RUPRI Center for Rural Health Policy Analysis Rural Policy Brief Brief No. 2023-1 JANUARY 2023 http://www.public-health.uiowa.edu/rupri/ COVID-19 Mortality Rates across Noncore, Micropolitan, and Metropolitan Counties by Community Characteristics, December 2020-January 2021 Whitney E. Zahnd, PhD; Khyathi Gadag, MHA,; Fred Ullrich, BA; Keith J. Mueller, PhD Purpose This policy brief examines differences in COVID-19 mortality rates across rural-urban designations and stratifications by geography, county-level sociodemographic factors, and county-level health care factors. Between December 2020 and January 2021, COVID-19 deaths were at their peak, hospital capacity was stretched, and COVID-19 vaccines were not widely available, making this a critical time period to examine. These findings may provide insights to ensure that the health care and social support system can manage the ongoing COVID-19 pandemic and prepare for future health emergencies. Key Findings e COVID-19 mortality rates between December 2020 and January 2021 were highest in noncore counties, followed by micropolitan and metropolitan counties. e The Midwest experienced the highest COVID-19 mortality rates, particularly in noncore counties. e Compared to more racially and ethnically homogeneous counties, those with at least 20 percent of the population being Hispanic or American Indian/Alaska Native (AI/AN) had higher COVID-19 mortality rates. Micropolitan counties with at least 20 percent AI/AN population reported the highest rates. ¢ Counties with the highest nursing home bed density (by quartile) reported the highest mortality rates, regardless of rurality. Across all nursing home bed density quartiles, micropolitan and noncore counties had the highest mortality rate. Background The COVID-19 pandemic hit nonmetropolitan populations hard. As of August 1, 2022, the cumulative mortality rate among nonmetropolitan counties exceeded that of metropolitan counties by 34 percent despite having lower cumulative case rates.?! Studies suggest that as many as half of rural residents are at risk for hospitalization or death if infected with COVID-19.2 At the intersection of race and geography, Cheng et al found that rural counties in the top quartile of Black and Hispanic populations had University of Iowa College This project was supported by the Federal ™ RUPRI Center for Rural R H R ' Office of Rural Health Policy (FORHP), Health Y Health Policy Analysis, Resources and Services Administration Rural Health Research (HRSA), U.S. Department of Health and of Public Health, & Policy Centers Human Services (HHS) under cooperative RURAL POLICY RESEARCH INSTITUTE ~ Department of Health Funded by the Federal Office of Rural Healln Poliey - agreement/grant #U1C RH20419. The Management and Policy, 145 Riverside Dr., Iowa City, IA www.ruralheafthresearch.org information, conclusions and opinions 52242-2007. (319) 384-3830 expressed in this policy brief are those of the authors and no endorsement http://www.public-health.uiowa.edu/rupri by FORHP, HRSA, HHS is intended or should be inferred. E-mail: cph-rupri-inquiries@uiowa.edu significantly higher than average daily increases in their COVID-19 mortality rates.34 Other studies have shown poverty, uninsured status, and other factors to be significant contributors.! However, many of these studies were reflective of the pandemic before the peak of cases and deaths in December 2020 and January 2021. It is important, therefore, to continue to assess the impact of various factors on COVID-19 mortality in rural areas during peak times to fully understand contributing factors. The first year of the pandemic was marked in rural areas by initial disparities in access to testing and more pronounced geographic hotspots of disease incidence and mortality due to high risk of exposure (e.g., counties with meatpacking plants, or adjacency to such counties) and high risk of severe disease among those living in institutional settings (e.g., nursing homes). 