Letters

CONSORT statement on the reporting standards of clinical trials

BMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7087.1126 (Published 12 April 1997) Cite this as: BMJ 1997;314:1126

Recommendations are inappropriate for trial reports

  1. T W Meade, Professora,
  2. Nicholas Wald, Professora,
  3. Rory Collins, Professorb
  1. a Wolfson Institute of Preventive Medicine, St Bartholomew's and the Royal LondonSchool of Medicine and Dentistry, London EC1M 6BQ
  2. b Clinical Trial Service Unit and Epidemiological Studies Unit, Radcliffe Infirmary, Oxford OX2 6HE
  3. c Department of Anaesthetics, Southmead Hospital, Bristol BS10 5NB
  4. d Medical Research Council, London W1N 4AL
  5. e ICRF Medical Statistics Group, Centre for Statistics in Medicine, Institute of Health Sciences, Oxford OX3 7LF
  6. f Clinical Epidemiological Unit, Ottawa Civic Hospital, Ottawa, Ontario K1Y 4E9, Canada
  7. g Institute for Health Policy Studies, 1388 Sutter Street, San Francisco, CA 94109, USA

    Editor—In his editorial on the CONSORT statement Douglas G Altman states that since “only randomised trials have the potential directly to affect patient care … it is … reasonable to require higher standards for papers reporting randomised trials than those describing other types of study.”1 But although randomised trials are usually the preferred approach for the unbiased evaluation of treatments, non-randomised observational studies also influence treatment. Because the correct interpretation of observational studies is much more difficult, largely due to uncontrollable biases, their description and analysis call for at least as much care as do randomised trials. Sometimes, of course, randomised comparisons are unnecessary–for example, to determine the value of α fetoprotein screening for neural tube defects.

    CONSORT is the product of deliberations conducted almost entirely through North American journals.2 Journal editors and statisticians are well represented among those involved, but clinicians and others with experience in actually conducting trials less so. Experience of the procedures recommended seems to be limited and not wholly positive. The authors of one trial report who had been persuaded to rewrite it3 in the structured format expressed mixed feelings: the paper was rendered less readable with “ideas that are logically linked together (such as point estimates with 95% confidence intervals) torn asunder.”4 Inflexibility and a lengthier manuscript were other reported disadvantages. The editor stated that a journal “should change the way it structures articles only when there is good agreement on the new format in the communities we serve”4; this includes writers and readers of trial reports as well as editors. Feedback from the test case3 or from other commentators does not seem to have materially altered the final version of …

    View Full Text

    Sign in

    Log in through your institution

    Subscribe