Epidemic prediction and control in clustered populations

Area of Science:

• Epidemiology
• Network Science
• Mathematical Biology

Background:

• Epidemic modeling on networks is crucial, especially with significant network clustering.
• Understanding early outbreak dynamics is key to predicting epidemic spread and control needs.

Purpose of the Study:

• To develop pairwise methods for predicting epidemic outcomes and intervention levels based on early outbreak observations.
• To investigate the impact of network clustering on epidemic prediction and control.

Main Methods:

• Development of pairwise modeling techniques for network epidemics.
• Analysis of early outbreak growth and its relationship to epidemic outcomes.
• Comparison of predictions across different levels of network clustering.

Main Results:

• Early epidemic growth accurately predicts the 'leaky' vaccine threshold and peak time, independent of clustering.
• The basic reproductive ratio predicts the random vaccination threshold, also independent of clustering.
• The relationship between other epidemic quantities and outcomes is highly sensitive to the level of network clustering.

Conclusions:

• Pairwise methods offer valuable insights into epidemic prediction and control on clustered networks.
• Network clustering significantly influences the predictability of certain epidemic outcomes and intervention effectiveness.
• Early outbreak data provides crucial information for managing epidemics, but clustering must be considered for accurate predictions.