Guidelines For Using Rural-Urban Classification Systems for Public Health Assessment

Purpose

Why a guideline on rural-urban classification systems?

What systems are commonly used to classify rural-urban character?

Which is the best system for identifying rural areas in Washington?

The Rural Urban Commuting Area (RUCA) system: a good choice

Is the RUCA system the best fit in all situations?

A suggested four-tiered consolidation of the RUCA system at the sub-county level

Is there a county-based version of RUCA codes?

How will the 2000 Census affect RUCA classifications?

Other considerations when making rural-urban comparisons

Guidelines: A recap

A summary of major rural-urban classification systems

Urban, Urbanized, and Rural Areas
Metropolitan and non-Metropolitan
Rural-Urban Continuum Codes (RUCC)
Goldsmith Modification to OMB Metropolitan – Non-Metropolitan County System
Frontier, Remote, Less Remote and Urban Counties
US Department of Agriculture (USDA) Urban Influence Codes
Rural Urban Commuting Area Codes
Four-Tiered Consolidation of RUCA Codes
Dominant RUCA County Codes
Metropolitan, Micropolitan and Outside Core-Based Statistical Areas

Other Variations

References

Guidelines For Using Rural-Urban Classification Systems for Public Health Assessment (Word Document)

Purpose

The Assessment Operation Group in the Washington State Department of Health is coordinating the development of guidelines related to data development and use in order to promote good professional practice among staff involved in assessment activities within the Washington State Department of Health and in Local Health Jurisdictions in Washington. While the guidelines are intended for an audience of differing levels of training related to data development and use, they assume a basic knowledge of epidemiology and biostatistics. They are not intended to recreate basic texts and other sources of information related to the topics covered by the guidelines, but rather they focus on issues commonly encountered in public health practice and where applicable, to issues unique to Washington state.

Why a guideline on rural-urban classification systems?

A review of recent Washington state health data and research (Schueler and Stuart, 2000) found differences in health status between residents of rural and urban Washington. The unique challenges facing rural health care and health care systems are getting more attention. Analysts looking at rural health disparities must choose from several classification systems. Guidelines are useful for promoting consistency and comparability among analyses that look at rural health. Local public health assessments might also benefit from a classification system that can be used to compare local health data to areas with similar population and settlement patterns.

This is uncharted territory. According to two of the country’s leading rural health researchers, Dr. Gary Hart at the University of Washington Rural Health Research Center and Dr. Thomas Ricketts at the Sheps Center at the University of North Carolina, no one has systematically addressed the question of how to best incorporate rural-urban classification systems into public health assessment.

What systems are commonly used to classify rural-urban character?

At least ten rural-urban classification systems are available for rural health assessment (Table 1). For detailed descriptions see A summary of major rural-urban classification systems.

Table 1: Common Rural-Urban Classification Systems

Which is the best system for identifying rural areas in Washington?

Washington state presents unique challenges in classifying rural areas because of the range in the size of its counties. The most common classification systems (for example, Metropolitan vs. Non-Metropolitan) use county geography (Figure 1). County-based systems can misclassify some areas. The likelihood of misclassification increases with the size of the county. Nationally, 14 percent of residents of Metropolitan counties, as defined by the US Office of Management and Budget, are classified as rural by Bureau of Census definitions (Ricketts et al., 1998). An analysis of Behavioral Risk Factor Surveillance System (BRFSS) data in Washington found that using a ZIP code-based classification system uncovered differences that were not apparent when responses were classified with county-based systems (Schueler and Simmons, 2000). Sub-county definitions using ZIP code or census geography are preferable to county-based systems, because they provide greater discrimination between rural and urban areas.

Washington’s rural areas are not homogenous.   There are significant demographic differences between remote, small-town rural areas (such as Republic), large towns (such as Wenatchee), and urban fringe areas.  A simple binary rural-urban classification can obscure important differences. However, the small populations in more remote rural areas often make it impractical to subdivide rural areas too finely.  This is a particular concern for sample-based data such as BRFSS and for rare health events. The ideal system would differentiate among different types of rural areas, but should be collapsible into a smaller number of classifications if needed.  For routine analyses, we recommend a three- to five- tiered system.

