Objective: To assess the number and quality of the reporting of randomised controlled trials (RCTs) published in Intensive Care Medicine.
Design: Systematic revision.
Setting: Randomised controlled trials published in Intensive Care Medicine.
Study selection: All RCTs published in this journal from its birth to December 2000 identified by MEDLINE and our own research.
Measurements and results: The Jadad scale and the individual assessment of key methodological components, namely the randomisation process, blinding and reporting and handling of loss to follow-up, were used to evaluate the quality of reporting. Other information was extracted regarding the design characteristics and the analytical approach. 173 RCTs, 63% of which were from European countries, were analysed. Adequately reported RCTs according to a Jadad scale score of more than 2 were 44 (25.4%). Analysis of individual methodological components revealed a variable percentage of adequate reporting ranging from 3.5% for randomisation to 10.4% for blinding and to 49.1% for loss to follow-up. Sample sizes were small with a median of 30 patients and rationale for its estimation was reported in 7.5%. Despite this, 81.5% of RCTs reported statistically significant results, suggesting that the treatment effects were strong or that a publication bias existed or that the uncertainty principle was not fulfilled.
Conclusions: Randomised controlled trials offer the best evidence of the efficacy of medical interventions, provided that high standards of transparent reporting are used. More resolute attention to the methodological quality of reporting and adherence to recently published guidelines (CONSORT II) may help to achieve this result.