Font Size: 

The SARS-CoV-2 pandemic emphasized the importance of mass testing for correct data collection and disease control. This study explores the challenges of optimizing testing, in particular with RT-qPCR and its alternatives. We introduce a population-level strategy that uses predictive mechanisms to assess individual contagion risk, considering factors related to the determinants of health. Using the “Actualidades 2021” survey, which sampled 2003 adults, we set classification models, including logistic regression, Random Forest, Gradient Boosting, and XGBoost. With a prevalence of 0.26 in the sample, we adjust the model to explain the outcome of whether the respondent had COVID-19 or not. The model shows sensitivity and specificity values of 0.79 and 0.76, respectively. Through Monte Carlo simulations, we evaluate the economic and epidemiological impacts of various testing strategies such as pooling, retesting and mixing technologies of RT-qPCR, Antigen and RT-LAMP.  In the talk we will discuss how these classification systems could help with the Costa Rican health public policies. The study is available at [1].

SARS-CoV-2 mass testing, Classification models, Determinants of health, Health public policies