ObjectiveTo evaluate the risk of bias and reliability of conclusions of systematic reviews (SRs) of lung cancer screening. MethodsWe searched PubMed, EMbase, The Cochrane Library (Issue 2, 2016), Web of Knowledge, CBM, WanFang Data and CNKI to collect SRs of lung cancer screening from inception to February 29th, 2016. The ROBIS tool was applied to assess the risk of bias of included SRs, and then GRADE system was used for evidence quality assessment of outcomes of SRs. ResultsA total of 11 SRs involving 5 outcomes (mortality, detection rate, survival rate, over-diagnosis and potential benefits and harms) were included. The results of risk of bias assessment by ROBIS tool showed:Two studies completely matched the 4 questions of phase 1. In the phase 2, 6 studies were low risk of bias in the including criteria field; 8 studies were low risk of bias in the literature search and screening field; 3 studies were low risk of bias in the data abstraction and quality assessment field; and 5 studies were low risk of bias in the data synthesis field. In the phase 3 of comprehensive risk of bias results, 5 studies were low risk. The results of evidence quality assessment by GRADE system showed:three studies had A level evidence on the outcome of mortality; 1 study had A level evidence on detection; 1 study had A level evidence on survival rate; 3 studies on over-diagnosis had C level evidence; and 2 studies on potential benefits and harms had B level evidence. ConclusionThe risk of bias of SRs of lung cancer screening is totally modest; however, the evidence quality of outcomes of these SRs is totally low. Clinicians should cautiously use these evidence to make decision based on local situation.
Previous methods of grading evidence for systematic reviews of diagnostic test accuracy have generally focused on assessing the certainty (quality) of evidence at the level of diagnostic indicators. When the question is not limited to follow the diagnostic test accuracy results themselves, the grading results may be inaccurate due to the lack of consideration of the downstream effects of the test accuracy in specific settings. To address these challenges, the GRADE working group conducted a series of studies focused on updating methods to explore or simulate important downstream effects of diagnostic test accuracy outcomes within a contextual framework. This paper aimed to introduce advances in the contextual framework of the GRADE approach to rate the certainty of evidence from systematic reviews of single diagnostic test accuracy.
GRADE要求明确说明相关的背景、人群、干预措施和对照，同时要求不论研究结果能否形成证据，均需详述所有重要结果。对某一特定管理问题，人群、干预措施及结果应在不同研究间足够类似，才能认为得到相似的效应量合乎情理。指南制定者在收集证据前应先详细说明各结局的相对重要性，同样地，证据总结完成时也需要详细说明这一点。考虑到替代结局的重要性，对采用替代指标描述且对患者很重要的结局，作者应评估其重要性，并进而降低这种间接结果的证据质量等级。
GRADE（Grades of Recommendation, Assessment, Development，and Evaluation）方法为卫生保健中的证据质量评价与推荐强度评级提供指导。对那些为系统评价、卫生技术评估及临床实践指南总结证据的人而言，GRADE具有重要意义。GRADE提供了一个系统而透明的框架用以明确问题，确定所关注的结局，总结针对某问题的证据，以及从证据到形成推荐或作出决策。GRADE方法的广泛传播与应用，获全球50余个组织认可，这些组织大多有很强的影响力（http://www.gradeworkinggroup.org/），足以证明该工作的重要性。本文介绍临床流行病学杂志将刊出的20篇系列文章，为如何使用GRADE方法提供指导。
在GRADE方法中，若多数相关证据来自高偏倚风险的研究，则起初被定为高质量证据的随机试验和低质量证据的观察性研究均有可能被降低质量等级。随机试验已确定的局限性包括：未进行分配隐藏、未实施盲法、未报告失访情况及未恰当考虑意向性治疗原则。最近提出的局限性包括：因明显获益而早期终止试验和基于结果选择性报告结局。观察性研究的主要局限性包括使用不合适的对照及未能充分调整预后的不平衡。偏倚风险可因不同结果而异（如全死因死亡率的失访远少于生命质量的失访），许多系统评价都容易忽略这一点。在决定是否因偏倚风险而降低质量等级时，不管是随机试验还是观察性研究，作者不应采用对各个研究取平均值的方法。相反，对任何单个结果，当同时存在高、低偏倚风险的研究时，则应考虑只纳入较低偏倚风险的研究。
ObjectiveTo systematically review the researches on grading criteria for quality of evidence and strength of recommendations of traditional Chinese medicine (TCM). MethodsPubMed, EMbase, Web of Science, CNKI, WanFang Data and VIP databases were electronically searched to collect researches on grading criteria for quality of evidence and strength of recommendations of traditional Chinese medicine from inception to June 2021. Two reviewers independently screened literature, extracted data; and then, descriptive analysis was performed using qualitative methods. ResultsA total of 18 studies were included. Specifically, 8 studies presented both the level of evidence and the strength of recommendations, 9 presented the level of evidence, and 1 presented the strength of recommendations. Thirteen studies considered both TCM evidence and modern medical evidence sources, 3 included only evidence from ancient literature, 1 focused only on post-marketing safety evaluation of Chinese medicine, and 1 focused only on real-world studies of TCM. ConclusionCurrently, there are numerous criteria for TCM related quality of evidence and strength of recommendations, and some are developed only for TCM characteristic evidence. Most researchers of TCM guidelines expect to fully value the significance of ancient literature and expert experience or opinion in guideline development.
