# Component A: A multi-level study of community context and type 2 diabetes and coronary heart disease using electronic health record and new primary data across nested geographies in Pennsylvania

> **NIH ALLCDC U01** · GEISINGER CLINIC · 2021 · $595,000

## Abstract

Project Summary
 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 pr...

## Key facts

- **NIH application ID:** 10201404
- **Project number:** 5U01DP006296-05
- **Recipient organization:** GEISINGER CLINIC
- **Principal Investigator:** Annemarie Gregory Hirsch
- **Activity code:** U01 (R01, R21, SBIR, etc.)
- **Funding institute:** ALLCDC
- **Fiscal year:** 2021
- **Award amount:** $595,000
- **Award type:** 5
- **Project period:** 2017-09-30 → 2022-09-29

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10201404

## Citation

> US National Institutes of Health, RePORTER application 10201404, Component A: A multi-level study of community context and type 2 diabetes and coronary heart disease using electronic health record and new primary data across nested geographies in Pennsylvania (5U01DP006296-05). Retrieved via AI Analytics 2026-05-23 from https://api.ai-analytics.org/grant/nih/10201404. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*