Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Feb 16.
Published in final edited form as: J Rural Health. 2020 Sep 7;36(4):602–608. doi: 10.1111/jrh.12511

COVID-19 Death Rates Are Higher in Rural Counties With Larger Shares of Blacks and Hispanics.

Kent Jason G Cheng 1, Yue Sun 2, Shannon M Monnat 2
PMCID: PMC7885172  NIHMSID: NIHMS1660938  PMID: 32894612

Abstract

Purpose:

This study compared the average daily increase in COVID-19 mortality rates by county racial/ethnic composition (percent non-Hispanic [NH] Black and percent Hispanic) among U.S. rural counties.

Methods:

COVID-19 daily death counts for 1,976 U.S. nonmetropolitan counties for the period March 2-July 26, 2020 were extracted from USAFacts and merged with county-level American Community Survey and Area Health Resource File data. Covariates included county percent poverty, age composition, adjacency to a metropolitan county, health care supply, and state fixed effects. Mixed-effects negative binomial regression with random intercepts to account for repeated observations within counties was used to predict differences in the average daily increase in the COVID-19 mortality rate across quartiles of percent Black and percent Hispanic.

Findings:

Since early March, the average daily increase in the COVID-19 mortality rate has been significantly higher in rural counties with the highest percent Black and percent Hispanic populations. Compared to counties in the bottom quartile, counties in the top quartile of percent Black have an average daily increase that is 70% higher (IRR=1.70, CI=1.48 to 1.95, P<0.001), and counties in the top quartile of percent Hispanic have an average daily increase that is 50% higher (IRR=1.50, CI=1.33 to 1.69, P <.001), net of covariates.

Conclusion:

COVID-19 mortality risk is not distributed equally across the rural U.S., and the COVID-19 race penalty is not restricted to cities. Among rural counties, the average daily increase in COVID-19 mortality rates has been significantly higher in counties with the largest shares of Black and Hispanic residents.

Keywords: COVID-19, coronavirus, mortality, racial disparities, rural America

Blacks and Hispanics have suffered a disproportionate burden of COVID-19 in the U.S.1–3 There has also been attention to rising rates in the rural U.S.4 However, with the exception of two recent commentaries in this journal,5,6 little attention has been paid to intersections between rurality and race/ethnicity in COVID-19 outcomes. Rural does not automatically equate to white. Racial/ethnic minorities account for 20% of the U.S. rural population, are geographically isolated, and face significant health challenges.7 Therefore, it is important to understand variation in COVID-19 outcomes across rural counties with greater racial/ethnic minority representation.

Thus far, COVID-19 infection rates have been lower on average in rural than in urban counties, but a third of rural counties are in the White House Coronavirus Task Force “red zone”8, and rural counties may experience higher COVID-19 case fatality rates if outbreaks occur.10 Rural populations are older and have higher rates of chronic health conditions that increase risk of COVID-19 mortality.10 Rural areas have also had lower testing rates and a less robust health care infrastructure to deal with cases.9,10 Since 2010, 126 rural hospitals have closed.11 Even when hospitals are available, most do not have the capacity to deal with a surge in cases,12 and only 1% of the nation’s ICU beds are located in rural areas.13

Beyond these general risk factors, there are several reasons to expect higher COVID-19 mortality rates in rural counties with larger shares of Blacks and Hispanics.3 Potential explanations include social determinants of health (e.g., low socioeconomic status, employment conditions, housing overcrowding, inadequate health care access) and comorbidities.6 Rural Blacks have lower SES14 and life expectancy13 than both rural Whites and urban Blacks. Rural Blacks also have higher rates of the very chronic diseases that increase risk of death should one contract COVID-19 (heart disease, diabetes, respiratory diseases)16–20 and shamefully low access to health care.21,22 Blacks are also more likely to work in service occupations that require face-to-face contact with customers23 and are more likely than Whites to live in multigenerational homes and with extended kin,24 conditions that reduce the ability to socially distance and increase risk of disease transmission.25 Moreover, during the early spread of COVID-19 in the U.S., we found that testing rates were lower in states with more Black and poor residents.26 These factors all increase risk of COVID-19 death.

