Publication bias still evident in US trials database

pharmafile | September 18, 2009 | News story | Medical Communications, Sales and Marketing ethics 

 

A study showing that two-thirds of clinical trials included in the US national register had no record of their primary outcome suggests selective publication of trials remains a major problem in the healthcare arena.

The authors of the study in PLoS Medicine conclude that the ability of the register to address “selective publication and better inform the public and professionals about the results from completed clinical trials is limited because critical information from trial registration … were not consistently reported”.

The US National Institutes of Health (NIH) set up the clinicaltrials.gov [http://www. clinicaltrials.gov] database in 2000 at a time when it was common practice for pharmaceutical companies to provide no information about the set up and conduct of clinical trials.

That environment raised concerns that pharmaceutical companies and other trial sponsors could publish only positive studies on their products, burying the results of those which painted their treatments in a less favourable light.

Since 2007, registration on clinicaltrials.gov has been mandatory for phase II-IV controlled clinical trials of US Food and Drug Administration-regulated drugs, biologics, and devices, and voluntary for phase I studies, with more than 70,000 trials around the world on the register.

The implementation of the 2007 FDA Amendment Act also introduced the requirement that results data or citations are added into the clinical trial record, including baseline characteristics of patients, as well as primary and secondary outcomes data, within either one or two years of the study ending.

The authors looked at 7,515 studies completed before the Act came into effect to see how

information was supplied in the absence of formal legislation demanding it. 64% of entries reported the primary outcome, but data quality “varied markedly”, according to the researchers, who were led by Joseph Ross of Mount Sinai School of Medicine in New York.

They also took a random sample of 10% of these trials and carried out a search of the literature (via MEDLINE) to see if they had been published. Less than 50% of the studies had published results, and only a third had their citations added to the clinicaltrials.gov record.

The authors also found that in this sample publication rates were lower for industry-sponsored studies (40% compared to those carried out by non-industry groups (56%).

implementation of the FDAA Act remains under negotiation, so any data on the current standard of practice is useful to inform that debate.

And while this situation may improve with full implementation of the Act’s requirements, the study also raises questions about whether there is any systematic overview of the clinicaltrials.gov database to make sure sponsors meet their filing obligations.

Related Content

Scientist who genetically modified babies faces punishment after acting illegally, Chinese authorities confirm

Chinese state media have confirmed that He Jiankui, the scientist who claimed to have genetically …

pillsphotos1

The pharmaceutical ethics of drug-price increases

What is an ‘ethical’ price for a drug?Last month Turing Pharmaceuticals, a start-up run by …

Watermeadow’s continued commitment to ethical publications

As part of Watermeadow Medical’s commitment to ethical publications, five more of its editorial staff …

Latest content