ANALYSISApril 13, 2026
Is the FDA too cautious in approving new treatments?
The FDA under Commissioner Marty Makary, appointed in 2025, is implementing sweeping changes to accelerate drug approvals, including a new National Priority Voucher program, a shift to single-trial standards, expanded use of real-world data, AI-assisted reviews, and cash bonuses for staff who complete reviews ahead of schedule. These reforms have sparked intense debate among public health experts, researchers, former FDA officials, and agency staff about the appropriate balance between speed and safety. The HHS Office of Inspector General is actively auditing the FDA's accelerated approval pathway amid concerns about political interference and patient safety.
When a dying patient is willing to accept any risk, does the FDA's insistence on lengthy trials protect them — or deny them the one shot they have left? And if we speed up approvals, who pays the price when the next thalidomide gets through?
Not familiar with this story? Get context →
Invisible deaths from approval delays
C
The FDA's bias almost never gets named. When a dangerous drug harms someone, the headlines follow. When a life-saving treatment sits in regulatory limbo for two extra years, nobody counts the dead — Louis Lasagna estimated thousands of Americans were dying annually from approval delays as far back as 1990. That asymmetry is not a neutral feature of the system; it is a structural incentive toward excessive caution that has never been adequately corrected.
L
The asymmetry you're describing cuts both ways. Yes, delay has a body count — ACT-UP knew that, and PDUFA in 1992 was a real response to a real problem. But your framing treats speed as the only dimension of patient welfare, which is exactly what Makary is doing when he accelerates approvals while the HHS Inspector General is finding drugs approved over reviewer objections already coming off the market.
C
The Inspector General finding proves too much for your argument: those drugs were approved over reviewer objections under the pre-reform regime, which means the vulnerability you're pinning on Makary was already present in the system you're defending.
L
That's the honest concession — the old system wasn't clean. But the reform environment makes that failure mode more likely and less correctable, because the personnel and structural changes underway reduce the institutional capacity to catch bad approvals after the fact.
Single-trial standard and confirmatory evidence
C
The 2025 finding that fewer than 20% of single-trial approvals from 2015–2023 had any confirmatory evidence is genuinely troubling, and I won't wave it away. But note what it actually proves: the confirmatory trial requirement, on the books since 2012, was not being enforced. The problem was not the evidentiary standard — it was the follow-through. Demanding more rigorous single trials while aggressively enforcing post-market commitments is a defensible position.
L
You're arguing the solution to unenforced confirmatory trial requirements is to make confirmatory trials optional — that's not fixing the enforcement gap, that's ratifying it. The 20% figure was the baseline before single-trial became official policy; the question is what that number looks like when the policy explicitly removes the expectation of a second trial.
C
Making confirmatory trials optional with aggressive post-market enforcement is meaningfully different from making them optional and walking away — your rebuttal only lands if you can show Makary's FDA is doing the latter.
L
The cash bonuses for early review completion are exactly that evidence: you cannot simultaneously reward speed and credibly threaten reviewers with consequences for flagging problems that delay approvals.
41% post-market safety warning rate
C
The 41% figure — drugs approved on limited data from 1998 to 2013 that later required safety warnings or withdrawals — is a real number, but it describes the pre-reform system. It does not prove that the current system without these reforms would have performed better; it proves that the old standard was already not delivering the protection it was credited with. The question is whether the proposed reforms make the tradeoffs better or worse, not whether the baseline was clean.
L
You're using the old system's failures to justify lowering the bar further, but the 41% figure is an argument for stricter enforcement, not weaker standards. Four in ten is not a tail risk — it's a structurally predictable outcome of approving on limited data, and Makary's reforms move in exactly the direction that produced it.
C
Stricter enforcement of what? The 2022 data showing 70% of rare disease approvals needed safety-related labeling changes came under the standard you want to preserve — 'enforce the existing system' is harder to defend when the existing system generated that number.
L
The rare disease context is the sharpest edge of this: those patients are the most desperate, the most willing to accept uncertainty, and the least positioned to demand accountability when the approval-stage evidence gaps catch up with them.
CNPV voucher program and review incentives
C
The CNPV voucher program deserves to be evaluated separately from the core reforms, and a market-oriented conservative should be deeply skeptical of it. Letting officials designate which drugs get 60-day priority reviews without transparent scientific criteria is not deregulation — it is discretionary favoritism in a different costume, and it creates perverse incentives for companies to lobby for priority status rather than produce superior evidence.