5% However, factors related to geographic differences in COVID-19 mortality rates during the period of peak cases and deaths may differ from rates earlier in the pandemic. Further, unique factors that affect COVID-19 mortality rates may also be in play, such as racial/ethnic composition and exposure-related and treatment-related health care access characteristics. Knowing that many hospitals, including those in rural areas, approached or exceeded their maximum capacity during the surge of cases in November-December 2020, we expect that this resource strain may have subsequently further contributed adversely to mortality rates in December 2020 through January 2021, especially as this time period pre-dates the widespread availability of vaccines. Data and Methods COVID-19 mortality data were extracted from the Johns Hopkins COVID-19 Data Repository by county for the period December 1, 2020, through January 31, 2021.7 Urban Influence Codes (UICs) were used to categorize counties as metropolitan, micropolitan, or noncore.® Measures of racial/ethnic composition were derived from 2015-2019 American Community Survey (ACS) five-year estimates.'" Data on county- level health care characteristics most relevant to COVID-19 mortality included ICU bed density as a measure of capacity to care for hospitalized patients with severe complications due to COVID-19, and nursing home bed density as a measure populations at high risk for severe outcomes (i.e., hospitalization and/or death). The number of ICU beds within each county was compiled from the 2019 American Hospital Association Survey, and nursing home bed counts were taken from 2020 CMS Nursing Home Compare Data.%!® Densities of both ICU beds and nursing home beds were calculated per 100,000 population within the county. We compiled data from multiple sources to examine crude mortality rates per 100,000 population aggregated across metropolitan, micropolitan, and noncore counties at national and state levels (crude mortality rates are calculated by dividing the number of deaths by the total population within an area of interest). Additionally, we performed stratifications based upon county racial/ethnic composition (using a cut point of 20 percent or higher of Black, Hispanic, or AI/AN populations, U.S. Census Region, and quartile of ICU bed and nursing home bed density. Results/Findings During this two-month time period, 178,739 COVID-19 deaths (crude mortality rate of 55.35 per 100,000) occurred in the U.S. Across geographies, 142,950 COVID-19 deaths occurred in metropolitan counties (crude mortality rate of 51.64 per 100,000) compared to 20,078 deaths in micropolitan (73.72 per 100,000), and 15,711 deaths (83.37 per 100,00) in noncore counties (Figure 1). In other words, if micropolitan and noncore counties had the same mortality rates as metropolitan areas, there would have been 11,992 fewer deaths in nonmetropolitan (micropolitan and noncore combined) counties. Crude mortality rates also varied across levels of rurality and U.S. Census Region. Overall, the Midwest experienced the highest mortality rate during this time (63.94 deaths per 100,000 population). The Midwest also had the highest mortality rates across each level of rurality, 94.24, 83.30, and 57.14 per 100,000 across noncore, micropolitan, and metropolitan counties, respectively. Within each region, noncore mortality rates were highest, except for the Northeast, which had the highest rate in micropolitan counties. Figure 1: Crude COVID-19 Mortality Rates by Rurality and U.S. Census Region 100 94.24 g 90 83.37 83.6 - . 3 T 79.83 o 833 § 80 73.72 73.31 & 70 672 63.94 64.18 W gp 55.35 57.86 56.4 7.14 g 1.64 50.79 53.1 2.252.91 S 50 5.55 E - < 40 S 30 S a 20 = o 10 0 All Northeast South Midwest West EAIl ENoncore [ Micropolitan @ Metropolitan Across states, the highest mortality rates tended to occur in the Midwest and South (Figure 2; specific rates are reported in Appendix 1). In metropolitan and micropolitan areas, the highest quartiles of mortality rates were scattered throughout the Midwest, South, and West. In noncore areas, the Midwest and South both had four states in the top quartile of deaths. Five states-Arizona, Arkansas, New Mexico, Pennsylvania, and South Dakota-were in the top