Transmission of infectious disease occurs across multiple scales with dynamical processes at both the individual and population levels. For example, heterogeneity in movement behavior and social interactions affect contact structure and transmission rates in populations. Likewise, individual-level disease dynamics within hosts determine population-level transmission rates. Using examples from carnivore rabies and influenza A, we show how understanding these heterogeneities can guide management planning and surveillance design for wildlife diseases.
The objectives of our first study were to examine the effects of variation in host movement on the spatial spread of raccoon rabies and vaccination effectiveness. To address these objectives we developed a spatially-explicit model of carnivore rabies that incorporates host movement at the individual level. Then, we examined how different levels of heterogeneity in movement affect spatial spread rates and the effectiveness of different vaccination strategies for preventing spatial spread. The objectives of our second study were to examine the potential for individual-level immunity measures within hosts to predict disease risk, inform surveillance design, and provide novel insights about the population-level disease dynamics. To address this objective we developed a hierarchical Bayesian model of serological data and applied the framework to influenza A surveillance data in wild birds.
Results/Conclusions
In our first example, when raccoon movements were more variable, faster rates of spatial spread of raccoon rabies occurred relative to movement distributions with lower variation and similar average magnitudes of movement. Also, modest increases in movement variation relative to data-based estimates led to large differences in the effectiveness of a vaccination barrier against rabies. We conclude that measures of variation in animal movement are critical for planning vaccination strategies - efforts should be made to measure individual-level variation appropriately. We also showed how our approach can be developed further to represent individual-level animal movement mechanistically as it relates to underlying habitat heterogeneity.
In our second example, we inferred population-level force of infection (FOI, disease risk) from individual-level serological data. Inference was least biased when variation in antibody titers was low to moderate, and when antibody decay rates were moderate. Although we developed the original model for one-time cross-sectional data, we also showed that serial cross-sectional surveillance data can provide better estimates of FOI. Using data from influenza A in mallards and plague in coyotes we show how collecting additional data to explain individual-level variation can improve estimates of FOI and our ecological understanding of the factors underpinning disease transmission.