Epidemic management and control through risk-dependent individual contact interventions

Area of Science:

• Epidemiology
• Computational epidemiology
• Public health informatics

Background:

• Traditional testing, contact tracing, and isolation (TTI) is difficult to scale due to manual contact tracing.
• Existing exposure notification apps offer limited binary exposure information.

Purpose of the Study:

• To demonstrate a scalable improvement to TTI and exposure notification apps using network data assimilation (DA).
• To provide users with continuously assessed individual risks of exposure and infection for targeted interventions.

Main Methods:

• Utilized data assimilation (DA) on a contact network, integrating diverse health data with mobile device proximity data.
• Simulated the early COVID-19 epidemic in New York City to establish proof-of-concept.

Main Results:

• Network DA identified up to twice as many infections as traditional contact tracing using the same data.
• Even with a small user fraction, network DA showed improved infection identification.
• With 75% adoption and compliance, network DA reduced deaths by up to 75% compared to TTI.

Conclusions:

• Network DA offers a scalable enhancement to TTI and exposure notification apps.
• It enables precise, effective epidemic control with minimized economic disruption.
• Network DA can be integrated into existing app infrastructure to significantly boost capabilities.