Assessing Vaccine Effectiveness in Observational Studies via Nested Trial Emulation

Abstract

Observational data are frequently used to evaluate real-world vaccine effectiveness (VE). For vaccines such as those developed against COVID-19, VE may vary over calendar time because of changes in circulating viral variants and over time since vaccination because of waning immunity. Nested trial emulation (NTE) provides a framework for estimating causal effects from observational data while reducing biases common to standard observational analyses. NTE involves emulating a sequence of hypothetical trials, each initiated from a different calendar date. This manuscript considers a NTE inverse probability weighted estimator of vaccine effectiveness that may vary over calendar time, time since vaccination, or both. The proposed approach characterizes trial-specific VE and uses those estimates to assess changes in vaccine protection across calendar time. As changes in VE estimates across trials may be attributable to variation in covariate distributions across trial-eligible populations, standardization of trial-specific estimates is considered. Statistical testing for evaluating heterogeneity in VE across trials is also considered. The methods are used to estimate vaccine effectiveness against COVID-19 outcomes using observational data on over 110,000 residents of Abruzzo, Italy during 2021.