At the Food and Drug Administration, the word evidence is supposed to behave like a well-trained lab animal: controlled, documented, reproducible, and not allowed to run around chewing through the furniture. Yet in 2025, under FDA Commissioner Marty Makary and top biologics regulator Vinay Prasad, the agency’s public conversation about evidence began to sound less like a marble courthouse and more like a jazz club. The beat was still science, but the rhythm suddenly became looser.
The controversy centers on a phrase critics borrowed from an old debate in alternative medicine: a “more fluid concept of evidence.” In plain English, it means being willing to consider signals beyond the traditional gold standard of randomized controlled trials, including mechanistic reasoning, surrogate endpoints, patient experience, real-world observations, and carefully interpreted anecdotes. That can be useful. It can also be dangerous. A hammer builds houses and breaks windows; the trick is not pretending it is a pillow.
Makary and Prasad entered the FDA with reputations as reformers. Both had criticized parts of American health policy during the COVID-19 era. Prasad, especially, was known for demanding rigorous evidence in oncology, vaccines, and public health. That is why observers noticed the contrast when the pair appeared willing to speak warmly about regulatory flexibility for cell and gene therapies while simultaneously demanding stricter clinical trial evidence for annual COVID-19 boosters in healthy adults.
This is the tension at the heart of the debate: Is the FDA modernizing evidence standards for an era of precision medicine, rare diseases, and individualized therapies? Or is it applying rigor selectively, asking vaccines to climb Mount Everest while allowing favored experimental treatments to take the scenic elevator?
Why the FDA Evidence Debate Matters
The FDA is not just a paperwork machine with fluorescent lighting. Its decisions shape what treatments reach patients, what risks society accepts, and how much uncertainty can be tolerated when lives are on the line. Every approval decision involves a trade-off. Move too slowly, and patients with serious diseases may die waiting. Move too fast, and patients may be harmed by products that looked promising in theory but failed in reality.
For decades, the agency has relied on a framework built around “substantial evidence” of effectiveness. In many cases, that means adequate and well-controlled clinical trials. Randomized controlled trials remain powerful because they help separate cause from coincidence. Without that separation, medicine becomes a very expensive guessing game wearing a white coat.
But modern medicine has complicated the old model. Rare diseases may affect only a few hundred patients. Gene therapies may be designed for small genetic subgroups. Personalized cell therapies may not fit neatly into the classic mass-market drug trial. When patient populations are tiny and diseases are severe, demanding two huge randomized trials may be scientifically tidy but practically impossible.
That is why the FDA already has flexible pathways, including accelerated approval, surrogate endpoints, expanded access, and special programs for serious conditions with unmet medical need. The debate is not whether flexibility exists. It clearly does. The debate is whether the new leadership’s version of flexibility is principled, transparent, and consistent.
The Makary-Prasad Moment
Marty Makary became FDA commissioner in 2025 after a career as a surgeon, researcher, author, and public health commentator. Vinay Prasad, an oncologist and health policy critic, was appointed to lead the FDA’s Center for Biologics Evaluation and Research, the division that oversees vaccines, blood products, cellular therapies, and gene therapies.
Their arrival signaled a philosophical shift. Both men had built public profiles around challenging medical consensus, questioning low-value care, and criticizing weak evidence. To supporters, that made them exactly the kind of disruptive thinkers the FDA needed. To critics, it raised concerns that skepticism might turn into selective skepticism: tough on interventions they disliked, forgiving toward those aligned with the administration’s priorities.
The controversy sharpened after an FDA roundtable on cell and gene therapies. Makary criticized rigid “levels of evidence” language as artificial and dogmatic, while Prasad spoke about making therapies available at early signs of biomedical promise and then following up with evidence on survival and quality of life. That sounded, to many ears, like an accelerated approval mindset expanded into a broader philosophy: approve sooner, monitor later, and let the evidence mature after patients gain access.
