Pharmaceutical companies routinely bury negative trial results. An analysis of registered trials found that approximately half are never published, systematically biasing the medical literature toward positive results.

The Publication Bias Problem

"We have no idea what the true effects of the drugs we prescribe are. Half the trials are never published. This is not a minor problem — it is a fundamental corruption of medicine."

"Ben Goldacre, MD"

Publication bias — the tendency to publish positive results and suppress negative results — is one of the most serious problems in medical research. It means that the published medical literature systematically overstates the benefits of drugs and understates their risks.

The mechanism is simple: pharmaceutical companies fund trials to test their drugs. When a trial shows that a drug works, they publish it. When a trial shows that a drug does not work or is harmful, they do not publish it. The result is a medical literature that looks much more favorable to drugs than the actual evidence warrants.

The Evidence

Ben Goldacre's analysis: In his 2012 book "Bad Pharma," Ben Goldacre documented systematic publication bias across multiple drug classes. For antidepressants, he found that the published literature showed 94% positive results, while the FDA data showed 51% positive results.

The AllTrials campaign: Goldacre's AllTrials campaign, launched in 2013, has documented that approximately half of all clinical trials are never published. This includes trials funded by governments and universities, not just pharmaceutical companies.

Approximately half of all clinical trials are never published. The published literature systematically overstates drug benefits and understates drug risks.

The Tamiflu example: Roche's antiviral drug oseltamivir (Tamiflu) was stockpiled by governments worldwide based on published evidence of efficacy. When the Cochrane Collaboration obtained the full clinical study reports (not just the published papers), they found that the drug's benefits were much smaller than the published literature suggested.

The Regulatory Solution

The FDA requires trial registration before trials begin, and requires reporting of results within 12 months of completion. However, compliance is poor: a 2015 study found that only 13% of trials registered with ClinicalTrials.gov reported results within 12 months.

The European Medicines Agency has made clinical study reports publicly available for drugs it has approved. This is a significant step toward transparency.

What This Means

Every drug that has been tested in multiple trials has unpublished negative results. The published evidence for any drug is systematically biased toward positive results. Physicians and patients should assume that the true evidence for any drug is weaker than the published literature suggests.