T. Schneider, O. Dunbar, J.-L. Wu, L. Böttcher, D. Burov, A. Garbuno-Iñigo, G. Wagner, S. Pei, C. Daraio, R. Ferrari, J. Shaman
PLOS Computational Biology 18 (6), e1010171
Testing, contact tracing, and isolation (TTI) is an epidemic management and control approach that is difficult to implement at scale. During the ongoing COVID-19 pandemic, exposure notification apps have been developed to scale manual contact tracing to larger groups of users. The apps use proximity data from mobile devices to automate finding and notifying direct contacts of an infection source. They suffer from lack of trust as they provide users only with rare and binary alerts.
Here we present network data assimilation (DA) as a new digital approach to epidemic management and control.
Network DA uses the same data as exposure notification apps but does much more effectively and provides frequently updated individual risk assessments to users. Network DA is based on automated learning about individuals’ risk of infectiousness from crowdsourced health data and proximity data. The data are aggregated with models of disease transmission and progression to produce statistical assessments of users’ risks.
In simulations of the early COVID-19 epidemic in New York City (NYC), network DA identifies up to a factor 2 more infections than contact tracing when harnessing the same diagnostic test data.
Targeting contact interventions with network DA reduces deaths by up to a factor 4 relative to TTI, provided compliance reaches around 75%.
In an extensive simulation study of the COVID-19 epidemic in NYC, we show that network DA with regular diagnostic testing could have achieved epidemic control with fewer than half the deaths that actually occurred during NYC’s lockdown, all while typically having only 5–10% of the population in isolation at any given time.
Network DA can be implemented by expanding the backend of existing exposure notification apps, thus greatly enhancing their capabilities. Implemented at scale, it has the potential to precisely and effectively control the ongoing or future epidemics while minimizing economic disruption.