Improved entropy-CRITIC population model based on temporal and spatial variability: Construction and application in wastewater epidemiology

Numerous factors contribute to the uncertainty inherent in conducting wastewater-based epidemiology (WBE), with shifting populations exerting a significant influence. However, traditional single- and multi-parameter population models suffer from certain limitations. This study employs an evaluation model framework to construct a model (EC model) based on data characteristics. Weight coefficients derived from 16 cities across seven regions of China are aggregated into a national model. In contrast to alternative models, the EC model exhibits a robust correlation (r2 = 0.98) with census population data, suggesting a potentially more precise depiction of population dynamics. The low variability (RSD = 9.73 %) indicates effective constraint of anomalous parameter fluctuations, yielding minimal Bias (−1.12 %) and SRMSE (14.75 %), thus ensuring reliable population estimation. The model is applied to estimate the consumption of lifestyle-related compounds and the prevalence of hypertension in China. Northern regions demonstrate higher consumption levels, alongside a significant disparity in hypertension prevalence (26.96 %) compared to the south (16.01 %). Hypertension exhibits positive correlations with lifestyle-related compounds such as alcohol and nicotine (r = 0.52, r = 0.55). Sensitivity analysis reveals that the EC model introduces an uncertainty of 24.48 % in population estimates. Through the incorporation of representative datasets and novel algorithms, this model has the potential to enhance the reliability of outcomes in WBE strategy implementation. © 2024 Elsevier B.V.