For a better experience, click the Compatibility Mode icon above to turn off Compatibility Mode, which is only for viewing older websites.

Retail Environment and Cardiovascular Disease

Evaluating how access to healthy food sources, physical activity and medical facilities affect racial disparities in cardiovascular disease.

Senior shopping for vegetables in grocery store

Research Summary

Environmental factors can impact individual and community health. Given trends toward urbanization and longer lifespans, evidence is needed on how neighborhood dynamics are connected to aging, cardiovascular health, and cognitive decline in diverse populations. Neighborhood businesses have the potential to affect health through the provision of products and services that affect individual dietary intake, physical activity patterns, and the management of chronic conditions. Research indicates local availability of potentially health-promoting businesses is lower for historically disadvantaged population subgroups, including African American adults. Previous studies linking place to health have been largely cross-sectional and limited in geographic scope. Yet, business densities change over time, with trends and fluctuations varying spatially. Analyses of these changes can be used to build more convincing causal evidence. In addition, moving to a new home represents another source of changing exposure. However, health problems may themselves precipitate relocation, requiring careful and innovative analytic strategies.

In this study, we assess how changes to access of healthy food sources, medical facilities, and physical activity venues relate to cardiovascular disease (CVD) and cognitive decline and subsequent residential relocation of older adults in the U.S, with a focus on addressing health disparities. The overarching goal is to generate and disseminate actionable knowledge on the role of the local neighborhood contexts  to inform strategies for tailoring local action to reduce  health disparities and support aging-in-place.


Reasons for Geographic and Racial Differences in Stroke (REGARDS) Project at University of Alabama Birmingham; Built Environment and Health Research Group at Columbia UniversityColumbia University Population Research CenterThe Collaborative Health Studies Coordinating Center (CHSCC) at University of Washington; Urban Form Lab (UFL) at University of Washington;  New York Academy of MedicineEnvironment and Policy Laboratory in the School of Kinesiology at University of Michigan; Multi-Ethnic Study of Atherosclerosis (MESA) at Department of Epidemiology, School of Public Health at University of Michigan; Leslie McClure PhD, Department of Epidemiology and Biostatistics, Dornsife School of Public Health at Drexel University; and Loni Tabb PhD, Department of Epidemiology and Biostatistics, Dornsife School of Public Health at Drexel University; Andrea Rosso, PhD,MPH, Brain,Environment, Aging and Mobility (BEAM) lab at University of Pittsburgh.

Research Methods

The Retail Environment and Cardiovascular Disease (RECVD) project uses individual health-relevant data from the two complementary population-based cohort studies of adults:  The Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study, and Cardiovascular Health Study (CHS). Both studies have recruited and followed-up with socioeconomically diverse populations, The 30,183 REGARDS participants were enrolled from a broad geographic region (48 contiguous U.S. states) in 2003-2007. CHS is a study of 5,888 older adults enrolled in 1990-1993 from four U.S. sites.

Health measures specific to cardiovascular disease and cognitive functioning are available longitudinally for both cohorts, including dementia and incident events such as myocardial infarction. Commercially available data on residential and social history available from LexisNexis has been linked to participants, and their address data has been georeferenced.

For longitudinal information on exercise places, walkable neighborhoods, food environments, and social destinations, the National Establishment Time Series (NETS) data was used. The NETS database is a set of longitudinal data files derived from Dun & Bradstreet’s (D&B) register in January of each year beginning in 1990 and available through 2014. Considered one of the most comprehensive establishment data sources available, NETS serves as a census of American businesses. Over 300 variables are available for each business, including company name, the establishment’s most recent location (city, state, ZIP, latitude and longitude), relocation history, industrial classification codes, sales volume, and number of employees.

Additional geographic data came from the U.S. Census and American Communities Survey, a national rail transit database, U.S. Geological Survey land cover data, and nationwide data relevant to safety. The linkage of individual and geographic data allows researchers to characterize residential environments experienced by participants as neighborhoods change over time, or as participants relocate.

