Considering Relative Rurality in Tobacco Regulatory Science

The overarching goal of our TCORS Center, AppalTRUST (Appalachian Tobacco Regulatory Science Team), is to evaluate the impact of the Federal Drug Administration (FDA) Center for Tobacco Products (CTP) regulatory policies in rural communities. Our Center focuses on a historically understudied population disproportionately affected by tobacco use. As of 2018-2019, Kentucky had a tobacco product use prevalence among adults that was nearly 10 points above the national average (24.8% vs. 15.4%).¹ More recently, the use rate of any tobacco product was estimated to be higher in rural Kentucky (32.4%) versus urban locations in the state (26.4%).²

To capture the variability of tobacco use in rural places, our Center purposely chose to include two rural catchment areas within Appalachian Kentucky; one inarguably rural and the other semi-rural/peri-urban, bordering on a metropolitan area. With data collection ongoing, the purpose of this commentary is to describe an adaptable method for quantifying rurality in Appalachian communities.

Rurality is challenging to define using solely quantitative or qualitative measures, since it exists in reality as a continuum, rather than simply urbanity’s binary opposite. In this commentary, we explore an alternative measure of rurality, the Index of Relative Rurality (IRR),³ that could provide novel insight into the relationship between rurality and tobacco use, which we suspect is more nuanced than suggested by previous research.

Background

In national measures, rurality has been variously defined with respect to population size, nearness to urban areas, commuting flows, and “remoteness.” The most widely used rurality measures are released decennially by the USDA Economic Research Service,⁴ including Rural-Urban Continuum (RUC) codes, Primary Rural-Urban Commuting Area (RUCA) codes, and Frontier and Remote Areas (FAR) codes. RUC codes are ordinal, based solely on proximity to the nearest metro area and the size of that area; RUCA codes extend RUC codes by also including commuting patterns, but these codes are categorical rather than ordinal; and FAR codes are not applicable to urban areas, so not all locations have them. The limitation of these measures in our context is that they are typically or necessarily dichotomized, which prevents gauging differences, such as tobacco use prevalence, among communities with varying rurality.

A Relative Measure of Rurality

In contrast to RUC, RUCA, and FAR codes, the IRR is a composite score capturing the rurality of a geographic area (e.g., census tract, zip code area, or county) relative to all other geographic areas at the same level in the United States. IRR codes are continuous, with scores closer to 0 indicating the least rural/most urban, and those closer to 1 suggesting the highest degree of rurality. IRR codes are the unweighted averages of four attributes, each rescaled to range from 0 to 1: log-transformed population size, log-transformed population density, network distance (i.e., degree of remoteness) to the nearest metropolitan county, and built-up area (urban development as a percentage of total land area).

The IRR has two immediate advantages for studying tobacco regulatory science (TRS) in rural areas: (1) it can be evaluated at any level of geography for which there are corresponding data available (whereas other measures of rural/urban status depend on fixed geographies, such as distance of a county to nearest metro); and (2) it is a continuous measure that does not rely on specific cutoffs and thus can be used as a scale variable in analysis.⁵ An additional benefit of the IRR is that it can be used to quantify rurality of the “activity space” of an individual, namely the geographic area bounded by places they visit on a regular basis. While the IRR has been used in a limited number of prior studies,^5,6 to our knowledge it has not previously been applied to TRS, despite its potential for better illustrating patterns of exposure to tobacco policies and marketing in routine crossing of administrative or geographic borders, such as access to more affordable tobacco products if one lives in a rural county but works in an urban community.

Evaluating Relative Rurality

Scholars have long debated how to conceptualize and measure rurality,^5,7 even valorizing the distinction between rurality and urbanity (e.g., as “backward” vs. “modern”).⁸ In the tobacco disparities literature, rural/urban differences have been characterized through a “cumulative disadvantage” framework, downplaying the strengths of rural communities.⁹ These stereotypical ways of framing rural and urban can render invisible the heterogeneity of experiences, identities, and inequalities.¹⁰

The 13-state Appalachian region of the United States has often been imagined as rural in public discourse, but many people in Appalachia live in cities, including Pittsburgh, PA, and Chattanooga, TN. Appalachian counties represented as “economically distressed”¹¹ may include wide economic inequalities within the county. Similarly, dominant narratives representing Appalachia through a racializing lens as white can erase the diverse identities of Appalachians.^10,12,13

“Peri-urban” is conceptualized as an ambiguous transition zone in flux on the urban fringe. Amirinejad and colleagues reviewed studies that used a variety of quantifiable markers to measure peri-urban status and behavior¹⁴; however, qualitative methods are valuable to further understand the heterogeneity within any zone and how people conceptualize their own identity and navigate the landscape in everyday ways.³

Rural, urban, and peri-urban categorizations are not fixed, nor can they be defined in uniform ways; they are relational ways to describe experience. We believe the relational aspect of the IRR helps to conceptually align it, even as a quantitative measure, with qualitative notions of rurality. That said, we will augment our understanding of this region by conducting qualitative structured interviews with participants about cues and choices related to tobacco use.

