Exposure to the natural environment is hypothesized to benefit health via three pathways: 1) reduction of harm such as heat, noise, and air pollution; 2) building capacity for health by promoting health-behaviors such as physical activity; and 3) restoration through stress reduction (Markevych et al., 2017). While extensive epidemiologic research demonstrates positive associations of exposure to greenspace (parks and other natural areas in a landscape) and greenness (green vegetation) with health outcomes (Fong, Hart, & James, 2018; Twohig-Bennett & Jones, 2018), the potential salutogenic effects of blue space (aquatic environments such as coasts, lakes, and rivers), remain comparatively underexplored (Gascon, Zijlema, Vert, White, & Nieuwenhuijsen, 2017; M. P. White, Elliott, Gascon, Roberts, & Fleming, 2020). Understanding the unique influence of blue space exposure on health could bolster health-promoting community development strategies.

Researchers increasingly recognize blue space as an important component of the therapeutic landscape (Völker & Kistemann, 2013) and that improving access to such spaces could play a role in preventing poor health outcomes (M. P. White et al., 2020). Since the 1980s, a trend of waterfront redevelopment in urban planning has reclaimed blue space for public access and enjoyment, and visits to such spaces evoke health-promoting experiences, including restoration, stress and anxiety reduction, physical activity and recreation (including land-based activities that occur near water, such as beach walks), and social interaction (de Bell, Graham, Jarvis, & White, 2017; Thomas, 2015; Völker & Kistemann, 2013; Völker, Matros, & Claßen, 2016; M. White et al., 2010). A 2017 review identified 35 epidemiologic studies on blue space and health, with most studies measuring proximity to blue space from individuals’ residential areas using spatial data to estimate availability (e.g., percentage of blue space within a defined area) or accessibility (e.g., distance to blue space) (Gascon et al., 2017). The strength of evidence of associations for blue space varied by health outcome, with “limited” evidence supporting an association with mental health or physical activity and “inadequate” evidence for associations with obesity, cardiovascular disease, or diabetes. Additional evidence has since emerged, including a large population-based Canadian cohort study that found protective associations between living near waterbodies and non-accidental causes of death, including cardiovascular disease, diabetes, stroke, and respiratory-related causes, with no evidence of confounding by socioeconomic status, greenness, or air pollution (Crouse et al., 2018). The nascent body of epidemiologic research on blue space and health remains limited by its predominant focus on coastal exposure; a geographic concentration in Europe and urban areas; and heterogeneity in blue space exposure metrics, which limits comparability across studies (Gascon et al., 2017; Völker et al., 2018). Few health studies have evaluated inland, freshwater blue space or blue space across diverse community types, gaps we aimed to address with the current study.

As part of an epidemiological study on determinants of geographic disparities in type 2 diabetes (T2D), we evaluated associations of residential proximity to inland freshwater blue space and new onset T2D within a geographically diverse range of communities in central and northeast Pennsylvania. Two prior studies analyzed prevalent diabetes as a covariate in evaluations of residence in coastal versus non-coastal regions and other health outcomes (Bergovec, Reiner, Milicic, & Vrazic, 2008; Modesti et al., 2013), but to our knowledge no studies have evaluated blue space and T2D onset. Diabetes is among the leading causes of mortality in the U.S., and it increases risk for other serious health conditions, including coronary heart disease, stroke, hypertension, and kidney disease (Centers for Disease Control and Prevention, 2020). Risk factors implicated in the development of T2D include physical inactivity and psychological stress (Centers for Disease Control and Prevention, 2020; Hackett & Steptoe, 2017), components of two of the primary pathways linking the natural environment and health (Markevych et al., 2017). Though limited, the strongest evidence for health-promoting effects of blue space relates to greater physical activity and improved mental health via restoration and stress reduction (Gascon et al., 2017). We therefore hypothesized that closer residential proximity to blue space would be associated with lower odds of T2D onset. Given the impact of flooding in the region and the potential implications of flooding on health (Fernandez et al., 2015; Tapsell & Tunstall, 2008), we also evaluated residence within the 100-year floodplain in association with T2D onset. Herein, we describe the findings that emerged from this analysis—which led us to explore residential property value data to help interpret the findings—as well as implications for the development of inland waterfront areas and future research on blue space.

This study was conducted by Geisinger-Johns Hopkins, one of four academic research centers in the Diabetes LEAD (Location, Environmental Attributes, and Disparities) Network (http://diabetesleadnetwork.org), a collaboration funded by the Centers for Disease Control and Prevention dedicated to providing scientific evidence to develop targeted community-based interventions and policies to prevent incident T2D and related health outcomes across the United States (Hirsch et al., 2020). The Geisinger Institutional Review Board approved the study and waived informed consent.

