Estimating Conditional Vaccine Effectiveness

Abstract

Vaccine effectiveness for COVID-19 is typically estimated for different outcomes that often are hierarchical in severity (e.g. any documented infection, symptomatic infection, hospitalization, death) and which may be subsets of each other. Conditional effectiveness for a more severe outcome conditional on a less severe outcome is the protection offered against the severe outcome (e.g. death) among those who have already sustained the less severe outcome (e.g. documented infection). The concept applies also to the protection offered not only by vaccines but also by previous infection. Formulas and a nomogram are provided here for calculating conditional effectiveness. Illustrative examples are presented from recent vaccine effectiveness studies, including situations where effectiveness for different outcomes changed at different pace over time. E(death | documented infection) is the percent decrease in the case fatality rate (CFR) and E(death | infection) is the percent decrease in the infection fatality rate (IFR). One can thus calculate the evolving CFR and IFR in a population as more people become vaccinated and/or infected. Conditional effectiveness depends on many factors including availability and use of better treatment options, lethality of predominant viral strains, background immunity, increased exposure of disproportionately low risk people (due to fewer precautions or/and more active epidemic waves) and more aggressive testing of asymptomatic people. Conditional effectiveness should be used for better personalized communication of the benefits of vaccination, considering also IFR and epidemic activity. Patients, physicians, and policy makers would benefit to know this information for shared decision making.

Note:
Funding: None.

Declaration of Interests: None.

Keywords: COVD-18, effectiveness, conditional effectiveness, vaccine, death, infection

JEL Classification: I, I12, I18, I19