Group Testing-Based Robust Algorithm for Diagnosis of COVID-19.

Area of Science:

• Epidemiology
• Medical Diagnostics
• Computational Biology

Background:

• The rapid spread of COVID-19 necessitates prompt and accurate diagnostic methods.
• Existing diagnostic approaches may face challenges with large-scale screening and identifying low prevalence infections.
• Group testing (GT) offers a potential strategy for efficient screening of large populations.

Purpose of the Study:

• To develop and evaluate a robust algorithm for COVID-19 diagnosis using group testing.
• To enhance the accuracy of detecting infected individuals while minimizing false positives and negatives.
• To assess the algorithm's performance under various noise levels and infection prevalence rates.

Main Methods:

• A robust algorithm (RA) based on maximum a posteriori (MAP) probability was developed.
• The RA utilizes iterative detection and belief propagation by exchanging marginal probabilities.
• The proposed method was tested on both noiseless and noisy group testing schemes.

Main Results:

• The robust algorithm demonstrates resilience against noise in group testing schemes.
• The method successfully identified infected samples across different COVID-19 incidence rates.
• Performance evaluation confirmed the algorithm's effectiveness in accurate case detection.

Conclusions:

• The proposed robust algorithm offers an effective solution for accurate COVID-19 diagnosis via group testing.
• This approach can aid in early detection and control of infectious disease outbreaks.
• The algorithm's robustness to noise makes it suitable for real-world public health surveillance.