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

While this approach works well when the total number of surveilled localities is known 748 (see Fig 3A) , localities often only appear in the dataset if they have reported cases; as a result we 749 may not know the total number of localities under surveillance. Ignoring localities with zero 750 cases can lead to biased parameter estimates (see Fig 3B) . We explored several alternative 751 approaches to account for these silent localities; the preferred approach rescales the likelihood 752 function to reflect that localities with zero cases are not included in the data. Several Table) . Simulation parameters were chosen to approximate the monkeypox dataset, with σ set at 818 0.75, R ranging from 0.2 to 0.6, and λ z ranging from 0.0001 to 0.1. Unless otherwise specified, 819 simulations were performed assuming the district-level model. Details on the models used for 820 sensitivity analyses that use the exact spatial location of cases or allow highly structured and 821 non-homogenous spillover patterns are provided in S1 Text. 822 The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/677021 doi: bioRxiv preprint S3 The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/677021 doi: bioRxiv preprint S1 Data. Case records. For all individuals included in the analyses, records the case 1178 identification number, the locality identification number, the day of surveillance when disease 1179 onset occurred (the first day of fever when known, otherwise the first day of the rash), the names 1180 of the district and region where the case occurred, and masked GPS coordinates of the locality. 1181