Dai X, Acosta N, Lu X, et al. A Bayesian framework for modeling COVID-19 case numbers through longitudinal monitoring of SARS-CoV-2 RNA in wastewater. Statistics in Medicine. 2024; 1–17. doi: 10.1002/sim.10009

 

Wastewater-based surveillance (WBS) has become an important tool for research groups and public health agencies investigating and monitoring the COVID-19 pandemic. While there is an emerging body of evidence exploring the possibility of predicting COVID-19 infections from associated viral signals in wastewater samples, there remain significant challenges for statistical modeling. Recorded numbers of viral copies in municipal wastewater can contain measurement errors and missing values due to irregular or sparse sampling of municipal wastewater. This work develops a Bayesian statistical framework to predict daily positive cases from irregular wastewater observations via advanced statistical techniques. The proposed statistical framework can achieve two goals: 1) estimating a smooth and continuous function (of time) from the irregularly observed and noisy SARS-CoV-2 RNA wastewater data; 2) combining the smooth function with recorded COVID-19 cases to showcase the idea of predicting future positive cases from SARS-CoV-2 RNA in wastewater. Most importantly, the developed framework is an integrative statistical framework, in which these two goals are accomplished simultaneously. The results of the simulated data and real data analysis demonstrate that this framework can perform daily predictions of the number of new positive cases in the city of Calgary from leading indicators in wastewater samples and a vaccination covariate one week prior. The proposed framework is flexible as it accounts for other factors such as vaccination rates, weather patterns, and the self-selection bias in the population tested for COVID-19, thus making the framework applicable to other WBS studies.