Which doctors die first?

Abstract

CONSORT statement on the reporting standards of clinical trials Recommendations are inappropriate for trial reportsEditor-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 fetoprotein screening for neural tube defects.CONSORT is the product of deliberations conducted almost entirely through North American journals. 2Journal 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 it 3 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 case 3 or from other commentators does not seem to have materially altered the final version of CONSORT.By announcing its decision, less than a fortnight after publication of CONSORT, the BMJ pre-empted input from many with experience of planning, conducting, and reporting trials.Few will dispute the need for improvements in the reporting of all studies with implications for patient care, whatever their design.But without appropriate consultation, the BMJ seems to have made its decision on CONSORT prematurely.The requirement that randomised trials report the numbers of patients excluded is generally pointless, 5 and making authors indicate where a large number of checklist items are dealt with would be unduly prescriptive.The CONSORT recommendations should not be imposed unilaterally.Others with experience of trial conduct should be consulted on the nature of any guidelines for the reporting of randomised trials and, more importantly, of observational studies of treatments.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 16 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.