Research Results

We have georeferenced all of the national NETS addresses between 1990 and 2014 (81 million locations documented for the 58 million businesses) to improve consistency and validity throughout the period of interest. Total establishment counts for selected categories for the same timeframe are displayed in Figure 1, which shows an upward trend for some of the categories of interest.

Figure 1: Selected categories of business environments relevant to behavioral pathways

The longitudinal nature of NETS introduces new challenges and possibilities, which we incorporated into our classification and analytic strategies. First, we have improved on georeferencing accuracy and efficiency. Our approach to georeferencing annual addresses for each establishment provided more high quality matches, particularly for older records that NETS had located only by ZIP code. Second, we have developed strategies informed by our earlier work and other research to identify specific types of businesses that can be combined to understand the stability of businesses in a neighborhood and its role associated with “aging in place”.

Third, we developed a strategy to address over-counting of businesses (e.g., where a single known business appeared at the same location in the same year multiple times). Finally, we developed a naming and data documentation system that has facilitated linkage and data sharing with multiple ongoing health studies. The resulting dynamic information on the changing neighborhood environment can be visualized in a variety of formats (e.g. static, animation).

We have linked all of the longitudinal neighborhood data to CHS and REGARDS cohorts. We have characterized 72,246 census tracts in the continental U.S., and circular buffers with respect to sociodemographic, transit, connectivity, land cover, and retail data.  Several of the variables of interest draw on multiple data sources as well, such as for a categorization relevant to gentrification, and a continuous measure of walkability. The unified data dictionaries have been updated to allow other investigators to work with linked and derived variables.

We have a growing team of collaborators. If you are interested in collaborating with the RECVD study data, please contact the RECVD Research Coordinator at

The following papers have been published:

Kaufman T.K., Rundle A., Neckerman K.M., Sheehan D.M., Lovasi G.S., Hirsch J.A. Neighborhood Recreation Facilities and Facility Membership are Jointly Associated with Objectively Measured Physical Activity. Journal of Urban Health.

Rundle, A.G., Chen, Y, Quinn, J.W., Rahai, N., Bartley, K., Mooney, S., Bader, M.D., Zeleniuch-Jacquotte, A., Lovasi, G.S., Neckerman, K.K. Development of a Neighborhood Walkability Index for Studying Neighborhood Physical Activity Contexts in Communities Across the U.S. Over the Past Three Decades. Journal of Urban Health.

Berger, N., Kaufman, T.K., Bader, M.D., Rundle, A.G., Mooney, S.J., Neckerman, K.M., Lovasi, G.S. Disparities in trajectories of changes in the unhealthy food environment in New York City: a latent class growth analysis, 1990-2010. Social Science & Medicine.

Rhodes-Bratton, B., Rundle, A., Lovasi, G.S., Herbstman, J. The relationship between childhood obesity and neighborhood food ecology explored through the context of gentrification in New York City. International Public Health Journal. (Accepted January 2019)

Goh, C.E., Mooney, S.J., Siscovick, D.S., Lemaitre, R.N., Hurvitz, P., Sotoodehnia, N., Kaufman, T.K., Zulaika, G., Lovasi, G.S. Medical facilities in the neighborhood and incidence of sudden cardiac arrest. Resuscitation, 2018. 130: p. 118-123. PMCID: 6245544

We anticipate many of our 30 “in progress” papers will be submitted for conference presentations or for publication in the coming months, describing methods innovations, dynamics of changing neighborhood disparities and individual–level associations between the retail environment and healthy aging.


Thank you to the participants of REGARDS and CHS for their participation in the parent studies and helping scientists understand how to prevent chronic disease.

Research Team


Please contact us for more information about the specific initiatives or collaboration.

Funding provided by the National Institute on Aging [R01AG049970: PI Lovasi], Columbia University Robert N Butler Center for Aging and a generous gift from Dana and David Dornsife to the Drexel University Dornsife School of Public Health.