Assessing Simultaneous Infection with Multiple Pathogens via Group Testing with Imperfect Multiplex Assays

Area of Science:

• Epidemiology
• Biostatistics
• Infectious Disease

Background:

• Pooled testing combines samples to reduce screening costs, especially for low-prevalence infections.
• Multiplex assays enable simultaneous detection of multiple pathogens, enhancing efficiency.
• Imperfect assay sensitivity and specificity can lead to false positives and negatives in pooled testing.

Purpose of the Study:

• Develop a statistical method to estimate co-infection prevalence from imperfect multiplex pooled testing data.
• Determine optimal pool sizes to minimize estimation variance.
• Provide a hypothesis test for infection independence.

Main Methods:

• Expectation-maximization (EM) algorithm for estimating infection probabilities.
• Louis's method for estimating the variance-covariance matrix.
• Simulation studies and real-world data application for validation.

Main Results:

• Accurate estimation of marginal and co-infection prevalence from imperfect pooled data.
• Identification of pool sizes that minimize estimation variance.
• A validated hypothesis test for assessing infection independence.

Conclusions:

• The developed statistical approach effectively handles imperfect multiplex pooled testing data.
• This method optimizes resource allocation in infectious disease surveillance.
• Applicable to various pathogens and pool sizes, including tick-borne diseases.