7.质量评价
纳入文献的研究设计不同，其所需要使用到的文献评估工具也不同。针对RCT，可以参考Cochrane的Risk of Bias 2 Tool（RoB 2）。该工具从5个领域对某一研究的质量进行评价，包括：Bias arising from the randomization process，Bias due to deviations from intended interventions，Bias due to missing outcome data，Bias in measurement of the outcome，Bias in selection of the reported result。每个领域内包含一系列的“信号”问题，通过回答这些问题对相关风险进行判断。基于以上答案，对各个领域的偏倚进行判定，包括“低”“高”及“不清楚”。该工具的具体使用方法，可以参看Cochrane网站：https：//methods.cochrane.org/bias/resources。
类似地，针对非随机试验，也可采用Cochrane的工具进行质量评价：Risk of Bias in Non-Randomized Studies of Interventions（ROBINS-I）。该工具从7个领域对某研究的质量进行评价，包括：confounding，selection bias，Bias in measurement classification of interventions，Bias due to deviations from intended interventions，Bias due to missing data，Bias in measurement of outcomes，Bias in selection of the reported result.每个领域内包含一系列的“信号”问题，通过回答这些问题对相关风险进行判断。基于以上答案，对各个领域的偏倚进行判定，包括“低”“中”“中高”“高”及“不清楚”。该工具的具体使用方法，可以参看Cochrane网站：https：//methods.cochrane.org/bias/resources。
针对非实验性研究（横断面研究、队列研究及病例对照）的质量评价，可采用Ottawa-Newcastle Scale（NOS）或者Joanna Briggs Institute（JBI）中的工具。NOS及其修订版可从3个领域对各研究质量进行评价，包括：研究对象的选择、可比性、结果测量的准确性。根据研究设计，每个领域总分（使用“star”表示）有所不同。分数越高，表明研究质量越高。针对不同的非实验性研究，JBI包含不同的工具。例如，对于横断面研究，该工具采用了8个问题，对该研究的质量进行评价。每个问题相应的回答包括：是、否、不清楚、不适用。根据对8个问题的综合评估，确定是否将该研究纳入最终的文献综述中。该工具的具体使用方法，见相关网站：https：//joannabriggs.org/critical-appraisal-tools。
在使用以上研究质量评价工具时，需要注意：①多数情况下，质量评价需要由两名研究人员独立开展，以减少主观偏倚。②质量评价结果可以用来指导数据分析、研究结果的解释。此外，有些系统文献综述，会将“研究质量”作为纳入和排除标准。当然，有些综述会把所有符合标准的研究均纳入分析。不管采用何种方案，只要给出合理的原因和解释即可。
如示例5-17所示，本综述采用了不同的工具，对不同类型的研究质量进行了评价。并根据研究的质量，确定是否将该研究纳入综述中：对于干预性和观察性研究，不论其质量如何，均将其纳入；对于病历报告，仅纳入了高质量研究。
示例5-17 将研究质量作为纳入标准
Quality assessments were conducted by two independent reviewers（Y.M.B.and D.S.）using modified versions of the Critical Appraisal Skills Programme quality assessment tool for randomized controlled trials（52）and cohort studies（53）and the Joanna Briggs Institute Checklist for case reports.（54）High quality was stated as at least 14 out of 17 points for interventional studies，at least 11 out of 14 points for observational studies，and at least 7 out of 8 points for case reports.These cut-points were established by the review team considering the most relevant items of the checklists.All interventional and observational studies were included irrespective of quality score.Only high-quality case reports were included.
如示例5-18所示，本综述对所有质性研究进行了回顾。但是，并未针对研究质量设置纳入或排除标准。该综述纳入了所有符合条件的文献（不管质量如何）。
示例15-18 纳入所有研究（不论其质量如何）(https://www.daowen.com)
Eligible studies were appraised for methodological quality using the JBI Critical Appraisal Checklist for Qualitative Research.（40）The checklist consists of 10 criteria concerning the methodology，methods and findings of qualitative studies.Three members of the review team participated in this step，and each paper was appraised independently by two of them.Discrepancies that arose between the reviewers were resolved through discussion.No studies were excluded based on methodological quality.
如示例5-19所示，本综述纳入了所有符合条件的文献（不管质量如何），但是，在数据提取和分析阶段，对研究质量进行了限定：仅对质量较好的研究进行了数据提取和分析。
示例5-19 数据提取和分析阶段考虑研究质量
The critical appraisal of original papers was conducted independently by two reviewers（JK and MW）based on evaluation criteria adapted from Fortin et al.［12］，used to assess methodological quality（Supplementary data，Supplement 2）.Each article was scored out of 30，based on criteria on originality，population studied，definition and measurement of multimorbidity and HrQoL，and limitations of the study.Risk of bias was assessed using an adapted version of the Newcastle-Ottawa Scale（NOS）［21］（Supplementary data，Supplement3）suitable for observational studies，with good reliability and face validity［22］.Studies with scores of at least 15 out of 30 using the Fortin checklist，and 4out of 9 stars using the NOS were eligible for data extraction and synthesis.
如示例5-20所示，本综述纳入了所有符合条件的文献（不管质量如何），但是，在数据分析阶段，考虑了研究质量。该综述采用了敏感性分析，检验低质量研究对研究结果稳定性的影响。
示例5-20 数据分析阶段考虑研究质量
The quantitative meta-analyses were conducted using the Review Manager software（RevMan 5.3）.The first included all studies and a within-groups calculation.The second involved studies with follow-up measures.They were subgrouped and the effect size was calculated considering any long-lasting effects for DMT with this client population.The third calculation included only studies with RCT designs and between-group scores.Finally，studies with low risk of bias were included in the last calculation that involved a sensitivity analysis.
在撰写质量评价这一部分时，可以通过以下几个方面介绍：介绍不同类型的研究所使用到的质量评价工具；对各个评价工具的主要评估内容及评分标准进行介绍，同时说明所得分数代表的含义；说明质量评价由几个人开展，遇到意见不一致时，是如何处理的。