Evidence synthesis serves as a bridge between clinical practice and the best available evidence. Evidence synthesis based on high-quality randomized controlled trials is generally considered the highest level of evidence, but its external validity is limited. In some scenarios, the inclusion of non-randomized intervention studies (NRSI) in evidence synthesis may further supplement or even replace randomized controlled trial evidence, such as assessing intervention effectiveness and rare events in a broader population to provide more information for health care decision-making. With the rapid development of real-world data and the improvement of statistical analysis methods, real-world evidence, as an important source of evidence for NRSI, has accelerated the development of high-quality NRSI. However, there are numerous challenges in integrating evidence from randomized and non-randomized intervention studies due to selection and confounding biases caused by the lack of randomization. Based on previous studies, this paper systematically examines the current status of integrated randomized and non-randomized intervention studies, including integration premise, timing, methods, and result interpretation, in order to provide references for researchers and policy-makers to correctly use non-randomized research evidence and further promote optimal evidence generation and clinical practice translation.
直接证据来自直接比较我们关注的干预措施用于我们关注的患者人群，并测量患者重要结局的研究。间接证据可由以下4种方式之一产生。第一，患者可能与我们关注的患者不同（适用性一词常用于这类间接性）。第二，所检验的干预措施可能与我们关注的干预措施不同。有关患者和干预措施间接性的决策取决于对生物或社会因素差异是否大到可能使效应尺度出现预期的较大差异的考虑。第三，结果可能有别于最初设定的结局指标——如替代结果本身不重要，但测量之是基于替代结果的变化反映患者重要结局变化这一假设。第四类间接性在概念上与前三类不同，发生于临床医生必须在未经直接比较的两种干预措施间做出选择时。这种情况下比较治疗方案需要特定的统计方法，并根据患者人群、联合干预措施、结局测量指标及备选干预措施试验方法的差异程度，将证据级别降低1或2级。
The methodology of conducting systematic review of prognostic studies has received a great deal of interest in recent years. Using GRADE for systematic review of prognostic studies, five aspects should be considered:risk of bias, indirectness, inconsistency, imprecision and publication bias. The methods of using GRADE system in systematic review of prognostic studies are similar to systematic review of interventional studies, meanwhile, there are differences. Not only the uniqueness of prognostic study but also the repeating downgrade should be taken into consideration in the GRADE process. Applying GRADE to systematic review of prognostic studies would be widely accepted along with the methodology development and quality improvement of systematic review of prognostic studies.
GRADE建议通过检查95%可信区间（CI）为决定不精确性的最佳方法。在指南实际运用中，如果CI的上、下限值代表了真实效应，而临床实际情况与之不符时，必须降低证据质量级别（即对效应估计值的把握度）。除外当效应值很大且可信区间提示效应稳健，而总样本量不大且事件数很少的情况，其他应考虑因不精确性而降低证据质量级别。作此决定时，可计算有足够检验效能的单个试验所需的病例数（定义为“最优信息样本量”，即optimal information size，OIS）。对连续型变量，我们建议用类似方法，首先考虑可信区间上、下限值，再计算OIS。系统评价（SR）所需方法略有不同。如果95%CI不包括相对危险度（RR）为1，且总事件发生数或病例数超过OIS标准，则精确性良好。如果95%CI包括了明显获益或危害（我们建议以RR值lt;0.75或gt;1.25作粗标准），即使达到OIS要求，因不精确性而降低证据质量级别较恰当。