L
I'll take the concession, but it doesn't isolate the program from the broader pattern. Eighteen vouchers awarded under opaque criteria, cash bonuses for staff who complete reviews early, and former commissioners warning in the NEJM that vaccines face a higher subjective bar than other drugs — these aren't separate problems, they're a common mechanism: replacing transparent scientific criteria with discretionary pressure.
C
The vaccine standard concern is serious precisely because if true it represents politically-motivated regulatory inconsistency — but twelve former commissioners agreeing in the NEJM is a warning, not a finding, and current leadership not rebutting it is damning only if we assume silence equals guilt.
L
When twelve former commissioners publish a warning in the New England Journal of Medicine and current leadership offers no substantive answer, the institutional burden of proof has shifted — silence in that context is not a neutral data point.
Post-market enforcement as structural load-bearing wall
C
The 2025 approval count of 46 novel drugs — well above the historic average of 36 — cannot simply be dismissed, but the liberal concession that lagged harms appear in post-market data years later, not in approval counts today, contains the real answer. That means aggressive post-market surveillance and enforcement is not optional window dressing under any accelerated approval framework — it is the structural load-bearing wall. If Makary's FDA is not strengthening post-market enforcement with the same energy it applies to acceleration, the approval count is a promissory note the system may not honor.
L
You're describing what a responsible single-trial framework would require — and then asking us to trust that this FDA will build it. But the same reform environment that produced 46 approvals also produced the bonus structure that rewards early completion, the personnel changes reducing institutional capacity, and the silence on the vaccine standard question. You're asking for faith in enforcement from an administration that has systematically weakened the enforcers.
C
That's a legitimate concern about implementation, but 'this administration won't enforce it well' is an argument against this FDA, not against the evidentiary framework itself — and conflating the two is how we end up defending a pre-reform system that was already failing its own enforcement requirements.
L
The framework and the enforcers aren't separable in practice — a post-market enforcement regime is only as strong as the institutional capacity and political will behind it, and both are being actively dismantled by the same people proposing to rely on it.
Conservative's hardest question
The 2025 HHS Inspector General finding that three drugs were approved over the objections of FDA's own reviewers and advisory committees — with two subsequently withdrawn — is difficult to dismiss as bureaucratic inertia, because it suggests the current reform environment may already be producing approvals that internal science does not support. If political pressure is overriding scientific review rather than improving it, the conservative case for these reforms collapses into something more troubling than excessive caution.
Liberal's hardest question
The strongest challenge to this argument is that the FDA approved 46 novel drugs in 2025, well above the historic average of 36 per year, which reform advocates will cite as proof that higher throughput and safety are not mutually exclusive. This is genuinely difficult to dismiss in the short term, because the harms from lowered standards are lagged — they appear in post-market safety data years later, not in the approval count today.
Both sides agree: Both sides accept that FDA approval delays carry real mortality costs — the liberal argument explicitly validates the ACT-UP protests and the Lasagna estimate, conceding that regulatory caution is not costless.
The real conflict: The core factual-interpretive conflict is about what the 41% post-market safety warning rate proves: conservatives read it as evidence that the old system was not reliably safe anyway, making reform defensible; liberals read it as a structurally predictable consequence of reduced evidentiary thresholds that the new regime will worsen.
What nobody has answered: If the confirmatory trial requirement codified in 2012 was already failing — fewer than 20% of single-trial approvals ever produced confirmatory evidence — what specific enforcement mechanism exists today that would prevent the same non-compliance from hollowing out any new post-market safety commitments, and has either side identified one?
Sources
- FDA CDER 2025 novel drug approvals data
- FDA Commissioner Makary public statements and policy announcements, 2025
- HHS Office of Inspector General report on FDA accelerated approval pathway, 2025
- New England Journal of Medicine letter from twelve former FDA commissioners, 2025
- Ross et al. 2025 study on single-trial drug approvals and confirmatory evidence
- 2022 study on rare disease drug approvals and safety labeling changes
- Retrospective analysis of 27 fast-tracked drugs approved 1998–2013
- Harvard/Brigham and Women's PORTAL program research on FDA advisory panels
- Historical records of ACT-UP FDA protest, October 11, 1988
- Louis Lasagna presidential advisory panel statements, August 1990
- Milton Friedman economic arguments on FDA regulatory bias
- Industry analyst reports on FDA instability and review timelines, 2025