Rural counties with large shares of Hispanics also appear to be at risk of high COVID-19 infection rates. A large share of rural Hispanics live in counties with meatpacking plants, which employ a disproportionate share of Hispanic workers.27 A recent study showed that rural counties with meatpacking plants with outbreaks of COVID-19 have average infection rates five times higher than the rest of rural America.28 Despite lower chronic disease rates than Whites, rural Hispanics have high poverty rates,29 are often residentially segregated,30 and face significant health care access challenges, including lower insurance rates31 and deportation fears that may reduce willingness to seek healthcare.32

Methods

Analyses are restricted to the U.S.’s 1,976 nonmetropolitan counties based on the USDA Economic Research Service’s Rural-Urban Continuum Codes (RUCCs).33 Counties with RUCCs 4–9 were classified as nonmetropolitan. COVID-19 deaths from March 2-July 26, 2020 came from USAFacts.34 Our dependent variable is daily increase in deaths.

We merged COVID-19 mortality data with data from the American Community Survey, 2014–1835 and Area Health Resource Files, 2018–19.36 Our independent variables were county percent non-Hispanic (NH) Black and percent Hispanic. Because these variables are highly skewed, we recoded them to quartiles for regression. Covariates included adjacency to a metropolitan county (RUCCs 4, 6, 8), percentage aged 65+, median household income quartiles, county designation as a health professional shortage area, quartiles for physicians and hospital beds per 100,000 population, and state fixed effects.

Because the outcome of interest is a longitudinal count (daily increase in deaths), we used mixed-effects negative binomial regression. Binomial models were preferred over Poisson because the outcome is overdispersed and has a high zero count.37 The log of the population served as the offset variable in our models. Random intercepts accounted for repeated observations within counties. Model results are presented as incidence rate ratios (IRRs).

Results

There were 9,431 documented COVID-19 deaths across all U.S. rural counties between March 2 and July 26. The mean is 4.77 deaths, with a mean rate of 17.8 per 100,000 population, and 42% of rural counties have had no documented deaths. As shown in Figure 1, the highest rural COVID-19 mortality rates are clustered throughout the south and southwest – regions with large shares of blacks and Hispanics. Of the 20 rural counties with the highest COVID-19 mortality rates, all of them are in the top quartile of percent black or percent Hispanic. The five highest rates are in Glenn County, CA; Hancock, Randolph, and Early counties, GA; and McKinley County, NM.

Figure 1. COVID-19 Mortality Rates (deaths per 100,000 persons) in Nonmetro Counties.

Figure 1.

Open in a new tab

Note: N=1,976 nonmetro counties. The map represents total deaths as of 07/26/20.

Figure 2 shows a graded relationship between county percent Black and COVID-19 deaths. Both counts and rates are highest and have grown the fastest in the highest percent Black counties (Quartile 4). Counties in Q4 of percent Black have an average COVID-19 mortality rate of 39.3 per 100,000 population compared to 7.0 in Q1 counties. Mortality rates much more comparable across the percent Hispanic quartiles. Although mortality rates are highest in Quartile 4, the gap between the highest and lowest percent Hispanic counties is much smaller than the gap between NH Black quartiles.

Figure 2. Cumulative Mean COVID-19 Death Counts and Rates by Percent Non-Hispanic Black Quartile and Percent Hispanic Quartile for Nonmetro Counties.

Figure 2.

Open in a new tab

N=1,976 nonmetro counties; Data are current to July 26, 2020.

Model results predicting the daily increase in COVID-19 deaths as a function of county racial composition are presented in Table 1. IRRs represent the average percentage increase in the daily COVID-19 death rate.

Table 1:

Mixed-Effects Negative Binomial Regression Predicting Daily Increase in COVID-19 Mortality Rate, Nonmetro Counties

Model 1 Model 2
VARIABLES IRR CI P-value IRR CI P-value
% Non-Hispanic Black (Ref: Lowest Quantile)
 2nd 1.25 (1.11 – 1.40) <.001 1.20 (1.07 – 1.35) <0.001
 3rd 1.16 (1.03 – 1.30) 0.02 1.08 (0.95 – 1.21) 0.23
 4th Quantile 1.88 (1.64 – 2.15) <.001 1.70 (1.48 – 1.95) <0.001
% Hispanic (Ref: Lowest Quantile)
 2nd 1.04 (0.96 – 1.14) 0.34 1.06 (0.97 – 1.16) 0.21
 3rd 0.99 (0.90 – 1.09) 0.85 1.01 (0.91 – 1.11) 0.92
 4th Quantile 1.51 (1.35 – 1.69) <0.001 1.50 (1.33 – 1.69) <0.001
COVARIATES
Adjacent to Metro (Ref: Not Adjacent to Metro) 1.22 (1.13 – 1.30) <0.001
% 65+ 0.97 (0.96 – 0.98) <0.001
Median Household Income (Ref: Lowest Quantile)
 2nd 0.77 (0.70 – 0.84) <0.001
 3rd 0.83 (0.74 – 0.92) <0.001
 4th Quantile 0.68 (0.60 – 0.76) <0.001
Health professional shortage area (Ref: no) 0.92 (0.82 – 1.03) 0.14
Physicians per 100,000 population (Ref: Lowest Quantile)
 2nd 0.86 (0.78 – 0.95) <0.001
 3rd 0.89 (0.81 – 0.98) 0.02
 4th Quantile 0.84 (0.75 – 0.94) <0.001
Hospital beds per 100,000 population (Ref: Lowest Quantile)
 2nd 0.94 (0.85 – 1.03) 0.18
 3rd 1.12 (1.01 – 1.24) 0.03
 4th Quantile 1.03 (0.92 – 1.15) 0.60
lnalpha 1.94 (1.88 – 2.00) <0.001 6.78 (6.39 – 7.19) <0.001
Open in a new tab

