The social vulnerability metric (SVM) as a new tool for public health

Abstract

Objective To derive and validate a new ecological measure of the social determinants of health (SDoH), calculable at the zip code or county level. Data Sources and Study Setting The most recent releases of secondary, publicly available data were collected from national U.S. health agencies as well as state and city public health departments. Study Design The Social Vulnerability Metric (SVM) was constructed from U.S. zip‐code level measures (2018) from survey data using multidimensional Item Response Theory and validated using outcomes including all‐cause mortality (2016), COVID‐19 vaccination (2021), and emergency department visits for asthma (2018). The SVM was also compared with the existing Centers for Disease Control and Prevention's Social Vulnerability Index (SVI) to determine convergent validity and differential predictive validity. Data Collection/Extraction Methods The data were collected directly from published files available to the public online from national U.S. health agencies as well as state and city public health departments. Principal Findings The correlation between SVM scores and national age‐adjusted county all‐cause mortality was r = 0.68. This correlation demonstrated the SVM's robust validity and outperformed the SVI with an almost four‐fold increase in explained variance (46% vs. 12%). The SVM was also highly correlated (r ≥ 0.60) to zip‐code level health outcomes for the state of California and city of Chicago. Conclusions The SVM offers a measurement tool improving upon the performance of existing SDoH composite measures and has broad applicability to public health that may help in directing future policies and interventions. The SVM provides a single measure of SDoH that better quantifies associations with health outcomes.

Social Vulnerability and Risk of Suicide in US Adults, 2016-2020

“…We used 2 measures of social vulnerability, the SVI and the SVM . The SVI ranks US counties in terms of their social vulnerability on 17 SDOH indicators drawn from the domains of education access and quality, health care and quality, neighborhood and built environment, social and community context, and economic stability .…”

“…The SVI was designed as a relative measure of the degree to which a community is at risk during a public health emergency, highlighting the needs of socially vulnerable populations in emergency response and recovery efforts. The SVM was developed to provide a zip code–level or county-level social vulnerability measure based on multidimensional item response theory applied to the domains of demography, education, economic context (eg, unemployment rate), physical infrastructure (eg, housing and transportation), and health care (eg, health insurance coverage) . The SVM is based on zip code–level measurements of 94 SDOH variables collected directly from publicly available government survey data from the Agency for Healthcare Research and Quality (AHRQ) SDOH Database and state and city public health departments.…”

“…The SVM is based on zip code–level measurements of 94 SDOH variables collected directly from publicly available government survey data from the Agency for Healthcare Research and Quality (AHRQ) SDOH Database and state and city public health departments. To construct county-level SVM measures, Zip Code Tabulation Area (ZCTA) SVM estimates were aggregated to the county level using the US Census Bureau relationship files ZCTA-county and ZCTA-Metropolitan Statistical Area crosswalks based on the 2010 census . Aggregation was done by taking a population-weighted average of SVM scores across all ZCTAs assigned to a particular county.…”

Efficient Determination of Social Determinants of Health From Clinical Notes for Timely Identification of Suicidality Among US Veterans

“…Additionally, it might be helpful to compare these results with traditional estimates of SDOH and the recently reported Social Vulnerability Metric. 10…”

“…Maybe we can think about frameworks where we can have more verifiability, while not losing individual privacy. Additionally, it might be helpful to compare these results with traditional estimates of SDOH and the recently reported Social Vulnerability Metric …”