Links Between Health and Place in Durham

In Durham, North Carolina, the home of Duke University, crossing from one neighborhood to the next brings many differences, some of which you can see, like the number of trees or bus stops, and others that you can’t, like a high rate of evictions or a low level of voting in primary elections.

As with many cities, some of Durham’s neighborhood boundaries date back to policies in place in the 1920s or 1930s. For instance, maps that were used in approving or denying federally subsidized mortgages or insurance indicated so-called risky neighborhoods in red, and these were often historically Black or integrated neighborhoods. Through this “redlining” practice, nationwide less than 2% of homes insured by the FHA between 1946 and 1959 were available to people of color, according to Bull City 150, a public project that traces Durham’s history.

Such policies still affect neighborhoods today. For instance, the NC Central and College View neighborhood (as defined by the U.S. Census) includes parts of a historically redlined neighborhood, and part of what was once Hayti, a center for Black-owned businesses and banks that was displaced in the 1960s by “urban renewal” and the creation of a highway. Today, most of the residents of this area (78%) are African American, and 62% are cost-burdened, meaning they pay more than is considered affordable for housing based on their income. This area had 71 eviction notices per square mile in 2020, higher than the County average of 14. Among its residents, 14% have kidney disease, and 18% live with diabetes, both much higher than the Durham County averages.

Taking a close look at data about this and 149 other neighborhoods in Durham County, researchers from Duke University School of Medicine and Icahn School of Medicine at Mt Sinai documented in a first-of-its-kind study that some of these differences, which are indicators of structural racism, are linked to health.

Neighborhoods with the highest levels of three chronic diseases — diabetes, high blood pressure, and kidney disease — tended to be those with the lowest proportion of white residents, and vice versa. The study also found that Durham neighborhoods with greater burdens of 12 indicators of structural racism, including lower income and education, lower voter participation in primary elections, higher levels of reported violent crime, and higher rates of eviction and uninsurance, also had higher prevalence of those three chronic diseases.

“The results were not surprising, but I think they’re still very important to share,” said Clemontina Davenport, PhD, co-author of the study published in JAMA Network Open in December 2023. “Structural racism has impacted this country for centuries. But we’re just finally putting the story out there.” Formerly a biostatistician at Duke, Davenport is now an assistant professor in the Department of Biostatistics and Data Science at the Wake Forest University School of Medicine.

The study made national headlines as lead study author Ebony Boulware, MD, MPH, former Duke internist and clinical epidemiologist now dean of Wake Forest University School of Medicine, was invited to share the findings. But the researchers say that it was just the beginning. Ultimately, they want to identify which elements contribute most to these clusters of disease, help other researchers conduct similar studies in their own cities, and use that data to inform policy efforts.

Why Durham?

Durham has some unique characteristics that made this analysis possible, including a diverse population with extremes of wealth and poverty, a large middle class, and a deeply engaged community that often works with Duke and other entities to improve community health, said the first author of the study, Dinushika Mohottige, MD, MPH. Now an assistant professor of population science and policy and medicine at the Icahn School of Medicine at Mount Sinai, Mohottige earned an undergraduate degree at Duke and was a nephrology fellow and chief resident at Duke University School of Medicine. She lived in Durham for 21 years.

The social data for Durham came from multiple sources, including the Durham Neighborhood Compass, an interactive website that anyone can use to track the health and other characteristics of neighborhoods (or census block tracts). The Durham County Public Health Department created the compass in 2014, and in 2016 they partnered with Duke Health and Lincoln Community Health Center to add data from their electronic health records. The two health centers provide summary reports that aggregate data at the neighborhood level and track a list of health conditions identified as high priorities in Durham County.

“This is a really special thing that Durham was able to build, through years of engagement with multiple stakeholders that preceded me,” Mohottige said. “In many other cities or regions, it’s nearly impossible to account for health at a neighborhood level.”

Defining Structural Racism

Residential and economic segregation by race is the general definition of structural racism, but no uniform measure exists. In conducting this study, the researchers identified 17 indicators pf structural racism, including “global indicators,” as well as “discrete indicators.” Davenport explained: “Global measures capture many different measures of structural racism in one indicator. Discrete measures are the smaller pieces.”

“If gentrification breaks up social networks that people use for support, that can have a detrimental impact on people's health.”

Nrupen Bhavsar, PhD, epidemiologist

For example, one of the global measures used in the study was the area deprivation index, a composite number that considers levels of education, poverty, and income in a region. Neighborhoods in the study with a higher area deprivation index showed higher rates of chronic disease.

The 2023 study also found links between disease and several discrete indicators of structural racism, including lower voter participation in primary elections and higher rates of evictions. Those two indicators have not been studied much when it comes health, Mohotttige said, but she is excited about exploring them further because they are factors that can be influenced by changing current policies. For example, cities or counties can address disenfranchisement by tackling such issues as unequal policing which can lead to higher rates of felony conviction for certain groups, or insufficient staffing at polling places.

Beyond Durham

Inspired by the Neighborhood Compass, Mohottige, study co-author Nrupen Bhavsar, PhD, associate professor in the Department of Surgery, and colleagues have built a similar site that aims to expand much of the data to the entire country.

The SEED Health Atlas started with data for Durham County as well as multiple counties surrounding it, and it now includes data for the entire country for many of the variables, said Bhavsar, associate professor surgery at Duke. “Can we collaborate with folks at other health systems to do similar work across different areas of the United States to understand whether or not we see similar associations of structural racism and outcomes?”

The atlas draws on the 2010-2019 census and other contextual data to track social, environmental, and climate indicators of health, and Bhavsar is adding data after the 2020 census. That will make the tool useful for tracking change over time, such as gentrification (the influx of healthier, wealthier, and younger people into an area).

“If gentrification breaks up social networks that people use for support, that can have a detrimental impact on people's health,” Bhavsar said. For example, if a church or even a barbershop must move out of a particular neighborhood, residents may lose that built-in support system. “Maybe someone they knew there took them to their physician appointment or was a friend they could talk to when they were stressed. Now they’re no longer there,” he said.

In addition, researchers can use the SEED atlas to target interventions to where they are most needed. For instance, data from SEED as well as Duke Health’s electronic health record and trauma registry were used to create a dashboard that Duke Health executives and researchers used to monitor patterns of gun violence in Durham and target interventions such as a program aimed at increasing enrollment in GED classes for high-risk youth. This effort was detailed in the journal JAMIA Open, and a description of the atlas was published in September 2024 in the Journal of Clinical and Translational Science.

To make sure the atlas included input from Durham community members, Bhavsar worked with Mina Silberberg, PhD, a professor in the Duke Department of Family Medicine and Community Health, and Jessica Sperling, PhD, director of the Office of Evaluation and Applied Research Partnership, a collaboration between the Duke Clinical and Translational Science Institute and the Duke Social Science Research Institute.

“We really wanted the community to be able to use it to understand their own neighborhood,” Bhavsar said.

Sperling conducted focus groups with 10 community members and leaders who told them the data could be useful when applying for funding for projects to improve neighborhoods and when speaking with policymakers, she said. They also suggested finding a way to include data about community resources such as food pantries, Black and Latino-owned businesses, and churches.

Residents of Durham understandably have strong opinions about Duke University and Duke Health, as do many people who live in cities that house large, sometimes wealthy, institutions, Bhavsar said. “That’s part of what drives the research that I do:  understanding how to improve the health of individuals and develop a shared goal moving forward.”

Angela Spivey is a senior science writer and managing editor for the School of Medicine’s Office of Strategic Communications.

Eamon Queeney is assistant director of multimedia & creative for the School of Medicine’s Office of Strategic Communications.