An Examination of Geographic Proximity to Outpatient Cardiac Rehabilitation in Rural Versus Urban Tennessee Counties

Abstract

Background:

Although outpatient cardiac rehabilitation (OCR) is associated with improved outcomes post myocardial infarction (MI), authors of limited US studies report OCR travel distance/time estimates with fewer providing rural-urban comparisons.

Objective:

We examined travel distance/time to the closest OCR facility for each Tennessee county.

Methods:

We identified n = 61 Tennessee OCR facilities through a Tennessee Association of Cardiovascular and Pulmonary Rehabilitation list and a data scraping process using cardiac rehabilitation–related keywords. County-level mean travel distance/time to the closest OCR facility was determined using geospatial analysis. We conducted Kruskal-Wallis tests to examine whether mean travel distance/time varied by rural/urban county status and also by MI hospitalization rate status (low, 0 to <33.33 percentile; medium, 33.33 to <66.66 percentile; high, ≥66.66 percentile).

Results:

Of Tennessee’s 95 counties, 62.3% of facilities were in its 42 urban counties. Mean (SD) county-level travel distance to the closest OCR was 16.6 (10.0) miles, and mean (SD) county-level travel time was 27.9 (13.6) minutes. Travel distance/time did not significantly differ by rural/urban county status (rural: 15.4 miles, 28.9 minutes; urban: 12.9 miles, 23.2 minutes) or by MI hospitalization rate status (low: 10.4 miles, 19.6 minutes; medium: 14.5 miles, 24.1 minutes; high: 18.8 miles, 30.7 minutes).

Conclusions:

Our findings indicate that overall mean travel distance was around half an hour, but lack of OCR facilities in rural Tennessee counties did not correspond with significantly greater travel distance/time to OCR in rural versus urban counties. Additional efforts are warranted to help patients post-MI with actual or perceived high travel burden navigate personal and structural factors precluding OCR receipt.

What’s New and Important

• Cardiologists may not consistently prescribe outpatient cardiac rehabilitation to rural patients because of an assumption bias that their longer travel distance/time will deter participation in outpatient cardiac rehabilitation.

• Compiling information about outpatient cardiac rehabilitation facilities through scraping Google Maps data can help supplement existing facility lists and help healthcare providers and patients access information on expanded provider options.

• Distance/time to an outpatient cardiac rehabilitation facility can be similar between rural and urban patients, even in rural areas with limited outpatient cardiac rehabilitation facilities.

Cardiac rehabilitation (CR) is a key component of the American College of Cardiology and American Heart Association secondary cardiovascular disease (CVD) prevention guidelines.¹ An American College of Cardiology/American Heart Association evidenced quality measure for CR is adherence with a dose-response relationship shown between greater adherence to attending CR sessions and larger reductions in myocardial infarction (MI) morbidity and mortality.¹ Nonetheless, attendance in CR programs remains poor in the United States.^2,3 In nationwide studies from the mid-2000s, researchers showed that only 19% to 34% of US patients with MI attended CR.² Investigators conducting a rural-urban examination of CR attendance among US patients with MI between 2011 and 2017 found that CR attendance was <40% for both rural and urban patients.³ Longer travel distance and time to a CR facility have been shown to be significantly associated with lower CR attendance among US patients with travel posing a larger issue for outpatient CR (OCR), which requires attending multiple sessions.^4–7

Despite evidence supporting the association between travel distance and time and CR attendance, there are few US studies where authors report actual estimates of travel time and distance to a CR facility, with these study authors either using data from the 1990s that may not reflect current CR availability,^4,5,8 providing estimates of either travel time or distance but not both,^4–7 or focusing on travel to Veterans Affairs–operated CR faculties only.⁷ One of these studies was on patients 62 years and older who were hospitalized for a coronary bypass surgery or an MI in Vermont and surrounding areas during 1992, with study authors finding that individuals who participated in OCR had a mean (SD) travel time of 17 (11) minutes and those who did not participate in OCR had a mean (SD) travel time of 25 (15) minutes.⁴ In another of these studies, investigators focused on female cardiac patients in Ohio hospitalized between April and September 1995, finding that 21% of these patients had to travel 20 or more miles to reach their closest CR facility.⁸ Investigators conducting a more nationwide examination of OCR among fee-for-service Medicare beneficiaries hospitalized for a coronary bypass surgery or an MI in 1997 noted that the distance to the closest OCR facility ranged from 0.3 to 231.0 miles.⁵

