I watched something a bit back and was thinking about one of the claims this morning as I listened to SGU on the way to work. I unfortunately can't recall the source anymore, but am fairly confident that it was either Food, Inc., Sweet Remedy, or perhaps this vaccine class with a nearby chiropractor.

In any case, the claim was that the Food and Drug Administration (FDA) does not actually conduct any of its own tests in order to approve foods or drugs as safe for public consumption/treatment.

The implication is that the companies seeking approval can "doctor" the data and test methods however they want in order to pass the FDA approval process, and I'm blanking on what the specific was at this point -- it might have been MSG, aspartame or a particular vaccine... I can't recall.

My questions are:

  • Is it the case that the FDA does no actual testing itself, only reviews of other studies?
  • If so, is it left up to the company to conduct it's own studies, and are these studies verified by independent and/or trusted testing agencies?
  • If the above are true, have there been cases of substance/drugs being recalled because falsified data, improper test methods, and/or suppression of potential negative effects allowed a drug to pass the FDA approval process when they were actually unsafe?
  • 5
    As someone with a second-hand close view of this (family member deals with FDA as one of the primary job responsibilities); I can tell you that "companies seeking approval can 'doctor' the data and test methods however they want" is utter B$. There's so many levels of review of this (and so much crap to drop on someone who violates any detail of numerous protocols), that doctoring data, while theoretically possible, is a highly unlikely event.
    – user5341
    Jun 6, 2011 at 16:26
  • I had a friend who once worked for the FDA in animal nutrition. She wanted three experiments that showed effectiveness with p < .05 before she recommended anything for approval, and examined the experiments carefully. She also told me that most food and supplements with FDA approval are actually on a "we do not object to the sale of this substance at this time" letters, which can be revoked at any time. Jun 7, 2011 at 2:07
  • Are we evaluating their crieteria or their process here? Jun 7, 2011 at 5:29
  • @DVK What are those "many levels of review" and who employs the reviewers? I understand that no single person can easily forge data, but what if the whole company is in on it: does the article referenced in Christian's answer change your opinion?
    – ChrisW
    Oct 9, 2013 at 14:09
  • as to where you say this: you may be thinking of a recent anti-vaccine facebook meme about the flu vaccine not being tested before being given out. Does the FDA have a separate process for vaccines vs. regular drugs?
    – KutuluMike
    Oct 10, 2013 at 2:26

4 Answers 4


The first two parts are easy to address. From the FDA's own website:

The mission of FDA's Center for Drug Evaluation and Research (CDER) is to ensure that drugs marketed in this country are safe and effective. CDER does not test drugs, although the Center's Office of Testing and Research does conduct limited research in the areas of drug quality, safety, and effectiveness.


Drug companies seeking to sell a drug in the United States must first test it. The company then sends CDER the evidence from these tests to prove the drug is safe and effective for its intended use. A team of CDER physicians, statisticians, chemists, pharmacologists, and other scientists reviews the company's data and proposed labeling. If this independent and unbiased review establishes that a drug's health benefits outweigh its known risks, the drug is approved for sale. The center doesn't actually test drugs itself, although it does conduct limited research in the areas of drug quality, safety, and effectiveness standards.

The testing is done by the company that wants to sell the drug, and the results verified by the FDA.

  • Nice! I'm nearly satisfied -- any take on the third bit? For example, drugs approved and then recalled or later found to not have checked out? Are there a significant number or is it negligible? Any example FDA responses when the CDER approves something later found to be harmful (if that's happened)? Thanks for the response.
    – Hendy
    Jun 6, 2011 at 18:33
  • 1
    Every time there is a problem with a drug there are allegations that the testing was "inadequate". There are certainly allegations of drug companies suppressing negative results. However I don't think there are many as part of the formal FDA testing procedure. But without references I'm reluctant to make that statement a part of the answer. Jun 6, 2011 at 18:37
  • 6
    It's standard for the onus of testing to be placed upon the company wishing to market the drug; otherwise it would be impossible for the FDA (or CDC, EPA, etc) to perform the testing needed to ensure a safe product. The same procedure occurs for any new chemical compound which hits the market.
    – Darwy
    Jun 6, 2011 at 19:06
  • @DJ: I would agree with that -- the FDA is simply reviewing and thus only gets one side of the picture. @Darwy: agreed that the onus is on the company. But I also see the point that a regulatory agency is supposed to be regulating. Thus, there's at least something to be said if perhaps due diligence isn't done in the case of some substance by those examining the test methods and reported results. The FDA's "misses" are probably much less than attackers want the public to think, though.
    – Hendy
    Jun 6, 2011 at 19:26
  • 6
    @Hendy the FDA sets forth a rigorous set of guidelines which must be adhered to by the drug company. The amount of testing which must occur prior to clinical trials is immense (as is the cost) in vivo, in vitro - expensive! Plus, the FDA has to approve the clinical trial (how it's designed, etc) before it's allowed to proceed. fda.gov/drugs/developmentapprovalprocess/…
    – Darwy
    Jun 6, 2011 at 20:36

