Premium
Nonpublication of trial results for new neurological drugs: A systematic review
Author(s) -
Hakala Amanda K,
Fergusson Dean,
Kimmelman Jonathan
Publication year - 2017
Publication title -
annals of neurology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 4.764
H-Index - 296
eISSN - 1531-8249
pISSN - 0364-5134
DOI - 10.1002/ana.24952
Subject(s) - medicine , clinical trial , confidence interval , drug trial , licensure , family medicine , medline , alternative medicine , hazard ratio , trial registration , pathology , nursing , political science , law
Objective To evaluate nonpublication rates among trials of new successful and unsuccessful neurological drugs. Methods “Licensed” drugs consisted of all novel agents receiving US Food and Drug Administration (FDA) licensure 2005–2012 inclusive in seven neurological disorders. “Stalled” drugs included all experimental agents tested in the same domains that had at least one completed phase III trial in the same time frame, but failed to receive FDA approval. Trials of these drugs were included in our sample if their primary outcome collection occurred before October 1, 2010. We determined the publication status of eligible trials using searches of clinicaltrials.gov , Google Scholar, PubMed, Embase, sponsor websites, and direct electronic query of trial contacts and sponsors. The primary outcome was time to journal publication (or results reporting in other media) after study completion. Results The adjusted hazard ratio for publication was 1.79 (95% confidence interval, 1.20–2.67) in favor of licensed drugs. Based on the criteria for nonpublication in this report, 14,092 and 33,882 volunteers participated in unpublished trials of licensed and stalled neurological drugs, respectively. Result data were not publicly available in any form for 10% (16 of 163) and 46% (94 of 203) of trials of licensed and stalled drugs, respectively. Interpretation Results of trials for stalled drugs are heavily under‐reported. This deprives research and care communities of evidence about pathophysiology, drug class effects, and the value of surrogate endpoints in trials. Ann Neurol 2017;81:782–789