Furthermore, comparisons of CR travel distance and time between rural and urban residents using contemporary data are limited,⁷ despite the known poorer cardiac outcomes among rural residents.^9–11 Authors of 1 study in this area examined OCR access among cardiac patients in rural Nebraska during 2003, reporting that the mean (SD) distance to the nearest OCR facility was 10.7 (14.1) miles for those who participated in OCR and 27.3 (26.6) miles for those who were referred to but did not participate in OCR.⁶ Authors of another study assessed travel time to the nearest Veterans Affairs facility offering CR among patients hospitalized for ischemic heart disease between 2016 and 2019 and found that 63.1% of urban patients had travel times less than or equal to 30 minutes compared with only 13.1% of rural patients.⁷

As such, it is difficult to determine with the existing literature whether there are geographic disparities in post-MI care services and the extent of these disparities. Cardiac rehabilitation providers and patients can use such information in aiding discussions on alternatives to in-person OCR for patients with large travel distance and times with the goal to reduce this CR attendance barrier. Accordingly, we sought to examine travel distance and time to the closest OCR facility for each county in Tennessee, a largely rural state with high CVD burden.^12,13 We also assessed whether there were differences in travel distance and time to the closest OCR facility between rural and urban Tennessee counties and between low–, medium–, and high–MI hospitalization rate counties within the state.

Methods

Outpatient Cardiac Rehabilitation Facilities

We identified OCR facilities in Tennessee through a combination of an established list compiled by the Tennessee Association of Cardiovascular and Pulmonary Rehabilitation (TACVPR) (https://www.tacvpr.life/home) and data scraping using CR-related keywords to cross-reference this list. The TACVPR list is available upon request from the authors. The TACVPR list contains n = 56 currently operating and verified rehabilitation programs in Tennessee and n = 17 in surrounding states and is updated regularly through program-to-program referral of patients.

To account for the possibility of CR facilities closing/opening since the last time the TACVPR list was updated, we cross-referenced facilities in the list with those identified through Google Maps Platform’s Places Application Programming Interface scraping process. Briefly, the scraping process works by searching through all business websites indexed in Google Maps with a Python script and retrieving the name, phone number, address, and latitude and longitude of businesses whose pages have any mention of the keywords “cardiac,” “rehabilitation,” and “outpatient.”¹⁴ These keywords were chosen because they were used in a feasibility study on the application of data scraping techniques in CVD research.¹⁵ For data-scraped results distinct from the TACVPR list, we conducted validation by calling and checking facility websites. As part of a sensitivity analysis, we considered that some individuals near Tennessee’s border may be closer to an out-of-state OCR facility and adjusted the Application Programming Interface settings to include facilities within 25 miles of Tennessee’s border.

The data scraping process yielded n = 5 currently operating OCR facilities in Tennessee and n = 4 currently operating facilities in surrounding states distinct from the TACVPR list facilities. In total, there were n = 61 unique Tennessee OCR facilities identified from the combined list and data scraping process. The study’s main analyses focused on the n = 61 unique facilities in Tennessee with n = 70 unique combined in-state and out-of-state facilities used for a sensitivity analysis expanding to facilities within 25 miles of the Tennessee border. Note that the number of out-of-state facilities is less than the combined number from the TACVPR list and the data-scraped results as not all out-of-state TACVPR facilities were within 25 miles of Tennessee’s border.

We conducted the data scraping and validation processes between October 1, 2023, and June 19, 2024. Additional detail on the data scraping and validation process can be found in Supplementary Table 1 (http://links.lww.com/JCN/A354).

