The FDA Will Now Approve Drugs After One Trial. The Last Time It Relaxed Standards, 57% of Cancer Drugs Failed Confirmation.

Ten percent.

That is the share of new drugs the FDA approved between 2018 and 2021 — one in ten — that did not meet their own primary pivotal trial endpoint, according to a JAMA Internal Medicine analysis of 210 novel drug approvals. Twenty-one drugs cleared the FDA's bar despite failing the statistical test their trials were specifically designed to pass. Two-thirds of those received expedited or priority review. The most common justification? Success in other studies, cited 61.9% of the time, followed by post hoc effectiveness analysis at 33.3%, a method that lets researchers hunt for positive signals in data subsets after the primary test has already failed.

Now the agency wants to make that much easier.

In a New England Journal of Medicine editorial, FDA Commissioner Marty Makary and physician Vinay Prasad announced the end of the two-pivotal-trial standard that has governed drug approval for decades. Under the new policy, one adequate, well-controlled trial is sufficient when supplemented by "confirmatory evidence" that can include real-world data drawn from electronic health records, mechanistic studies demonstrating a plausible biological pathway, or biomarker endpoints that serve as proxies for clinical benefit without requiring a second randomized controlled trial. No statute ever mandated two trials. It was a regulatory norm, born from decades of post-thalidomide caution and codified through institutional habit rather than congressional mandate. Makary argues the convention has outlived its usefulness, inflating development costs by hundreds of millions of dollars per drug program and delaying therapies by years for patients with serious unmet medical needs who cannot afford to wait for a second confirmatory trial.

He is partially right. But the numbers tell a more complicated story, and the closest historical precedent for looser evidence standards carries a warning that deserves more attention than it is receiving.

What the Policy Actually Changes

The FDA is not eliminating the requirement for rigorous clinical evidence; it is redefining what "sufficient" means. Under the previous norm, most new drug applications included two Phase III trials, each independently demonstrating efficacy and safety. Under the Makary standard, one well-designed trial can suffice if accompanied by supporting evidence from a separate domain: a biomarker study that confirms the drug's mechanism, a large observational dataset from insurance claims or electronic health records, or a pre-specified subgroup analysis demonstrating consistent treatment effects across patient demographics.

This is less radical than headlines suggest. A JAMA Network Open analysis covering 1995 to 2017 found that the agency had already been drifting toward single-trial approvals for years, with increasing use of Priority Review, Accelerated Approval, and Breakthrough Therapy designations. What Makary has done is formalize the drift: single-trial programs are not merely tolerated exceptions but the new expected norm.

Practical consequences depend entirely on what counts as "confirmatory evidence," a term whose definition will determine whether this policy change is a modest efficiency gain or a fundamental restructuring of the evidence requirements that have protected American patients for half a century. If it means a rigorous Phase II study with a preregistered protocol, the evidence bar drops modestly. If it means a post-marketing observational study or a mechanistic plausibility argument, the bar drops considerably. Makary's editorial leaves this deliberately ambiguous.

Cost Savings Are Real but Smaller Than You Think

The pharmaceutical industry's pitch for single-trial approvals centers on cost. Phase III clinical trials are expensive, often requiring thousands of patients across dozens of sites over three to five years. The Tufts Center puts the median cost at $19 million, but that figure masks enormous variance: large cardiovascular outcomes trials routinely exceed $100 million, and oncology trials can push past $50 million.

Eliminating one Phase III trial per approved drug would generate real savings, but the arithmetic is slippery: roughly 40% of recent FDA approvals already relied on a single pivotal trial, which means the new policy affects only the remaining 60% of programs that were still running two full Phase III trials under the old expectation, and aggregate savings are considerably smaller than the headline figures suggest.

Running the math: approximately 50 novel drugs receive FDA approval each year. If 60% eliminate one Phase III trial at a median cost of $19 million, the aggregate annual savings are roughly $570 million; at the higher end, using a $30 million average for the more complex therapeutic areas where trials are genuinely expensive, the savings reach $900 million. The likely range: $380 million to $900 million per year across the entire pharmaceutical industry, which sounds impressive until you realize it is less than 1% of the $600 billion U.S. prescription drug market.

Time savings matter more. Dropping a second Phase III trial can shave two to four years off a development program. For patients with aggressive cancers, degenerative neurological conditions, or rare diseases with no approved alternatives, that acceleration is not an abstraction but the difference between accessing a potentially life-saving treatment and dying on a clinical development waiting list.

Oncology's Warning Signal

Critics' strongest argument against loosening the evidence bar comes not from theory but from the FDA's own track record with accelerated approvals, the closest historical analog to single-trial approval, where drugs are cleared based on surrogate endpoints from a single study with the requirement that confirmatory trials be completed afterward.

The results are sobering: an OncLive analysis of accelerated cancer drug approvals found that over a five-year span, only 43% demonstrated clinical benefit in confirmatory trials. Put differently: 57% failed. Some were eventually withdrawn, while others remained on the market for years, prescribed to patients on the basis of preliminary data that subsequent studies could not replicate.

