Assessment of methodological quality and outcome measures of acute stroke randomized controlled trials in China in recent 15 years

BACKGROUND: Randomized controlled trials (RCTs) are the most important evidence to guide clinical practice in the treatment of acute stroke. This study aims to evaluate the changes of quantity and methodological quality of acute stroke RCTs in Mainland China published in recent 15 years. METHOD: We included un-confounded RCTs on acute stroke from eight databases published in Chinese or English from 1996 to 2010. General characteristics, design methodology, and outcome measures of studies were assessed. RESULTS: Totally, 9061 RCTs were identified. The number of acute stroke RCTs had increased by years, the total of trials published in 2010 was 20 times of that published in 1996. Three thousand four hundred eighty-eight trials (38.5%) used western drugs in the treatment, 3026 (33.4%) trials used traditional Chinese medicine (TCM), and 1933 (22.0%) trials used physical therapy. Ischemic stroke was the most common research subject among all trials (65.1%, 5,989). There were 541 (6.0%) RCTs using adequate randomization methods, 34 (0.4%) RCTs using adequate allocation concealment, and 195 (2.2%) using adequate blinding methods. Thirty-three (0.4%) RCTs adopted both adequate randomization methods and allocation concealment. Only 23 (0.3%) trials used all three methods of adequate randomization methods, allocation concealment, and blinding. During the 15 years, only the number of trials using adequate randomized methods and reporting adverse events had significantly increased (both P < 0.001). As for these RCTs, only the number of trials using adequate blinding method in pharmaceutical intervention was statistical differences compared with that of nondrug intervention trials (P = 0.043). Outcomes were assessed blindly in 72 trials. Death was reported by 14.2% of trials, impairment by 85.1%, disability by 22.5%, and handicap or quality of life by 0.4%. 99.7% trials reported positive results. Larger trials were more likely to use adequate randomized methods, allocation concealment and blinding methods, as well as to measure disability. Only 14 large-sample high-quality RCTs were found. CONCLUSION: During 15 years, the number of acute stroke RCTs has increased dramatically, but the quality of trials improves slowly. Most acute stroke trials used inadequate outcome measures in terms of their content, reliability, validity, blinding assessment.