Rural/Urban Status

We used Rural-Urban Continuum Codes to classify Tennessee counties as either rural or urban.¹⁶ The Rural-Urban Continuum Codes comprise a system developed by the US Department of Agriculture that categorizes US counties based on population size and proximity to metropolitan areas.¹⁷ The US Department of Agriculture considers counties with codes 1 to 3 to be metropolitan counties and those with codes 4 to 9 as nonmetropolitan counties.¹⁷ Following the classification scheme used by the US Department of Agriculture and existing health disparities literature, we categorized counties with codes 1 to 3 as urban counties and counties with codes 4 to 9 as rural counties.^17–20 With this approach, n = 53 out of 95 Tennessee counties were classified as rural (55.8%); and n = 42 out of 95, as urban (44.2%).

High, Medium, and Low Myocardial Infarction Hospitalization Status

We used county-level 2019–2021 MI hospitalization rate data from the Centers for Disease Control and Prevention’s Interactive Atlas of Heart Disease and Stroke to classify Tennessee counties as either low (0 to <33.33 percentile), medium (33.33 to <66.66 percentile), or high (≥66.66 percentile) MI hospitalization rate counties.¹³ Using these previously mentioned percentiles, we determined the rate cutoffs for each of the categories from the Interactive Atlas data as follows: low (0 to <7.4 MI hospitalizations/1000 Medicare beneficiaries), medium (7.4 to <8.7 MI hospitalizations/1000 Medicare beneficiaries), and high (≥8.7 MI hospitalizations/1000 Medicare beneficiaries). Determining travel distance/time to the closest OCR facility based on county MI hospitalization status is important because MI is a leading cause of death in addition to being the foremost used diagnosis in CR and one of the first diagnoses approved by Medicare for CR coverage.^21–26 Patients, clinicians, and policymakers can use such travel distance/time estimates to provide them with insights on whether counties with high MI burden experience greater travel barriers to post-MI care than low–MI burden counties.¹³

Ethical Approval

With regard to ethical approval, in this study, we did not use any human subject data or involve animals. All data used in this study are publicly available. Thus, after consultation with the University of Tennessee’s institutional review board, our study was deemed not subject to institutional review board review following the institution’s institutional review board guidelines.

Geospatial and Statistical Analyses

Because we did not have patient addresses, county-level mean travel distance and time to the closest OCR facility were determined with ArcGIS Pro’s Closest Facility tool in a 2-step process based off a hypothetical individual residing in a census block’s center. Additional detail about this process can be found in Supplementary Table 1 (http://links.lww.com/JCN/A354). We created 2 separate maps, one overlaying rural/urban county and MI hospitalization rate status onto mean travel distance and the second with these overlays onto mean travel time. With data visualizations and statistical tests indicating average travel distance and average time being nonnormally distributed (right-skewed) in rural/urban county status categories and in MI hospitalization rate status categories, we conducted Kruskal-Wallis tests to examine whether mean travel distance and time differed significantly by rural/urban county status and whether mean travel distance and time varied significantly by MI hospitalization rate status.

As part of a sensitivity analysis, we determined county-level mean distance and time, created maps with rural/urban county and MI hospitalization rate overlays, and conducted the same Kruskal-Wallis tests on the data set that included both in-state and out-of-state facilities within 25 miles of Tennessee’s border. All analyses were carried out in ESRI 2022 (ArcGIS Pro: Release 3; Environmental Systems Research Institute, Redlands, California) and SAS version 9.4 software. Statistical testing was 2-sided and conducted at α = .05.

Results

Of the 61 unique OCR facilities identified in Tennessee, n = 38 facilities (62.3%) were located in urban counties and n = 34 facilities (55.7%) were located in low–MI hospitalization rate counties (Table). We determined that the mean (SD) county-level travel distance to the closest OCR was 16.6 (10.0) miles and the mean (SD) county-level travel time to the closest OCR was 27.9 (13.6) minutes (Figure; Supplementary Table 2, http://links.lww.com/JCN/A354). When examined by rural/urban county status, we found that rural counties had slightly higher median of mean travel distance and time (15.4 miles, 28.9 minutes) compared with urban counties (12.9 miles, 23.2 minutes). However, the P values for the distance and time comparisons between rural and urban counties were not statistically significant, being .182 and .284, respectively.

