Vaccine Effectiveness Studies

Abstract

Vaccine effectiveness studies are observational studies, and thus at risk of confounder bias. This chapter therefore opens with a discussion of what confounding is and how it can be eliminated. Many vaccine effectiveness studies have a so-called case-referent design. The key idea of these designs is clarified, and the major case-referent designs—including the popular test-negative design—are inspected. The important but difficult to grasp difference between cumulative reference sampling and incidence sampling is explained. Next, adjusted data analysis to control for confounding, either by means of stratification or regression, is discussed. The chapter concludes with considerations on how to represent and select confounders in a regression model.