Most people assume that FDA drug approval standards require convincing proof that a medicine actually helps patients. Doctors often make the same assumption. After all, the agency has long reassured the public that the drugs it approves are “safe and effective.”
Former FDA Commissioner Scott Gottlieb, MD, once put that message into a playful rhyme:
“Some pills are red,
Some pills are blue,
All are safe and effective,
If they undergo FDA review.”
That is comforting. But an investigation published in JAMA Psychiatry (June 24, 2026), raises an uncomfortable question: What does “effective” really mean to the FDA?
FDA Drug Approval Standards and the Exxua Story
The drug at the center of this controversy is gepirone extended release (ER), sold under the brand name Exxua. It was approved in 2023 to treat major depressive disorder (MDD).
Researchers led by former FDA medical reviewer Erick Turner, MD, examined the drug’s long and complicated regulatory history. Aaron Kesselheim, MD, JD, MPH, a respected Harvard expert on drug regulation, was another author. Quick confession: I am a huge admirer of Dr. Aaron S. Kesselheim. His ethics and research are impressive.
The Regulatory History of Gepirone ER
The FDA evaluated 13 trials as evidence of effectiveness: 12 relatively short studies of acute depression treatment and one longer relapse-prevention trial.
Only two of the 12 acute-treatment trials produced statistically significant results favoring gepirone. In the other studies, the drug did not demonstrate superiority to placebo. I love to hate how researchers use language. In their description of the research, the authors of the JAMA Psychiatry study describe this result:
“Ten of these 12 trials had nonsignificant results. The sponsor meta-analyzed the data from these trials, yielding a statistically nonsignificant overall gepirone-placebo difference of approximately one-half of 1 point on the HAMD-17.”
The Hamilton Depression Rating Scale
What, you may ask, is the HAMD-17? That stands for the Hamilton Depression Rating Scale. It is a tool that psychiatrists have used for decades to assess how depressed a patient is. I won’t get into the weeds on this because it can be way too complex. There are 17 aspects of depression that are evaluated. The HAMD-17 runs from 0 to 52 points. A drug generally has to improve depression between 3 to 6 points over placebo to be considered clinically beneficial. Moving the needle 0.5 points is not impressive.
In three trials, FDA reviewers found that gepirone performed statistically worse than an established antidepressant used for comparison.
The FDA rejected applications for the drug four times. In 2015, an FDA advisory committee voted 9 to 4 that its effectiveness had not been demonstrated.
Here is how the authors describe the process in JAMA Psychiatry:
“Overall, the FDA rejected the New Drug Application [NDA] from the sponsor 4 times (Organon in 1999, 2002, and 2004, and Fabre-Kramer Pharmaceuticals, Inc in 2007). The FDA rejected the first and second NDA submissions from the sponsor related to gepirone ER in 1999 and 2002 for inadequate evidence of efficacy. The sponsor resubmitted the NDA in December 2003, including additional data from a (long-term) maintenance or relapse prevention trial. However, the FDA’s analysis of this trial yielded statistically nonsignificant results, and it was unimpressed by the results of the (short-term) acute treatment trials; the FDA issued its third rejection in June 2004.”
Senior agency officials ultimately concluded, however, that the two favorable trials were unlikely to have occurred by chance and that the drug met the legal standard for approval.
That does not mean anyone acted illegally or that Exxua cannot help certain patients. It does mean that an FDA finding of “effectiveness” may not carry the meaning many doctors and patients attach to that word.
“If At First You Don’t Succeed, Try, Try Again!”
You no doubt have heard that expression before.
Some people attribute a parody version of the proverb to W.C. Fields:
“If at first you don’t succeed, try, try again. Then quit.
There’s no point in being a damn fool about it.”
Others doubt that Fields ever said that…but it might apply to some drug trials.
A Drug Trial Is Not Supposed to Be a Best-of-13 Tournament
Imagine a soccer team playing 13 games. It wins two, fails to win ten others and loses three matchups in which the opposing team clearly outperforms it. At the end of the tournament, officials display only the two victories and declare that the team has proved it can win.
No one is suggesting that clinical research is like a soccer tournament. Trials can fail for legitimate reasons. A study may be too small, poorly designed or unable to distinguish an effective treatment from an ineffective one.
But when many reasonably designed studies point in one direction and only a few point in another, shouldn’t the entire record matter?
The researchers call the FDA’s approach “trial counting.” Individual studies are judged largely on whether they cross the conventional line for statistical significance. But this can give two “positive” studies extraordinary influence even when the full body of evidence is far less impressive. The authors found that when all 12 acute-treatment trials were combined, gepirone’s advantage over placebo was less than half a point on the 17-item Hamilton Depression Rating Scale and was not statistically significant.
Statistical Significance Is Not the Same as Meaningful Improvement
I know this is complicated, but statistical significance is not the same thing as clinical significance.
