Linking Lyme disease ecology and epidemiology: reservoir host identity, not richness, determines tick infection and human disease in California

Abstract

Understanding the community ecology of vector-borne and zoonotic diseases, and how it may shift transmission risk as it responds to environmental change, has become a central focus in disease ecology. Yet, it has been challenging to link the ecology of disease with reported human incidence. Here, we bridge the gap between local-scale community ecology and large-scale disease epidemiology, drawing from a priori knowledge of tick-pathogen-host ecology to model spatially-explicit Lyme disease (LD) risk, and human Lyme disease incidence (LDI) in California. We first use a species distribution modeling approach to model disease risk with variables capturing climate, vegetation, and ecology of key reservoir host species, and host species richness. We then use our modeled disease risk to predict human disease incidence at the zip code level across California. Our results suggest the ecology of key reservoir hosts—particularly dusky-footed woodrats—is central to disease risk posed by ticks, but that host community richness is not strongly associated with tick infection. Predicted disease risk, which is most strongly influenced by the ecology of dusky-footed woodrats, in turn is a strong predictor of human LDI. This relationship holds in the Wildland-Urban Interface, but not in open access public lands, and is stronger in northern California than in the state as a whole. This suggests peridomestic exposure to infected ticks may be more important to LD epidemiology in California than recreational exposure, and underlines the importance of the community ecology of LD in determining human transmission risk throughout this LD endemic region of far western North America. More targeted tick and pathogen surveillance, coupled with studies of human and tick behavior could improve understanding of key risk factors and inform public health interventions. Moreover, longitudinal surveillance data could further improve forecasts of disease risk in response to global environmental change.