There is a reasonable version of this argument. Patients with fatal rare diseases do not always have the luxury of waiting for textbook-perfect studies. Families facing rapidly progressive conditions may reasonably accept more uncertainty than regulators would tolerate for a common, mild illness. In those situations, a rigid evidence hierarchy can feel less like scientific discipline and more like bureaucratic cement.
But there is also a troubling version. If anecdotes and early biological signals become too persuasive, the door opens to expensive products that promise miracles without proving meaningful benefits. In regenerative medicine, that door has already been kicked open by unapproved stem cell clinics selling treatments for conditions ranging from arthritis to autism to neurodegenerative disease. The warning sign is large enough to be seen from space, assuming the space telescope is not also awaiting accelerated approval.
Evidence-Based Medicine Is Not the Enemy
Evidence-based medicine is sometimes caricatured as cold, rigid, and allergic to common sense. That caricature is convenient but unfair. Good evidence-based medicine does not mean ignoring patient experience or worshiping randomized trials like ancient tablets. It means using the best available evidence, understanding its limits, and updating decisions when better data arrive.
Randomized trials are not perfect. They can be expensive, slow, narrow, and sometimes ethically difficult. They may exclude the very patients who later use the treatment. They can measure endpoints that matter more to statisticians than to patients. Anyone who has read a dense clinical trial protocol knows that somewhere between page 47 and the third appendix, the human being can begin to disappear.
Still, randomized trials exist for a reason. Human beings are spectacularly good at seeing patterns that are not there. A patient improves after a treatment, and the treatment gets credit. A disease naturally fluctuates, and the intervention looks magical. A dramatic case goes viral, and suddenly a weak hypothesis gets a fan club. Without controls, randomization, blinding, and careful follow-up, medicine can confuse hope with proof.
That is the core objection to a “more fluid concept of evidence.” Fluidity can mean flexibility, but it can also mean slipperiness. If the standard changes depending on the product, the politics, or the preferred narrative, public trust erodes. The FDA’s credibility depends not only on approving good products but also on explaining why similar cases are treated similarly.
The COVID Vaccine Contrast
The most striking contrast came from Makary and Prasad’s approach to COVID-19 vaccines. In a New England Journal of Medicine article, they argued for a risk-based framework. Annual COVID shots could continue for older adults and people at higher risk, supported by immune-response data. For healthy people under 65, however, they wanted new randomized controlled trials to show clinical benefit before broader approval.
That position is not automatically unreasonable. The risk-benefit calculation for annual boosters in a healthy 25-year-old is different from that for an immunocompromised 75-year-old. Public health recommendations should change as population immunity, viral severity, prior infection, and vaccine performance change. A one-size-fits-all recommendation can become outdated.
But critics saw a double standard. When the topic was COVID vaccination, Makary and Prasad emphasized uncertainty, clinical outcomes, and the need for better trials. When the topic shifted to cell and gene therapies, they sounded more open to early signals, surrogate endpoints, and post-market learning. In other words, the evidence bar seemed to grow taller or shorter depending on which product was standing in front of it.
Supporters would respond that different products deserve different standards. A booster for millions of healthy people should require a different evidentiary threshold than a one-time therapy for a fatal rare disease. That is true. The problem is not difference; the problem is opacity. If the FDA uses different evidence standards, it must explain the principles clearly: disease severity, population size, alternative treatments, biological plausibility, magnitude of effect, reversibility of harm, and feasibility of trials.
Cell and Gene Therapies Need Flexibility, Not Fairy Dust
Cell and gene therapies are not ordinary pills. They can involve living cells, viral vectors, genetic modification, individualized manufacturing, and long-term safety questions. The promise is enormous. CAR-T therapies have transformed parts of cancer treatment. Gene therapies have opened possibilities for conditions once considered untreatable. Scientific progress here is real, not science fiction with a lab badge.
Yet the field also attracts hype. Stem cells are especially vulnerable to marketing abuse because the phrase sounds futuristic, natural, and vaguely miraculous. Bad actors can sell “regenerative” treatments with little evidence, high prices, and emotional advertising aimed at desperate patients. When someone is facing a progressive disease, a polished clinic website can feel like a lifeboat. Sometimes it is just a very expensive raft with a hole in it.
