Abstract
OBJECTIVE: To evaluate the accuracy of a 2015 cross-sectional analysis published in the BMJ Open which reported that pharmaceutical industry compliance with clinical trial registration and results reporting requirements under US law was suboptimal and varied widely among companies. DESIGN: We performed a reassessment of the data reported in Miller et al to evaluate whether statutory compliance analyses and conclusions were valid. DATA SOURCES: Information from the Dryad Digital Repository, ClinicalTrials.gov, Drugs@FDA and direct communications with sponsors. MAIN OUTCOME MEASURES: Compliance with the clinical trial registration and results reporting requirements under the Food and Drug Administration Amendments Act (FDAAA). RESULTS: Industry compliance with FDAAA disclosure requirements was notably higher than reported by Miller et al. Among trials subject to FDAAA, Miller et al reported that, per drug, a median of 67% (middle 50% range: 0%-100%) of trials fully complied with registration and results reporting requirements. On reanalysis of the data, we found that a median of 100% (middle 50% range: 93%-100%) of clinical trials for a particular drug fully complied with the law. When looking at overall compliance at the trial level, our reassessment yields 94% timely registration and 90% timely results reporting among the 49 eligible trials, and an overall FDAAA compliance rate of 86%. CONCLUSIONS: The claim by Miller et al that industry compliance is below legal standards is based on an analysis that relies on an incomplete dataset and an interpretation of FDAAA that requires disclosure of study results for drugs that have not yet been approved for any indication. On reanalysis using a different interpretation of FDAAA that focuses on whether results were disclosed within 30 days of drug approval, we found that industry compliance with US statutory disclosure requirements for the 15 reviewed drugs was consistently high.