quartile of overall mortality rates as well as metropolitan, micropolitan, and noncore rates. The Pacific Northwest and northern New England were in the lowest quartile across all maps. Figure 2: COVID-19 Crude Mortality Rate per 100,000 by State and Rurality, December 2020-January 2021. (NOTE: States in gray either had no counties with the respective designation or had fewer than 16 deaths among that designation and rates were unstable.) Crude Mortality Rate per 100,000 by State <4266 ) 4z6s-534 [ S341-7435 =7435 Metropolitan Crude Mortality Rate per 100,000 by State <3958 [0 3958-4962 [ 4962-6698 2608 Micropolitan Crude Mortality Rate per 100,000 by State «51 Wl sz-o5 [les-ss zas Noncore Crude Mortality Rate per 100,000 by State 1 61-81 [l s-100 =101 Assessing crude mortality rates at the intersection of rurality and racial/ethnic composition yields mixed findings (Table 1). With one exception (AI/AN populations of 20% or greater), there is a trend of higher crude mortality rates with increasing rurality (i.e., noncore rates are higher than micropolitan, micropolitan are higher than metropolitan. The difference based on percentage of the population is more complicated, with mostly higher mortality rates in areas with a higher percentage of AI/AN and Hispanic populations, but mostly lower mortality rates in areas with a higher percentage of Black population. As such, rates were highest among micropolitan counties with 20 percent or more of the population being AI/AN (103.72 per 100,000) and among noncore counties in which 20 percent or more of the population is Hispanic (99.40 per 100,000). Table 1: Crude COVID-19 Mortality Rates per 100,000 by Rurality and Racial/Ethnic Composition, December 2020-January 2021 All Noncore | Micropolitan | Metropolitan (n=3,142 | (n=1,335 (n=641 (n=1,166 counties) | counties) counties) counties) All 55.35 83.37 73.72 51.64 Counties where 20 percent or more of the 47.25 79.02 69.98 44.34 population is Black {(n=497 counties) Counties where less than 20 percent of the 57.61 84.42 74.33 53.81 population is Black {n=2,645 counties) Counties where 20 percent or more of the 83.24 75.05 103.72 72.22 population is American Indian/Alaska Native (n=63 counties) Counties where less than 20 percent of the 55.21 83.63 73.14 51.60 population is American Indian/Alaska Native (n=3,079 counties) Counties where 20 percent or more of the 53.21 99.40 79.86 51.82 population is Hispanic (n=396 counties) Counties where less than 20 percent of the 56.52 82.04 72.89 51.53 population is Hispanic (n=2,746 counties) Counties with no ICU beds had higher mortality rates (70.03) than both the third (52.63) and fourth (55.66) quartiles of ICU bed density (Table 2). However, there was no apparent pattern between the density of ICU beds within the county and the crude mortality rate across rurality groupings. With increasing density of nursing home beds, the crude COVID-19 mortality rate increased overall and within noncore, micropolitan, and metropolitan counties. Noncore counties with the highest density of nursing home beds had the highest mortality rate (106.90 per 100,000). Table 2: Crude COVID-19 Mortality Rates per 100,000 by Rurality and ICU and Nursing Home Bed Density, December 2020-January 2021 All Noncore Micropolitan Metropolitan (n=3,142 (n=1,335 (n=641 counties) (n=1,166 counties) counties) counties) ICU Bed Density per 100,000 No Beds* (n=1,590 counties) 70.03 83.08 70.05 57.75 Quartile 3 (>0-18.02 per 100,000) 52.63 78.72 71.46 50.74 (n=766 counties) Quartile 4 ( 18.03+ per 100,000) 55.66 86.62 77.06 52.24 (n=786 counties) Nursing Home Bed Density per 100,000 Quartile 1 (<416.38 per 100,000) 45.61 58.23 47.09 45.33 (n=785 counties) Quartile 2 (416.38-697.42 per 53.69 70.05 68.25 51.92 100,000) (n=786 counties) Quartile 3 (697.43-1097.5 per 75.98 81.62 87.05 72.42 100,000) (n=785 counties) Quartile 4 (1097.6 per 100,000) 94.33 106.90 91.17 82.46 (n=786 counties) *The median number of ICU beds per 100,000 across counties is 0, and thus quartiles 1 and 2 are represented by the "no beds" category. 