The FDA has repeatedly warned that most regenerative medicine products, including many stem cell and exosome products, are not approved for broad uses. The agency has emphasized that approved stem cell products are limited and that many advertised uses for neurological, orthopedic, cardiac, pulmonary, and autoimmune conditions lack FDA approval. That warning is essential because patient vulnerability is part of the business model for unproven clinics.
A modern FDA should absolutely improve pathways for legitimate cell and gene therapies. It should support adaptive trials, Bayesian designs, external controls when appropriate, natural history studies, patient-centered endpoints, and real-world follow-up. But none of that means replacing evidence with enthusiasm. The answer to outdated rigidity is not regulatory astrology.
Surrogate Endpoints: Useful Tools With Sharp Edges
One reason the debate gets complicated is that the FDA already accepts surrogate endpoints in certain contexts. A surrogate endpoint is a marker believed to predict clinical benefit. Tumor shrinkage, viral load reduction, biomarker changes, or immune response can sometimes provide early evidence that a product is working.
Surrogates can be lifesaving when waiting for long-term outcomes would take years. The accelerated approval pathway was built around this logic for serious diseases with unmet need. The bargain is simple: patients get earlier access, but companies must complete confirmatory studies to verify real clinical benefit.
The bargain breaks down when confirmatory trials are delayed, weak, or inconclusive. A surrogate is not a patient. A lab value does not hug its grandchildren, return to work, avoid hospitalization, or live longer. Surrogates are useful only when they reliably connect to outcomes that matter.
Prasad previously criticized overreliance on surrogate endpoints in oncology, arguing that some drugs won approval without clear evidence of improved survival or quality of life. That history made his openness to early promise in cell and gene therapy especially notable. Has he changed his view? Is he applying the same logic to a harder category of medicine? Or is the FDA selectively rediscovering flexibility when the political mood favors speed?
What a Fair Evidence Framework Should Look Like
A credible FDA evidence framework should not be rigid, but it should be disciplined. Flexibility needs rules. Otherwise, “case-by-case judgment” becomes a fog machine, and everyone pretends the room is mysterious on purpose.
First, the FDA should define when randomized trials are necessary and when alternative designs may be acceptable. Rare diseases, rapidly fatal conditions, and ultra-small populations may justify external controls or single-arm studies, but the agency should clearly explain why.
Second, surrogate endpoints should be graded by strength. Some are well validated. Others are educated guesses wearing a tie. Sponsors, patients, and physicians need to know the difference.
Third, patient experience should inform endpoints, risk tolerance, and trial design, but it should not substitute for proof of efficacy. Patients can tell regulators what outcomes matter. They cannot, by personal testimony alone, establish that a product caused improvement.
Fourth, post-approval commitments must have teeth. If a company receives early approval based on uncertainty, confirmatory evidence should not become optional homework eaten by the corporate dog.
Finally, the FDA should apply its principles consistently across vaccines, oncology drugs, rare disease therapies, stem cell products, and gene therapies. Different categories can have different thresholds, but the reasoning should be visible. Trust depends on transparency.
Public Trust Is the Real Endpoint
The FDA’s ultimate product is not approval letters. It is trust. Patients trust that approved products have been seriously evaluated. Doctors trust that labels reflect real evidence. Companies trust that standards are predictable. Investors trust that regulatory risk is not a roulette wheel wearing a stethoscope.
When leaders publicly challenge old evidence hierarchies, they may energize innovation. They may also unsettle the people who rely on the FDA to be the adult in the room. That does not mean the agency must be slow or timid. It means the agency must be able to explain itself in a way that survives political weather.
Makary and Prasad’s approach exposed a real issue: the FDA needs better tools for evaluating modern therapies. Traditional randomized trials are not always feasible, especially in rare disease and individualized medicine. But the solution cannot be a vague romance with “common sense” and early promise. Common sense is useful for crossing the street. It is less reliable for approving a therapy that may alter immune function, gene expression, or long-term cancer risk.
