Advancing evaluation of respiratory virus vaccine effectiveness in real time data

Background Accurate real-world estimates of vaccine effectiveness (VE) are crucial to evaluate the impact of vaccination programmes, promote confidence in vaccination, and support cost-effective use of NHS resources. Current methods for estimating VE in the UK rely on laboratory tests performed in a subset of general practices (GPs) and hospitals. Detailed electronic health record (EHR) data can now be analysed at sufficient scale and frequency to provide rapid estimates of VE and waning VE within epidemic seasons. However, uncertainty remains regarding optimal methods for estimating VE and how best to communicate findings to promote vaccine uptake. Research question How can estimation of respiratory virus VE be optimised to provide accurate and timely estimates that facilitate optimal uptake and impact, overall and in granular population subgroups? Aims Establish optimal methods for estimating VE against influenza, COVID-19, and respiratory syncytial virus (RSV). Use new population-scale EHR data to derive rapid (within-season) VE estimates across granular demographic and clinical subgroups. Develop and apply methods to estimate waning VE during epidemic seasons. Quantify disparities in vaccine coverage and determine whether respiratory vaccines contribute synergistically to protect against disease. Co-produce multifaceted interventions that use insights from aims 1-4 to promote confidence in and uptake of vaccination among underserved communities. Each aim will be addressed by a work package (WP). Data sources This project will use the NHS England OpenSAFELY EHR research platform. This highly secure platform enables analysis of detailed primary care records and linked data (e.g. hospital episodes) for >99% of GPs in England. Analyses will also be done by the UK Health Security Agency using laboratory test data from GPs and hospitals. Methods WP1: We will evaluate three methods for VE estimation (target trial emulation, test-negative case-control, and regression discontinuity), comparing findings across vaccines, seasons, and outcomes of differing severity to establish best practice for different scenarios. Simulation studies will be used to compare methods under controlled conditions. WP2: We will apply optimal methods from WP1 to estimate VE in population subgroups defined by age, ethnicity, socioeconomic status, and clinical factors (e.g. diabetes, immunosuppression). We will extend cost-effectiveness analyses based on these estimates. WP3: We will develop and apply methods to quantify waning VE within epidemic seasons, and how this varies across population subgroups. We will extend cost-effectiveness analyses to accommodate waning protection. WP4: We will assess individual- and GP-level risk factors for reduced vaccine uptake, and adapt VE estimation methods to explore synergistic effects of multiple vaccines. WP5: We will hold workshops with key populations to discuss how insights from WP1-4 can be used to support vaccination uptake. Using person-based approach tools, we will combine evidence, behavioural science theory, and community insight to inform intervention development. Impact and Dissemination We will provide robust, granular estimates of influenza, COVID-19, and RSV VE. We will estimate cost-effectiveness to support vaccine policy decisions, and identify risk factors for low vaccine coverage. We will disseminate findings to policy-makers, and work with the public and practitioners to develop interventions that promote improved and more equitable vaccine coverage.