Estimating influenza vaccine effectiveness using the Test-negative Design

Abstract

OBJECTIVES: Vaccination is generally recognised as aneffective interventionagainst influenza. Monitoring of IVE helps confirm if the vaccine is providing adequate protection and identify factors affecting IVE. However, the standard design of randomised controlled trial (RCT) cannot be regularly conductedwhile cohort study or traditional case-control study designs are prone to confounding by health-care seeking behaviour and misclassification bias. METHODS: In test-negative design (TND), subjects tested positive are defined as test-positive cases while those meet with the same inclusion criteria but test negative are test-negative cases. Subjects can be recruited in various settings including primary care clinics, sentinel surveillance networks and hospitalised patients when presenting for medical attention. IVE can be estimated as in traditional case-control studyusing one minus adjusted odds ratio of vaccination among test-positive versus test-negative cases with comparable accuracy to RCT but of more feasible logistics and acceptable cost. For example, TND has been used for monitoring IVE in different age groups and influenza viruses annually in the I-MOVE multicentre case-control study since 2007,involving eight European countries recruiting cases with acute respiratory infection (ARI) or influenza-like illnessthrough general practitioners in an influenza surveillance network and confirming influenza in laboratory. RESULTS: In Hong Kong, IVE was estimated among hospitalised children in 2009-2013 using TND and found an overall IVE estimate of 61.7%, with substantial variability among different years, age groups and influenza viruses. Starting in 2014, IVE study in the community will be conducted annually using TND, in which people presenting with ARIto selected out-patient clinics will be recruited and tested for influenza virus by RT-PCR. CONCLUSION: By using TND, we can monitor IVE in the population and investigate the effects of different factors and matching of vaccine strains with epidemic strains on IVE using a method feasible in the community setting with reasonable cost.