SEPIA

SEPIA evaluates prioritization algorithms for people living with HIV by simulating contact networks and transmission histories to quantify their accuracy in predicting future HIV transmissions.

Key Features:

• Simulation-Based Framework: Creates realistic simulated scenarios that model individuals' transmission risks over time using contact networks and transmission histories.
• Comparison of Algorithms: Compares multiple prioritization methods, including phylogenetic approaches such as ProACT and distance-based approaches such as growth of HIV-TRACE transmission clusters.
• Effectiveness Metrics: Implements multiple metrics derived from simulated contact networks and transmission histories to evaluate algorithm performance.
• Generalizability: Adapts to various simulation conditions and datasets, including datasets produced by FAVITES.
• Quantitative Analysis: Computes the Kendall Tau-b correlation coefficient between a user's ordering of individuals and an optimal ordering based on actual transmission data.

Scientific Applications:

• Benchmarking prioritization algorithms: Enables quantitative benchmarking of HIV transmission risk prioritization methods, including phylogenetic and distance-based approaches.
• Informing public-health interventions: Quantifies algorithm accuracy to support targeted interventions for people living with HIV aimed at reducing transmission rates.

Methodology:

Simulate contact networks and transmission histories, evaluate algorithms by comparing users' individual orderings to optimal orderings based on actual numbers of transmissions over a specified period, and compute the Kendall Tau-b correlation coefficient as the measure of accuracy.