Quest for Optimal Regression Models in SARS-CoV-2 Wastewater Based Epidemiology

Area of Science:

• Environmental science
• Epidemiology
• Public health

Background:

• Wastewater-based epidemiology (WBE) is a valuable tool for monitoring infectious disease trends.
• SARS-CoV-2 (the virus causing COVID-19) surveillance in wastewater complements individual testing data.

Purpose of the Study:

• To investigate the correlation between SARS-CoV-2 wastewater signals and COVID-19 incidence.
• To evaluate regression models for predicting viral incidence using wastewater data.
• To identify optimal data pre-processing and modeling strategies for WBE.

Main Methods:

• Collected wastewater and COVID-19 incidence data from four Austrian treatment plants over five months.
• Applied eight regression models with varying data inputs and pre-processing techniques.
• Analyzed the impact of population normalization, smoothing, and time-lag on model performance.

Main Results:

• Population-based normalization and smoothing significantly influenced regression model fitness.
• A time lag of 2-7 days was observed between wastewater signals and reported COVID-19 cases.
• Multivariate modeling incorporating time-lag improved prediction accuracy.

Conclusions:

• Wastewater-based epidemiology provides reliable SARS-CoV-2 surveillance data.
• Effective data pre-processing and multivariate modeling are key for accurate WBE predictions.
• The choice of regression model structure is less critical than data handling and time-lag incorporation.