|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||
Members of the Journal's staff routinely check the database registration of trials that are being considered for publication. The registration provided by Dr. Motzer — NCT00083889 [ClinicalTrials.gov] in ClinicalTrials.gov — failed to meet the standards established by the International Committee of Medical Journal Editors (ICMJE).1 This registration had been filed by Pfizer, the study's sponsor. On August 26, 2005, the sponsor had entered the following in the outcome-measures field: "Primary (Secondary) outcome information was omitted due to [its] commercial sensitivity and will be revealed at a later date." Since this entry regarding the study's outcome was clearly not informative, the decision was made, on this basis alone, not to consider the manuscript further.
Given that the article appears in this issue of the Journal,2 there is clearly more to this story. Shortly after Motzer was informed that his manuscript could not be considered, the Journal learned that one of his coauthors had registered the trial through one of the cancer centers, rather than through Pfizer. This June 2005 registration in ClinicalTrials.gov, NCT00098657 [ClinicalTrials.gov] , included an informative entry in each of the fields required by the ICMJE. Thanks to this second registration, the manuscript met the requirements to be considered for publication.
In the past year, Pfizer has changed its posture on the registration of clinical trials. None of the database entries for the 115 trials that Pfizer has registered since the beginning of 2006 are missing information in either the intervention or the outcome-measures field, though there are still a few Pfizer records created before 2006 that are missing some of this information. Pfizer is not alone: since the start of 2006, other pharmaceutical companies have also improved the way they register trials. As of December 1, 2006, of the 2983 pharmaceutical-industry registrations filed in ClinicalTrials.gov in the past year, 8% were missing information on outcome measures (down from 26% of the 5355 industry registrations entered before January 1, 2006), and none were missing the name of the intervention. (A small number of trials that were registered before 2006 were still missing intervention names.) Although more can be done, this improvement in registration quality is to be praised.
In the current case, the author's inadvertent duplicate registration allowed the article to be published in an ICMJE-compliant journal, but a better approach for authors in the future would be to coordinate with study sponsors to ensure that a single, informative registration record is provided. Multisite studies, including those conducted in multiple countries, should also be registered under one record. Ensuring that registration is completed properly requires close cooperation among investigators, sponsors, and other study officials.
Investigators can contact the registry database if they are not certain whether their study has already been registered. Inadvertent duplication creates difficulties for the registry and the research community by making it impossible to determine how many and which studies are being conducted, thereby partially undermining the purpose of registration. (Once duplicates are found, ClinicalTrials.gov suppresses one record and cross-references both NCT numbers.)
We urge investigators, including those who receive funding from the National Institutes of Health or other nonindustry sources, to ensure that the studies they conduct are registered with complete information and to check the registration records for accuracy. Specific information about how to register a study in ClinicalTrials.gov can be found at http://prsinfo.clinicaltrials.gov and is described in a recent article.3 Investigators should avoid participating in trials if they are not confident that an accurate and complete record of the trial will be maintained in an acceptable trials registry.
The sunitinib study is a case in which the willingness of one of the investigators to disclose key information about the protocol allowed a study to be considered for publication by one of the many medical journals that adhere to the ICMJE trials-registration policy. The message should be clear to all investigators participating in clinical trials: before you enroll a patient in a study, be sure that there is a full and appropriate registration of the trial in a public database approved by the ICMJE (www.icmje.org). It could salvage a study report that otherwise would not be published.
No potential conflict of interest relevant to this article was reported.
Source Information
From ClinicalTrials.gov, National Library of Medicine, National Institutes of Health, Bethesda, MD (D.A.Z.).
References
| |||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Trial Registration
Lott J. P., Drazen J. M.
Extract |
Full Text |
PDF
N Engl J Med 2007;
356:1581, Apr 12, 2007.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