In this study of residential proximity to inland freshwater blue space and diabetes, we hypothesized that living closer to blue space would be associated with lower risk of T2D onset. Closer proximity to blue space is associated with more frequent visits (Elliott et al., 2020)—providing nature-based opportunities for recommended behavioral approaches for preventing diabetes (e.g., physical activity, stress reduction) (Centers for Disease Control and Prevention, 2020; Hackett & Steptoe, 2017)—and prior studies suggest a beneficial effect of blue space on health (Crouse et al., 2018; Gascon et al., 2017; M. P. White et al., 2020). Contrary to our hypothesis, we observed higher odds of new onset T2D among individuals living closer to blue space in these central and northeast Pennsylvania communities. Among residents of townships and boroughs, living within 0.25 miles of blue space was associated with 8% and 17% higher odds of T2D, respectively, as compared to living more than 1.25 miles away; among city residents, odds were 38%-39% higher between 0.25 to 0.75 miles from blue space. These associations were consistent when adjusting for Medical Assistance status, an indicator of individual-level socioeconomic status, and CSD, an indicator of community-level disadvantage. We postulate that our unexpected finding is not a direct effect of blue space exposure, but rather explained by geographic patterns of socioeconomic disadvantage and a lack of human and economic capital to transform waterfronts from landscapes of risk into salutogenic spaces conducive to physical activity and restoration.

Although analyses controlled for Medical Assistance and CSD, a post hoc evaluation of residential property values suggested our findings may be partly explained by residual confounding by individual- or neighborhood-level socioeconomic status, known risk factors for T2D (Agardh et al., 2011; Bilal et al., 2018). Residential property values from Lycoming County, located in the heart of the Susquehanna River basin, demonstrated a trend of decreasing mean property values with nearer distance to blue space in townships and boroughs, with the lowest mean property values closest to blue space. The exception to this pattern was observed in cities, where the lowest mean property values were 0.25 to 0.50 miles from blue space, similar to the pattern of T2D onset risk. This pattern of lower property values closer to waterbodies such as the Susquehanna River reflects historic and economic development in the region. Many towns in the study region developed along the Susquehanna River due to its historic importance as a transportation corridor—via a canal system—for raw goods through rugged and isolated terrain, and as a driver of industrial development, providing waterpower to sawmills, iron mills, and coal machinery. In addition to devaluation that may have arisen from flooding, railways and major roads that are co-located along the transportation corridor originally established by the Susquehanna and other rivers may further degrade property values, and could also reduce opportunities for physical activity and stress reduction in blue space through physical barriers and the creation of air and noise pollution. Additionally, economic decline in the region following dual waves of de-industrialization in the past half-century has reshaped communities (Keil & Keil, 2015; MacGaffey, 2013) and their relationship with the river, and left many towns lacking the human and economic capital required to develop waterfronts into desirable places to live. With many towns originally settled in close proximity to the river, newer homes were constructed away from rivers—in the outer areas of towns and on historically agricultural land, as evidenced by the loss of farmland to development in the Northeast states during the second half of the twentieth century (Hellerstein et al., 2002). The legacy of these trends appears to have created a socioeconomic gradient that remains evident in the region’s geography today, highlighting the importance of understanding historic and economic context when evaluating the health impacts of blue space.

Individuals with lower incomes may also be more likely to reside in areas subject to flooding due to price structures that constrain their residential options, further exacerbating the geographic patterns of socioeconomic disadvantage described above. The impact of flooding—a relatively common occurrence for the Susquehanna River and its tributaries—may be most detrimental to economically vulnerable populations, who are less resilient in recovering from floods and lack the financial resources to move or modify their property (Tapsell & Tunstall, 2008), particularly as the base flood elevation area designated by the Federal Emergency Management Agency has expanded. With new floodplain studies conducted in the early 2000’s, more properties have become subject to flood insurance and elevation requirements, leading to reduced property values and making properties more difficult to sell. Our findings demonstrated that living within the 100-year floodplain was associated with 16% and 14% higher odds of T2D onset in townships and boroughs, respectively, after adjusting for individual and community socioeconomic factors. Observations in townships also suggested that compared to living nearest a stream or lake, living nearest a river or tributary—the most flood-prone waterbodies—increased the odds of T2D. The lack of association in cities may reflect greater flood mitigation efforts such as the construction of levees in larger metropolitan areas and fewer homes located in the floodplain. These findings provide further evidence that the socioeconomic context, rather than blue space exposure, may explain elevated T2D odds with closer proximity to the regions rivers and tributaries. That said, the accumulated psychosocial stress of living through floods and anticipating future floods could have detrimental effects on long-term health, including chronic conditions such as T2D (Hackett & Steptoe, 2017). Experiencing flooding has been shown to negatively impacts mental health (Fernandez et al., 2015), with even relatively small events having long-term effects on mental health and well-being (Tapsell & Tunstall, 2008).

