Speakers
Description
Contact tracing (CT) is a targeted intervention that can reduce transmission during infectious disease outbreaks without requiring population-wide restrictions. Most mathematical models evaluate CT using transmission-tree frameworks that assume independent infection events and don’t explicitly account for the underlying contact network. These approaches ignore that traced susceptible contacts may temporarily quarantine and that, in clustered networks, CT may identify cases infected by individuals other than the index case.
We develop a stochastic event-driven epidemic simulation model that represents transmission and interventions directly on contact networks. Infection, detection, isolation, testing, and CT events are scheduled in continuous time and processed through a priority queue, allowing transmission and interventions to interact dynamically on the network. The framework allows flexible generation-time and detection-time distributions without restrictive Markovian assumptions.
Using this model, we compare the effectiveness of isolation, one-step CT, and multi-step CT across unclustered and clustered network structures. Our simulations show that network clustering can enhance the effectiveness of CT at moderate transmission levels by enabling quarantining of traced contacts to interrupt transmission within overlapping neighborhoods. Tracing both susceptible and infected contacts also yields a small but consistent improvement over tracing infected contacts alone.
