Industry Funding Linked to Increased Efficacy in Psychiatric Drug Trials
核心洞察
A new study reveals that psychiatric drugs appear approximately 50% more effective in manufacturer-funded clinical trials compared to independently sponsored trials.
The "sponsorship effect" highlights potential bias, where manufacturer-funded trials are more likely to report favorable outcomes for their drugs.
Publication bias, where positive results are more likely to be published, significantly contributes to the observed sponsorship effect in drug trials.
New research indicates a significant bias in clinical trials for psychiatric drugs, with manufacturer-funded studies reporting approximately 50% greater efficacy compared to trials funded by other sources. This "sponsorship effect," identified by Tamar Oostrom, assistant professor of economics at The Ohio State University, raises concerns about the reliability of efficacy data for antidepressants (搜索) and antipsychotics (搜索).
The study, published in the Journal of Political Economy, compared head-to-head trials of the same drugs, differing primarily in funding source. Oostrom's analysis of 509 published clinical trials revealed a dramatic difference in reported efficacy based on sponsorship. For example, in trials comparing Effexor (funded by Wyeth Pharmaceuticals (搜索)) to Prozac, Effexor was more frequently found to be more effective in Wyeth-funded trials (12 of 14) compared to independently funded trials (1 of 3).
Impact of Publication Bias
The research identified that publication bias plays a crucial role in the sponsorship effect. Trials with positive outcomes for the manufacturer's drug were more likely to be published, while those with less favorable results often remained unpublished. Oostrom identified 77 unpublished trials, and incorporating just one of these trials into the analysis reduced the sponsorship effect by 20%. "The addition of unpublished trials reduces the effect of sponsorship, and most of the sponsorship effect can be explained by publication bias," Oostrom stated.
Mitigation Through Preregistration
To combat publication bias, preregistration of clinical trials has been implemented, requiring researchers to register trials and report results as conditions for publication or funding. Oostrom's analysis suggests that the sponsorship effect has declined since 2005, coinciding with the increasing adoption of preregistration. However, she notes that preregistration is not a complete solution, as only about 25% of preregistered trials currently report results.
Implications for Existing Drugs
Oostrom emphasizes that many currently available antidepressant and antipsychotic drugs were approved before preregistration requirements were in place. This raises concerns about the potential for biased evidence influencing the approval and use of these medications. Further research and increased transparency are needed to address these concerns and ensure the reliability of clinical trial data.