Risk-aware Temporal Cascade Reconstruction to Detect Asymptomatic Cases : For the CDC MInD Healthcare Network

This paper studies the problem of detecting asymptomatic cases in a temporal contact network in which multiple outbreaks have occurred. For many infections, asymptomatic cases present a major obstacle to obtaining a precise understanding of infection-spread. We show that the key to detecting asymptomatic cases well, is taking into account both individual risk as well as the likelihood of disease-flow along edges. Most related research has ignored the interplay between these dual aspects influencing disease-spread. We take both aspects into account by formulating the asymptomatic case detection problem as a Directed Prize-Collecting Steiner Tree (DIRECTED PCST) problem. We present an approximation-preserving reduction from this problem to the Directed Steiner Tree problem and use this reduction to obtain scalable algorithms for the DIRECTED PCST problem. Using these algorithms, we solve instances with more than 1.5M edges obtained from both synthetic and actual fine-grained hospital data. On synthetic data, we demonstrate that our detection methods significantly outperform various baselines (with a gain of $3.6 \times$). As an application of our methods, we use a measure of exposure to detected asymptomatic Clostridioides difficile (C. diff) infection (CDI) cases as an additional feature for the important task of predicting symptomatic CDI cases. In this application, our method outperforms all baselines, including those that don’t use asymptomatic CDI cases as a feature and those that use other methods for detecting asymptomatic CDI cases. We also demonstrate that the solutions returned by our approach are clinically meaningful by presenting a case study.