A risk-varying and perturbed self-controlled case series design for assessing the safety of COVID-19 vaccines in a large health care system

PROJECT SUMMARY/ABSTRACT Despite the success of the unprecedented COVID-19 vaccine rollout in the U.S., vaccine hesitancy is evident, partly due to safety concerns about severe adverse events (SAEs) surrounding the novel technology of mRNA COVID-19 vaccines and reports of blood clots following receipt of the Janssen COVID-19 vaccine. While rigorous safety monitoring may help support COVID-19 vaccination, it is methodologically challenging to thoroughly evaluate the safety of the two-dose mRNA COVID-19 vaccines and the one-dose Janssen COVID- 19 vaccine. Existing approaches can produce false positive and false negative signals when 1) risk windows after vaccination are incorrectly specified, 2) a constant risk of SAEs during the risk window is wrongly assumed, 3) factors that may influence receipt of the second dose of mRNA COVID-19 vaccines are not accounted for, and 4) the nature of the risk of SAEs during potential overlapping risk windows of the first and second doses of mRNA COVID-19 vaccines is not assessed. In response to the FOA, PA-18-873, this proposal addresses the specific objective: "creation/evaluation of statistical methodologies for analyzing data on vaccine safety, including data available from existing data sources such as passive reporting systems or healthcare databases." We propose to develop novel statistical models to properly measure the risk of new COVID-19 vaccines by allowing the risk level to vary during unknown risk windows and using a data-driven approach to define these risk windows. We will also create a new metric for measuring the risk of SAEs considering both the risk level and the length of the risk window, address the potential overlap of risk windows of two doses, and employ a propensity score model approach to account for factors that may influence receipt of the second dose of mRNA COVID-19 vaccines. We will establish these novel approaches to evaluate COVID-19 vaccine safety and will apply them to existing data from members of Kaiser Permanente Southern California, a large, racially, and socio-economically diverse population. Through this research, we will detect SAEs of concern, better inform the public and policymakers about the safety of COVID-19 vaccines, and generate vaccine safety information that may be helpful for clinicians to deliver appropriate care to those at risk.