drafting this column. My primary tool is Claude AI. I review and verify all findings, interpretations, and final wording.