ABSTRACT
Background: Mortality statistics are fundamental to understand the magnitude of the COVID-19 pandemic. Due to limitation of real-time data availability, researchers had used mathematical models to estimate excess mortality globally during COVID-19 pandemic. As they demonstrated variations in scope, assumptions, estimations, and magnitude of the pandemic, and hence raised a controversy all over the world. This paper aims to review the mathematical models and their estimates of mortality due to COVID-19 in the Indian context. Methods: The PRISMA and SWiM guidelines were followed to the best possible extent. A two-step search strategy was used to identify studies that estimated excess deaths from January 2020 to December 2021 on Medline, Google Scholar, MedRxiv and BioRxiv available until 0100 h, 16 May 2022 (IST). We selected 13 studies based on a predefined criteria and extracted data on a standardised, pre-piloted form by two investigators, independently. Any discordance was resolved through consensus with a senior investigator. Estimated excess mortality was analysed using statistical software and depicted using appropriate graphs. Results: Significant variations in scope, population, data sources, time period, and modelling strategies existed across studies along with a high risk of bias. Most of the models were based on Poisson regression. Predicted excess mortality by various models ranged from 1.1 to 9.5 million. Conclusion: The review presents a summary of all the estimates of excess deaths and is important to understand the different strategies used for estimation, and it highlights the importance of data availability, assumptions, and estimates.