Quantifying Transmission Between Wild and Domestic Populations

Disease modeling at the wildlife–livestock interface aims to predict disease transmission, evaluates spill-over risk and control strategies, and quantify the consequences for animal health and conservation, among other objectives. Different types of approaches have been extensively used in epidemiology to evaluate pathogen transmission among individuals in one or more populations. However, considering wildlife–livestock or wildlife–livestock–human host groups is less common because of the complexity of gathering detailed/quality data on wildlife and livestock (and human) populations simultaneously as well as other methodological challenges. This chapter aims to provide a brief overview of the main modeling approaches available to quantify the multi-host disease transmission at the wildlife–livestock interface, illustrated with specific case studies. We focus on what we have classified into three groups of approaches: (1) correlative approaches; (2) mechanistic approaches; and (3) molecular approaches. All those approaches can be used alone or in combination to study disease transmission at the wildlife–livestock interface across different spatio/temporal scales. The most appropriate method and scales to consider will depend on feasible/available data streams and objectives (e.g., designing surveillance at a national level or proposing protective measures regarding farms). “Correlative approaches” (data-driven) make use of data obtained through observational (genetic data, surveillance) or experimental (sentinel studies, intervention studies) studies for the purpose of estimation (i.e., calculating unknown parameters) or prediction (approximating outcomes for unseen data or future time periods). However, we usually need to combine data into knowledge-driven or mechanistic models to obtain a more holistic understanding of the magnitude and dynamics of a problem (to evaluate the risk of disease introduction and/or spread as well as to quantify the magnitude and economic impact of an epidemic at the wildlife–livestock interface. Finally, molecular approaches are useful for identifying the source of transmission of infections (i.e., contact tracing at the wildlife–livestock interface) through analyzing genetic relationships in a set of samples from different species and also allow inference about the spatial and temporal dynamics of multi-host diseases. Underreporting of disease in livestock may occur and for several reasons, but it is usually a bigger problem in the wildlife side of the interface. In addition, wildlife sources of data are usually skewed to nonrandom samples due to convenience, or only based on detected cases. Wildlife surveillance may be too sparse and limited to particular diseases and hosts, and the accuracy of diagnostic tests may present some concerns since they are often only validated for livestock. All this hinders estimation of disease dynamics in the real population affected. Today, there is still a significant need to increase our knowledge about how some wild populations are distributed and structured (also social structure and movements and interactions with other species) to infer disease spread. Models that explicitly include the spatial structure and contact networks of the population have become increasingly used. However, estimating contacts is challenging and usually only possible at small scales, and factors influencing interactions may vary from one area to another. Nevertheless, we believe that technological advances and the expansion of interdisciplinary teams with expertise in the wildlife, livestock, and human side of the interface will allow better characterization of disease transmission at the interface using these modeling approaches, thus providing better prevention and control of emerging infectious diseases locally and globally.