Notes: Deaths are captured from March 2 to July 26, 2020; IRR=Incident Rate Ratios; CI=95% Confidence Interval N= 290,325 observations in 1,975 nonmetro counties (one county was excluded due to missing information on median household income). Both model control for state fixed effects and include random intercepts for counties.

Model 1 controls only for state fixed effects. We find that counties with the highest percent Black and percent Hispanic populations have significantly higher average daily increases in COVID-19 cases. The IRRs change very little with the introduction of covariates (Model 2). Compared to counties with the lowest percent Black (Q1), counties with the highest percent Black (Q4) had an average daily increase that was 70% higher (IRR=1.70, CI=1.48–1.95, P<0.001). Compared to counties with the lowest percent Hispanic (Q1), counties with the highest percent Hispanic (Q4) had an average daily increase that was 50% higher (IRR=1.50, CI=1.33–1.69, P<0.001).

Adjacency to a metropolitan county was associated with a 22% higher average daily increase in the mortality rate, and average daily increases were significantly lower in counties with higher median household income and more physicians per capita.

Discussion

COVID-19 mortality is not distributed equally across the rural U.S., and the COVID-19 race penalty is not restricted to cities. In this national study of rural counties, we show that the average daily increase in COVID-19 mortality since early March 2020 has been significantly higher in rural counties with the largest shares of Black and Hispanic residents. This study did not examine potential explanations, but as with urban areas, the rural U.S. has a long legacy of structural and systemic racism that influences access to socioeconomic and health care resources, resulting in severe health inequities. COVID-19 is the latest in a long line of inequities that disproportionately affects racial minority populations.

Several interventions are needed to reduce these geographic and racial/ethnic disparities. First, we must increase free COVID-19 testing in rural areas generally, but especially in rural areas with vulnerable population groups. Surveillance is important both because governments need to know where to allocate limited treatment resources and because individuals may be more likely to self-isolate if they know there are positive cases in their community, thereby reducing virus spread.38 Second, local governments should work with trusted community-based organizations, including the faith-based community, to help educate, test and trace, and provide necessary resources to Black and Hispanic residents. Thanks to a long history of racist medical practices, there is understandable black community distrust in the health care system. Hispanic immigrants may fear deportation should they seek services. Working with trusted community partners may facilitate community access to conduct testing and contact tracing and provide education and services. Ultimately, any policy intervention that aims to prevent or mitigate COVID-19 in rural America must prioritize places with the least resources, the most vulnerable populations, and the worst health outcomes.5

Limitations

Analyses are ecological and do not examine individual mortality risk. We are also unable to calculate race-specific mortality rates with data currently available. It is possible that whites also have higher COVID-19 mortality rates in counties with larger shares of Blacks and Hispanics if the conditions in these counties increase risk of underlying health conditions that increase risk of transmission and death (e.g., insufficient testing, poor health care access, social determinants). We encourage state and county health departments to release testing, infection, and mortality data by race/ethnicity so researchers can identify intersections between geographic and racial/ethnic inequities.

Conclusion

The COVID-19 landscape changes rapidly. These preliminary findings reflect significantly higher mortality rates in high percent Black and Hispanic rural counties during the first five months of virus spread in the U.S. These vulnerable populations have less access to health care and have higher rates of the very chronic health conditions that increase risk of complications and death from COVID-19. As rates continue to rise, federal, state, and local governments must act now to properly educate residents about the severity of COVID-19 and provide adequate supplies and resources to these high-risk places.

References:

RESOURCES