Type 2 diabetes is the sixth leading cause of death in the U.S., largely as a result of cardiometabolic complications such as coronary heart disease (CHD). T2D ranges in prevalence from under 4% to almost 18% in counties across the country. The geographic variation can be explained in part by several place-level contextual factors, but the proportion of variance explained by county-level indicators differs regionally. For example, while a combination of nine county-level measures of mainly socioeconomic, race/ethnicity, and built environmental features explain up to 94% of the variation in T2D prevalence in the Midwest, these same factors explain very little variation in Mid-Atlantic counties, including those in Pennsylvania (PA). Our study focuses on four contextual domains that we hypothesize impact T2D and CHD outcomes in PA: chronic environmental contamination, social environment, food environment, and physical activity environment (both utilitarian [walkability] and recreational physical activity). We will evaluate multiple mechanisms through which these factors could impact T2D and CHD. Potential pathways include the influence of physical and social environmental contextual factors on stress, sleep quality, mental health, health care system effectiveness, and obesity-related behaviors. Multilevel studies using nested geographies at several scales are necessary to disentangle the role of contextual and individual factors associated with T2D and its consequences. Thus, we will conduct two nested multilevel studies of T2D and CHD onset and control of T2D. The setting for each is the large, diverse region of PA served by the Geisinger Health System, allowing us to link contextual factors to the wealth of individual-level data contained in electronic health records (EHR). Each will evaluate contextual main effects and then mediation and moderation of these effects. The two connected studies represent an efficient, big data approach to research while simultaneously offering a deep contextual examination of T2D and CHD. The first (38-county EHR study [38co-EHR]) will use existing data on 30,000 patients with T2D and 200,000 patients without T2D. This longitudinal study will be conducted in 38 counties comparing associations in counties and smaller, nested geographies (multi-scale) in relation to T2D onset and control and CHD onset. The second (4-county Behavior and Biomarker Study [4co-BBS]) will study T2D control in four of these counties with high and low T2D burden, evaluating cross-sectional associations in smaller geographies. 4co- BBS will supplement EHR measures and secondary contextual data with primary data collection using saliva cortisol, direct observation of communities, and patient questionnaires. The 4co-BBS study will broaden the scope of pathways examined, collecting data on health behaviors, health system distrust, and community perceptions. Identifying the contextual factors with the greatest influence on the prevalence of T2D and T2D cardiometabolic complications in PA is critical to understanding regional differences and will inform the development of targeted primary and secondary prevention strategies.
The sixth leading cause of death in the U.S., type 2 diabetes (T2D) ranges in prevalence from under 4% to almost 18% in counties across the country. While a combination of nine county-level measures of mainly socioeconomic, race/ethnicity, and built environmental features explain up to 94% of the variation in T2D prevalence in the Midwest, these same factors explain very little variation in Mid-Atlantic counties, including those in Pennsylvania. Identifying the contextual factors with the greatest influence on T2D prevalence in Pennsylvania is critical to understanding regional differences and will inform the development of targeted T2D prevention strategies.
