Changes to registration elements and results in a cohort of Clinicaltrials.gov trials were not reflected in published articles

J Clin Epidemiol. 2016 Feb;70:26-37. doi: 10.1016/j.jclinepi.2015.07.007. Epub 2015 Jul 29.

Abstract

Objectives: To assess effectiveness of legislative initiatives to stimulate public registration of trial results, we assessed adherence to protocol and results reporting, changes to registry, and publication data for randomized controlled trials (RCTs) after introduction of Food and Drug Administration Amendment Act (FDAAA).

Study design and setting: Observational study of a cohort of ClinicalTrials.gov registered FDAAA-covered RCTs found through ClinicalTrials.gov between 2009 and 2012 and data from corresponding publications. WHO Minimum Data Set items were abstracted by one author and verified by the other author.

Results: Among 81 eligible trials, most were industry-funded, with a drug intervention in parallel assignment. Secondary outcomes at the initial and last registration were omitted for 17% and 19.7% of RCTs, respectively. RCT registration changes mostly involved scientific title (18.8%). Inclusion criteria omission was most common (88%) in publications. Inferential statistical methods for primary and secondary outcomes matched between registry and publication for 53.4% and 28.6% of RCTs, respectively. Serious and other adverse events (AEs) that were absent for 23.8% and 4.8% of RCTs, respectively, were published as nonoccurring.

Conclusion: Discrepancies remain relatively high between registered and published outcomes, particularly regarding registered omissions in publications and concomitant reporting, nature of statistical method used, and reporting of AEs. This seriously undermines transparency of clinical trials and needs immediate attention of all stakeholders in health research.

Keywords: Bias (epidemiology); Databases; Drug side effects; Factual; Randomized controlled trials as topic; Reporting.

Publication types

  • Observational Study

MeSH terms

  • Drug Evaluation
  • Editorial Policies
  • Financing, Government
  • Guideline Adherence
  • Humans
  • Information Dissemination / methods*
  • Internet
  • Periodicals as Topic*
  • Publication Bias
  • Publishing / standards*
  • Publishing / statistics & numerical data
  • Randomized Controlled Trials as Topic / standards*
  • Registries / standards*
  • Registries / statistics & numerical data
  • Research Report / standards*
  • Research Support as Topic
  • United States
  • United States Food and Drug Administration