Background: Eviction represents a traumatic form of housing instability and is an important social determinant of health. Growing wealth inequality and a housing crisis have resulted in high rates of eviction since 2000, with women renters and those of Black race disproportionately affected. Prior studies demonstrate the negative impact of eviction on self-reported mental and physical health. However, the association of eviction with all-cause mortality in the US is not known.
Methods: In this cross-sectional study, we evaluated the association between 2016 county-level eviction rates and all-cause mortality. We obtained eviction rates from the Eviction Lab at Princeton, county characteristics from the American Community Survey and CDC Behavioral Risk Factor Surveillance System, and age-adjusted all-cause mortality from the CDC Wonder database. We categorized counties into tertiles by eviction rate ranging from lowest (Tertile 1) to highest (Tertile 3) and compared county characteristics across groups using Jonckheere-Terpstra tests. We constructed a multivariable linear regression model to evaluate associations of eviction and county characteristics with all-cause mortality. We tested interaction between eviction rate and proportion of Black residents, women residents, and proportion of renters for risk of all-cause mortality.
Results: We evaluated all-cause mortality for 686 counties with available eviction data. Counties with higher eviction rates had higher proportion of residents of Black race, women, and greater prevalence of poverty and comorbid health conditions (Table 1). In unadjusted comparison, counties with higher eviction rates had higher age-adjusted rates of mortality (Tertile 1 vs. 2 vs. 3 [per 100,000 people]: 739 vs. 780 vs. 837: p-trend< 0.0001). In adjusted analysis, higher county-level eviction rates were significantly associated with higher mortality (Table 2). Furthermore, the association between county-level eviction rate and mortality was modified by county-level proportion of Black individuals, women individuals, and renters. In stratified analyses by proportion of each, the association between eviction rates and mortality was stronger among counties with higher proportion of Black individuals (Pinteraction=0.04), women (Pinteraction=0.003), and renters (Pinteraction=0.03) (Table 2).
Conclusions: Our study demonstrates a strong association of eviction on mortality risk and suggests that communities with high proportions of Black and women residents are more likely to experience death in association with eviction. Eviction Lab data has previously shown that Black renters make up more than 30% of eviction filings despite representing less than 20% of renters nationally, with Black women representing the highest risk group for eviction, highlighting the complex relationship between housing instability, race, and gender in the US. Policies aimed at increasing affordable housing by means of regulatory approaches and cushioning the stressors that contribute to housing instability may translate to improvements in downstream health disparities. This study is limited in its ability to determine causality due to its cross-sectional design. We capture only 22% of US counties of which 2% of residents were of Black race due to missingness in available eviction data, limiting the generalizability of our findings. Future studies may elucidate mechanisms by which housing instability contributes to adverse health and identify opportunities for intervention.