Abstract
This cross-sectional study analyzes county-level eligibility, participation, adherence, and completion rates for cardiac rehabilitation services among Medicare beneficiaries.
Among the myriad of barriers to center-based cardiac rehabilitation (CR) use in the US, geography is highlighted as an important factor in rural area residents’ reduced CR use as travel distances to CR centers increase.1,2,3 However, data reported in US studies are decades old, and CR use has yet to be described nationally across modern urbanization classifications.1,3 Characterizations of urban vs rural areas and CR use in non–US countries are also not generalizable to the US.1,3 Clarifying modern urban-rural disparities in center-based CR use throughout the US may generate the information needed to develop solutions for increasing CR use.1,2,3 We aimed to characterize geographical and urban-rural patterns in CR eligibility and center-based CR use throughout the US.
Methods
In this cross-sectional study, we queried the Centers for Disease Control and Prevention Interactive Atlas of Heart Disease and Stroke database for county-level CR eligibility rates (per 1000 beneficiaries), participation rates (percent eligible who completed ≥1 session), adherence rates (mean number of sessions completed ≤365 days of qualifying event), and completion rates (percent completing ≥36 sessions) for Medicare fee-for-service beneficiaries 65 years or older who met at least 1 primary or secondary CR qualifying indication from 2017 to 2018.4 Numbers of hospitals offering center-based CR per county were also queried from the database. This study was exempt from review and the informed consent requirement per the Common Rule because only publicly available, deidentified data were used. We followed the STROBE reporting guideline.
The National Center for Health Statistics defines 6 urbanization classifications characterizing a county as either 1 of 4 urban metropolitan statistical area types or 1 of 2 rural area types.5 The most to least urban areas included large central metropolitan, large fringe metropolitan, medium metropolitan, and small metropolitan.5 The least to most rural areas included micropolitan and noncore.5
We reported data as medians (IQRs). Generalized linear models were used to test the differences between urbanization classifications using a negative binomial or Poisson distribution. Multivariable models included the interaction between urbanization classification and geographical region (Midwest, Northeast, South, or West) with or without the center-based CR availability covariable.
Two-tailed P < .05 indicated significance. Analyses were performed using SAS, version 9.4 (SAS Institute), from June to September 2022.
Results
Residents of large urban areas had access to the greatest number of CR centers per county, yet CR eligibility, participation, and completion were significantly lower than the highest levels observed among noncore rural area residents with access to the least number of centers (Table). However, significant interactions between geographical region and urbanization classification signaled urban-rural associations with CR eligibility (χ2 = 39.48; P < .001), and participation varied significantly nationwide (χ2 = 162.44; P < .001) (Figure). Unlike rural patterns in Midwestern, Northeastern, and Western regions, high CR eligibility rates among Southern rural area residents were contrasted by participation rates that were among the lowest observed nationally. Adjusting multivariable models for CR center density per county did not yield associations that differed significantly from those of unadjusted models (Figure).
Table. Cardiac Rehabilitation Eligibility, Center-Based Use, and Hospital-Based Cardiac Rehabilitation Availability by Urbanization Classificationsa.
| Variable | Urbanization classification | |||||
|---|---|---|---|---|---|---|
| Urban areas | Rural areas | |||||
| Large central metropolitan | Large fringe metropolitan | Medium metropolitan | Small metropolitan | Micropolitan | Noncore | |
| Eligibility rate per 1000 beneficiaries, median (IQR) | 13.6 (12.1-15.1)b | 15.4 (13.4-18.0)c | 15.2 (13.1-18.1)c | 16.6 (13.6-19.9)d | 17.3 (14.6-20.8) | 17.6 (14.4-21.1) |
| Counties estimated, No. (%) | 68 (100) | 366 (99) | 368 (99) | 352 (98) | 633 (99) | 1244 (93) |
| Participation rate, median (IQR), % | 25.5 (14.9-32.8)b | 31.2 (22.8-37.6)e | 29.6 (21.3-39.2)f | 33.2 (22.4-44.4) | 33.3 (22.2-43.2) | 36.2 (23.0-52.5)g |
| Counties estimated, No. (%) | 68 (100) | 336 (91) | 333 (90) | 320 (89) | 540 (84) | 756 (57) |
| Adherence rate: No. of sessions completed, median (IQR) | 25.8 (24.2-27.1) | 25.5 (23.2-27.7) | 25.2 (22.6-27.3) | 25.2 (21.7-27.6) | 25.8 (22.0-28.2) | 24.4 (20.6-28.2)h |
| Counties estimated, No. (%) | 68 (100) | 336 (91) | 333 (90) | 320 (89) | 540 (84) | 756 (57) |
| Completion rate, median (IQR), % | 27.5 (21.4-37.6)i | 31.0 (22.0-41.7)j | 31.0 (20.4-41.2)j | 34.1 (23.8-44.3)j | 38.6 (27.3-49.9) | 41.2 (30.0-53.2) |
| Counties estimated, No. (%) | 67 (99) | 247 (67) | 243 (65) | 191 (53) | 299 (47) | 221 (17) |
| Hospitals offering CR per county, median (IQR), No. | 4.0 (2.5-6.0)b | 1.0 (0.0-1.0)c | 1.0 (0.0-2.0)c | 1.0 (0.0-1.0)k | 1.0 (0.0-1.0) | 0.0 (0.0-1.0)g |
| Counties estimated, No. (%) | 68 (100) | 366 (99) | 368 (99) | 352 (98) | 633 (99) | 1244 (93) |
| Counties with no hospitals offering CR, No. (%) | 0 | 150 (41) | 148 (40) | 130 (37) | 273 (43) | 833 (67)g |
| Beneficiary per CR center ratio per 1000 beneficiaries, median (IQR) | 15.6 (10.4-22.1) | 4.7 (0.0-14.0) | 5.4 (0.0-12.8) | 4.8 (0.0-10.4) | 2.7 (0.0-5.3)l | 0.0 (0.0-1.3)g |
| Counties estimated, No. (%) | 68 (100) | 366 (99) | 368 (99) | 352 (98) | 633 (99) | 1244 (93) |
Abbreviation: CR, cardiac rehabilitation.
