![[ Visit Client Website ]](images/banner.gif)
Schedule:
Sunday, January 24, 2027
Continental 1-4 (Hilton San Francisco Union Square)
* noted as presenting author
Background and Purpose: Health insurance coverage is a critical dimension of health equity because lack of coverage can limit access to preventive care, treatment, and other essential services. Although prior research has linked insurance gaps to socioeconomic disadvantage and racial inequities, less is known about how these patterns cluster within urban space. Using a health equity lens and a social determinants of health perspective, this study investigates how neighborhood poverty, median household income, and the percentages of Black and White residents are spatially associated with uninsurance across St. Louis City census tracts. The goal is to identify place-based disparities that can inform more targeted and equitable health and social policy responses.
Methods: This quantitative spatial study used census tract-level data from the 2017–2021 American Community Survey 5-year estimates for St. Louis City, Missouri. The dependent variable was no health insurance, and the independent variables were poverty, median household income, percent Black, and percent White. Tract-level patterns were visualized in ArcGIS. Global Moran’s I statistics assessed overall spatial autocorrelation, while BiLISA maps in GeoDa identified local clustering between uninsurance and each socioeconomic and racial variable. Together, these methods were used to examine whether uninsurance was randomly distributed or spatially concentrated in relation to neighborhood disadvantage and racial composition.
Results: Uninsurance showed significant global spatial autocorrelation across spatial weight specifications, with Moran’s I values between 0.242 and 0.299 (all p ≤ .001), indicating nonrandom clustering across the city. Local spatial analyses showed clearer clustering between uninsurance and poverty and between uninsurance and percent Black than for income or percent White. High-high clusters of uninsurance and poverty appeared in 11 census tracts concentrated in the north and northeast parts of the city, while low-low clusters appeared in 20 tracts in the southwest. High-high clustering between uninsurance and median household income appeared in only 2 tracts, and high-high clustering between uninsurance and percent White appeared in only 1 tract. By contrast, high-high clustering between uninsurance and percent Black was concentrated across much of the upper-middle portion of the city.
Conclusions and Implications: Taken together, the results indicate that uninsurance in St. Louis is spatially concentrated rather than randomly distributed, with stronger local clustering in relation to poverty and percent Black than to income or percent White. For social work research, health policy, and community-based outreach, these findings underscore the value of place-based approaches to identifying neighborhoods where insurance access gaps may be most pronounced. Spatially grounded evidence can help social work and public health stakeholders target outreach, strengthen attention to coverage inequities, and guide future research on neighborhood context and access to care.
Methods: This quantitative spatial study used census tract-level data from the 2017–2021 American Community Survey 5-year estimates for St. Louis City, Missouri. The dependent variable was no health insurance, and the independent variables were poverty, median household income, percent Black, and percent White. Tract-level patterns were visualized in ArcGIS. Global Moran’s I statistics assessed overall spatial autocorrelation, while BiLISA maps in GeoDa identified local clustering between uninsurance and each socioeconomic and racial variable. Together, these methods were used to examine whether uninsurance was randomly distributed or spatially concentrated in relation to neighborhood disadvantage and racial composition.
Results: Uninsurance showed significant global spatial autocorrelation across spatial weight specifications, with Moran’s I values between 0.242 and 0.299 (all p ≤ .001), indicating nonrandom clustering across the city. Local spatial analyses showed clearer clustering between uninsurance and poverty and between uninsurance and percent Black than for income or percent White. High-high clusters of uninsurance and poverty appeared in 11 census tracts concentrated in the north and northeast parts of the city, while low-low clusters appeared in 20 tracts in the southwest. High-high clustering between uninsurance and median household income appeared in only 2 tracts, and high-high clustering between uninsurance and percent White appeared in only 1 tract. By contrast, high-high clustering between uninsurance and percent Black was concentrated across much of the upper-middle portion of the city.
Conclusions and Implications: Taken together, the results indicate that uninsurance in St. Louis is spatially concentrated rather than randomly distributed, with stronger local clustering in relation to poverty and percent Black than to income or percent White. For social work research, health policy, and community-based outreach, these findings underscore the value of place-based approaches to identifying neighborhoods where insurance access gaps may be most pronounced. Spatially grounded evidence can help social work and public health stakeholders target outreach, strengthen attention to coverage inequities, and guide future research on neighborhood context and access to care.