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 sampled transmission trees attributed an average of 60.8% (standard deviation of 2.2%) of 438 cases to zoonotic spillover, 28.5% (standard deviation of 0.9%) of cases to within-locality 439 human-to-human transmission, and 10.7% (standard deviation of 2.1%) of cases to between-440 locality human-to-human transmission. . The highest success was seen for pairs of epidemiologically-linked cases whose dates of 475 symptom onset were between 7 and 25 days apart ( Fig 7B) . Although it is generally believed that 476 the generation interval for human-to-human transmission of monkeypox is between 7 and 23 477 . CC-BY 4.0 International license is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the author/funder.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 days [43,53], several case pairs that occurred more than 23 days apart were epidemiologically 478 linked through contact-tracing. It is possible that these links, which were often missed in the 479 sampled transmission trees, are not true instances of human-to-human transmission. Cases that 480 occurred in different localities were also less likely to be linked in a sampled transmission tree, 481 though even for these inter-locality pairs, the district-level model tended to perform better than 482 the other three models (S6 Fig The copyright holder for this preprint (which was not peer-reviewed) is the author/funder.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 Comparison of the transmission tree generated using only contact-tracing data with the 500 trees created using the district-level and locality-level models highlights how much our 501 perception of the transmission dynamics depends on assumptions about spatial spread (Fig 8) . 502