Methodology and Data Sources

Rural literature uses differing definitions of “rural” or “non-metro” areas and different data sources also define “rural” in different ways. The United States Department of Agriculture’s (USDA) “What Is Rural” page includes definitions for the most common—Office of Management and Budget definitions and U.S. Census Bureau definitions. An explanation of each is here. We also recommend the Center on Rural Innovation’s excellent analysis explaining how the different definitions of rural can lead to different conclusions about the composition and conditions in rural communities.

The Pew Survey Research report cited throughout this report used respondents’ self-descriptions of their communities for their survey data and U.S. Census Bureau Data on county-level classifications created by the National Center for Health Statistics at the Centers for Disease Control and Prevention for the demographic analysis.

The Caregiving in the United States 2020 study used respondents’ self-descriptions of their communities and the community of their care recipient to identify rural respondents. The USDA sources cited throughout this paper use rural and non-metro interchangeably and use the metro and non-metro definition defined by the Office of Management and Budget.

Our original data analysis of the 2017-2018 American Time Use Survey Leave and Job Flexibilities Module, the 2019 National Health Interview Survey and the 2021 Behavioral Risk Factor Surveillance System (BFRSS) survey used census data to code individual respondents to non-metropolitan (used interchangeably with rural) and metropolitan counties. We used the data within the survey to identify access and experiences among subgroups of interest (e.g., poverty level, race/ethnicity).

Our original data analysis of distances traveled to different types of hospital-based health care services used data from the American Hospital Association Survey to determine which services were available at each hospital and to determine each hospital’s location. Each census tract’s location was defined as the geographic center of the census tract, with data coming from the 2019 U.S. Census Gazetteer files. Census tracts were defined as urban or rural using Rural Urban Commuting Area (RUCA) Codes, which assign a value of 1 (Metropolitan Core) to 10 (most rural), with RUCA codes of 1-3 being considered urban and codes of 4-10 being considered rural. Demographic information (such as race/ethnicity, poverty level) on all census tracts was obtained from the U.S. Census and from the 2019 American Community Survey 5-year estimates.

Information on skilled nursing facility locations was obtained from LTC Focus. Geographic files used in maps were obtained from IPUMS National Historical Geographic Information System at the census tract level. LTCFocus is sponsored by the National Institute on Aging (1P01AG027296) through a cooperative agreement with the Brown University School of Public Health. Travel distances were calculated using the Open Source Routing Machine package in Stata 17 and R.

There are multiple ways to define a high concentration of minoritized populations, including using fixed cutoffs (like 20 percent) or using a percentile of the distribution like the 90th or 95th, which has been done in prior studies. For this study we defined highly concentrated Black and Hispanic/Latine areas as being in the 90th or 95th percentile by proportion of Latine or Black population in the census tract. The 90th and 95th percentile cutoffs were calculated separately for urban and rural areas. Analysis using both the 90th and 95th percentiles yielded similar findings, so we reported only the 90th percentile figures in the report. Census tracts classified as having persistent poverty were census tracts where 20 percent or more of the population was below the poverty line in the 1980, 1990, 2000 censuses, and 2007-2011 American Community Survey five-year average.¹ To date, Rural Urban Commuting Area (RUCA) codes are only available for the census tracts as they were defined from 2010-2019 with new RUCA classifications for the 2020 revised census tracts forthcoming, so we used 2019 data for all census-related data, including estimates from the American Community Survey on racial/ethnic composition and census tract locations.