5 Discussion We examined COVID-19 mortality between December 2020 and January 2021 across nonmetropolitan (micropolitan and noncore) and metropolitan counties by community characteristics. During this time, the highest COVID-19 crude mortality rates were seen in noncore counties across the U.S. and regionally, except for the Northeast. Regionally, the Midwest experienced the highest mortality rates. At the intersection of race/ethnicity and rurality, the highest rates in micropolitan areas occurred among counties with more than 20 percent AI/AN population, followed by noncore areas with 20 percent or more Hispanic populations. Lastly, crude COVID-19 mortality rates were higher with increases in nursing home bed densities in all three geographies. Overall, we did not observe any pattern between crude COVID-19 mortality rates and ICU bed densities in noncore, micropolitan, and metropolitan counties. The finding of the highest mortality rates in noncore counties is aligned with earlier studies assessing rates in the spring and early summer of 2020.12 Despite being more geographically isolated, a potentially protective factor against the spread, most rural counties experienced notable increases in deaths. The higher mortality rates in noncore counties may reflect the limited health care infrastructure, higher proportions of the elderly population, and higher rates of comorbidities. We found that the Midwest experienced particularly high rates of COVID-19 mortality across county rurality, but especially in micropolitan and noncore counties. Regional differences may, in part, reflect the Northeast experiencing a peak surge early in 2020 and the South experiencing a surge during the summer of 2020. Further, analysis from the Kaiser Family Foundation shows vast swaths of counties in the Upper Midwest lacking ICU beds.!3 However, our analysis did not necessarily show higher COVID-19 mortality rates in noncore or micropolitan counties lacking ICU beds, which may reflect the ability of rural hospitals to be adaptive and collaborative during a crisis.* The highest COVID-19 mortality rates were observed in micropolitan counties with 20 percent or more AI/AN and noncore counties with greater than 20 percent Hispanic population. These results align with a recent study by the Kaiser Family Foundation that evaluated COVID-19 mortality rates by racial and ethnic composition.!3 The increased COVID-19 mortality rates in AI/ANs and Hispanic populations residing in micropolitan and noncore counties could be due to living circumstances such as holding a job that does not allow working from home, or lack of availability of public health services. In particular, many rural Hispanic communities may be more at risk due to holding jobs that put them at high risk for exposure (e.g., meatpacking), which have been shown earlier in the pandemic to be associated with higher rates of COVID- 19 cases and deaths. Both Hispanics and American Indian/Alaska Natives also are more likely to live in multigenerational living arrangements which also may contribute to spread of the virus, 16-18 Areas of higher nursing home bed density also had higher COVID-19 mortality rates during this time period. Although this was true regardless of rurality, noncore areas with the highest quartile of nursing home bed density also reported the highest mortality rates. While we report mortality rates only by county, other research reports that nursing home deaths were also highest in December 2020 and January 2021.° Previous work has also shown that staffing shortages in rural nursing homes accelerated COVID-19 deaths.?