Our paper published in the Journal of Clinical Epidemiology clearly outlines why doing this is not just inaccurate but actually wrong because it introduces bias. This is regularly done and advocated by Cochrane [1]. I think this erroneous practice should now stop - see link to the paper [2]
https://authors.elsevier.com/a/1YCRZ3BcJPyirp
[1] Higgins JPT, Altman DG, Gøtzsche PC, J€uni P, Moher D, Oxman AD, et al. The Cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928.
[2] Stone J, Gurunathan U, Glass K, Munn Z, Tugwell P, Doi SAR. Stratification by quality induced selection bias in a meta-analysis of clinical trials. J Clin Epidemiol. 2018 Nov 17. pii: S0895-4356(18)30744-3.
Yes, Sir. High-dimension data analysis by DIFFEOMORPHISM [by the team under my supervision] has actually proven that :
we must abandon stratification of meta-analyses by quality
Hi Suhail Doi,
I would also not advocate for the stratification of meta-analysis by study quality.
However, if the objective is to attempt to quantify bias, through a sensitivity analysis, stratification seems acceptable but this may better be reported as additional analysis but not the final results.
Regards,
PS: Your book on clinical epidemiology is a good read.
Thanks Godfrey. Even a sensitivity analysis by quality seems questionable based on our results in the paper. So the Cochrane recommendations do not seem to be valid and we should attempt to bias adjust using methods that include all studies.
Thanks for reading my book :-)
Regards
Suhail
I checked it but could not see any relevance to the question asked - please be more specific
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis