Abstract

Clinical Trials and Tribulations: “The Registration Haze”

Denise M. Goodman,1 Karen E. Gutzman,2 William F. Balistreri3

Objective

As editors, we have encountered author confusion regarding which studies require trial registration and at what point in time this should be done. We hypothesized that this may be attributable to discrepancies between the instructions for authors of individual journals and requirements found on registry websites (eg, ClinicalTrials.gov) and on ICMJE.org and sought to describe these differences.

Design

To be included, a journal had to have published at least 10 high-impact clinical trials from 2015 to 2016. A sample of more than 30 journals was identified for examination. Each journal title was searched using Google and the term <name of journal> instructions for authors to emulate author behavior. Journal websites were inspected for information on trial registration. Two investigators independently compared instructions for authors language with guidelines published at ClinicalTrials.gov and ICMJE.org and then coded the language as “conservative” (replicating ClinicalTrials.gov), “expanded” (replicating ICMJE.org), “other” (parochial, hybrid), or “silent” (no language provided). Discrepancies between investigators were resolved by discussion.

Results

For the 32 journals inspected, the Impact Factor ranged from 10.95 to 59.56, and the journals published a median of 47.5 trials (range, 11-300). The κ value between the 2 investigators was 0.513, with disagreement on 10 of 32 journals. None of the instructions for authors were judged to be conservative, 15 as expanded, 11 as other, and 6 as silent. Of those coded as “other,” 1 journal differed from ICMJE in 2 aspects and all others in 1. Differences between ICMJE instructions and those from individual journals included timing of trial registration (before enrollment vs up to 6 months after, n=1), interventions requiring registration (n=1), and study design (requiring registration only for randomized clinical trials or a comparison or control group, n=3). The largest group (n=7) were nonspecific, stating only that the author should provide a registration number or listing approved registries. There was no association between code for registration instructions and number of citations or Impact Factor.

Conclusions

More than one-third of high-impact journals did not follow either ICMJE or ClinicalTrials.gov guidelines, with one-fifth offering no guidelines on clinical trial registration. Harmonization of journal requirements with ICMJE could ameliorate this problem.

1Northwestern University Feinberg School of Medicine, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL, USA, dgoodman@luriechildrens.org; 2Galter Health Sciences Library, Northwestern University Feinberg School of Medicine, Chicago, IL, USA; 3University of Cincinnati School of Medicine, Cincinnati, OH, USA

Conflict of Interest Disclosures:

Denise Goodman is an associate editor at the Journal of Pediatrics, Elsevier, for which she receives an honorarium; the journal was not involved in this work. William Balistreri is editor in chief of the Journal of Pediatrics, for which he receives an honorarium; the journal was not involved in this work. Karen Gutzman reported no financial disclosures.

Funding/Support:

This work was not funded.