Letter to the Editor: “Measles outbreak in the Philippines: epidemiological and clinical characteristics of hospitalized children, 2016–2019”

As readers of The Lancet Regional Health – Western Pacific, along with the fact that measles outbreaks have been warned of perfect storm of conditions for within 2022 by UNICEF and WHO, we pay a significant interest to the content of the paper entitled “Measles outbreak in the Philippines: Epidemiological and clinical characteristics of hospitalized children, 2016–2019” published earlier within the same journal.1World Health Organization UNICEF and WHO warn of perfect storm of conditions for measles outbreaks, affecting children.https://www.who.int/news/item/27-04-2022-unicef-and-who-warn-of%2d%2dperfect-storm%2d%2dof-conditions-for-measles-outbreaks%2d%2daffecting-childrenDate accessed: December 2, 2022Google Scholar,2Domai F.M. Agrupis K.A. Han S.M. et al.Measles outbreak in the Philippines: Epidemiological and clinical characteristics of hospitalized children, 2016-2019.Lancet Reg Health West Pacific. 2022; 19: 100334Summary Full Text Full Text PDF PubMed Scopus (2) Google Scholar However, upon reproducing the results, we suspect that the authors has made several computational errors. First, in Table 1, the authors did not state the referenced group within two characteristics, vitamin A supplementation and clinical information. This prevents readers from understanding the related context of the odds ratios (ORs), while also obstructs us from re-evaluating the calculation of the crude ORs within these sub-groups.Table 1Re-calculated odds ratios (ORs) for the association between socio-demographic and clinical characteristics and deaths. The newly calculated and original ORs are considered to be significantly different if and only if the difference is greater than or equal to 0.05, and are marked in bold.VariableReported number and percentages in the original articleCalculated percentages (if different) by the authors of this letterReported crude ORS within the original articleRe-calculated crude ORs by the authors of this letterOR (95% CI)P valueOR (95% CI)P valueAge group (months) <31.07 (0.29–3.90)0.9190.87 (0.21–3.66)0.926 3–51.82 (1.13–2.93)0.0131.82 (1.13–2.93)0.016 6–81.29 (0.83–2.01)0.2561.29 (0.82–2.02)0.261 9–111.14 (0.67–1.94)0.6221.13 (0.66–1.93)0.648 12–241.72 (1.12–2.66)0.0141.72 (1.12–1.66)0.015 >24RefRefSex MaleRefRef Female0.96 (0.71–1.29)0.7650.95 (0.71–1.29)0.761Region of residence In NCRRefRef Outside NCR1.55 (1.04–2.31)0.0321.53 (1.02-2.29)0.046Admission timing Non-epidemicRefRef Epidemic3.52 (1.22–10.20)0.0204.09 (1.30–12.88)0.003 Vaccine statusVaccinated (≥1 doses)RefRefNon-vaccinated1.75 (1.05–2.93)0.0321.80 (1.07–3.03)0.019Duration between fever onset and admission (days) 0 – 3d48 (239)48 (2.3)RefRef 4 – 6d1.44 (1.01–2.05)0.0441.44 (1.01–2.06)0.04 7–14d2.45 (1.58–3.78)<0.0012.44 (1.57–3.78)<0.001 >14d1.81 (0.35–9.53)0.4821.24 (0.17–9.24)0.759Duration between rash onset and admission (days) 0 –3dRefRef 4 –6d1.88 (1.25–2.81)0.0021.85 (1.24–2.78)0.005 7–14d3.84 (2.04–7.23)<0.0013.70 (1.94–7.05)<0.001 >14d1.03 (0.06–17.19)0.9860.00 (0.00-inf)0.629 Open table in a new tab Apart from the two characteristics, using the package epitools v0.5–10.1 available within R statistical software v4.1.3, we were able to re-calculate 15 remaining ORs using the unconditional maximum likelihood estimation (Wald) method, with the p-value calculated using the mid-p method.3R Core Team R: A language an environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria2022https://www.R-project.org/Google Scholar,4Aragon T.J. Epitools: epidemiology tools.https://CRAN.R-project.org/package=epitoolsDate: 2020Google Scholar Among them, 6 (40%) are significantly different from the original results (more information can be found in the highlighted details in the attached table below). Besides, there is also one typographical error in the calculation of the percentage of children that have an interval of 0–3 days between fever onset and hospital admission. In this letter, we do not investigate the calculation of the adjusted ORs since the detailed information of the patients involved in the research are not publicly available. From the provided information, we encourage the authors, reviewers and editors to revisit and/or further elaborate on the methods for the calculation and conclusion of any OR within the paper, and make changes wherever applicable. Yours sincerely. H.A.N: Conceptualisation, Formal Analysis, Supervision, Writing – Original Draft Preparation; N.T.H.P: Formal Analysis, Validation, Writing – Review & Editing; P.H.P: Supervision, Validation, Writing – Review & Editing; M.H.T: Formal Analysis, Writing – Review & Editing; H.N.V: Formal Analysis, Writing – Review & Editing. The information analysed within this letter are available in the GitHub repository, https://github.com/hoanganhngo610/recalculate-ORs-measles-Philippines-LRHWP. The authors declare that they have no conflict of interest regarding the publication of this letter to the editor. Funding: No specific grant from funding agencies in the public, commercial or not-for-profit sectors supported the publication of this letter to the editor. Author's reply – Measles outbreak in the Philippines: epidemiological and clinical characteristics of hospitalized children, 2016–2019We thank Ngo and colleagues for raising potential computational errors with our publication. Full-Text PDF Open AccessMeasles outbreak in the Philippines: epidemiological and clinical characteristics of hospitalized children, 2016-2019The Philippines remains at risk of future measles epidemics. Routine immunization needs to be strengthened and earlier timing of MCV1 requires further evaluation to reduce measles incidence and mortality. Full-Text PDF Open Access