A Smartphone Magnetometer-Based Diagnostic Test for Automatic Contact Tracing in Infectious Disease Epidemics

Area of Science:

• Epidemiology
• Mobile Sensing Technology

Background:

• Close human contact is a primary transmission route for infectious diseases.
• Traditional contact tracing methods are labor-intensive and can be inaccurate.
• Smartphone sensors offer potential for automated and scalable contact detection.

Purpose of the Study:

• To evaluate the diagnostic accuracy of smartphone magnetometer readings for detecting close human proximity.
• To assess the feasibility of using magnetometer data as a proxy for contact events in epidemic scenarios.
• To determine optimal parameters for magnetometer-based contact detection algorithms.

Main Methods:

• Analyzing linear correlation of magnetometer readings between co-located smartphones.
• Evaluating diagnostic performance using sensitivity, specificity, likelihood ratios, and diagnostic odds ratios.
• Determining appropriate trace segment sizes and correlation cutoff values for contact detection.

Main Results:

• High linear correlation observed between magnetometers of nearby smartphones.
• Magnetometer-based detection demonstrates significant discriminative and predictive power for contact events.
• Specific ranges for trace segment sizes and correlation cutoffs were identified for accurate testing.

Conclusions:

• Smartphone magnetometer data can serve as an accurate diagnostic test for close human contact.
• This technology offers a promising tool for modernizing contact tracing in public health.
• Optimized parameters enhance the reliability of magnetometer-based epidemic surveillance.