We elaborated the reasons why systematic reviews need to use GRADE based on a couple of specific examples. Aiming to provide references to understand and use GRADE correctly, we also answered some frequently-asked questions and concerns about GRADE as follows: a) differentiating the uses of GRADE between its application in guidelines and in systematic reviews; b) how to determine the overall quality of evidence? c) can GRADE be used to access the quality of single study or not? d) different uses of GRADE between randomized controlled trials (RCTs) and observational studies; e) weight of GRADE items; and f) factors that might influence the results of GRADE and the balance between upgrading and downgrading.
本文介绍证据质量分级的GRADE方法。GRADE将证据质量分为高、中、低和极低四个级别。该分级应用于证据群，而非针对个别研究。系统评价中，质量反映了我们认为效应估计值正确的把握度。对推荐意见而言，质量反映了我们认为效应估计值足以支持某特定推荐的把握度。随机对照试验初定为高质量证据，观察性研究初定为低质量证据。GRADE中所使用的质量一词不只是意味着偏倚风险，还可能受研究结果的不精确性、不一致性和间接性，及发表偏倚的影响。此外，若干因素可增加我们对效应估计值的把握度。GRADE提供了一种系统方法来思考并报告各因素。GRADE将评估证据质量的过程与给出推荐建议的过程分开。推荐强度的判断不仅依赖于证据质量。