The Rural Urban Commuting Area (RUCA) system: a good choice

No systematic study or standards identify which definitions are most appropriate for analyzing specific types of public health data. The Office of Community and Rural Health recommends the Rural Urban Commuting Area (RUCA) system, because it is more flexible and precise than available alternatives.

The RUCA system is a ten-tiered classification system based on census tract geography. Both population size and commuting relationships are used to classify census tracts. First, urbanized (continuously built up areas of 50,000 or more), large town (10,000-49,999), and small town (2,500 to 9,999) cores areas are identified. Next, the primary (largest) and secondary (second largest) commuting flows of remaining tracts are examined using the most recently available commuting data. High commuting tracts are those where the primary or largest commuting flow is greater than 30% to a core area. Low commuting or influence area tracts are those where the largest flow to core areas is 5-30%.

The RUCA system provides a great deal of flexibility as the codes can be collapsed or combined in several different ways. Washington state RUCA codes using census tract geography are mapped in Figure 2. A ZIP code approximation is also available and is mapped in Figure 3. See Rural Urban Commuting Area Codes for a more detailed discussion of the system and information on where to obtain the codes.

Is the RUCA system the best fit in all situations?

Although the RUCA system is a good all around system, the analyst also should take into account how the classification system relates to the health phenomena being studied. Rural classification schemes are most commonly based on

• population density or clustering and/or
• degree of connection to or isolation from a urban core, as measured by commuting patterns or proximity to urban areas.

These approaches measure different population characteristics. For example, the prevalence of AIDS is closely linked to population centers. In studying AIDS migration to rural areas, rural classification systems that identify population clusters (including small and large town cores) are most appropriate and likely to identify changes over time. For health analyses where incidence may be tied to access to treatment, definitions using proximity or degree of connection to urban core areas are more appropriate. In this second case, it may be best to use county Urban Influence Codes or a combination of RUCA codes that emphasizes commuting relationships.

A suggested four-tiered consolidation of the RUCA system at the sub-county level

Many data sets will not support analysis using a ten-tiered classification system. The RUCA system can be collapsed in several ways. For general analyses of sub-county data, we suggest a four-tiered system.

• Urban Core Areas - continuously built up areas 50,000 persons or more. These areas correspond to US Bureau of the Census defined Urbanized Areas.
• Suburban Areas - areas with high commuting relationships with Urban Core Areas. Suburban areas include Large Town, Small Town and Isolated Rural Areas with high commuting levels to Urban Core Areas.
• Large Town Areas - towns with populations between 10,000 and 49,999 and surrounding rural areas with high commuting levels to these towns.
• Small Town and Isolated Rural Areas - towns with populations below 10,000 and their commuter sheds and other isolated rural areas.

The census tract version (Figure 4) is slightly more precise than the ZIP code version (Figure 5), but the ZIP code version is readily used with a greater number of public health data sets. The specifications for this consolidation are found in Table 3.

Is there a county-based version of RUCA codes?

The RUCA system was expressly developed at the census tract level to solve misclassification problems with county-based systems. In some cases, data may only be available at the county level. The Office of Community and Rural Health has developed a classification system based on the percentage of total county population residing in specific RUCA codes. There are five possible classes: Urban Dominated counties (includes Urban Core Areas and Suburban Areas), Large Town Dominated counties, Small Town/Isolated Rural Dominated counties, and two classes of counties with a mix of urban or rural characteristics. The rules for assigning dominant RUCA codes to counties are found in Table 4. Dominant RUCA codes by county are mapped in Figure 6 and listed in Table 5.

How will the 2000 Census affect RUCA classifications?

