LETTER TO EDITOR Response to: A statistical commentary on ‘Does vitamin D supplementation reduces COVID-19 severity? A systematic review’ K. Shah, V.P. Varna, U. Sharma and D. Mavalankar From the Department of Public Health, Indian Institute of Public Health Gandhinagar, Gandhinagar 382042, Gujarat, India Address correspondence to Dr K. Shah, Department of Public Health, Indian Institute of Public Health Gandhinagar, Opp. Air Force Head Quarters, Nr. Lekawada Bus Stop, Gandhinagar-Chiloda Road, Gandhinagar 382042, Gujarat, India. email: kshah@iiphg.org Dear Editor, We thank Dr Zhou for sharing his view on our article studying role of vitamin D supplementation in reducing COVID-19 severity. 1 Herewith, we are addressing his comments in point-by- point manner. Meta-analysis of meta-analysis is not an uncommon form of research especially when multiple primary outcomes are con- sidered. 2,3 Regarding this particular study, we wanted to assess role of vitamin D supplementation on mortality, Intensive care unit (ICU) stay and mechanical ventilation requirement in COVID patients. Upon preliminary analysis, it was observed that not all the eligible reviews provide details of the outcomes. Similarly, not all the primary study data were available for con- ducting meta-analysis of individual studies. In that scenario, outcome assessment using already published meta-analysis provided good opportunity for quantitative analysis. Regarding statistical analysis, duplication and overlapping we have been extremely transparent about the methodologies being opted for pooled data analysis. Explicit analysis of over- lapping primary studies was conducted and was provided as supplementary material. Though we agree that there is an over- lap among the primary studies included in meta-analysis, we considered pooled data analysis as an important exercise due to multiple reasons—(i) degree of overlapping might be misleading due to the possibility that a primary study might have been included in more than one review for different outcome assess- ment. So, there might be no overlap between the outcomes however primary studies might be overlapping. (ii) As per updated guideline of Cochrane, 4 if the purpose is to present and describe the current body of systematic review evidence on a topic, it may be appropriate to include the results of all relevant systematic reviews, regardless of topic overlap. It also states that in case when authors are not able to avoid double-counting outcome data for methodological or logistical reasons may still opt to include all relevant Cochrane and non-Cochrane system- atic reviews in the overview and provide a documentation of ex- tent of the primary study overlapping (which have been provided by us as overlapping matrix). (iii) Bias and heterogen- eity are inherent limitations of the reviews conducted on COVID-19 due to obvious reasons such as lack of robust randomized controlled trials and studies with smaller sample size. These have been duly acknowledged in the study, and we highlighted the need for more robust primary data from randomized controlled trial to substantiate the findings further. In summary, we want to emphasize again that the study has included most critical aspects of methodologies such as PRISMA guidelines, AMSTAR and GRADE assessment for risk of bias and quality, I 2 for heterogeneity assessment, publication bias through Begg’s and Egger’s test, overlapping matrix presenta- tion, and CCA assessment and hence contribute significantly in current body of evidence. Author contributions Komal Hasmukh Shah (Conceptualization [equal], Formal ana- lysis [equal], Investigation [equal], Methodology [equal], Writing—original draft [equal], Writing—review & editing [equal]), V.P. Varna (Data curation [equal], Formal analysis [equal], Project administration [equal], Writing—review & Submitted: 20 March 2023 V C The Author(s) 2023. Published by Oxford University Press on behalf of the Association of Physicians. All rights reserved. For permissions, please email: journals.permissions@oup.com 1 QJM: An International Journal of Medicine, 2023, 1–2 https://doi.org/10.1093/qjmed/hcad046 Advance Access Publication Date: 27 March 2023 Letter to Editor Downloaded from https://academic.oup.com/qjmed/advance-article/doi/10.1093/qjmed/hcad046/7087105 by guest on 06 April 2023