Handling missing data in a study of Covid-19 vaccine effectiveness on risk reduction of severe outcomes

closed

Missing data pose a potential source of bias in statistical analyses of surveillance data and there is a growing need to support disease experts to address missing data issues. Therefore, understanding the different mechanisms behind the missing data and the ways to address them is crucial. The objectives of the present request are to use a case study on Covid-19 vaccine effectiveness on risk reduction of severe outcomes to provide: • An overview of the mechanism behind the missing data; • Recommendations for reporting missing data; • A review of the statistical methods for handling missing data. The above-mentioned study aimed at assessing the risk reduction in hospitalization and death among fully vaccinated individuals compared to unvaccinated individuals after adjusting by sex, age group, underlying medical conditions and reporting country. However, both the outcomes of interest (hospitalization and death), the vaccination status and or/dates of vaccination and the confounding variables used in the regression model had many missing values. For achieving the objectives, the contractor should develop a technical report containing an overview of what are the missing data and the mechanism behind them, a list of statistical methods for handling missing data according to different epidemiological analyses, an analysis plan for reporting and handling missing data applied to this study, statistical scripts (format R) covering the analysis performed to impute missing data, the outputs of statistical analysis performed. Moreover, one online training session is expected to be provided by the contractor. estimated total amount: 60,000 EUR