Statistical approaches for estimating vaccine effectiveness in a real world setting 

Tarylee Reddy, Unit Director

Dept. of Biostatistics, South African Medical Research Council

Francie van Zijl Drive, Parowvallei, Cape Town; South Africa

Location: ZOOM

   https://cesnet.zoom.us/j/92685451604?pwd=TUJzZTV0V1JXaXNLTlNVaWMvTS9Idz09

Meeting ID: 926 8545 1604

Passcode: 358199

Date:   16 March 2023

Time: 13:00 CET

Abstract: 

Real-world effectiveness (RWE) studies are generally difficult to design, analyse and interpret due to their susceptibility to biases and methodological complexities.  This is largely due to the lack of randomization in their design. In this presentation, the advantages and disadvantages of different designs will be discussed. A detailed case study of the Sisonke trial will be presented.

The Sisonke trial was a  single-arm, open-label, phase 3B  implementation study which was conducted in  health-care workers across all regions of South Africa. This real-world effectiveness study was designed to  assess the effectiveness of a single dose of the Ad26.COV2.S vaccine (Johnson & Johnson) against COVID-19 related hospitalization and death. The primary effectiveness analysis utilized  the computationally intensive approach of simulating a clinical trial using a matched cohort design. This study involved three sets of analyses: two  independent health insurance schemes and the Western Cape healthcare worker database. The statistical approaches for combining estimates, while accounting for the variability in observations and the matching process will be discussed. We also discuss current challenges in the evaluation of vaccine effectiveness including: assessing durability of vaccines, effectiveness of booster doses, the importance of sensitivity analysis and the availability of data. 

Keywords: 

Vaccine, effectiveness, COVID-19, matching, cohort, Poisson

References: