Quality of Reporting Randomized Controlled Trials in Five Leading Neurology Journals in 2008 and 2013 Using the Modified "Risk of Bias" Tool

BACKGROUND: To examine the risk of bias of methodological quality of reporting randomized clinical trials (RCTs) in major neurology journals before and after the update (2011) of Cochrane risk of bias tool. METHODS: RCTs in 5 leading neurology journals in 2008 and 2013 were searched systematically. Characteristics were extracted based on the list of the modified Cochrane Collaboration's tool. Country, number of patients, type of intervention, and funding source also were examined for further analysis. RESULTS: A total of 138 RCTs were enrolled in this study. The rates of following a trial plan were 61.6% for the allocation generation, 52.9% for the allocation concealment, 84.8% for the blinding of the participants or the personnel, 34.8% for the blinding of outcome assessment, 78.3% for the incomplete outcome data, and 67.4% for the selective reporting. A significant setback was found in "the selective reporting" in 2013 than that in 2008. Trials performed by multi-centers and on a large scale had significantly more "low risk of bias" trials. Not only the number of surgical trials (5.8%) was much less than that of trials using drugs (73.9%), but also the reporting quality of surgical trials were worse (P = 0.008). Finally, only 17.4% trials met the criterion of "low risk of bias." CONCLUSIONS: The modified "risk of bias" tool is an improved version for assessment. Methodological quality of reporting RCTs in the 5neurology journals is unsatisfactory, especially that for surgical RCTs, and it could be further improved.