Characterizing the impacts of disease on behavior across scales: Policy, perception, and potential for infection

BackgroundThe COVID-19 pandemic highlighted the importance of incorporating human behavior into infectious disease models. Yet to do so requires an ability to predict how individuals will respond to novel outbreaks and government policies, which remains challenging. Key questions that limit the integration of dynamic behavior into disease models are: To what degree is individual behavior change driven by policy, objective information on disease, or subjective risk perception? Can objective disease data (more easily measured) approximate subjective risk perception? Is mitigation behavior influenced more by policy or disease information at broad or fine spatial scales? MethodsTo examine the determinants of pandemic mitigation behavior, we leverage US survey data on the number of non-household contacts (a measure of social distancing behavior) and concern about COVID-19 infection (a measure of perceived risk), alongside public data on COVID-19 cases (i.e., measured risk) and mitigation policies. Using a county-level spatiotemporal regression model focused on Sept. 2020 through Jan. 2021, we evaluate whether social distancing is driven by policy, perceived risk, or measured risk, while controlling for differences in demographics and environment. By including predictors at multiple spatial scales, we assess whether individuals use broad or fine-scale information sources to guide their behavior. We then use transmission dynamics models to demonstrate how disease outcomes differ when mitigation behavior is driven by information sources with different spatial scales and accuracy. ResultsWe find that perceived risk and measured risk are both meaningfully predictive of changes in mitigation behavior. State-level variables are more predictive of changes in social distancing compared to county-level conditions, especially for policy covariates. However, conditions in socially adjacent counties are better predictors of behavior than conditions in spatially adjacent counties. In our transmission model, objective and subjective risk yield similar epidemic dynamics, though broader spatial scales of information explain more behavioral variation. ConclusionsThese results indicate that the US population modified their social distancing behavior during the COVID-19 pandemic in response to case incidence and policy at broader spatial scales, potentially reflecting the lack of localized surveillance or mandates. Going forward, it is reasonable for US disease models to assume rational human responses to disease incidence at larger spatial scales.