A finding can be statistically significant, meaning that it is unlikely to be due to chance under the study’s assumptions, without being large enough to make a noticeable difference in a patient’s life.
Suppose a blood pressure drug lowers a dangerously high reading from 180/120 to 175/115. With enough participants, that change might be statistically convincing. But the patient’s blood pressure would still be dangerously high and might lead to a stroke.
Here is how the authors of the JAMA Psychiatry article sum up their analysis:
“Gepirone ER was approved despite serious doubts about the drug’s efficacy due to 2 clinical trials that supported efficacy by reaching statistical significance vs 11 trials that failed to do so. After 3 decades of FDA-sponsor interactions, including 4 rejections and a negative vote by the Advisory Committee (which leads to nonapproval in ~60% of cases), the FDA determined that the statutory standard for substantial evidence of effectiveness was met and granted approval; however, full information about this extended process was not included in the final labeling.”
What Did the Exxua Label Tell Doctors?
The prescribing information for Exxua described the “two eight-week randomized, double-blind, placebo-controlled, flexible-dose studies in adults” that produced favorable results. According to the JAMA Psychiatry analysis, it did not describe the other unsuccessful efficacy trials or tell prescribers that the combined advantage over placebo was very small.
That matters because busy clinicians rarely have time to reconstruct a drug’s entire regulatory history. They rely on the FDA-approved label, drug-information databases and medical references that often draw heavily from that label.
A doctor reading about two successful studies could reasonably assume that those studies fairly represent the complete evidence. The JAMA Psychiatry authors argue that labels should report all adequate and well-controlled trials, not just the favorable ones.
We agree. A restaurant could not give you an honest picture of customer satisfaction by displaying only its five-star reviews. Drug labeling should not operate that way either.
FDA Drug Approval Standards May Become Even More Flexible
This story is especially timely. In June 2026, the FDA released revised draft guidance explaining how a company might establish effectiveness with one adequate and well-controlled clinical investigation plus confirmatory evidence. The document is still a draft, is open for public comment and is not yet binding agency policy.
That is an important qualification. The FDA is not simply proposing that any drug can be approved after one favorable experiment. The single trial would need to be persuasive, and there would need to be additional evidence supporting it.
Nevertheless, “confirmatory evidence” may not always be a second independent randomized controlled trial. The proposed approach therefore places even more weight on how the FDA evaluates the totality of evidence—including conflicting or unsuccessful studies.
A single excellent trial may sometimes be appropriate, particularly for a rare or devastating disease in which conducting two large studies would be impractical or unethical. But greater flexibility demands greater transparency. Otherwise, patients and prescribers may have no easy way to know whether a favorable result has been repeatedly confirmed or whether it is the lone bright spot in a dismal record.
Why Should You Care About FDA Approval Standards?
This is not merely an argument about one antidepressant.
An FDA approval does not necessarily mean that:
- The drug helps most people who take it.
- Its benefit is large enough for patients to notice.
- It works as well as older or less expensive treatments.
- Every well-designed clinical trial produced favorable results.
- The FDA-approved label presents the complete trial history.
Surveys cited by the researchers suggest that many physicians misunderstand these distinctions. Seventy percent reportedly believed FDA approval requires both statistical and clinical significance. Nearly three-fourths thought a new medicine had to be at least as effective as drugs already available for the same condition. Neither belief accurately describes the general approval standard.
They go on to state:
“This lack of insight into the data underlying FDA-approved drugs may help explain why clinicians tend to overestimate interventions’ benefits and underestimate their harms.”
Here’s the kicker:
The authors of the article also note:
“Among surveyed US adults, almost half mistakenly believe that the FDA approves only extremely effective drugs.”
The People’s Pharmacy Bottom Line
FDA approval remains critical to what does and does not make it to your local pharmacy. It should mean that objective FDA scientists have reviewed extensive information and concluded that the medicine satisfies both safety and effectiveness standards. It does not mean the drug is a miracle, that it is superior to existing treatments or that its benefits will be meaningful for most patients.
The gepirone story reveals a gap between what “FDA approved” means under the law and what doctors and patients often think it means in the real world.
The answer is not to prevent companies from conducting additional research after an unsuccessful trial. Better-designed studies can correct genuine problems and identify patients who may benefit. But all the scientifically credible results should remain on the scoreboard.
When the FDA allows two victories to establish effectiveness, doctors deserve to know about the eleven games the drug did not win. Anything less turns informed consent into a highlights reel. Statistical significance does not always lead to meaningful clinical significance. In other words, a drug may be slightly better than placebo but only 1 out of 50 patients will benefit. If your shiny new toaster only worked 1 out of 50 times, you might return it and demand your money back.
What do you think? We would love to read your feedback on this complex subject. If you think it might be helpful for someone you know, please pass it along. Thank you for supporting our work.