Table 1.

Distribution of Outpatient Cardiac Rehabilitation Facilities in Tennessee

	Rural/Urban County Status		Myocardial Infarction Hospitalization Rate (per 1000 Medicare Beneficiaries) County Status
	Rural Counties (N = 53)	Urban Counties (N = 42)	Low-Rate Counties (0 to <7.4) (N = 31)	Medium-Rate Counties (7.4 to <8.7) (N = 32)	High-Rate Counties (≥8.7) (N = 32)
No. facilities	23	38	34	16	11

FIGURE.

Map of proximity to outpatient cardiac rehabilitation facilities in Tennessee with rural/urban county status and myocardial infarction hospitalization rate (MIHR) overlays. A, Travel distance. B, Travel time.

The medians of mean travel distance and time to the closest OCR were 10.4 miles and 19.6 minutes for low–MI hospitalization rate counties, 14.5 miles and 24.1 minutes for medium–MI hospitalization rate counties, and 18.8 miles and 30.7 minutes for high–MI hospitalization rate counties (Figure; Supplementary Table 2, http://links.lww.com/JCN/A354). Travel distance (P = .268) and travel time (P = .244) did not significantly differ between low–, medium–, and high–MI hospitalization rate counties. Sensitivity analyses yielded similar answers to the main analysis (Supplementary Figure 1, Supplementary Table 3, http://links.lww.com/JCN/A354).

Discussion

We identified OCR facilities in Tennessee, a state with a large proportion of rural residents and high CVD prevalence. Overall, mean county-level travel distance to the closest OCR was almost half an hour. Less than 40% of the state’s OCR facilities were located in rural counties, which typically also were classified as a high–MI hospitalization rate county. Despite a lack of OCR facilities in rural areas, travel distance and time to the closest OCR facility did not significantly differ by rural/urban county status. Furthermore, travel distance and time to the closest OCR facility did not significantly vary between low–, medium–, and high–MI hospitalization rate counties. The lack of significantly longer distance/time to OCR in rural areas that we observed, potentially explained by higher traffic in urban areas, is encouraging. However, there were still a dozen Tennessee counties where travel time to the nearest OCR was >45 minutes, which can result in a large travel burden considering multiple OCR sessions are needed.

Greater travel burden to OCR may deter patients from OCR receipt altogether or attending fewer than the requisite 36 sessions due to existing financial and time constraints from work or family.^1,27,28 With the proportion of individuals 40 years or younger who experience MI increasing by 2% annually over the last decade, an increasing number of MI survivors needing CR will still be in the workforce and may find it difficult to receive time off to attend multiple OCR sessions.²⁹ Depending on the line of work an individual is in, time off to attend OCR sessions may mean a loss of wages for the day.³⁰ Furthermore, some MI survivors with adverse effects that preclude driving may find it even more difficult to access OCR if it is far away. To reach an OCR facility, these individuals may have to rely on family members or caregivers for rides, access public transportation that could require multiple transfers, or use ride shares that can be costly.

The use of telerehabilitation has been spreading to help overcome OCR access barriers.^31,32 Furthermore, authors of a review of cardiac telerehabilitation indicated that, overall, patients with access to the broadband needed for this care viewed telerehabilitation favorably in terms of the quality of care provided, ease of use, and convenience.³³ Despite patients' general acceptance of cardiac telerehabilitation, they may experience financial constraints when trying to use this care option because audio-only telerehabilitation is not covered by Medicare³⁴ and many of the Medicare COVID-era policies allowing for coverage of some telerehabilitation services are set to expire in December 2024.³⁵ In addition, telerehabilitation may not be appropriate for all OCR services and is still reliant on broadband, which is not available or reliable in all rural areas.³⁶

