Introduction to WBE case estimation: A practical toolset for public health practitioners.
Journal:
The Science of the total environment
Published Date:
Apr 28, 2025
Abstract
Public health practitioners can use wastewater data to grasp disease dynamics, including incidence, prevalence, and potential disease trajectory. The expertise required to analyze and interpret wastewater data exceed those of most entry-level epidemiologists. We aimed to predict cases from wastewater and detail a customizable, low-resource toolkit for local public health jurisdictions. Using literature-accessible data (viral shedding rate and human fecal production rate), municipality data (flow rate), and wastewater-derived viral concentrations, we evaluated four approaches to predict COVID-19 infections in a U.S. university. The probabilistic gradient boosted decision tree model demonstrated most accurate results (mean absolute error (MAE) = 0.09 %-0.25 %); however, the simpler deterministic calculation was only slightly less accurate (MAE = 1.2 %). The probabilistic method produced the poorest results (MAE = 3.6 %). Jurisdictions should weigh benefits of each model, including data availability, technical familiarity, and relevance. Priorities are community-dependent, thus each model should be tailored to meet a community's specific needs.
