Proposed study design for estimating COVID-19 vaccine effectiveness in post-pandemic Taiwan

Post-licensure observational studies are used to assess vaccine effectiveness (VE) when clinical randomized controlled trials are not feasible. Although vaccination engenders various effects at the individual and population levels, few studies have specifically examined the types of effects estimated in the VE study. The test-negative design (TND) is more commonly used to evaluate direct vaccine performance against the influenza virus, rotavirus, and Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). It is preferably used within a standard sentinel surveillance platform, and certain constraints may affect its validity. After the World Health Organization declared that coronavirus disease 2019 (COVID-19) is no longer a public health emergency, developing robust methodologies for evaluating VE is critical. Accordingly, this study aims to (1) clarify the types of vaccine effects typically captured by current observational VE studies, and (2) propose a scientifically rigorous and operationally feasible study design for estimating seasonal COVID-19 VE against severe outcomes after the pandemic. After reviewing various VE methodological issues, including vaccine effect, study design, outcomes, and exposures of interest, key covariates, and Taiwan's post-pandemic context. We propose a refined population-based retrospective cohort design. This design enables estimating direct VE against severe COVID-19 outcomes in the post-pandemic era by utilizing Taiwan's integrated national databases and incorporating validated health-seeking behavior and healthcare access markers, evidence-based underlying conditions, and advanced statistical models. The proposed design provides a scientifically valid and feasible alternative to the TND for direct COVID-19 VE assessment in the absence of a sentinel surveillance platform. It is also applicable to other respiratory-pathogen-related vaccines, such as the seasonal influenza vaccine, and to countries with similar settings.