Letter to the Editor in Response to Khazanchi et al.

To the Editor,

In a recent research report (published online on June 23, 2020), Khazanchi et al.¹ describe the association of county-level Social Vulnerability Index (SVI)² with positive tests and deaths from COVID-19 per capita as of April 19, 2020. In an analysis unadjusted for testing frequency, age, and comorbidity burden of county residents, they found that more socially vulnerable counties (SVI in the top quartile) experience higher COVID-19 incidence and deaths compared to the least socially vulnerable (SVI in the bottom quartile), driven by the minority status and language domain. These results are interesting and similar to a prior analysis by our group (published online on April 14, 2020)³ utilizing publicly available data from counties with at least 50 confirmed COVID-19 cases as of April 4, 2020 (433 counties accounting for 283,256 cases and 6644 deaths included). After adjustment for county-level proportion of residents aged ≥ 65 years, and comorbidities (utilizing Hierarchical Condition Category (HCC) risk scores acquired from the Centers for Medicare and Medicaid Services),⁴ we determined that greater SVI is associated with higher COVID-19-related Case Fatality Rate (CFR, deaths/100 cases), driven by the socioeconomic status, minority status/language, and housing type/transportation domains. Unlike Khazanchi et al. (1), after adjustment for covariates, we did not find that overall SVI associated with COVID-19 incidence as of April 4, 2020; although two sub-components of the SVI, socioeconomic status and minority status/language, were associated with higher incidence. In addition, understanding that a major limitation of our study was our inability to account for the contribution of county-level COVID-19 testing, we performed sensitivity analyses in the 6 states with the highest level of testing (New York, New Jersey, Washington, Massachusetts, Vermont, and Louisiana), and were able to replicate our findings. In our study, we determined that more than a quarter of the counties studied had both high social vulnerability and high CFR, with potentially devastating effects from an ensuing “disease-driven poverty trap,” a vicious cycle of disease and worsening social disparity in these counties.⁵ Consequently, to aid public health resource allocation to these vulnerable counties, we have developed an interactive online dashboard to track COVID-19 outcomes stratified by social vulnerability (http://www.eccri.emory.edu/covid/index.html), updated weekly since April 16, 2020.