Objective measures for sentinel surveillance in network epidemiology

Area of Science:

• Network epidemiology
• Public health surveillance

Background:

• Sentinel surveillance aims to detect emerging disease outbreaks early or reliably by probing specific network nodes.
• The choice between early or reliable detection depends heavily on the specific scenario and public health goals.

Purpose of the Study:

• To investigate and compare three objective measures for quantifying sentinel surveillance performance: time to detection or extinction, time to detection, and frequency of detection.
• To analyze how these measures rank nodes differently under various conditions within network structures.

Main Methods:

• Utilized the susceptible-infectious-recovered (SIR) model on both static and temporal human contact networks.
• Evaluated node performance based on the three distinct objective measures for early and reliable disease detection.

Main Results:

• Demonstrated that the three objective measures can yield substantially different rankings of nodes, depending on the parameter space.
• Observed similar conclusions across both static and temporal network analyses.
• Found no single network structure that consistently predicts surveillance performance across different datasets and SIR parameter values.

Conclusions:

• Sentinel surveillance is a multifaceted problem requiring careful selection of an objective measure tailored to the specific scenario.
• Network structure alone does not reliably predict the effectiveness of different surveillance strategies.