Surveillance of zoonotic, wildlife diseases is critically important from public-health, economic and conservation perspectives. However, estimating disease prevalence is challenging for multi-tissue diseases, where multiple tissues can be infected, and a patient can be in the one of several possible states of infection depending on the types of tissues affected. Furthermore, infection of multiple tissues may be inter-dependent. Additional complications in multi-tissue disease surveillance arise from the lack of a biologically meaningful definition of infection, an absence of clear guidance regarding which or how many tissues are to be sampled for testing, and imperfect detection of the pathogen when the infection status is not independent among affected tissues.
We developed a general multi-state occupancy model which accounts for imperfect detection, and provides unbiased estimates of disease state-specific prevalence parameters even when the infection process is interdependent among tissues being affected The general multistate modeling framework developed here is broadly applicable to the surveillance of pathogens that infect multiple tissues and where the infection status or detection of pathogen in one tissue may depend on infection status or detection of pathogen in other tissue(s) within a host. We believe that the modeling framework developed here would be applicable to many wildlife, livestock and human diseases where multiple tissues within a host would be affect.