Essays on Behavioral Epidemic Modeling: Theoretical, Empirical, and Methodological Contributions

Abstract

Understanding the interplay between infectious disease transmission and human behavior is essential for developing epidemic models that can reliably inform public health policy. Behavioral epidemic models capture the dynamic feedback between human behavior and disease spread. Models that incorporate this relationship endogenously have better long-term forecasting accuracy and can anticipate shifts in epidemic waves. However, many existing models are theoretical, rely on simplistic behavioral assumptions, and lack empirical validation. Strengthening the integration of human behavior into epidemic models is crucial for improving forecasting accuracy and designing policies that are adaptive to behavioral change in future pandemics. This dissertation advances the science and methods of behavioral epidemic modeling through three essays. The first essay addresses challenges in parameter estimation, showing that early-pandemic data often produces biased estimates, and that data spanning at least one full epidemic wave are necessary for reliable inference. The second essay develops a behavioral epidemic model incorporating response to perceived risk, policy shifts, adherence fatigue, and societal learning. Tested against data from eight countries, the model replicates trends in deaths, mobility, pandemic fatigue, and policy changes, offering a comprehensive depiction of the feedback loops between behavior and disease spread. The third essay empirically tests theoretical assumptions about societal response to perceived risk in epidemics using established psychological theories. It finds that behavioral sensitivity to epidemiological indicators varies with cumulative deaths, vaccination coverage, deviations from shifting reference points, and indicators of pandemic fatigue or shifting social norms. Collectively, these essays highlight the value of data-driven, feedback-rich models that incorporate a broad range of human behavioral dynamics rooted in social and psychological theory, and are tested against diverse data, over extended periods, and multiple regions. These contributions enhance our capacity to accurately forecast disease trajectories and design policies that remain effective as public behavior evolves over the course of an epidemic.