Exploring the Interplay Between Disease Avoidance Behavior and the Dynamics of Measles

Abstract

Disease avoidance is characterized by a set of behaviors thought to have evolved for the purpose of reducing the likelihood of encountering pathogens. Though many studies in the existing literature have examined disease avoidance behavior in relation to outbreaks of infectious disease, not much scholarly work has been done with regard to measles, a highly contagious respiratory infection that may result in fatal complications. Though eliminated from the United States in 2000, measles remains present in the country because international travelers sometimes import the virus from places where it is still endemic and subsequently infect unvaccinated individuals. This was exemplified by the multi-state outbreak of measles that took place between December 2014 and April 2015, which was linked back to Disneyland Resort Theme Parks in California. In the midst of this nationwide outbreak, a case of measles was confirmed on Princeton’s own campus.

This provided a prime opportunity to explore how disease avoidance behavior might change when students are faced with a highly contagious infection, as well as how their behaviors might also change the dynamics of that infection. The former was gauged in a survey distributed to a sample of the Princeton undergraduate population. The latter was explored via an individual-based discrete-time model designed to simulate the spread of a measles-like infection among a population of students, in which certain parameters were manipulated to incorporate survey responses. Our main hypotheses predicted that (1) degree of disease avoidance and disease-associated anxiety would be lower in students that reside in non-central locations on campus or in dormitories that comprise many single-occupancy rooms, and (2) the majority of students would predict themselves to exhibit at least a moderate degree of disease avoidance behavior in the event of a measles outbreak on campus, but that this would not significantly influence the dynamics of an infectious disease as highly transmissible as measles.

The results indicate that degree of disease avoidance and various measures of disease-associated anxiety tend to be lower among students that reside non-centrally or in locations with more singles, as shown by distribution of the relevant responses per residential location (Figures 2-6). Moreover, the majority of students responded with moderate or high avoidance. When disease avoidance was incorporated into simulations of an outbreak on campus, unexpectedly significant differences from a control condition were found in the timings of peak of infection when minimum distance for transmission (p = 0.002) and probability of transmission (p = 0.009) were manipulated, though the overall time course of infection remained fairly consistent with or without avoidance.

This suggests that disease avoidance behavior is neither a generalized nor a static response, but can vary based on contextual factors, such as residential location on a college campus. Furthermore, while this type of behavioral response does not appear to have the potential to stop a highly contagious infection from spreading altogether, it is able to slow the rate at which the infection spreads through a susceptible population. Future research hopes to address the major limitations of the model, namely by allowing for more realistic representation of daily students’ activities, accounting for the complexities of individual variation, and improving the design so that disease avoidance can be dynamic over the course of an infection.