Quantifying transmission of emerging zoonoses: Using mathematical models to maximize the value of surveillance data

Understanding and quantifying the transmission of zoonotic pathogens is essential for directing public health responses, especially for pathogens capable of transmission between humans. However, determining a pathogens transmission dynamics is complicated by challenges often encountered in zoonotic disease surveillance, including unobserved sources of transmission (both human and zoonotic), limited spatial information, and unknown scope of surveillance. In this work, we present a model-based inference method that addresses these challenges for subcritical zoonotic pathogens using a spatial model with two levels of mixing. After demonstrating the robustness of the method using simulation studies, we apply the new method to a dataset of human monkeypox cases detected during an active surveillance program from 1982-1986 in the Democratic Republic of the Congo (DRC). Our results provide estimates of the reproductive number and spillover rate of monkeypox during this surveillance period and suggest that most human-to-human transmission events occur over distances of 30km or less. Taking advantage of contact-tracing data available for a subset of monkeypox cases, we find that around 80% of contact-traced links could be correctly recovered from transmission trees inferred using only date and location. Our results highlight the importance of identifying the appropriate spatial scale of transmission, and show how even imperfect spatiotemporal data can be incorporated into models to obtain reliable estimates of human-to-human transmission patterns.\n\nAuthor SummarySurveillance datasets are often the only sources of information about the ecology and epidemiology of zoonotic infectious diseases. Methods that can extract as much information as possible from these datasets therefore provide a key advantage for informing our understanding of the disease dynamics and improving our ability to choose the optimal intervention strategy. We developed and tested a likelihood-based inference method based on a mechanistic model of the spillover and human-to-human transmission processes. We first used simulated datasets to explore which information about the disease dynamics of a subcritical zoonotic pathogen could be successfully extracted from a line-list surveillance dataset with non-localized spatial information and unknown geographic coverage. We then applied the method to a dataset of human monkeypox cases detected during an active surveillance program in the Democratic Republic of the Congo between 1982 and 1986 to obtain estimates of the reproductive number, spillover rate, and spatial dispersal of monkeypox in humans.