Correspondence on “Electrocardiographic findings and prognostic values in patients hospitalised with COVID-19 in the World Heart Federation Global Study” by Pinto-Filho et al

Abstract

The manuscript by Martins Pinto-Filho et al., a multicentre study on the prognostic value of electrocardiographic findings, was evaluated. However, certain critical issues were highlighted. First, the authors didn't verify the randomness of missing data before implementing the Multiple Imputation by Chained Equations (MICE), which assumes randomness. Second, there was no use of external or internal validation despite the multicentre nature of the study. The study also lacks essential diagnostic measures, like pseudo-R2 statistics and the Hosmer-Lemeshow test. The time-dependent nature of ECG findings and the lack of consistent treatment strategy across different centers raises concerns. Furthermore, the study does not clarify how the imbalanced sample size influenced the conclusion. Lastly, the paper does not provide essential outputs and parameters like Receiver Operating Characteristic (ROC) and Harrell's C-index, which are important to avoid overfitting and ensure global applicability.