Adaptive risk-based pooling in public health screening

Pooled testing is commonly used in public health screening for classifying subjects in a large population as positive or negative for an infectious or genetic disease. Pooling is especially useful when screening for low-prevalence diseases under limited resources. Although pooled testing is used in various contexts (e.g., screening donated blood or for sexually transmitted diseases), a lack of understanding of how an optimal pooling scheme should be designed to maximize classification accuracy under a budget constraint hampers screening efforts. We propose and study an adaptive risk–based pooling scheme that considers important test and population level characteristics often over looked in the literature (e.g., dilution of pooling and heterogeneous subjects). We characterize important structural properties of optimal subject assignment policies (i.e., assignment of subjects, with different risk, to pools) and provide key insights. Our case study, on chlamydia screening, demonstrates the effectiveness of the proposed pooling scheme, with the expected number of false classifications reduced substantially over policies proposed in the literature.