The right question is not whether evidence should be rigid or fluid. The right question is whether the FDA can be flexible without becoming arbitrary. Water is fluid, yes, but so is spilled coffee. One supports life; the other ruins the keyboard.
Experience Notes: What This Debate Looks Like Outside Washington
For patients, the FDA evidence debate is not abstract. It shows up in exam rooms, family group chats, fundraising pages, and late-night searches typed by people who are scared enough to click anything. A parent whose child has a rare genetic disorder may not care whether a study design satisfies every academic preference. They want a chance. They want a regulator that understands urgency. They also want protection from false hope, even if they do not phrase it that way.
In real life, desperation changes how evidence feels. A modest early signal can look enormous when the alternative is decline. A single dramatic recovery story can travel faster than a peer-reviewed paper. Families may ask, “What if this is the one?” That question is emotionally powerful, and any humane system has to respect it. But regulators must also ask the colder question: “How do we know?” Without that second question, the market fills with confident sellers and vulnerable buyers.
Physicians experience the debate differently. Many doctors have sat with patients asking about treatments advertised online as regenerative, personalized, cutting-edge, or “not yet accepted by mainstream medicine.” The marketing often sounds scientific enough to be persuasive. It may mention stem cells, exosomes, immune modulation, mitochondrial repair, or inflammation. The language is polished. The evidence is sometimes thinner than a hospital gown in January.
When FDA standards are clear, physicians can guide patients with confidence. They can say, “This is approved for your condition,” or “This is experimental,” or “This marker improved, but we do not yet know whether people live longer or feel better.” When standards become murky, counseling becomes harder. Patients may suspect doctors are being paternalistic or outdated. Doctors may feel trapped between empathy and evidence.
Biotech innovators also live inside this tension. Small companies developing therapies for rare diseases often face impossible math. They may not have enough patients for large trials. They may have promising biology, strong animal data, natural history comparisons, and early human signals. For them, FDA flexibility can be the difference between progress and bankruptcy. A smarter evidence framework could help serious innovators move faster without pretending uncertainty has vanished.
But serious innovators are not helped by a market crowded with weak claims. When unproven clinics sell expensive treatments under the same broad banner of “regenerative medicine,” they damage the reputation of the legitimate field. Good science needs sunlight. Bad science loves fog. That is why the FDA must modernize without lowering its guard.
The best practical experience from this debate is simple: patients need hope, doctors need clarity, companies need predictability, and regulators need humility. No single study design answers every question. No anecdote should carry the weight of approval by itself. The FDA should be flexible enough to recognize extraordinary circumstances and firm enough to prevent wishful thinking from becoming national policy.
Makary and Prasad’s “more fluid” evidence moment should not be dismissed as pure recklessness, nor celebrated as pure reform. It should be treated as a stress test. If the FDA can define flexible standards clearly, enforce confirmatory evidence, protect patients from unproven products, and apply rigor consistently, then a more modern evidence system could be valuable. If not, the agency risks turning evidence into a mood ringand medicine has enough accessories already.
Conclusion
The debate over Marty Makary, Vinay Prasad, and a “more fluid concept of evidence” at the FDA is really a debate about trust. Modern medicine needs regulatory flexibility, especially for rare diseases, cell therapies, gene therapies, and areas where traditional trials are difficult. But flexibility must be transparent, consistent, and anchored to patient-centered outcomes.
Makary and Prasad were right to question whether old evidence categories always fit new science. Critics were right to worry that selective flexibility can weaken standards, invite hype, and expose patients to poorly proven interventions. The FDA’s challenge is to avoid both extremes: no fossilized bureaucracy, no free-range miracle market.
Evidence can evolve. It can adapt. It can incorporate real-world data, biomarkers, patient priorities, and innovative trial designs. But it cannot become whatever leaders need it to be on Tuesday afternoon. At the FDA, evidence should be flexible like a suspension bridge: engineered to move under pressure, but strong enough not to collapse when everyone starts crossing.