Most rural-urban classification systems rely on 1990 census data for population counts and commuting information to establish adjacency. Areas that have grown rapidly in the last decade may be misclassified. The greatest opportunity for misclassification is on the urban/rural fringe and between the urban and large town classifications.  Rural-urban classification systems will be updated when the 2000 census data on commuting behavior become available in 2002.  The US Bureau of the Census anticipates updated RUCA codes will be available by fall 2002.

Other considerations when making rural-urban comparisons

All population-based health indicators comparing urban and rural areas should be age-adjusted, as the proportion of elderly residents in rural areas is higher than in urban areas.  (See Rates guideline.)  Analysts should also keep in mind that, in general, the residents of rural Washington have lower incomes and have completed fewer years of formal education than those in other areas. Differences in health status between rural and urban Washingtonians may reflect underlying differences in demographics.

Guidelines: A recap

• If data are available at the census tract or ZIP code level, use the RUCA system.
• All rural-urban classification systems currently depend on 1990 commuting data.  Updated codes are not likely to be available until fall 2002.  Until the updated codes are released, the potential for misclassification should be noted in technical notes.
• For routine analyses we suggest collapsing the ten RUCA codes into four categories,
  • Urban Core Areas
  • Suburban Areas
  • Large Town Areas
  • Small Town and Isolated Rural Areas
• If data are only available at the county level, we recommend using the Office of Community and Rural Health’s Dominant RUCA codes.  The potential for misclassification should be discussed.
• Rural-urban differences may reflect underlying differences in demographics.  In general, rural-urban comparisons of health indicators should be age-adjusted, as the proportion of elderly residents in rural areas is higher than in urban areas.    Analysts should also keep in mind that the residents of rural Washington have lower incomes and have completed fewer years of formal education than those in other areas.
• Document your choice of a rural-urban classification system and be sensitive to each system’s limitations.

A summary of major rural-urban classification systems

This summary of rural classification methods draws heavily on Definitions of Rural: A Handbook for Health Policy Makers and Researchers (Ricketts et al., 1998), which is available at http://www.shepscenter.unc.edu/research_programs/rural_program/ruralit.pdf.

Urban, Urbanized, and Rural Areas:  The US Bureau of the Census maintains definitions of Urban, Urbanized, and Rural Areas for classifying populations. Urban populations are those residing in incorporated areas or Census Designated Places with 2,500 or more or an Urbanized Area.  An Urbanized Area (subset of Urban) is a continuously built up area of 50,000 people or more. A built up area is an area with a population density of more than 1,000 persons per square mile.  This is calculated at the census block level. Rural populations are all those not classified as Urban or Urbanized. The definition of Urban population is overly inclusive because it includes very small towns. The definition for Urbanized is not inclusive enough. Areas with a population density of 999 persons per square mile are considered Rural. See http://www.census.gov/population/censusdata/urdef.txt for a detailed definition. The US Bureau of the Census expects that updated definitions based on 2000 census data will be available in fall 2002.

Metropolitan and Non-Metropolitan: The US Office of Management and Budget (OMB) has maintained this national classification system since the 1940s. The federal government uses this system extensively for statistical reporting and allocating funds. In this system, counties with cities or urbanized areas over 50,000 are classified as Metropolitan.  Outlying counties meeting a complex set of conditions based on commuting patterns and population density are also designated Metropolitan.  All other areas are designated Non-Metropolitan. Non-Metropolitan counties are not differentiated. The low commuting thresholds (in some cases 15%) used to tie outlying counties to core Metropolitan counties result in some counterintuitive classifications.  For example, in Washington state, Island county is classified as a Metropolitan county because of commuting patterns from southern Whidbey Island and Camano Island. See Table 5 for the current list and Figure 1 for a map of Metropolitan and Non-Metropolitan counties in Washington. OMB revised this standard in December 2000. The new standard classifies counties as Metropolitan, Micropolitan and Outside Core-Based Statistical Areas.

