Selective reporting and bias are undermining the integrity of the clinical evidence system say the authors of new analysis who suggested changes to ClinicalTrials.gov would have a positive impact.
The analysis – published in Nature this month – is based on a review that suggests publication bias is the root cause of the issue.
“Clinical researchers tend to favour positive outcomes while neglecting negative results, thereby contributing substantially to an unhealthy clinical evidence ecosystem and presenting significant challenges to high-quality, unbiased clinical decision-making,” the authors write.
In theory the website Clinicaltrials.gov - a repository of data from completed and in-progress clinical trials – is designed to help address these bias issues. But, the authors say there are problems with how data is arranged, which has limited its use to date.
“The unique value inherent in CT.gov data has been demonstrated through various comparisons and analyses. When compared with results from PubMed, it has been noted that CT.gov often contains a more comprehensive report of adverse events. In CT.gov, safety results were reported at a similar rate as in peer-reviewed literatures, yet with more thorough reports of certain safety events.
However, the authors add “the current storage of CT.gov reported results is limited to web-based or raw XML format. The absence of automated processing tools and a lack of structured reporting results dataset constitute one of the major barriers to the widespread utilization of CT.gov.”
Knowledge graph
To address this, the authors propose a curated dataset that links efficacy with safety results at the experimental arm group level within each trial, and connects them across all trials through a “knowledge graph,” a network of data in which the relationship between each point is illustrated.
The potential benefits of using CT.gov in this way are significant according to the authors, who say it “bridges the gap between generally described searchable information and specifically detailed yet underutilized reported results, and promotes a dual-faceted understanding of interventional effects.
“Adhering to the “calculate once, use many times” principle, the structured dataset will enhance the reuse and interpretation of ClinicalTrials.gov results data.”
DepositPhotos/realinemedia