雷公藤制剂治疗类风湿关节炎随机对照试验文献方法学质量评价

目的 调查并评价雷公藤制剂治疗类风湿关节炎(RA)随机对照试验的方法学运用情况，为改进和提高临床试验水平及进行进一步的系统评价提供参考。方法 检索中国生物医学文献光盘数据库(CBMdisc，1978—2000年)及中医药科技文献数据库(TCMLRS，1984—2000年)中收入的以雷公藤制剂为主要干预措施治疗RA的随机对照试验，采用广州中医药大学循证医学方法研究课题组制定的《中医药治疗性文献系统性评价质量评价及信息采集表》对纳入文献进行方法学质量评价，同时运用日前国际上较通用的Jadad量表对文献质量计分。结果 27篇随机对照文献被纳入评价。有26％的文献描述了随机方法，仅有1篇文献有随机分配方案隐匿的描述。6篇文献采用了盲法。89％的文献Jadad计分低于3分，属于低质量的研究。另外，虽然所有文献均有疾病的诊断标准，但只有37％的文献提供了相应的纳入或(和)排除标准。21％的研究未提供药物安全性方面的情况。结论 本次调查的雷公藤制剂治疗RA随机对照试验在方法质量方面存在许多不足之处，因此仍有必要强测随机、对照、盲法等原则在随机对照试验研究中的正确运用。

Methodological quality evaluation of randomized controlled trials in rheumatoid arthritist related with preparations of Tripterygium wilfordii

Objective To assess the methodological quality of randomized controlled trials (RCTs) of rheumatoid arthritis (RA) treated with preparations of Tripterygium wilfordii to provide a scientific basis for the improvement of clinical trials and systematic review. Methods CBMdisc (1978-2000) and TCMLRS (1984-2000) were searched for RCTs on rheumatoid arthritis treated with preparations of Tripterygium Wilfordii. Jadad scale and the evaluating scale designed by the EBM methodological research group of Guangzhou University of TCM were used to evaluate the methodological quality of the trials. Results Of the 27 randomized trials included in this study, seven described the methods for randomization and only one had randomized concealment. Blindness was used in six of the trials. Twenty-four of the trials were found to be of low methodological quality with Jadad scores below three. Although all of the trials had their diagnostic standards for subject selection, only 37% described the inclusion and (or) exclusion criteria. Information about drug safety was not reported in 21% of the trials. Conclusions In general, the methodological quality of the studies is disappointing. Many shortcomings exist in the methodology of the included literatures, which calls for future attention to the correct use of the principles of randomization, control and blindness in randomized control trials.