Rural-Urban Continuum Codes (RUCC): The US Department of Agriculture’s Economic Research Service developed the RUCC system, also known as the Beale code system, in the mid-1970s.  The system uses OMB’s Metropolitan and Non-Metropolitan classifications as a starting point. Metropolitan counties are classified into four population categories. Non-Metropolitan counties are classified into six categories on the basis of total population in US Bureau of Census defined Urbanized Areas. Non-Metropolitan communities are further classified by adjacency to Metropolitan counties. Adjacent counties must be physically adjacent to a Metropolitan county and have at least 2% of the resident labor force commuting to a central Metropolitan county.  This system better differentiates between central and fringe metropolitan areas.  RUCC’s have not been developed at the sub-county level as US Bureau of Census Urbanized Areas definitions are not readily transferable to census tract and ZIP code geography. The most recent update was in 1994 using 1990 census data. For more information see http://www.ers.usda.gov/emphases/rural/data/index.htm#Beale.

Goldsmith Modification to OMB Metropolitan – Non-Metropolitan County System: The Federal Office of Rural Health Policy developed the Goldsmith Modification to OMB’s Metropolitan and Non-Metropolitan system in the 1980s to target funding to isolated rural areas in large Metropolitan counties.  The Goldsmith method has two steps. First, Metropolitan counties over 1,225 square miles are identified. Several criteria are applied within these counties to identify whether individual census tracts are isolated from large cities in the county. Because of the 1,225 square mile threshold, isolated areas in smaller Metropolitan counties, for example Whatcom county, are not identified.  This method was last updated in 1980s. The US Federal Office of Rural Health Policy which developed this classification system has discontinued use of the Goldsmith Modification in May 2002 and no longer publishes or updates the list. It has been replace with the ZIP Code version of the RUCA system. For more information on the Goldsmith Modification contact Vince Schueler at vince.schueler@doh.wa.gov.

Frontier, Remote, Less Remote and Urban Counties: The Office of Community and Rural Health developed this classification system in the mid 1990s for the “Rural Health Data Book.” The system is an amalgamation of three different approaches. Frontier counties include all counties with population density less than six persons per square mile. Remote counties are defined as not having population centers of more than 10,000 and the majority of the population is more than 30 minutes travel time from such population centers. Less Remote counties are all other Non-Metropolitan counties. Urban counties are all Metropolitan counties with the exception of Island county which was classified as Less Remote. This method has not been used outside of Washington state.

US Department of Agriculture (USDA) Urban Influence Codes: The USDA Economic Research Service developed this classification scheme in the mid-1990s to emphasize the tendency of economic systems to centralize around very large metropolitan counties.  Metropolitan counties are classified as Large Metropolitan (population >= 1 million) or Small Metropolitan (population < 1 million). Non-Metropolitan counties are classified as whether or not they are adjacent to these Large or Small Metropolitan counties using the same definition as RUCC (link paragraph above). This method is most useful for looking at the structure of health care systems and whether care or outcomes may be related to the complexity of the medical community or threshold levels of institution size (Rickets et al., 1998). This scheme is only available at the county level. It was calculated in 1993 using 1990 census data. For more information see http://www.ers.usda.gov/Briefing/Rurality/.

Rural Urban Commuting Area Codes: The RUCA system is a ten-tiered classification system based on census tract geography. Both population size and commuting relationships are used to classify census tracts. First urbanized (continuously built up areas of 50,000 or more), large town (10,000-49,999), and small town (2,500 to 9,999) core tracts are identified. Next, the primary (largest) and secondary (second largest) commuting flows of remaining tracts are examined using the most recently available commuting data. High commuting tracts are those where the primary or largest commuting flow is greater than 30% to a core area. Low commuting tracts are those where the largest flow to core areas is 5-30%. Isolated rural areas are those with no town greater than 2,500 where the primary commuting flow is local. This yields the following scheme:

Table 2: Full Rural Urban Commuting Area (RUCA) Classification System

The University of Washington’s Rural Health Research Center and Geography Department have developed a ZIP code approximation.  The ZIP code approximation and other related research and tools are available at http://www.fammed.washington.edu/wwamirhrc/.

