From wastewater to epidemiological insights: A systematic review of modeling strategies for infectious disease surveillance

Wastewater-based epidemiology (WBE) has emerged as a promising complementary tool in infectious diseases surveillance systems, offering real-time insights into the disease dynamics across various spatial coverage. By leveraging wastewater data, a wide range of epidemiological metrics can be estimated and predicted effectively. To date, diverse modeling approaches have been employed and evolved utilizing wastewater surveillance data, including compartmental models, regression-based models, and more flexible machine learning (ML) and deep learning (DL) techniques. In addition, other statistical approaches have been proposed to incorporate complex biological processes such as viral shedding dynamics and decay patterns. While many models have demonstrated promising performance, several critical challenges remain for robust and practical implementation. These challenges range from model-related limitations to inherent issues within wastewater data itself, as well as broader considerations throughout the analytical pipeline, including selection and availability of clinical outcomes and explanatory variables, temporal alignment, data preprocessing strategies, evaluation of model performance and result interpretability. Enhancing the transferability of these models across diverse epidemiological and geographical context remains a key concern. This paper provides a comprehensive review of existing modeling efforts utilizing wastewater surveillance data for the estimation and prediction of various infectious disease indicators. We systematically classify and discuss the modeling approaches, critically evaluate the limitation, and highlight key methodological considerations for future development. Our aim is to offer a well-organized reference framework to support the development of a robust and generalizable system using wastewater data for epidemiological surveillance.