Intended for healthcare professionals

Letters

Journals should see original protocols for clinical trials

BMJ 2001; 323 doi: https://doi.org/10.1136/bmj.323.7324.1309 (Published 01 December 2001) Cite this as: BMJ 2001;323:1309
  1. C J Hawkey (cj.hawkey{at}nottingham.ac.uk), professor of gastroenterology
  1. School of Medical and Surgical Sciences, Division of Gastroenterology, University Hospital, Nottingham NG7 2UH

    EDITOR—Gottlieb reports the criticism of the editorial in JAMA mentioning celecoxib, an arthritis drug, 1 2 but surely more censure should be reserved for the journal that published the article and its reviewers. It is their responsibility to ensure that published articles reflect the truth.

    It was widely known that the celecoxib long term arthritis safety study (the CLASS study) lasted for a median of nine (maximum 13) months, as the data had been extensively presented at meetings for six months before the JAMA publication. It was therefore surprising that JAMA chose to publish partial censored data from a completed trial. When I tried to begin some correspondence about this in JAMA the journal's response was that there was not sufficient interest.

    Journals could protect themselves against such problems if they insisted that the original protocol was submitted with any publication describing a clinical trial. This simple expedient would protect against publication of partial data or the retrospective identification of end points that were not prospectively stated; its adoption is long overdue.

    References

    1. 1.
    2. 2.