Recreation, Rainfall, and Waterborne Disease: How Social, Environmental, and Climatic Factors Influence Waterborne Disease Risk in Areas Impacted by Combined Sewer Overflows

Abstract

Background: Rainfall is a major driver of waterborne disease. One mechanism by which rainfall can introduce pathogens into environmental waterways is by causing combined sewer overflows (CSOs). Exposure to contaminated surface water may occur during recreational activities that result in accidental ingestion of enteric waterborne pathogens. These pathogens can cause acute gastrointestinal illness (AGI), which is particularly harmful for certain vulnerable groups. Due to projected increases in precipitation in certain areas of the United States, it is important to better understand these relationships and identify populations that may be impacted. Objectives: This dissertation aims to explore the relationships between rainfall, CSOs, and waterborne disease in Philadelphia and in Pennsylvania and to determine how environmental, social, and climatic factors impact these relationships. To do so, this dissertation will 1) identify demographic and economic characteristics of populations that are more likely to be affected by CSO-impacted waters in Philadelphia, 2) quantify the risk of AGI due to recreational exposure to CSO-impacted and non-impacted waters in Philadelphia, and 3) assess the relationships between precipitation and potentially waterborne reportable diseases in Pennsylvania. Methods: To accomplish the first aim, survey data were collected from adults recreating at or near CSO-impacted sites in Philadelphia. Surveys measured participant demographics, recreation behavior, and travel behaviors. Travel measures reported in surveys were used to develop network buffers around both CSO-impacted and non-impacted sites where recreational exposures may occur in Philadelphia. Modified Poisson regression models compared populations living within the reported travel distances to CSO-impacted sites to those living within the same distances to non-impacted sites. To accomplish aim 2, the concentrations of five pathogens were determined by quantifying the concentration of the HF183 marker for human Bacteroides in water samples collected at three sites in Philadelphia during CSO-impacted and non-impacted conditions. Observational data that measured recreational exposures to these waterbodies were collected along with water samples. These data were used to develop quantitative microbial risk assessment (QMRA) models that measured the risk of illness due to these five pathogens during CSO-impacted and non-impacted conditions. Finally, for aim 3, mixed-effects negative binomial regression models were used to measure associations between precipitation, temperature, and the monthly incidence of campylobacteriosis, salmonellosis, cryptosporidiosis, and giardiasis reported per patient zip code in Pennsylvania from January 2014-December 2018. Additional models assessed the impact of seasonality, explored lagged relationships, and determined whether these relationships differed in areas that did or did not contain a combined sewer outfall. Results: Survey data collected for aim 1 found that participants walk an average of 1.4km to get to recreation sites. Populations that reside near CSO-impacted waterways were more likely to be minority and low-income than those that reside near non-impacted water bodies in Philadelphia, presenting a potential environmental justice issue. Results of QMRA models found that several recreational exposure scenarios resulted in a mean risk of illness above the United States Environmental Protection Agency’s acceptable threshold of 30 illnesses per 1,000 exposed. For several exposure scenarios, the risk was still above this threshold during non-impacted conditions. Finally, a 1cm increase in rainfall per month was associated with a 1% increase in the incidence of campylobacteriosis and salmonellosis (IRR: 1.01, 95% CI: 1.00-1.01, for both) and every additional day with at least 2.54mm of rainfall was associated with 1% increase in the incidence of salmonellosis (IRR: 1.01, 95% CI: 1.00, 1.03) and a 3% increase in cryptosporidiosis (IRR: 1.02, 95% CI: 1.01, 1.05). Associations were also found between the daily maximum temperature per month and the incidence of campylobacteriosis, salmonellosis, and cryptosporidiosis. These relationships were strongest during summer months. No differences were observed when comparing results among areas that did or did not contain a combined sewer outfall. Conclusions: Findings from this dissertation suggest that recreation in CSO-impacted waters in Philadelphia presents a health risk. In addition, minority and low-income communities may be more impacted by this potential health risk as they are more likely to reside near potential recreation sites that are impacted by CSOs in Philadelphia. Finally, relationships between precipitation and waterborne disease incidence in Pennsylvania demonstrate how projected increases in precipitation events due to global climate change may increase waterborne disease incidence in Pennsylvania, although the drivers of this relationship are unclear and likely diverse.