Two Former FDA Chiefs, Two Very Different Warnings

By Dan Schell, Chief Editor, Clinical Leader

I’ve now sat through two very different conversations with two former FDA commissioners — Scott Gottlieb and Robert Califf — in the span of a few weeks. Gottlieb, who now serves as a partner at venture capital firm New Enterprise Associates (NEA) and a senior fellow at the American Enterprise Institute, continues to publish policy commentary in national outlets like The Washington Post. His focus these days is on how regulation, innovation, and market forces intersect and what that means for the life sciences economy.

Califf, on the other hand, is channeling his energy into reforming the broader health system. The former commissioner and longtime cardiologist has been writing regularly on his Substack page, Califf’s Commentary, where he examines how U.S. medicine continues to fail as a “learning health system” — one that should be integrating evidence generation into everyday care rather than treating research as a separate industry.

Gottlieb warned that a leaner, under-resourced FDA is about to make life a lot harder for clinical operations teams, because guidance will slow down, reviews will get tougher, and the bar for evidence isn’t getting any lower. Califf, meanwhile, basically said: Sure, FDA is stressed, but the real problem is bigger. The American health system is structurally incapable of learning from itself, and unless we fix that, it won’t matter how fast you can answer an FDA query — you’ll still be running trials in a country that spends $5 trillion a year on healthcare and delivers the worst survival in the entire high-income world.

That’s the tension. Gottlieb is warning about the gatekeeper. Califf is warning about the game.

Slower, Harder, And Less Useful Clinical Trials

Califf’s first message is that clinical research in the U.S. is going the wrong way operationally, even though the theory of how to fix it has never looked better. “The theory has become far advanced,” he said. We know what good trial design should look like. We know what ‘learning health system’ means. We know how to use electronic health records, automation, and even AI to identify patients and capture data. But in practice, “it’s actually taking longer. Our protocols are more complicated and we’re sort of gummed up in terms of the speed we need to move.”

That’s not an abstract efficiency rant. He’s saying it’s now harder to actually run a trial that matters. Contracts take longer. Consents are longer. Sites are spending months just on CDAs. He literally asked why on earth a human experiment — a clinical protocol — should even be confidential in the first place.

He also says CRO layering is making things worse, not better. Sponsors and CROs routinely bolt three, four, five extra levels of process on top of what FDA and ICH GCP actually require, which turns routine execution into a compliance obstacle course and drives cost. Califf has receipts. He talks about running a 40,000-patient heart attack trial across 14 countries with a three-page case report form and a $300-per-patient site payment — and publishing 160 papers out of it. His point is not nostalgia. His point is that we’ve allowed complexity, paranoia, and billable tasks to replace judgment.

Now compare that to Gottlieb. When he spoke at DPHARM, his warning was that FDA itself is stretched thin. He described a leaner FDA that’s losing people, falling behind on policy work, and slowing its ability to generate clarity for industry. ClinOps teams should expect more scrutiny with less guidance, not less scrutiny with more flexibility.

So, you’ve got Califf saying sponsors and CROs are overbuilding process out of fear of FDA, and you’ve got Gottlieb saying FDA is so resource-constrained that sponsors shouldn’t expect hand-holding. Put those together and here’s the ugly math: Companies will continue to over-engineer trials to “protect” themselves from the FDA, but the FDA isn’t staffed to tell you which 60% of that work is pointless. That’s how you get a system where trials cost more, take longer, ask narrower questions, and still don’t give clinicians usable answers.

QbD Was Supposed To Save Us. We Ignored It.

Califf is very direct here. We are addicted to collecting junk data because we’re scared of being cited. He said sites treat FDA Form 483 like a horror movie monster. In his view, the 483 was never supposed to be a scarlet letter. It was supposed to be “someone noting that there's an issue that needs to be resolved with the quality of the work that's being done.” Yes, sometimes a 483 points to a legitimate disaster — he mentioned generic facilities with animals literally running through production. But more often it’s nitpicky documentation and signature minutiae that don’t change patient safety or answer the scientific question.

That fear, he says, drives all the waste at the back end of a trial. Teams dump staggering time and money into “the last 10% of stuff” that has nothing to do with whether the drug works, just to avoid getting publicly dinged.

Califf’s alternative is not new, but he’s mad that we still haven’t actually done it: True quality by design. Start with the real question. Decide which data actually drive that question. Define which fields must be pristine and which can tolerate noise. Stop treating every datapoint like it’s dosing in a first-in-human oncology study. He pointed out that FDA reviewers themselves don’t even look at most of what’s collected. In anti-infective work, he said 94% of site-collected data was never touched by FDA reviewers.

Here’s where Gottlieb’s and Califf’s messages rhyme. Gottlieb basically told the room that the FDA you’re dealing with now is lean, stressed, and politically exposed. When that happens, risk tolerance drops. You will feel that. Califf is describing what that “feeling” becomes downstream: CROs selling process, sponsors buying process, sites drowning in process, and patients either walking away or never being asked in the first place.

Research Can’t Sit Outside Of Care Anymore

Califf draws a hard line between two kinds of trials. On one side you’ve got early-phase work (first-in-human, dose-finding, most phase 2). That belongs in a controlled research environment. Measure everything. Time every blood draw. Nobody argues with that.