Ranbaxy is a good example of a company that simply faked a lot of their data. Katherine Eban wrote a very good article in Fortune that summarizes the episode:

Ranbaxy pleaded guilty to seven federal criminal counts of selling adulterated drugs with intent to defraud, failing to report that its drugs didn't meet specifications, and making intentionally false statements to the government.


Despite the agency's finding of fraud and misconduct, it granted Ranbaxy lucrative rights to sell new generic drugs. In the most high-profile example, in November 2011 the FDA allowed the company to maintain its exclusive first dibs on making the generic version of a medicine taken by tens of millions of Americans: Lipitor. In the first six months, this privilege allowed Ranbaxy to generate $600 million in sales of generic atorvastatin, as nonbranded Lipitor is known.


In November 2012, Ranbaxy had to recall millions of pills after tiny glass particles were discovered in some of them. Even that, it turns out, was enough for only a temporary suspension, and the FDA permitted the company to resume sales in March.

Basically companies who want to cheat the system can easily get away with it:

As the Ranbaxy story makes vividly clear, generic-drug makers intent on breaking the rules -- especially those operating abroad -- can easily do so. Drug applications work on the honor system: The FDA relies on data provided by the companies themselves. "We depend on that information to be truthful," Gary Buehler, who headed the FDA's office of generic drugs for 10 years, said in December 2009.

  • 7
    Basically companies who want to cheat the system can easily get away with it. You show one example of one company which tried to cheat the system, and was trialled for that. I fail to see how this implies that anyone can do as they please and easily get away with it...
    – nico
    Oct 9, 2013 at 13:03
  • @nico: (1) If a senior FDA official says "We depend on the information to be truthful", that means people can cheat the system. (2) Nobody of the leadership of Ranbaxy got into prison. The amount of fines wasn't high enough for them to avoid simply repeating the same thing with a new drug and the FDA simply approved the new drug which had faked evidence. (3) I would expect that most people who cheat just tweak the numbers in a way that's hard to detect. It seems to take a lot of wrongdoing to be sued by the FDA for fraud.
    – Christian
    Oct 9, 2013 at 13:22
  • 1
    @nico One thing I found surprising in that article is that the FDA eventually placed 'Application Integrity Policy' notices (No longer would the FDA have the burden of proving fraud if it wanted to block a Ranbaxy product. The onus had flipped, and now the company would have to prove its products weren't fraudulent in order to get them approved.) only on specific facilities (factories), instead of on every facility run by the company. Other surprises are how long it takes for the legal action, the extent to which they don't recall products, and how very fraudulent Ranbaxy allegedly were/are.
    – ChrisW
    Oct 9, 2013 at 13:46
  • I would expect that most people who cheat just tweak the numbers in a way that's hard to detect I can only hope that most people/companies don't want to cheat: that the FDA's policies/advice on how and what to test are useful, because drug manufacturers want to use them.
    – ChrisW
    Oct 9, 2013 at 13:54
  • @ChrisW: still, one company being able to (partially) cheat the system does not imply that everyone else can or does.
    – nico
    Oct 9, 2013 at 14:41

The book "Bad Pharma" by Ben Goldacre covers this in great detail. In theory the FDA has lots of processes, but in practice drug companies have devised ways to get around them, which Goldacre documents in this book.

Ben Goldacre is not some fringe conspiracy theorist. He is currently a Senior Clinical Research Fellow at the Centre for Evidence-Based Medicine, part of the University of Oxford's Nuffield Department of Primary Care Health Sciences. His first book, "Bad Science" tackled everything from the bafflegab in cosmetics adverts to Andrew Wakefield's MMR vaccine scare. He was sued for libel by Matthias Rath, vendor of magic vitamin pills that cured AIDS, but Rath eventually dropped the suit and paid costs.

