The quality of reporting of randomized clinical trials published in Ophthalmology

Purposes To evaluate the quality of reporting of all the new randomized cli nical trials published during 1999 in Ophthalmology using the CONSORT state ment parameters. Design: Descriptive series of published studies. Studies Included: Randomized clinical trials. Methods: Eligible studies were all randomized clinical trials published in Volume 106 (1999) of Ophthalmology that reported the results of prospective ly conducted studies in humans comparing two or more therapeutic options wi th regard to their effect on one or various outcomes, in which the authors stated that a process of randomization or quasi-randomization was performed before the interventions. Each study was evaluated by two independent obse rvers establishing the specific report in the published manuscript of 57 CO NSORT statement descriptors, and consensus was obtained between evaluators. Main Outcome Measures: Overall presence of CONSORT statement descriptors in published reports. Results: The mean number of positive CONSORT descriptors for the 24 eligibl e randomized clinical trials was 33.42 (standard error of the mean [SEM] 1. 57) of 57 possible. This score was significantly different (P < 0.001) from the score of 16.6 (standard error of the mean 0.6) reported for Ophthalmol ogy in the 1991 to 1994 study by Scherer and Crawley. Of the nine descripto rs measuring the quality of abstract reporting, a mean of 6.25 (SEM 0.3) wa s included in the 24 study abstracts evaluated, significantly different fro m the Scherer and Crawley score (5.2, SEM 0.2. P < 0.001). Four and six des criptors increased and decreased, respectively, their reporting in more tha n 20%. Key descriptors such as hypothesis postulation, primary outcome dete rmination, sample size calculations, and a discussion of the limitations of the study such as bias, imprecision, and internal validity are still repor ted in less than 42% of clinical trials. Conclusions: This study describes for the first time after the adoption of the CONSORT statement recommendations the status on the quality of reportin g of recently published randomized clinical trials in Ophthalmology. Althou gh not a direct measure of the intrinsic quality of a study, reporting qual ity provides the reader with useful tools for the evaluation of its validit y. The quality of reporting of randomized clinical trials shows an overall improvement when compared with the period between 1991 and 1994. However, t here is still significant room for improvement, especially in descriptors e ssential for the validation of clinical trial results that are still widely underreported. Ophthalmology 2001;108:410-415 (C) 2001 by the American Aca demy of Ophthalmology.