Smart Pooling: AI-powered COVID-19 testing

Abstract

Massive molecular testing for COVID-19 has been pointed as fundamental to moderate the spread of the disease. Pooling methods can enhance testing efficiency, but they are viable only at very low prevalences of the disease. We propose Smart Pooling, a machine learning method that uses sociodemographic data from patients to increase the efficiency of pooled molecular testing for COVID-19 by arranging samples into all-negative pools. We show efficiency gains of 42% with respect to individual testing at disease prevalence of up to 25%, a regime in which two-step pooling offers marginal efficiency gains. Additionally, we calculate the possible efficiency gains of one- and two-dimensional two-step pooling strategies and present the optimal strategies for disease prevalences up to 25%. We discuss practical limitations to conduct pooling in the laboratory.

RETRACTED ARTICLE: Modeling the progression of COVID-19 deaths using Kalman Filter and AutoML

“…The study trained a machine learning model on samples from more than 8,000 patients tested for SARS-Cov-2 from April to July in Bogot, Colombia (Escobar et al. 2020 ). Ribeiro et al use the autoregressive integrated moving average (ARIMA), cubist regression (CUBIST), random forest (RF), ridge regression (RIDGE), support vector regression (SVR), and stacking ensemble learning are evaluated in the task of time-series forecasting with one, three, and six days ahead the COVID-19 cumulative confirmed cases in ten Brazilian states with a high daily incidence (Ribeiro et al.…”

Sample pooling: burden or solution?

“…Similarly, a dynamic pool testing with an adaptive design moving between pooling strategies or selecting optimal pool size could be used. Similar to smart pooling [ 20 ], one can incorporate (dynamic) pre-test likelihood to increase efficiency and reduce false-negative rates. Other alternatives—such as the prevalence spiralling method in conjunction with non-hierarchical matrix-based pooling strategies—can also be used, where high-risk individuals are clustered on one side of the array.…”

“…There are several modifications to these classical pool-testing strategies to further improve efficiency and minimize dilution effect and misclassification of the results. Escobar et al [ 20 ] employed machine learning, using clinical or demographic data, to rearrange samples into pools based on their probability of testing positive for COVID-19. Samples with high probabilities are tested individually, and low-probability samples undergo the S2 strategy.…”

Test Groups, Not Individuals: A Review of the Pooling Approaches for SARS-CoV-2 Diagnosis

“…This type of grouping can greatly improve the efficiency of any pooling scheme. Similarly, Escobar et al [ 18 ] recently developed a machine learning method that uses clinical and sociodemographic data from patients to order samples according to the predicted probability of yielding a positive result in the test, thus increasing the efficiency of pooled molecular testing. In Figure 6 , we show a graphic comparison of pooling (upper panel) vs informed pooling (lower panel).…”