Industry trial registrations indicate positive publication bias

Clinical trials that are funded by industry are much more likely to publish positive results about the drugs they study than trials funded primarily by the government or other sources, according to a new analysis.

Researchers used the Web-based registry ClinicalTrials.gov to collect information about safety and efficacy trials of anticholesteremics, antidepressants, antipsychotics, proton-pump inhibitors and vasodilators conducted between 2000 and 2006. The analysis included 546 drug trials, 63% of which were primarily funded by industry, compared to 14% by the government and 23% by nonprofit or nonfederal organizations. Overall, the researchers were able to find published results for 66% of the registered trials. The study was published in the Aug. 3 Annals of Internal Medicine.

Of the industry-funded trials that were published, 85% showed positive outcomes, compared to 50% of government trials and 72% of the nonprofit/nonfederal group. Some of the nonprofit/nonfederal trials used contributions from industry, and those that did so were significantly more likely to report positive outcomes than those that didn't get money from industry. Industry-funded trials were also less likely to be published within two years of completion than the other types—only 32% compared to 56% of the nonfederal/nonprofit trials.

Noting that industry trials are more likely to be in phase 3 or 4, and therefore potentially more likely to be positive than early-stage research, the researchers controlled for trial phase and still found an association between industry funding and positive outcomes. The study authors also noted that many trials were not registered until after the study got under way or, in some cases, was published. Many registrations also failed to link to resulting publications, making it more difficult to track the results of trials.

The study authors concluded that more accurate, timely and expanded reporting in trial registrations is needed to maximize the transparency of the clinical trial process. They also listed some potential changes to ClinicalTrials.gov, including incorporating study protocols and amendments into registrations and making the information from FDA drug reviews public.