Prior research suggests that definitions of rurality are so contingent that it is useful to define rurality in relation to the topic being studied.^10,15 In AppalTRUST, the focus on access to tobacco products for purchase and interactions with tobacco marketing materials in everyday life makes our use of the terms rural and peri-urban salient for the two catchment areas purposively chosen. It becomes important, for example, whether residents have Internet access; the media markets to which they are exposed; and what might shape their patterns of crossing county lines to shop, work, or run errands for kin or those in neighboring networks—thus our emphasis on activity spaces.

Underscoring the need for evaluating relative rurality, there are myriad shifting stories of how people “place” themselves and one another.¹⁶ Just as a county categorized overall as rural may have internal urban-rural cultural distinctions between a county seat of 2000 people and an unincorporated community, a person residing in the county seat may identify at one moment as urban (e.g., talking with someone from an unincorporated community at a gas station) and as rural (e.g., speaking with a coworker who lives in the city of 350 000 where they both work).¹⁶

IRR and TRS

The IRR is useful in TRS because it allows for greater heterogeneity than binary designations of “urban” and “rural,” or even an ordinal classification system, by providing a continuous quantitative measure that incorporates multiple dimensions of rurality in assigning a value to places. This enables tobacco regulatory scientists to pose different questions about the relationship between rurality and tobacco use and employ a wider variety of analytic approaches.

Regrettably, the people who live in these places may be labeled “rural” and “urban.” But just as a suburban municipality might possess some features common to urban areas and other features common to rural locations, there are individuals who similarly exist across these classifications. Examining individuals’ movements in their daily travels, as well as the environmental features in the spaces through which they move (i.e., activity spaces), extends our understanding of rurality to reflect the individual’s unique experiences with tobacco, as conditioned by the places where they live their lives. Individuals’ activity spaces can be quantified by the mean or median IRR scores of the places they visit regularly. Figure 1 demonstrates the use of IRR to measure places visited and activity spaces for two individuals. While the relative rurality of their homes is identical, the average IRRs of their activity spaces are quite different. Future work with this measure may allow us to differentiate individuals’ unique activity spaces using other quantitative measures, such as the number or density of tobacco retailers, or coverage by smoke-free/flavor policies or tobacco advertising regulations.

Figure 1.

Index of Relative Rurality (2020) by census tract and activity space.

AppalTRUST describes places and individuals’ activity spaces in a robust way that captures the nuance of experiences with tobacco in areas often characterized monolithically as “rural.” Prior research has established higher use of cigarettes and smokeless tobacco in rural areas, based on rurality defined as binary.^17–19 Use of the IRR in AppalTRUST communities will allow us to capture the heterogeneity of rural areas and evaluate tobacco use indicators relative to this variable.

Conclusion

The IRR is a quantitative method that can distinguish nuances in rural geographical areas that have been broadly generalized previously. As we use IRR in rural TRS, we will be able to not only characterize the relative rurality of specific locations, but also that of individuals’ activity spaces. Having the ability to quantify locations and activity spaces will allow tobacco regulatory scientists to connect tobacco use behaviors and access to tobacco products with social and environmental factors to better represent the heterogeneity of people and places traditionally characterized as only “rural.”

Author Contributions

Mary Kay Rayens (Conceptualization [Equal], Writing—original draft [Equal], Writing—review & editing [Lead]), Ann Kingsolver (Conceptualization [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), W. Jay Christian (Conceptualization [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), Ellen Hahn (Writing—review & editing [Equal]), Teresa Waters (Writing—review & editing [Equal]), Shyanika Rose (Writing—review & editing [Equal]), Mikhail Koffarnus (Writing—review & editing [Equal]), and Seth Himelhoch (Writing—review & editing [Equal])

Declaration of Interests

There are no conflicts of interest to report.

Funding

Research reported in this publication was supported by grant number U54DA05825601 from the National Institutes of Health (NIH) and the Food and Drug Administration (FDA) Center for Tobacco Products. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the FDA.

Data Availability

The IRR values as described here are available for 2000 and 2010 at the following link and cited in the text. https://purr.purdue.edu/publications/2960/1.