Meta-Analysis of Vaccine Effectiveness Studies
- 35 Downloads
This chapter explores the meta-analysis of vaccine effectiveness studies. These meta-analyses tend to be non-comparative, although a simple statistical test to compare vaccine effectiveness estimates exists, as well as a powerful regression method known as meta-regression. It is demonstrated how these comparative analyses, as well as non-comparative analyses according to, for example, the DerSimonian–Laird method, can be conducted using simple SAS codes. It is shown that analytical confidence intervals for the difference of two vaccine effectiveness estimates do not exist, and that the problem of finding such a confidence interval comes down to finding a confidence interval for the difference of the medians of two lognormal distributions, and that such an interval can be obtained by means of parametric bootstrapping.