Refining a traditional urban-rural classification approach to better assess heterogeneity of treatment effects in patient-centered outcomes research

Satya Surbhi; Elizabeth A Tolley; Ronald E Cossman; Ankur A Dashputre; James E Bailey

doi:10.1016/j.mex.2021.101299

Refining a traditional urban-rural classification approach to better assess heterogeneity of treatment effects in patient-centered outcomes research

MethodsX. 2021 Mar 8:8:101299. doi: 10.1016/j.mex.2021.101299. eCollection 2021.

Authors

Satya Surbhi^{1

2}, Elizabeth A Tolley^{1

3}, Ronald E Cossman⁴, Ankur A Dashputre^{1

5}, James E Bailey^{1

2

3}

Affiliations

¹ Center for Health System Improvement, College of Medicine, University of Tennessee Health Science Center, Memphis, TN, United States.
² Division of General Internal Medicine, Department of Medicine, University of Tennessee Health Science Center, Memphis, TN, United States.
³ Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, TN, United States.
⁴ Social Science Research Center, Mississippi State University, Starkville, MS, United States.
⁵ Institute for Health Outcomes and Policy, College of Graduate Health Sciences, University of Tennessee Health Science Center, Memphis, TN, United States.

Abstract

This article describes a rationale and approach for modifying the traditional rural-urban commuting area (RUCA) coding scheme used to classify U.S. ZIP codes to enable suburban/rural vs. urban core comparisons in health outcomes research that better reflect current geographic differences in access to care in U.S. populations at risk for health disparities. The proposed method customization is being employed in the Patient-Centered Outcomes Research Institute-funded Management Of Diabetes in Everyday Life (MODEL) study to assess heterogeneity of treatment effect for patient-centered diabetes self-care interventions across the rural-urban spectrum. The proposed suburban/rural vs. urban core classification scheme modification is based on research showing that increasing suburban poverty and rapid conversion of many rural areas into suburban areas in the U.S. has resulted in similar health care access problems in areas designated as rural or suburban.•The RUCA coding scheme was developed when a much higher percentage of U.S. individuals resided in areas with very low population density.•Using the MODEL study example, this study demonstrates that the RUCA classification scheme using ZIP codes does not reflect real differences in health care access experienced by medically underserved study participants.•Both internal and external validation data suggest that the proposed suburban/rural vs. urban core customization of the RUCA geographic coding scheme better reflects real differences in healthcare access and is better able to assess the differential impact of clinical interventions designed to address geographic differences in access among vulnerable populations.

Keywords: African American; Diabetes; Medically underserved populations; Multiple chronic conditions; Patient-centered outcomes research; Urban-rural classification.