COVID-19 vaccination and all-cause and non-COVID-19 mortality. A revaluation of a study carried out in an Italian Province

The COVID-19 vaccination prevents COVID-19 specific mortality. Well planned population-based studies, however, are necessary to evaluate the overall effectiveness of vaccination programmes. A study carried out in the province of Pescara is used to illustrate the potential biases that may affect such studies. The Pescara study analysed total and non-COVID-19 mortality and the occurrence of Potentially Vaccine-Related Serious Adverse Events (PVR-SAEs) in vaccinated and unvaccinated people, from January 2021, when vaccines became available, to July 2022. The study reported a lower probability of both total and non-COVID-19 death in vaccinated people. However, the authors did not include in the denominator of the unvaccinated cohort the population experience of the vaccinated cohort before vaccination (immortal time bias). Correcting the denominator of the unvaccinated cohort, the crude death rate of vaccinated and unvaccinated persons becomes the same. For the same reason, the unvaccinated non-COVID-19 mortality was overestimated, as was the mortality of people receiving only one or two vaccine doses. Confounding by indication and the healthy vaccinee bias will also be discussed, as well as the bias deriving by not considering the evolution of risk over time.