JAMA Netw Open. 2026 Apr 1;9(4):e266019. doi: 10.1001/jamanetworkopen.2026.6019.
ABSTRACT
IMPORTANCE: Several integrative indices at the neighborhood level have been developed in the US, but direct comparisons across these indices for the prevalence of cardiovascular-kidney-metabolic (CKM) conditions are limited. Moreover, it is not known whether certain indices better capture place-based variability in CKM conditions or provide additional information when compared with a single measure of income.
OBJECTIVE: To determine how much variability is explained by neighborhood indices in the prevalence of CKM conditions at the census tract level and the incremental value of each index when added to median household income.
DESIGN, SETTING, AND PARTICIPANTS: This cross-sectional study of US national surveillance data (Behavioral Risk Factor Surveillance System and American Community Survey) included all census tracts with complete data for exposures, covariates, and outcomes from 2010 to 2022. Analyses were completed between October 2024 and July 2025.
EXPOSURES: Seven neighborhood indices available at the census tract level (Area Deprivation Index, Child Opportunity Index, Environmental Justice Index, Neighborhood Deprivation Index, Social Deprivation Index, Social Vulnerability Index, Structural Racism Effect Index) and median household income at the census tract level from the American Community Survey.
MAIN OUTCOMES AND MEASURES: The primary outcome was prevalence of CKM conditions (coronary heart disease [CHD], stroke, and chronic kidney disease [CKD]) at the census tract level. Differences in the exposures and outcomes were visualized by mapping index scores, median household income, and prevalences of CKM conditions for all census tracts. To assess index agreement, pairwise correlations were computed with Spearman rank correlation coefficients, and to assess the incremental variability explained by each index when added to median household income, change in r2 was calculated.
RESULTS: Of the 65 476 US census tracts included, median (IQR) prevalence was 5.9% (4.7%-7.4%) for CHD, 3.3% (2.6%-4.1%) for stroke, and 2.9% (2.5%-3.4%) for CKD. The indices and income were modestly to highly correlated (range, 0.46-0.94), with some discordance in how each measure classified census tracts by quartiles of index scores. All indices and income were significantly associated with CKM condition prevalence at the census tract level in multivariable linear regression models, adjusted for median population size and age. The r2 values of the indices with CHD, stroke, and CKD ranged from 0.379 (SE = 0.033) for the correlation between the Environmental Justice Index and stroke to 0.688 (SE = 0.002) for the correlation between the Structural Racism Effect Index and stroke. Additionally, the improvement in variability for CHD, stroke, and CKD (change in r2) explained by the addition of each index to income ranged from 0.014 (SE = 0.001) for the correlation between the Environmental Justice Index and CHD to 0.195 (SE = 0.002) for the correlation between Structural Racism Effect Index and stroke.
CONCLUSIONS AND RELEVANCE: In this cross-sectional study, there were similar associations across neighborhood indices and income with the prevalence of CKM conditions. These findings inform how different place-based measures can be applied in public health research and policy.
PMID:41954933 | DOI:10.1001/jamanetworkopen.2026.6019