This is the authors' abstract. We don't add key points for this paper.
The COVID-19 pandemic underscored the importance of mass testing to mitigate the spread of the virus. This study presents mass-testing strategies developed through machine learning models, which predict the risk of COVID-19 contagion based on health determinants. Using data from the 2021 “Actualidades” survey in Costa Rica, we trained models to classify individuals according to risk of contagion. After theorizing four possible strategies, we evaluated these using Monte Carlo simulations, analyzing the distribution functions for the number of tests, positive cases detected, tests per person, and total costs. In addition, we introduced the metrics efficiency and stock capacity, to assess the performance of different strategies. Our classifier achieved area under the Receiver Operating Characteristic Curve (AUC-ROC) of 80.48 % and area under the Precision-Recall curve (AUC-PR) of 58.98 %, considering the prevalence of the disease of 26.91 %. We evaluated four mass testing strategies: (1) antigen testing without low-risk attention, (2) antigen testing with pooling, (3) antigen testing with multiple testing, and (4) a hybrid strategy with targeted testing according to its risk. In particular, using traditional tests like reverse transcription polymerase chain reaction (RT-PCR) tests, antigen tests, and alternative options like Reverse Transcription Loop-mediated Isothermal Amplification (RT-LAMP) tests. Our main finding is that a machine learning model to triage the population combined with a targeted mass testing strategy is a cost-effective option, offering valuable insights for scalable and adaptable pandemic response.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Pasquier et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: