Identifying Safety-Vaccine Association for COVID-19 Vaccines from VAERS

- Vaccine Adverse Event Reporting System (VAERS) is a national early warning system to detect possible safety issues in US licensed vaccines. However, the data mining task for VAERS is quite challenging mainly due to the high dimensionality of the data set and complex confounding presented among adverse events and vaccines. In addition, the inconsistent data quality and the rarity of cases and events add more difficulties to the analysis of VAERS data. Tan et. al. (2022+) [1] proposed a 3-step strategy to detect safety associated vaccines by using zero-inflated Poisson regression model with variable selection from VAERS. In this paper, I discuss the safety issue related with COVID-19 vaccines that were reported to VAERS from year 2020 to current with the implementation of a modified version of 3-step strategy.