Good day fellow researchers! I wish to conduct a meta-analysis of the sensitivity and specificity of a diagnostic method and compare them to those of another method. I have already collected, evaluated and included studies in my paper, currently a systematic review. What characteristics need to be comparable among the studies for them to be included in a meta-analysis as well? Sample size, study design, population? And how is "comparable" even defined? Your insight and assistance will be greatly appreciated by this rookie researcher. Thank you all in advance!
Emmanouil M. Xydias, I think that your questions are best answered in their proper context, rather than in relative isolation by point-by-point responses here, through comprehensive resources on systematic review and meta-analysis of diagnostic test accuracy.
These two thematic articles can give proper overview of your meta-analysis goal of interest:
Article Systematic Review and Meta-Analysis of Studies Evaluating Di...
(Part 1)Article Systematic Review and Meta-Analysis of Studies Evaluating Di...
(Part 2)Cochrane also has a good introduction to systematic reviews of diagnostic test accuracy that can be found here: https://training.cochrane.org/diagnostic-test-accuracy-dta-reviews.
Your specific question on "comparability" for meta-analysis pooling is a universal concern across different designs of studies included for systematic review. For guidance, please read this particular part of the Cochrane Handbook: https://training.cochrane.org/handbook/current/chapter-10. If you have free time, I recommend that you also read all preceding chapters. Take note that the Cochrane Handbook deals with systematic reviews of interventional studies, not diagnostic accuracy studies, but the issues on pooling studies remain relevant for both. Specific guidance for diagnostic test accuracy systematic reviews can be found here, but it is yet to be finalized/decisively codified by Cochrane: https://methods.cochrane.org/sdt/handbook-dta-reviews. Nevertheless, the current resources there are already helpful as they are.
The following online course gives a comprehensive and digestible overview of the endeavor (however, it focuses on interventional studies too): https://www.coursera.org/learn/systematic-review.
Good luck!
Aedrian Abrilla thank you for your quick, informative response as well as for the plethora of resources you provided. I greatly appreciate it!
I totally agree with Aedrian Abrilla answer. The sources he has provided you are excellent to perform the best possible analysis.
Of course that including best designed studies (RCTs) will give you the most accurate results in a meta-analysis.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis