Back

A County-Level Susceptibility Index and Coronavirus Disease 2019 Mortality in the United States: A Socioecological Study

Khan, S.; McCabe, M.; Krefman, A.; Petito, L. C.; Yang, X.; Kershaw, K.; Pool, L.; Allen, N. B.

2020-07-06 public and global health
10.1101/2020.07.04.20146084 medRxiv
Show abstract

As of June 2020, the United States (US) has experienced the highest number of deaths related to coronavirus disease 2019 (Covid-19) in the world, but significant geographic heterogeneity exists at the county-level. Therefore, we sought to classify counties in the United States across multiple domains utilizing a socioecological framework and examine the association between these county-level groups and Covid-19 mortality. We harmonized and linked county-level sociodemographic, health, and environmental metrics associated with increased susceptibility for Covid-19 mortality. Latent class analysis defined a county-level susceptibility index (CSI) based on these metrics (n=2701 counties). Next, we used linear regression models to estimate the associations of the CSI and Covid-19 deaths per capita and initial mortality doubling time (as of 6/2/20), adjusted for days since first Covid-19 case. We identified 4 groups classified by the CSI with distinct sociodemographic, health, and environmental profiles and widespread geographic dispersion. Covid-19 deaths per capita were significantly higher in the group consisting of rural, vulnerable counties (55.8 [95% CI 50.3-61.2] deaths per 100,000) compared with the group with diverse, urban counties (32.2 [27.3-37.0]) at similar points in the outbreak (76 days since first case). Our findings can inform equitable resource allocation for Covid-19 to allow targeted public health preparedness and response in vulnerable counties.

Matching journals

The top 10 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.