The effect of the modifiable areal unit problem (MAUP) on spatial aggregation of COVID-19 wastewater surveillance data.

Large wastewater-based epidemiology (WBE) projects often have wide coverage and multiple sampling sites, necessitating spatial aggregation for data reporting and interpretation. However, the outcome may be impacted by a type of statistical bias called the modifiable areal unit problem (MAUP). In this study, we examined the presence and extent of the MAUP scaling effect on a New York State COVID-19 wastewater surveillance project. Specifically, we investigated three metrics: 1) the difference in wastewater SARS-CoV-2 concentrations between sampling at city-level site (i.e., city's primary wastewater treatment plant influent stream) and at upstream sampling sites; 2) the correlation between WBE data and clinical indicators at the WWTP-level and the more aggregated county-level; and 3) the proportion of population affected by misalignment of COVID-19 community risk levels at different spatial scales. The results showed that the MAUP can have a negative impact on risk perception by masking regions with high wastewater viral load or COVID-19 community risk level. On the other hand, the MAUP improved the correlation between wastewater surveillance and clinical measures by an average of 26.02 %. This is the first study to investigate the MAUP in the context of WBE and may encourage future WBE projects to consider the implications of the MAUP when interpreting and reporting spatial data, ultimately leading to better data representativeness and accuracy.