Eliminating Bias in Epidemic Surveillance and Enhancing Public Health Interventions through Physics-Based Methodological Advances in Epidemic Modeling

ABOUT THE SPEAKER
Dr. Eugenio Valdano

ABSTRACT
Controlling infectious disease epidemics in populations requires two capabilities: accurately estimating epidemic trends—whether an epidemic is growing or subsiding—based on available surveillance data, and enhancing public health interventions, such as immunization campaigns, to drive disease elimination. In this seminar, we will demonstrate that the complex spatial and behavioral patterns inherent to human populations may bias both epidemic assessment and the measurement of intervention effectiveness. Using methods from complex networks coupled with branching processes, we will then address both challenges. Specifically, we will show that spatial structure of contact patterns can introduce significant bias into epidemic trend estimates by misrepresenting transmission dynamics. This bias can lead to underestimations of epidemic spread, possibly mistaking growing epidemics for subsiding ones. Then, we will examine how complex behavioral profiles and imperfect immunization tools may reduce the effectiveness of campaigns designed according to current theories, which advocate for prioritizing individuals at high risk of infection and transmission. We will illustrate that, contrary to these traditional approaches, a more uniform, non-selective distribution can often achieve better population-level outcomes. Together, these studies underscore the critical need for physics-based methodologies that leverage high-resolution data to produce practicable public health insights.