This ten-tiered classification system was developed in the late 1990s and is rapidly gaining wide use.  It is the only system available at the census tract or ZIP code level.  For more details see http://www.ers.usda.gov/briefing/Rurality/RuralUrbanCommutingAreas/. 

Four-Tiered Consolidation of RUCA Codes: Many data sets will not support analysis using a ten-tiered classification system.  The Washington state Office of Community and Rural Health developed a Four-Tiered Consolidation of RUCA codes in 2001 for general analyses of sub-county data.

• Urban Core Areas - continuously built up areas 50,000 persons or more. These areas correspond to US Bureau of the Census defined Urbanized Areas.
• Suburban Areas - areas with high commuting relationships with Urban Core Areas. Suburban areas also include Large Town, Small Town and Isolated Rural Areas with high commuting levels to Urban Core Areas.
• Large Town Areas - towns with populations between 10,000 and 49,999 and surrounding rural areas with high commuting levels to these towns.
• Small Town and Isolated Rural Areas - towns with populations below 10,000 and their commuter sheds and other isolated rural areas.

The census tract version (Figure 4) is slightly more precise than the ZIP code version (Figure 5), but the ZIP code version is readily used with a greater number of public health data sets.

Table 3:  Four-Tiered Consolidation of RUCA Codes

Dominant RUCA County Codes:  For cases where sub-county data are not available, the Office of Community and Rural Health has classified counties by dominant RUCA codes.  To do this, we aggregated the population of census tracts within counties by RUCA code.  Counties are classified as predominantly Urban, Large Town, or Small Town Rural, using the following rules:

Table 4: Rules for Assigning Dominant RUCA Codes to Counties

Counties with less than 75% of the population residing within Urban Core, Suburban RUCAs, Large Town, or Small Town and Isolated Rural RUCAs as defined in the Four-Tiered Consolidation of RUCA Codes system are classified as mixed counties. There are currently no Mixed Urban counties in Washington.

Metropolitan, Micropolitan and Outside Core-Based Statistical Areas:  The US Office of Management and Budget posted a revision of the Metropolitan and Non-Metropolitan system in the Federal Register 12/27/00.  The Office of Management and Budget will not implement the revised system until 2003. The revised Metropolitan standard has three tiers based on the number of persons residing in Urbanized Areas within a county.

• Metropolitan - Over 50,000
• Micropolitan - 10,000 to 49,999
• Outside a Core-Based Statistical Area (CBSA) - all other counties.

In addition, any county in which at least 50% of the population resides in an Urbanized Area will be designated as Metropolitan or Micropolitan.  Any outlying county in which at least 25% of the residents commute to a Metropolitan or Micropolitan area will be designated with the core area designation.  Although the commuting thresholds used to tie in outlying counties are higher than those in the original classification system, in some cases, they are low enough so that some counties adjacent to Metropolitan counties with distinctly rural characteristics are included with Metropolitan counties.  For example, Skamania county will be considered part of the Portland Metropolitan area, although there is no community over 1,200 residents in the county.  Nonetheless, the 2000 revision is an improvement over the prior Metropolitan and Non-Metropolitan classification.  For more information on existing and proposed definitions see http://www.census.gov/population/www/estimates/masrp.html.

Other variations

Several other classification systems that apply to subsets of areas, activities, or populations may be useful for rural public health assessment.  These include several competing definitions of frontier areas and county-based typologies of primary economic activity.  See Ricketts et al. (1998) or the USDA Economic Research Service Website at http://www.ers.usda.gov/Briefing/Rurality/ for more detail.

Table 5: Rural Urban Classifications for Washington Counties

References

Ricketts TC, Johnson-Webb KD, Taylor P. Rural definitions for health policy makers. Bethesda (MD): Dept. of Health and Human Services (US), Federal Office of Rural Health Policy; 1998 July.

Schueler V, Stuart B. Recent research and data on rural health in Washington State. Olympia (WA): 2000 October.

Schueler V, Simmons K. Defining rural for public health assessment. Washington State Department of Health , Office of Community and Rural Health; 2000 October.