Department of Hygiene, Social-Preventive Medicine & Medical Statistics, School of Medicine, Faculty of Health Sciences, Aristotle University of Thessaloniki, University Campus, 54124 Thessaloniki, Greece.
Diabetes Research Centre, Leicester General Hospital, University of Leicester, Leicester, UK; National Institute for Health Research Applied Research Collaboration-East Midlands, Leicester Diabetes Centre, Leicester LE5 4PW, UK.
J Clin Epidemiol. 2021 Apr;132:34-45. doi: 10.1016/j.jclinepi.2020.12.004. Epub 2020 Dec 9.
To introduce potential static tabular and graphical techniques for visually presenting overlap between systematic reviews (SRs) included in overviews of systematic reviews (OoSRs).
The graphical approaches described include Venn and Euler diagrams, as well as matrix-based, node-link, and aggregation-based techniques. We used fundamental concepts of mathematics from set and network theory to develop our novel graphical approaches. The graphical displays were created using R.
Overview authors have the flexibility to choose from a variety of visualizations, depending on the characteristics of their study. If the OoSRs include few SRs, a Venn or an Euler diagram can be used. In case of OoSRs with more SRs, Upset plots, heatmaps, and node-link graphs are more appropriate for visualizing overlapping SRs. Stacked bar plots constitute an aggregation-based technique of illustrating overlap. Strengths and limitations of each graphical approach are presented.
The degree of overlap should be explored for the entire study and for specific outcomes of interest. The proposed graphical techniques may assist methodologists and authors in identifying overlap, which in turn may improve validity and transparency in OoSRs. More research is needed to understand which technique would be most useful and easiest to understand.
介绍潜在的静态表格和图形技术,用于直观呈现系统评价综述(OoSR)中包含的系统评价(SR)之间的重叠情况。
描述的图形方法包括 Venn 和 Euler 图,以及基于矩阵、节点链接和聚合的技术。我们使用集合论和网络理论的基本数学概念来开发我们的新图形方法。图形显示是使用 R 创建的。
综述作者可以根据研究的特点灵活选择各种可视化方法。如果 OoSR 包含的 SR 较少,可以使用 Venn 或 Euler 图。对于包含较多 SR 的 OoSR,可以使用 Upset 图、热图和节点链接图更适合可视化重叠的 SR。堆叠条形图构成了一种用于说明重叠的聚合技术。介绍了每种图形方法的优缺点。
应该探索整个研究和特定感兴趣结果的重叠程度。所提出的图形技术可以帮助方法学家和作者识别重叠,从而提高 OoSR 的有效性和透明度。需要进一步研究以了解哪种技术最有用且最容易理解。