Variations in Seasonal Patterns of Gastrointestinal Infections within a Watershed

Epidemiologic analysis of waterborne diseases typically considers socio-economic, demographic, and pathogen-specific characteristics. However, hydrological parameters may need to be considered as well. Increases in waterborne diseases have been associated with water quality parameters such as turbidity and river flow. Each watershed has unique characteristics, such as streamflow, which may affect the rates of waterborne diseases in populations residing within the watershed. We hypothesized that watershed attributes predict the rates of waterborne disease in these populations.

This preliminary analysis explores seasonal patterns of waterborne diseases along the Mississippi River and Ohio River watersheds in the U.S. We considered diagnoses of various gastrointestinal infections using elderly hospitalization records from the Center for Medicare and Medicaid Services (CMS) from 1991 to 2004. Cases were aggregated according to diagnosis code, location, and date of admission.

We assessed seasonal patterns of disease rates using an annual harmonic regression controlling for streamflow. We found streamflow to be more strongly correlated with outcome rates in cities along the Ohio River compared to the Mississippi River. Controlling for streamflow did not significantly affect the estimate of peak timing of outcome rate for cities in either watershed. This suggests that there is an interaction between the seasonal pattern of GI illness and river discharge. This is a first attempt at modeling rates of waterborne diseases associated with river characteristics within a watershed. An understanding of the relationship between watershed characteristics to health outcomes will allow for appropriate public health protection to be implemented.