Urbanization classifications reflect 2010 US Census Bureau geography and urban-rural code structure defined in 2013 by the National Center for Health Statistics. The maximum number of counties where any of the 4 CR metrics could be estimated in 2017 to 2018 was 3031 of a possible 3142. Beneficiary per hospital center–based CR availability ratio was ascertained by dividing the per county total number of residents who were Medicare fee-for-service beneficiaries aged 65 years or older by the number of hospitals per county offering center-based CR.
Large central metropolitan results were different from each urbanization classification after Tukey-Kramer post hoc testing; all comparisons had P < .002.
Large fringe and medium metropolitan areas results were each different from small metropolitan, micropolitan, and noncore after Tukey-Kramer post hoc testing; all comparisons had P < .02.
Small metropolitan results were different from noncore after Tukey-Kramer post hoc testing, with P < .001.
Large fringe results were different from small metropolitan after Tukey-Kramer post hoc testing, with P = .045.
Medium metropolitan results were different from small metropolitan and micropolitan after Tukey-Kramer post hoc testing; all comparisons had P < .02.
Noncore results were different from each urbanization classification after Tukey-Kramer post hoc testing; all comparisons had P < .002.
Noncore results were different from large fringe and micropolitan after Tukey-Kramer post hoc testing; all comparisons had P < .04.
Large central results were different from micropolitan and noncore after Tukey-Kramer post hoc testing; all comparisons had P < .001.
Large fringe, medium, and small metropolitan areas were different from micropolitan and noncore after Tukey-Kramer post hoc testing; all comparisons had P < .01.
Small metropolitan results were different from micropolitan and noncore after Tukey-Kramer post hoc testing, with P < .006.
Micropolitan results were different from each urbanization classification after Tukey-Kramer post hoc testing; all comparisons had P < .03.
Figure. Cardiac Rehabilitation Eligibility and Center-Based Use Among Medicare Fee-for-Service Beneficiaries Stratified by Urbanization Classification and Geographical Region.
Metro indicates metropolitan.
aBetween-region comparison at a given urbanization classification; significantly different from the South region.
bBetween-region comparison at a given urbanization classification; significantly different from the Midwest region.
cWithin-region comparison; significantly different from large central metro.
dWithin-region comparison; significantly different from large fringe and medium metro areas.
eBetween-region comparison at a given urbanization classification; significantly different from the Northeast region.
fWithin-region comparison; significantly different from noncore.
gWithin-region comparison; significantly different from micropolitan.
Discussion
The urban-rural disparities in center-based CR use that we observed do not support the generalization that rural area residents are less likely than urban area residents to initiate and complete CR.1,3 Noncore rural area residents demonstrated the highest participation rates in 3 of 4 regions, and their completion rates consistently exceeded those of large urban area residents.
A study limitation is an inability to explain why noncore rural area residents with access to the least number of CR centers exhibited patterns inconsistent with the generalization that CR center availability is a factor in CR use, particularly among financially vulnerable and underserved populations in either rural or urban areas.1,2,3,6 Although no causal associations can be established from the findings, this study presented contemporary observations that may help shape future discussions on urban-rural health policies and practice-level interventions to increase CR use throughout the US.
References
- 1.Mathews L, Brewer LC. A review of disparities in cardiac rehabilitation: evidence, drivers, and solutions. J Cardiopulm Rehabil Prev. 2021;41(6):375-382. doi: 10.1097/HCR.0000000000000659 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ades PA, Keteyian SJ, Wright JS, et al. Increasing cardiac rehabilitation participation from 20% to 70%: a road map from the Million Hearts Cardiac Rehabilitation Collaborative. Mayo Clin Proc. 2017;92(2):234-242. doi: 10.1016/j.mayocp.2016.10.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Valencia HE, Savage PD, Ades PA. Cardiac rehabilitation participation in underserved populations. Minorities, low socioeconomic, and rural residents. J Cardiopulm Rehabil Prev. 2011;31(4):203-210. doi: 10.1097/HCR.0b013e318220a7da [DOI] [PubMed] [Google Scholar]
- 4.Thomas RJ, Balady G, Banka G, et al. 2018 ACC/AHA clinical performance and quality measures for cardiac rehabilitation: a report of the American College of Cardiology/American Heart Association Task Force on Performance Measures. J Am Coll Cardiol. 2018;71(16):1814-1837. doi: 10.1016/j.jacc.2018.01.004 [DOI] [PubMed] [Google Scholar]
- 5.Ingram DD, Franco SJ. 2013 NCHS urban-rural classification scheme for counties. Vital Health Stat 2. 2014;(166):1-73. [PubMed] [Google Scholar]
- 6.Suaya JA, Shepard DS, Normand SL, Ades PA, Prottas J, Stason WB. Use of cardiac rehabilitation by Medicare beneficiaries after myocardial infarction or coronary bypass surgery. Circulation. 2007;116(15):1653-1662. doi: 10.1161/CIRCULATIONAHA.107.701466 [DOI] [PubMed] [Google Scholar]