High-fidelity injection detectability experiments: a tool for evaluating syndromic surveillance systems

Area of Science:

• Public Health
• Epidemiology
• Biostatistics

Background:

• CDC recommends evaluating public health surveillance systems for sensitivity, specificity, and timeliness.
• Characterizing these metrics is crucial for outbreaks of varying sizes, causes, and scopes.
• Existing tools lack comprehensive evaluation capabilities for real-world surveillance data.

Purpose of the Study:

• To develop a novel tool for health departments to estimate outbreak detection.
• To provide a method for calculating expected sensitivity, specificity, and timeliness.
• To enable robust validation of public health surveillance systems.

Main Methods:

• Developed High-Fidelity Injection Detectability Experiments (HiFIDE) tool.
• HiFIDE utilizes semisynthetic injection methods with real outbreak data.
• Replaced geometric injects with data-derived injects to maintain outbreak size-effect relationships.

Main Results:

• Demonstrated HiFIDE's utility in analyzing a waterborne Cryptosporidium outbreak in Washington, DC.
• Successfully estimated the detectability of the outbreak using the HiFIDE tool.
• The tool allows inference of the smallest detectable outbreak size.

Conclusions:

• HiFIDE empowers public health departments to perform CDC-recommended system validations.
• The tool enhances the ability to assess and improve public health surveillance.
• HiFIDE is freely available for non-commercial use.