° The particularly high COVID-19 mortality rates in noncore, Midwestern, and nonmetropolitan areas with higher proportions of Hispanic and AI/AN residents are concerning. While some of these mortality disparities may reflect the higher proportions of older adults in rural areas, Hispanic communities tend to be younger, and Hispanic populations are the fastest-growing populations in rural areas.?! It is important to ensure that these populations as well as AI/AN populations have adequate access to important acute care services. Our findings underscore the importance of a well-resourced rural health care infrastructure that can respond to surges in novel viruses and other health care emergencies. Similarly, it is imperative that nursing home settings have the resources to provide adequate infection monitoring, isolation procedures, and staffing policies to protect residents and staff. Future work should continue to monitor inequities in COVID-19 deaths in rural areas, particularly as uptake of vaccines has been lower in rural communities.?? References 1. Centers for Disease Control and Prevention. Trends in COVID-19 Cases and Deaths in the United States, by County-level Population Factors. Accessed August 1, 2022. https://covid.cdc.gov/covid-data-tracker/#pop-factors totaldeaths 2. Kauffman BG, Whitaker R, Pink G, Holmes GM. Half of Rural Residents at High Risk of Serious Iliness Due to COVID-19, Creating Stress on Rural Hospitals. J Rural Health. 2020; 36 (4):584-590. 3. Cheng KJG, Sun Y, Monnat SM. COVID-19 Death Rates Are Higher in Rural Counties With Larger Shares of Blacks and Hispanics. Journal of Rural Health. 2020;36(4):602-608. doi:10.1111/jrh.12511 4. Glance L, Thirukumaran CP DA. The Unequal Burden of COVID-19 Deaths in Counties With High Proportions of Black and Hispanic Residents. Medical Care. 2021;59(6):470-476. 5. Yang BK, Carter MW, Nelson HW. Trends in COVID-19 cases, deaths, and staffing shortages in US nursing homes by rural and urban status. Geriatr Nurs. 2021;42(6):1356- 1361. doi:10.1016/j.gerinurse.2021.08.016 6. Saitone TL, Aleks Schaefer K, Scheitrum DP. COVID-19 morbidity and mortality in U.S. meatpacking counties. Food Policy. 2021;101:102072. doi:10.1016/j.foodpol.2021.102072 7. Johns Hopkins Coronavirus Resource Center. COVID-19 Map - Accessed May 12, 2021. https://coronavirus.jhu.edu/map.html 8. United States Department of Agriculture. Urban Influence Codes. Access May 12, 2021. https://www.ers.usda.gov/data-products/urban-influence-codes.aspx 9. U.S. Census Bureau. American Community Survey. Accessed May 12, 2021. https://data.census.gov/cedsci/ 10.American Hospital Association. American Hospital Association Annua Survey Database Access May 12, 2021 https://www.ahadata.com/aha-annual-survey-database 11.Centers for Medicare and Medicaid Services. Accessed May 12, 2021. https://www.medicare.gov/care-compare/?providerType=NursingHome&redirect=true 12.Karim SA and Chen HF. Deaths From COVID-19 in Rural, Micropolitan, and Metropolitan Areas: A County-Level Comparison. Journal of Rural Health. 2021; 37(1): 124-132. 13.Kaiser Health News. Millions Of Older Americans Live In Counties With No ICU Beds As Pandemic Intensifies. Accessed May 12, 2021. https://khn.org/news/as-coronavirus- spreads-widely-millions-of-older-americans-live-in-counties-with-no-icu-beds/ 14 Atkinson MK, Hick JL, Singer SJ, Cagliuso NV, Biddinger PD. Do Rural Hospitals Have An Advantage During A Pandemic? Health Affairs Forefront. March 22, 2022. https://www.healthaffairs.org/do/10.1377/forefront.20220321.42134/ 7 15.Ndugga N, Hill L, Artiga S, Haldar S. Latest Data on COVID-19 Vaccinations by Race/Ethnicity. Accessed August 1, 2022. https://www.kff.org/coronavirus-covid- 19/issue-brief/latest-data-on-covid-19-vaccinations-by-race-ethnicity/ 16.Saitone TL, Aleks Schaefer K, Scheitrum DP. COVID-19 morbidity and mortality in U.S. meatpacking counties. Food Policy. 2021 May; 101: 102072. 17.Albrecht DE. Factors Explaining Variations in COVID 19 Deaths in Rural America. Journal of Rural Social Sciences.2022; 37(2) Article 2. Available at: https://egrove.olemiss.edu/jrss/vol37/iss2/2 18.Vandenberg A. The Impacts and Implications of COVID-19 on Household Arrangements. October 2021. https://aspe.hhs.gov/sites/default/files/documents/f8348b44ab9008397797e50935ec168 8/impacts-and-implications-covid-19-household-arrangements. pdf 19.Chidambaram P and Garfield R. Nursing Homes Experienced Steeper Increase In COVID- 19 Cases and Deaths in August 2021 Than the Rest of the Country. Accessed August 1, 2022 https://www.kff.org/coronavirus-covid-19/issue-brief/nursing-homes-experienced- steeper-increase-in-covid-19-cases-and-deaths-in-august-2021 -than-the-rest-of-the- country/ 20.Yang BK, Carter MW, Nelson HW. Trends in COVID-19 cases, deaths, and staffing shortages in US nursing homes by rural and urban status. Geriatr Nurs. 2021 Nov- Dec;42(6):1356-1361. doi: 10.1016/j.gerinurse.2021.08.016 21.Lichter DT and Johnson KM. A Demographic Lifeline? Immigration and Hispanic Population Growth in Rural America. Population Research and Policy Review. 