From the Introduction to Bad Pharma (as quoted by Wikipedia):

Drugs are tested by the people who manufacture them, in poorly designed trials, on hopelessly small numbers of weird, unrepresentative patients, and analysed using techniques which are flawed by design, in such a way that they exaggerate the benefits of treatments. Unsurprisingly, these trials tend to produce results that favour the manufacturer. When trials throw up results that companies don't like, they are perfectly entitled to hide them from doctors and patients, so we only ever see a distorted picture of any drug's true effects. Regulators see most of the trial data, but only from early on in a drug's life, and even then they don't give this data to doctors or patients, or even to other parts of government. This distorted evidence is then communicated and applied in a distorted fashion.

In their forty years of practice after leaving medical school, doctors hear about what works through ad hoc oral traditions, from sales reps, colleagues or journals. But those colleagues can be in the pay of drug companies – often undisclosed – and the journals are too. And so are the patient groups. And finally, academic papers, which everyone thinks of as objective, are often covertly planned and written by people who work directly for the companies, without disclosure. Sometimes whole academic journals are even owned outright by one drug company. Aside from all this, for several of the most important and enduring problems in medicine, we have no idea what the best treatment is, because it's not in anyone's financial interest to conduct any trials at all. These are ongoing problems, and although people have claimed to fix many of them, for the most part they have failed; so all these problems persist, but worse than ever, because now people can pretend that everything is fine after all.


This addresses the third question, which was about the recall of any drugs that had been misrepresented by company-sponsored research. Paroxetine has not been recalled, but calls for the retraction from the scientific record of its maker's clinical trials with adolescent subjects have yet to cease. The FDA warns of serious side effects including suicide that would not have occurred without the drug, but apparently doesn't believe that's enough of a reason to recall it.

Just a note about the FDA and the premise of the question: The FDA does not limit approvals to drugs that are universally helpful and harmless, and they do not necessarily recall drugs that have killed people. They do require warnings of adverse events when those are known, and highlight the most important ones by drawing a black box around them. They have yet to recall Serzone (nefazodone), which caused so many liver failures, transplants, and deaths that it can no longer be sold in Europe. Here, the manufacturer withdrew it voluntarily after the death of a 15-year-old girl followed 20 other deaths and dozens of liver failures, but the FDA doesn't object to the availability of generics; doctors can prescribe it and pharmacies can still sell it.

GSK (formerly Glaxo Smith Kline) was found to have mis-reported results from their Phase III trials of the safety and effectiveness of paroxetine for adolescents. Paroxetine is a selective serotonin re-uptake inhibitor (SSRI) sold as Paxil in the US and Seroxat in the UK (and many other names around the world). The FDA relied on the reports and the drug was approved for the expanded market of adults and adolescents. Also, there was a heavy contribution of GSK-funded ghostwriters on the original study, which listed high-profile academic psychiatrists as its authors.

Reports of unexpected suicides among adolescents on the drug began to surface. In 2002, the British series "Panorama" presented an investigative report called "The Secrets of Seroxat" that had a large impact on the British public. BBC received nearly 1400 emails and published all of them. The episode is currently viewable on youtube.

In 2004, New York State Attorney General Eliot Spitzer sued GSK for withholding negative information about the drug's effect on children. The company promised to post clinical trial data on their web site and paid $2.5M to the state.

Last year the British Medical Journal published a multi-author study that presented a re-analysis of the trial data, and new findings about the drug's safety and effectiveness, in a purportedly unbiased manner. In sum: not especially effective, and more likely to induce suicidality than placebo or the comparison drug, an older, non-SSRI antidepressant. From the study:

The efficacy of paroxetine and imipramine was not statistically or >clinically significantly different from placebo for any prespecified >primary or secondary efficacy outcome.

Suicidal adolescent test subjects on paroxetine in GSK's write-ups were described as emotionally labile, if their adverse events were even mentioned. This was explicitly stated in an email among several dozen that were leaked Emotional lability can mean nothing more than rapidly changing moods, none of which are depression or despair, and isn't interchangeable or closely linked with "suicidality" in mental health parlance. In the current Diagnostic and Statistical Manual, these related terms are offered as equivalent to emotional lability:

prone to mood fluctuations, emotions that are easily aroused and >intense, unstable emotional experiences and frequent mood changes, and >incapacitation by extremity of emotions

The authors of the BMJ article built a web site, Study329.org, that describes the process of obtaining access to the raw data, along with information about the original study, the original investigators, and press coverage of the BMJ article. The anonymized patient-level raw data they needed were not posted to the company's web site, and gaining access to the data and copying it over to the authors' computers, given the difficulties imposed by GSK, were among the largest challenges of the project.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .