The Community Vulnerability Compass: a novel, scalable approach for measuring and visualizing social determinants of health insights
Yolande Pengetnze, Venkatraghavan Sundaram, Yusuf Tamer, Albert Karam, Lance Rather, Olayide Adejumobi, Leslie Wainwright, Steve MiffAbstract
Objectives
To determine whether a novel digital tool, the Community Vulnerability Compass (CVC), built using large datasets, can accurately measure neighborhood- and individual-level social determinants of health (SDOH) at scale. Existing SDOH indexes fall short of this dual requirement.
Materials and Methods
Setting: A cross-sectional study by Parkland Health (Parkland) and Parkland Center for Clinical Innovation (PCCI) to design, build, deploy, and validate CVC in Dallas County/across Texas (2018-2024). Data Sources: Parkland Electronic Health Records; population-level data from diverse national datasets. Statistical Analysis: CVC’s Community Vulnerability Index (CVI), and 4 subindexes were used to classify all 18 638 Texas census-block groups as Very-High, High, Moderate, Low, and Very-Low social vulnerability. Individuals were assigned the vulnerability of their home address census-block group. CVC’s classifications were compared against 3 existing SDOH neighborhood tools (Area Deprivation Index [ADI], Social Vulnerability Index [SVI], or Environmental Justice Index [EJI]) and validated against individual-level SDOH screening tools or Z-code documentation. Spearman rank correlation was used for neighborhood-level comparisons and precision/recall, for individual-level comparisons.
Results
Neighborhood-level CVI measurement of social vulnerability strongly correlated with EJI (r = 0.83), SVI (r = 0.82), and ADI (r = 0.79). Individual-level CVI measurement had higher recall than ADI (68% vs 39%, respectively; P < .001) and high recall across self-reported SDOH (77%-79.6%). Precision was highest for food needs (75.1%); lowest for safety needs (1.2%).
Discussion
CVC measured a cross-cutting range of neighborhood social vulnerabilities and accurately approximated individual-level SDOH, outperforming existing indexes.
Conclusion
CVC can be leveraged as an accurate and scalable SDOH digital measurement tool.