But on the other side, you’ve got the questions that actually matter for how doctors treat real people — Which therapy works better in practice? What’s the right dose in the wild, not in an ICU with four sub-investigators hovering over a PK curve? How do I sequence these two approved agents for the next patient in front of me? Califf’s position is blunt: We do not have a functioning system to answer those questions in the U.S. “There is no way you can do that kind of research separately from integrating it with clinical practice,” he said.

Right now, the U.S. does the opposite. We peel research away from care. We build standalone sites and site networks that look efficient on paper. We throw private equity money at scaling them. Then we act shocked when trial populations don’t look like the patients actually walking into community clinics. Califf’s take is that in the short term, these for-profit site networks might impose some badly needed operational discipline. Long-term, he thinks walling off research from routine care is “bad,” because you stop learning from the real population and you lock trials inside a business model instead of inside medicine.

He also argues that U.S. health systems actively make this worse. Most clinicians are on RVUs (Relative Value Unit). They’re told to produce billable visits, not contribute to evidence. If a physician stops to say, “I don’t actually know the best option here, let’s discuss a trial,” they worry the patient will just go down the street to the doctor who pretends to be certain. Meanwhile, administrators are explicitly incentivized to chase the most profitable procedures and the most profitable payer mix. In his words, “as long as we run a health system where profit is king and administrators are more powerful than clinicians,” we will not fix enrollment, diversity, or access.

Gottlieb isn’t talking at that system level. His day job in venture capital and policy commentary keeps him focused on the machinery that governs regulation and innovation — how FDA staffing, legislation, and public perception shape the environment for biotech and pharma. He said a thinner FDA workforce, plus louder political pressure, means sponsors and ClinOps teams should expect slower guidance development and tougher expectations with fewer people to clarify them. He’s essentially saying we need to prepare for regulatory drag. But, Califf is saying that regulatory drag is almost a sideshow compared to the structural rot in how U.S. healthcare (and by extension U.S. clinical research) is financed and rewarded.

Trust, Politics, And The Next Patient

Califf is worried about public trust, but not in the lazy “people need more education” way. He says Americans have lost trust not just in the FDA but in the entire medical enterprise, and in a lot of cases, we earned that. Patients can’t get an appointment. Hospitals talk about being “patient centered” with a straight face while designing every pathway around margin and payer mix. Pharma spent years fighting even basic transparency around trials. Of course people get cynical.

Layer onto that a political environment where national leadership openly undermines mainstream science while pushing unproven products, and you get a public that’s not sure who to believe. Califf’s answer is not “do more patient recruitment ads.” Instead, he suggests (probably as no surprise coming from a physician) we support the frontline clinicians. Patients still trust their own doctors, nurses, and pharmacists more than they trust institutions. If those people feel resourced, protected, and proud to bring up research as a valid care option, enrollment follows. If those people are burned out, RVU-driven, and terrified of the FDA, enrollment dies.

This is one of the few places Califf and Gottlieb land in roughly the same political neighborhood, even if they’re attacking it from different angles. Gottlieb told an audience that today’s FDA is strained by staff loss and political pressure, which is going to make ClinOps work harder to navigate expectations. Califf is saying that same pressure cooker spills downstream — it feeds fear of 483s, which feeds overbuilt trials, which feeds burnout, which feeds distrust.

Where This Leaves ClinOps

Califf is not optimistic about the status quo, but he’s not fatalistic either. He sees clear opportunities emerging as stakeholders (e.g., sites, health systems) increasingly recognize their ability to drive change.

• Cut The Noise. He’s adamant that we can strip protocols down to the true decision points inside routine care, pay investigators fairly for that focused work, and stop treating every study like a moon landing.
• Integrate Research Into Care. He wants health systems to treat clinical trials as part of standard practice — not as an off-to-the-side revenue hobby, not as a marketing tagline, but as the way we learn what actually helps the next patient. That means executives have to stop worshipping RVUs and start tolerating “I don’t know, let’s find out.” (Yeah, I know, easier said than done.)
• Stop Weaponizing 483s. He’s pushing for a culture where identifying and fixing problems is rewarded, not punished, because hiding issues just to look “clean” is how quality actually gets worse.
• Use AI For The Right Thing. Califf is pro-AI, but not as a magic enrollment robot. He sees AI and EHR connectivity as a way to cheaply capture data we already generate, and to help us answer real outcomes questions faster, at scale. He does not want AI used to cherry-pick only the “profitable” patients and dump the rest.

Gottlieb, for his part, is essentially calling for resilience: Brace for a leaner regulator and a tighter environment, and don’t expect clarity anytime soon. His policy and investment work keeps him close to the boardroom and the Beltway, where efficiency is about pipeline, not protocol. Califf, writing from the perspective of medicine and research, keeps aiming at the underlying rot — the separation of care from learning, and the fear-based bureaucracy that has turned evidence generation into an endurance test.

If you’re in clinical operations, here’s the un-fancy summary. Gottlieb says the referee is tired and angry. Califf says the field is cracked and tilting. They’re both right. And unfortunately, we don’t have the luxury of waiting this out.