Vaccination is the most effective method to reduce influenza-associated morbidity and mortality. Accurate annual assessment of the effectiveness of influenza vaccines is very important for (a) understanding the relationship between antigenic match and mismatch and vaccine effectiveness, (b) evaluation of vaccination programs and strategies in terms of individual and population-wide benefits, (c) identifying risk factors for vaccine failure to assist in determining strategies to improve effectiveness in such groups (e.g., higher doses). As influenza vaccination is now recommended in the U.S. for almost everybody above 6 months of age, randomized clinical trials to evaluate influenza vaccine effectiveness (IVE) are no longer ethical, and health authorities have to rely on observational studies that are known to produce biased IVE estimates. The main objectives of the proposed research project are to (a) evaluate and compare existing observational study designs for estimating IVE, and (b) develop new study designs that are expected to result in improved IVE estimates. We will pay special attention to the test-negative-controls (TNC) study design, where patients with an influenza-like-illness (ILI) who test negative for influenza infection serve as controls. This simpe study design is now the most commonly used design in the U.S. and world-wide, though its bias and precision have not yet been fully evaluated. We propose to improve this design by combining it with a smaller cohort study where participants are constantly monitored and are tested for influenza infection once they develop an ILI. We also plan to develop guidelines for determining sample sizes for IVE studies and to explore the accuracy of mid-season (interim) estimates of IVE. To achieve these goals we will develop a detailed agent-based stochastic simulation model to (a) generate outbreaks of influenza infection/illness and cases of non-influenza ILI in a structured population, and (b) use data from these outbreaks to conduct observational IVE studies following specified study designs. Multiple simulations under fixed settings will be conducted to evaluate bias and precision of estimates from different study designs. This is the first project to evaluate IVE study designs from stochastic simulations accounting for various underlying contact processes, influenza transmission dynamics and other real-life factors. Results of the proposed research project will guide investigators conducting IVE studies in selecting the most appropriate study design in terms of reducing bias and improving precision of resulting estimates. This will help produce more robust estimates of vaccine effectiveness, which are essential for developing new and improved vaccines, for making the public aware of the benefits of influenza vaccination, and for targeting sub- populations in which IVE is low. While parameters of our model will be based on data from seasonal influenza outbreaks, we believe that our results will also be useful when designing studies to evaluate the effectiveness of vaccines against possible influenza pandemics. 1
This research project will evaluate existing and new study designs for estimating the effectiveness of influenza vaccines. Accurate annual assessment of the effectiveness of influenza vaccines is essential for (a) understanding the relationship between antigenic match and vaccine effectiveness, (b) evaluating vaccination programs and strategies, and (c) identifying risk factors for vaccine failure to assist in determining strategie to improve effectiveness in such groups. The results of this project will help produce more robust estimates of vaccine effectiveness, which are essential for developing new and improved vaccines and for making the public aware of the benefits of influenza vaccination. 1
