Search for: All records

Abstract Traditional health surveillance methods play a critical role in public health safety but are limited by the data collection speed, coverage, and resource requirements. Wastewater‐based epidemiology (WBE) has emerged as a cost‐effective and rapid tool for detecting infectious diseases through sewage analysis of disease biomarkers. Recent advances in big data analytics have enhanced public health monitoring by enabling predictive modeling and early risk detection. This paper explores the application of machine learning (ML) in WBE data analytics, with a focus on infectious disease surveillance and forecasting. We highlight the advantages of ML‐driven WBE prediction models, including their ability to process multimodal data, predict disease trends, and evaluate policy impacts through scenario simulations. We also examine challenges such as data quality, model interpretability, and integration with existing public health infrastructure. The integration of ML WBE data analytics enables rapid health data collection, analysis, and interpretation that are not feasible in current surveillance approaches. By leveraging ML and WBE, decision makers can reduce cognitive biases and enhance data‐driven responses to public health threats. As global health risks evolve, the synergy between WBE, ML, and data‐driven decision‐making holds significant potential for improving public health outcomes. 

Rapid and accurate detection of the pathogens, such as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) for COVID-19, is critical for mitigating the COVID-19 pandemic. Current state-of-the-art pathogen tests for COVID-19 diagnosis are done in a liquid medium and take 10–30 min for rapid antigen tests and hours to days for polymerase chain reaction (PCR) tests. Herein we report novel accurate pathogen sensors, a new test method, and machine-learning algorithms for a breathalyzer platform for fast detection of SARS-CoV-2 virion particles in the aerosol in 30 s. The pathogen sensors are based on a functionalized molecularly-imprinted polymer, with the template molecules being the receptor binding domain spike proteins for different variants of SARS-CoV-2. Sensors are tested in the air and exposed for 10 s to the aerosols of various types of pathogens, including wild-type, D614G, alpha, delta, and omicron variant SARS-CoV-2, BSA (Bovine serum albumin), Middle East respiratory syndrome–related coronavirus (MERS-CoV), influenza, and wastewater samples from local sewage. Our low-cost, fast-responsive pathogen sensors yield accuracy above 99% with a limit-of-detection (LOD) better than 1 copy/μL for detecting the SARS-CoV-2 virus from the aerosol. The machine-learning algorithm supporting these sensors can accurately detect the pathogens, thereby enabling a new and unique breathalyzer platform for rapid COVID-19 tests with unprecedented speeds.