This is not a safety catastrophe. These drugs did not cause the kind of mass poisoning that created emergency headlines and congressional hearings in the thalidomide and Vioxx crises, the kind where patients died in numbers large enough to force front-page retractions of regulatory decisions that should never have been made. What they caused was something more corrosive: the erosion of the evidence base that patients and physicians use to make treatment decisions. A drug approved on thin evidence that never confirms its benefit creates an information crisis, quiet, systemic, and extraordinarily difficult to reverse once approval creates a presumption of efficacy. Oncologists prescribe it because the FDA approved it. Insurers cover it because oncologists prescribe it. Patients take it because their doctor recommended it. Nobody in the chain has strong evidence that the drug works better than the available alternatives, and no individual actor in the system has the incentive or authority to challenge the original approval decision.

A Cost-Risk Tradeoff Nobody Has Calculated

The question that neither the NEJM editorial nor its critics have quantified is this: at what point do the costs of safety withdrawals exceed the savings from eliminated trials?

Drug safety withdrawals currently run at approximately 1.6%, down from 3.2% in the 1980s, according to the Tufts Center. That decline reflects genuine improvements in clinical pharmacology, better trial design, and stronger post-market surveillance, but critically, these improvements occurred under the two-trial regime.

If the single-trial standard increases the withdrawal rate by even half a percentage point, from 1.6% to 2.1%, the downstream costs are substantial. Conservative estimates put the cost of a major drug withdrawal at $1 billion to $5 billion per event, encompassing product liability litigation, recall logistics, lost development investment, pipeline credibility damage, and the downstream costs of switching patients to alternative therapies. At a 2.1% withdrawal rate and 50 annual approvals, the industry would face roughly one additional withdrawal per year, costing $1 billion to $5 billion.

Compare that to the $380 million to $900 million in annual trial cost savings. The fragility is obvious: expected savings from eliminating redundant trials are roughly the same order of magnitude as the cost of a single high-profile withdrawal, a margin that leaves almost no cushion for error in a system where the consequences of getting one drug wrong cascade through product liability, regulatory backlash, and lost patient trust in a way that can tighten approval standards well beyond the pre-Makary baseline. One bad withdrawal erases years of savings. Two in quick succession could harm the very patients the policy was designed to help.

Strongest Counterargument

The best case for single-trial approval is the Tufts Center's finding that faster FDA review times have not historically correlated with higher safety withdrawal rates. Drugs approved through priority review paths between 1980 and 2009 were withdrawn at statistically similar rates to standard-review drugs. Speed, by itself, did not kill patients.

This finding deserves serious weight, because if three decades of data show no speed-safety tradeoff, the single-trial standard may function similarly: faster approvals without meaningfully higher risk. Dismissing it on precautionary grounds alone would be intellectually dishonest.

One caveat: the Tufts data was collected under a regulatory regime where the two-trial standard existed as a floor beneath all approvals, meaning the historical withdrawal rates reflect a system in which drugs had already cleared a higher evidence bar before reaching patients. Removing that floor is a fundamentally different experiment than merely speeding up reviews within it. Accelerated approval in oncology suggests that when you reduce the evidence requirement below a certain threshold, you do not get more safety events but rather something harder to detect: drugs that remain on the market for years without confirmation that they deliver the clinical benefit that justified their approval in the first place.

What This Analysis Does Not Prove

The $19 million median Phase III cost figure from Tufts is disputed. Other researchers estimate significantly higher costs when accounting for site management fees, data monitoring, and regulatory filing expenses, placing a more realistic median between $30 million and $50 million, which would shift the savings calculation upward. This cost-risk calculation is directionally correct but depends on estimates with wide confidence intervals.

That 57% oncology failure rate cannot be extrapolated to all therapeutic areas. Oncology's accelerated approvals use surrogate endpoints like tumor shrinkage, which are inherently weaker predictors of clinical benefit than the hard endpoints used in cardiovascular or metabolic disease trials, so failure rates for single-trial approvals in diabetes or hypertension might be substantially lower.

And the biggest variable remains undefined: what counts as "confirmatory evidence"? If the FDA implements it rigorously, the effective evidence bar may not drop much at all. Real-world impact depends on guidance documents that have not been written yet.

What You Can Do

If you are a patient or caregiver: ask whether your medication was approved based on one or two pivotal trials. FDA's Drug@FDA database includes approval packages with this information. A single-trial approval is not a red flag, but it is relevant context for discussing alternatives with your physician.

If you invest in pharmaceutical companies: watch for pipeline acceleration announcements. Companies with drugs in late-stage development can now design single-trial programs, potentially saving $20 million or more per candidate and bringing approval forward by two to three years. That creates near-term value, but it also means the post-marketing evidence burden shifts from pre-approval to post-approval, increasing the tail risk of costly withdrawals.

If you work in healthcare policy: the most urgent need is investment in post-market surveillance infrastructure. Makary is moving the evidence burden downstream, from pre-market trials to real-world evidence collection after approval. Without commensurate investment in pharmacovigilance systems, electronic health record integration, and active safety monitoring programs, the single-trial standard becomes an evidence gap rather than an evidence shift.

Bottom Line

Makary's single-trial standard is not the deregulatory catastrophe that critics fear, nor the innovation breakthrough that proponents promise. It is a formalization of a trend already underway. Savings are real but modest: $380 million to $900 million annually, dwarfed by the pharmaceutical industry's overall expenditure. Risk is real but conditional: if post-market surveillance holds, the policy works. If it doesn't, the accelerated approval track record in oncology tells you exactly what happens. Not more dead patients. More drugs on the market that nobody can prove actually work.