Our findings highlight potential health implications of the planning, management, and development of freshwater areas in this region of Pennsylvania and similar regions where historical and economic trends have led to geospatial disparities in wealth and health. First, investments in sustainable flood prevention and mitigation and assisting low-income households to move out of flood-prone areas would likely have positive impacts on community health. This will be increasingly important if climate change results in more frequent major flood events. Second, purposeful waterfront development enhances the therapeutic value of blue space, facilitating experiences of restoration, contemplation, and a sense of place (Völker & Kistemann, 2013), and can provide areas for physical activity. Formative research we conducted in 2018 in our study region among key informants and recreational users of Pennsylvania’s rivers suggested that the Susquehanna River may not be perceived as an asset to some surrounding communities, particularly those that have not developed their riverfront. Interviewees reported that despite its scenic beauty and many recreational uses, the river remains “invisible” as a destination to many people, with recreation hampered by insufficient access (e.g., the major road depicted in Figure 2 was specifically noted to be a barrier to visiting the river), and that some communities located along the river perceive it as a liability due to flooding risk, likely remnants of destructive floods that occurred in 1972 and 2011 (publication in process). Interviewees also described perceptions of the river as “destructive and dirty,” which reflects its past and present pollution. These findings parallel research in the region that has highlighted the population’s concern about health threats from contamination of streams and soil by large companies (Silva, 2019). Many towns along the river have not developed their waterfront to encourage recreation, which, according to interviewees, is due to a lack of resources and, in some cases, political will. Without such development, this region of Pennsylvania may not see health-promoting benefits of its blue space.

Our study also has implications for epidemiologic research on blue space and health more broadly. To avoid spurious conclusions, such research must consider the historic and economic context that has shaped the development of communities, and researchers should exercise caution when making claims about the generalizability of study findings. As in Pennsylvania, many communities were established along water to facilitate transportation and industrial processes, and while the past several decades have brought significant investment in waterfront redevelopment, particularly in large cities (Völker & Kistemann, 2013), such redevelopment is not universal. Not considering such context may lead to inappropriate aggregation of blue space across large areas. While we found blue space to be negatively associated with our health outcome, the reverse is also plausible in other regions; for example, historic development of coastal areas that advantage the wealthy may lead to positive associations with health. In fact, homes in many countries that are closer to blue space tend to be more expensive (M. P. White et al., 2020). In such cases, adjustment for individual- and community-level socioeconomic status may be insufficient to control for structural confounding by socioeconomic status (Ahern et al., 2013).

Strengths of this study include evaluation of an understudied health outcome in relation to freshwater blue space across multiple community types and evaluation of greenness as a potential confounder, a shortcoming of some prior blue space and health research (Gascon et al., 2017). As is common in studies of nature and health (Hartig, Mitchell, de Vries, & Frumkin, 2014), the primary limitation is that our measure of blue space captured proximity to blue space but did not account for other exposure types (indirect, incidental, or intentional (M. P. White et al., 2020)), or quality of blue space, which likely influence the degree to which blue spaces are utilized for health-promoting experiences such physical activity and stress reduction. We did not have residential history information and so could not account for exposure duration, although our study region experiences low rates of out-migration (Casey et al., 2016). Due to the ubiquity of waterbodies in the region, the majority of the study population lived close to a waterbody and so categories of residential distance to blue space did not differ widely. We therefore may not have detected a difference in T2D onset that could occur if comparisons were made with individuals far from blue space, pointing to a need for research on blue space and health in additional regions with different landscape types. Finally, as discussed, our findings are subject to unmeasured confounding by socioeconomic status. The property value analysis provided support that distance to blue space is related to socioeconomic status in our study region, but with limited availability of tax parcel data and without the ability to link individual health data to property values, we were not able to evaluate property values in relation to T2D onset.

Contrary to hypothesized health-promoting effects of blue space, we found that higher odds of new onset T2D with closer proximity to freshwater blue space. This potentially spurious association appears to reflect the primary role that the Susquehanna and other rivers have historically played in shaping development in central and northeast Pennsylvania, creating a geographic gradient of socioeconomic disadvantage that emanates from their banks. Our findings indicate that relationships between proximity to blue space and health depend not only on the use of the space for healthy behaviors, but also on historic and economic context, and provide a cautionary tale for epidemiologic studies evaluating the influence of blue space on health. They also highlight the potential broad health benefits that could result from waterfront development and investment in flood-prone communities.