We detail hereinafter some existing and potential strategies to improve OCR travel-related barriers and telerehabilitation in the United States. Paratransit and, when available, other group-assisted transportation could be an option for patients living in rural communities throughout the United States who cannot drive to OCR.^36,37 However, these transportation services, with their required additional time allowances (ie, 1 hour before appointment), may not be suitable for time-constrained individuals.^36,37 Although not widely available, a potential option to offset the cost of travel to OCR is offering patients reimbursement to help cover gas or public transportation expenses.^36,37 Travel reimbursement policies such as those from the US Veterans Affairs could be a model for a similar plan to be implemented in Medicare but would require extensive discussion with and support from policymakers, patients, and professional OCR-focused organizations (ie, American Heart Association, American Association of Cardiovascular and Pulmonary Rehabilitation).³⁸ With consideration to strategies targeted toward cardiac telerehabilitation barriers, placing broadband access points in communal public spaces such as libraries and community centers in US rural areas has been somewhat successful in helping patients access telehealth.^39,40 Expanded Medicare coverage for telerehabilitation could also aid in increasing telerehabilitation access but, like with the proposed travel reimbursements through Medicare, would entail involvement from multiple policy, clinical, and patient-oriented stakeholders.⁴⁰

Additional structural- and policy-level efforts to help patients with large travel burden attend OCR are particularly relevant after the passing of the Bipartisan Budget Act of 2015, which significantly disincentivizes health systems placing CR programs further into the community through reduced Medicare payment rates for new or relocated OCR programs.⁴¹ Moreover, there has been no further action by the Centers for Medicare & Medicaid Services to remedy this via any subsequent budget acts.⁴¹ With that said, there have been ongoing efforts from the American Association of Cardiovascular and Pulmonary Rehabilitation and its joint affiliate members such as TACVPR to advocate for CR legislation to mitigate the impact of the Budget Act of 2015 on the expansion of CR facilities in underserved areas as well as increasing availability of telerehabilitation in the home. With regard to TACVPR specifically, whose advocacy goals closely align with that of the American Association of Cardiovascular and Pulmonary Rehabilitation, efforts have focused on (1) HR-955/S. 1849, SOS: Sustaining Outpatient Services Act, which would exempt certain hospital outpatient services, including CR, from a drastic reimbursement reduction that is based solely on the location of the hospital outpatient service, and (2) HR-1406/S. 3021 Sustainable Cardiopulmonary Rehabilitation Services in the Home Act, which will allow Medicare beneficiaries to receive CR services via real-time, audio/visual communication from their home.^42,43

We discuss the study’s limitations hereinafter. Because this is a single-state study focused on Tennessee, findings may extend to other largely rural sates in the southeast but less so to more urban states in other US regions. Additional multistate and international research on rural-urban travel burden to OCR can help to increase the generalizability of results. We did not examine travel distance and time to specific OCR services, which may reflect varying travel burden for individuals needing services that may not be offered at their closest OCR facility. In addition, we determined travel distance and time to the closest OCR facility based on a hypothetical individual in the middle of a census tract rather than actual patient addresses with this approach providing travel distance and time estimates averaged over patients at the census tract’s center and very edges. With the current study limitations, we recommend further investigations using data from multiple states or countries, examining proximity to specific OCR services using actual patient addresses, and determining whether perceptions of OCR travel burden rather than actual travel distance and time have a greater impact on OCR participation.

Conclusions

Although there remains a lack of OCR facilities in rural Tennessee counties, this did not translate into significantly greater travel distance and time to the nearest OCR in rural versus urban counties. This finding suggests that, in terms of travel burden, the current distribution of OCR facilities within Tennessee may help to offset the scarcity of OCR facilities in the state’s rural areas, a contrast to what is usually observed for rural-urban distribution of services throughout the MI care continuum. With that said, an overall mean travel time of nearly half an hour is still considerable as OCR requires multiple sessions. Moreover, patient perception of OCR travel burden can vary even for the same travel distance and time. Accordingly, additional efforts are needed to help patients with MI who forgo OCR because of large travel burden and those who perceive their OCR travel burden to be too great navigate the personal and structural factors contributing to this perception.