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

Monique R. Ambrose; Adam J. Kucharski; Pierre Formenty; Jean-Jacques Muyembe-Tamfum; Anne W. Rimoin; James O. Lloyd-Smith

The number of human cases that become symptomatic on each day in each locality caused by 695 zoonotic spillover is assumed to follow a Poisson distribution with mean λ z . For simplicity and 696 because reservoir disease dynamics are rarely well characterized, we assume the Poisson process 697 is homogenous through time and across localities, but this assumption could be modified for a 698 system where more information is available about the reservoir dynamics (e.g., The factor that describes the amount of transmission that occurs between localities v and 723 w (H(v,w)) could reflect Euclidean distance, travel time, inclusion in different spatial zones, or 724 any other available measurement. To accommodate the imperfect spatial information available 725 for many zoonotic surveillance systems, this study focused on developing methods for the 726 situation when only a locality name and an aggregated spatial zone (such as district or country) is 727 reported for cases, rather than an exact position. We assume that inter-locality transmission 728 occurs only among localities within the same broader contact zone (Fig 1A) . Because 729 transmission will be greater within a locality than between localities, a proportion σ of secondary 730 cases are assumed to occur in the same locality as the source case and a proportion (1-σ) of 731 secondary cases are assumed to occur amongst the outside localities that are within the same 732 broader contact zone as the source case. This outside transmission is assumed to be divided 733 equally among all localities within the index case's broader contact zone: 734 where Z v indicates the broader contact zone of locality v and v is the total number of localities 736 in the broader contact zone of locality v. For a given locality v, the sum of H(v,w) across all w 737 equals one. To observe the effect of assuming different broader contact zones, the monkeypox 738 case study was repeated under four different assumptions about the spatial scale of human-to-739 human transmission: locality, district, region, and country-level. 740