Because influenza pandemics occur with little warning, vaccine development and distribution take place at a slower timescale than transmission of the emergent strain. Similarly, although seasonal influenza epidemics occur annually, they are also notoriously difficult to predict, and necessitate rapid response to changing circumstances. While vaccination, antivirals and non-pharmaceutical interventions (NPIs) are available to mitigate these challenges, imperfect protection and coverage mean that their direct and indirect protective benefits are conditional on the state of immunity in the population. Therefore, the overall objective of this application is to develop a sustainable, scalable pipeline of analytic, predictive, and visualization tools to translate detailed clinical and cohort data to into timely population-level guidance on vaccination, antiviral use, and NPIs. We will accomplish these goals through the following specific aims:
Aim 1) We will use the extensive clinical and cohort data resources generated by the Michigan Influenza Center to identify and address key questions in influenza prevention and control;
Aim 2) We will integrate these multiple sources of data using statistical and simulation based models of infectious disease transmission. Specifically, we will Aim 2A) use robust models of longitudinal serologic data to characterize response to natural infection and vaccination, and then in Aim 2B) integrate this information into household-based transmission models to understand the impact of these immune responses on susceptibility to influenza infection. Using the predictions of these individual- level models parameterized using longitudinal cohort data, in Aim 2C) we will construct synthetic cohorts representative of the age-specific distribution of immunity in different populations, e.g. the State of Michigan, and use these data to develop targeted population-level strategies for influenza vaccination.
In Aim 2 D) we will then apply the insights of these models to the layered application of antivirals and NPIs in an influenza pandemic using a network-based simulation platform we have developed. All of these models will be designed, implemented and analyzed in collaboration with CDC and other partners to ensure clearly-articulated guidelines for modeling assumptions and inputs (Aim 3). This will be augmented by tools for automated model verification, validation and synthesis which will ensure adherence to these standards and integrate the findings of multiple modeling groups (Aims 4 & 5). All of these tools will be released publicly as open-source software and interactive tools. All of these products will be implemented with the goal of communicating key findings as well as uncertainty in model inputs, structure, and outcomes as clearly as possible to a wide array of scientific and policy-focused stakeholders using state-of-the-art tools for data visualization (Aims 6,7 & 8). The outcome of this project will be the development of a validated, systematic and collaborative modeling approach tailored for rapid evaluation of both pandemic and seasonal influenza mitigation strategies.
The proposed research is highly relevant to public health because of the need for modeling tools to guide more effective, rapid responses to both influenza pandemics and seasonal influenza epidemics. We propose to integrate data and models from the individual to the population level using highly-detailed information from longitudinal cohort and clinical data. These models will evaluate vaccination, antiviral distribution, and non- pharmaceutical intervention strategies to be employed for pandemic mitigation at the population level. This research is relevant to the mission of the CDC Influenza Division in that it will inform efforts to guide vaccination and treatment policies and improve initiatives to prevent and treat influenza infection.
