Primary Submission Category: Structural factors
Neighborhood-level Social Vulnerability and ED visits: Linking EHR and Area-level Metrics to uncover the relationship between neighborhood vulnerability and poor outcomes among patients with diabetes
Authors: Pricila Mullachery Huanmei Wu Jay Patel Ilene Hollin Gabriel Tajeu
Presenting Author: Pricila Mullachery*
Significance: Neighborhoods are shaped by past discriminatory policies, which created disparities in the distribution of resources such as healthcare services, transportation, and fresh food markets. Exposure to neighborhood disadvantage results in a disproportional burden of chronic conditions among individuals in these communities. Electronic health records (EHR) are being increasingly used to study population health, but the connection between neighborhood factors and population health is rarely explored in EHR data. We examined the association between neighborhood social vulnerability, measured by a composite score of 15 variables that capture disadvantage, and emergency department (ED) use among patients with diabetes. Data: We analyzed EHR from HealthShare Exchange, a regional health information system encompassing health systems and insurers in the Greater Philadelphia region. Data were from a sample of individuals (n=119,832) who had at least one healthcare encounter between 2018 and 2020. We extracted data on demographic variables, type of encounter, and diagnosis codes. The outcome was number of ED visits among patients with diabetes. Zip code area identifiers (n=210) were used to link social vulnerability data to EHR. Preliminary findings: A total of 10.3% of people had a diagnosis of diabetes; on average, they had 1.9 ED visits between 2018 and 2020. People with diabetes residing in high-vulnerability neighborhoods (top 20%) were three times more likely to have two or more ED visits than those in low-vulnerability neighborhoods (bottom 20%). In adjusted models with individuals nested in neighborhoods, a 10% increase in social vulnerability was associated with an 11% increase (p<0.001) in the likelihood of one additional ED visit. Implications: Our results highlight the need for structural approaches targeting upstream determinants of health in communities. Allocating resources to socially vulnerable communities is critical to achieving health equity.