Modeling Global Influenza Risk and Pandemic Severity

Introduction: Seasonal and pandemic influenza are major threats to public health. However, as the 2009 H1N1 pandemic has shown, there can be substantial uncertainty in the outcome of a pandemic, especially in the early stages. In addition, seasonal influenza can sometimes cause greater excess mortality than pandemic influenza. Methods: We developed a peer-reviewed, stochastic, metapopulation epidemiologic model to simulate the global spread of influenza. The model includes physical modeling of population movement, age-based contact and disease transmission parameters, and mitigation measures. We also developed a stochastic model to define a suite of realistic influenza scenarios, based on the spatial distribution of animal reservoirs and evolution of influenza viruses. Published data for seasonal and pandemic influenza historical events were used to validate the models, and we conducted sensitivity analyses on key model parameters. Additionally, the effects of pharmaceutical and non-pharmaceutical mitigation measures were investigated. Results: Simulations of historical influenza events, including pandemics of the 20th and 21st centuries, generally showed good agreement with published metrics such as incidence rates and case fatality ratios. The suite of stochastic scenarios ranged in intensity from mild to severe. For the sensitivity analyses, event parameters such as the basic reproduction number (R₀), start date, start location, and mitigation measures significantly affected the overall severity of simulated influenza events. Discussion and Conclusion: A metapopulation epidemiologic model can be used effectively to simulate global influenza spread. Such a model can be used to better understand historical influenza events and to plan for future events.