Interim seasonal influenza vaccine effectiveness estimates as proxy for final estimates: analysis of systematically identified matched pairs of interim/final estimates from test-negative design studies in outpatient settings from 2010/11 to 2018/19

Abstract

ABSTRACT Objectives Limited time for seasonal influenza vaccine development means that the World Health Organization has to consider interim (early) rather than final vaccine effectiveness (VE) estimates in deciding influenza vaccine composition. We assessed agreement between interim and final VE estimates, and factors that may determine a substantial difference (≥10%) between point estimates. Methods This was a mixed methods study. We systematically searched, identified, and matched interim/final VE studies of test-negative design (TND) type in outpatient settings after the 2009/10 influenza pandemic. The chi-square statistic (χ2) was used to assess the statistical significance of the difference between paired interim/final VE estimates. We calculated the difference between point estimates and used multivariable logistic regression to assess factors that may determine a substantial difference. Results We identified 68 interim/final VE pairs. There was no statistically significant difference between almost all compared pairs. An inconsistent statistical model for interim/final VE estimation and interim VE estimation before the epidemic peak increased the odds of having a substantial difference between estimates. Conclusion : Interim influenza VE appears to be sufficient for vaccine composition decision-making. Consistency in interim/final VE estimation, and interim VE estimation during/after epidemic peak may increase agreement between the VE estimates.