2020; 39, 785-803. 22.Ullrich F and Mueller K. COVID-19 Cases and Vaccination Rates. https://rupri.public- health.uiowa.edu/publications/policybriefs/2021/COVID%?20Vaccinations%20and%20Case s.pdf Preferred Citation: Zahnd, WE; Gadag, K; Ullrich, F; Mueller, K. COVID-19 Mortality Rates Across Noncore, Micropolitan, and Metropolitan Counties by Community Characteristics, December 2020 - January 20221. RUPRI Center for Rural Health Policy Analysis, Brief 2023-1. Appendix Table 1: Crude Rates of COVID-19 Mortality by State and County Rurality, December 2020-January 2021 State All Noncore Micropolitan Metropolitan Deaths | Crude |Deaths| Crude | Deaths| Crude | Deaths| Crude Mortality Mortality Mortality Mortality Rate per Rate per Rate per Rate per 100,000 100,000 100,000 100,000 Alabama 4200 83.34 602 97.47 574 108.55 3024 81.33 Alaska 144 19.50 40 20.71 ** ** 103 20.63 Arizona 6499 93.56 157 154.33 370 150.06 5972 90.51 Arkansas 2378 79.51 611 107.61 481 84.50 1286 69.38 California 21934 56.03 77 28.48 195 34.68 21662 56.54 Colorado 2883 52.12 282 91.80 244 62.17 2357 48.78 Connecticut 2027 56.60 92 50.27 1935 56.94 Delaware 485 51.08 485 51.08 District of 233 34.04 233 34.04 Columbia Florida 8166 39.64 225 64.93 200 55.17 7741 38.92 Georgia 4683 45.48 642 81.37 648 65.44 3393 39.83 Hawaii 165 11.60 19 7.06 146 12.66 Idaho 824 48.82 83 59.48 154 36.73 587 51.99 Illinois 7025 54.79 570 94.60 824 94.54 5631 49.62 Indiana 5026 75.22 401 87.44 1003 99.87 3622 70.00 Iowa 2585 82.52 1070 136.93 404 81.91 1111 59.80 Kansas 2426 83.40 571 146.90 614 112.76 1241 62.82 Kentucky 1919 43.22 654 66.67 477 56.04 788 30.21 Louisiana 2491 53.41 361 100.98 277 68.95 1853 47.46 Maine 398 29.86 122 28.86 27 22.21 249 31.58 Maryland 2527 42.09 75 91.64 34 48.94 2418 41.32 Massachusetts | 3827 56.03 *¥ *¥ 21 23.80 3806 56.55 Michigan 6169 61.95 670 100.96 782 68.96 4717 57.81 Minnesota 2623 47.45 369 64.02 444 67.12 1810 42.20 Mississippi 2250 75.28 668 101.02 756 79.72 826 59.89 Missouri 3077 50.52 547 66.00 438 61.22 2092 46.02 Montana 578 55.48 172 46.89 168 53.08 238 66.40 Nebraska 983 51.61 232 68.48 228 69.59 523 42.24 Nevada 2134 73.01 29 88.69 175 73.01 1930 72.82 New 530 39.45 194 42.98 332 39.32 Hampshire New Jersey 4511 50.79 4511 50.79 New Mexico 1721 82.24 121 131.68 620 103.54 980 69.91 New York 8946 45.60 233 59.54 679 68.7 8034 44.05 North Carolina 4150 40.86 398 62.99 863 54.84 2889 36.34 North Dakota 499 66.34 182 92.23 133 73.99 184 49.05 Ohio 9230 79.28 435 96.73 1814 94.94 6981 75.22 Oklahoma 1804 46.04 280 51.72 416 51.87 1108 43.03 Oregon 1047 25.65 30 30.21 191 33.94 826 24.15 State All Noncore Micropolitan Metropolitan Deaths | Crude |Deaths| Crude | Deaths| Crude | Deaths| Crude Mortality Mortality Mortality Mortality Rate per Rate per Rate per Rate per 100,000 100,000 100,000 100,000 Pennsylvania 11311 88.43 523 128.31 1472 138.57 9316 82.29 Rhode Island 783 74.10 783 74.10 South 2685 54.18 205 67.38 300 67.63 2180 51.80 Carolina South Dakota 844 97.65 341 156.51 206 89.37 297 71.41 Tennessee 4961 74.59 650 100.42 832 97.43 3479 67.56 Texas 15313 54.91 1519 107.12 1573 96.26 12221 49.21 Utah 788 25.88 121 83.48 37 20.86 630 23.14 Vermont 105 16.80 17 10.45 ** ** 67 30.73 Virginia 2498 29.69 459 58.98 172 64.20 1867 25.34 Washington 1596 21.88 32 20.50 120 20.86 1444 22.00 West Virginia 1378 75.18 278 69.54 166 55.24 931 82.48 Wisconsin 3000 51.92 478 66.35 482 62.04 2040 47.66 Wyoming 383 65.83 109 68.19 137 56.221 137 76.84 Cells that have been "grayed out" have no counties of the corresponding type; cells with ** indicate that there were fewer than 16 deaths within this category during this time indicating an unstable rate that is therefore suppressed. 10