The American College of Cardiology/American Heart Association guidelines recommend exercise training (typically in a cardiac rehabilitation [CR] setting) to improve functional status in patients with symptomatic heart failure with reduced ejection fraction (HFrEF) able to participate (Class I recommendation, level of evidence A). To date, few studies have determined CR referral and enrollment rates in heart failure (HF) patients.1-4 These studies have found that CR referral and participation rates for HF patients are low (10.4% and 2.6-41%, respectively).1-4 A major barrier to CR referral was the lack of insurance coverage.5,6 In 2014, the Centers for Medicare & Medicaid Services (CMS) initiated coverage of CR for selected patients with HFrEF (ie, those who meet CMS referral criteria). It would be expected that the introduction of this CMS coverage would result in greater CR referral and participation rates. To this point, a previous study reported that CR participation rates were 41% in patients with HFrEF following CMS coverage3; however, referral rates were not reported. Other studies reporting CR referral and participation rates following CMS coverage approval have found low CR participation rates1,4; however, it is unclear whether these patients met CMS entry criteria for CR. Thus, we sought to assess CR referral, enrollment, and graduation rates in patients with HFrEF following introduction of CMS coverage for CR.
METHODS
This was a historical cohort study that included 265 consecutive patients residing in Olmsted County, Minnesota, who had a primary diagnosis of HFrEF at any inpatient or outpatient visit from January 1, 2009, to October 31, 2012 (ie, 2009-2012 cohort) and January 1, 2014, to March 31, 2019 (ie, 2014-2019 cohort). The 2009-2012 cohort was included to establish a baseline estimate of CR referral, enrollment, and graduation prior to CMS approval of CR. Patients were initially identified using the International Classification of Diseases (ICD-9/10) code 428/I.50 (HF) and were excluded if they had a left ventricular ejection fraction (LVEF) >35%, history of stroke with persistent and debilitating symptoms (eg, paralysis), or lacked valid research authorization as required by the Minnesota statute. An LVEF >35% was used as exclusion criteria, as the CMS has approved coverage for CR for HFrEF patients with an LVEF <=35%. This study was approved by the Mayo Clinic Institutional Review Board, and all patients agreed to the use of their medical records for research.
Electronic medical record (EMR) review was performed to obtain all the patient-level data. The LVEF measurements from outpatient echocardiography visits that were closest to the discharge date were recorded (<180 d). Referral and enrollment (ie, attending >=1 outpatient CR session) were determined via the EMR. Graduation was defined as EMR documentation that a patient had a final CR graduation assessment or attended >=25 CR sessions.
All analyses were completed using R. Patient characteristics are reported as mean +/- SD or as n (%). Normality was assessed using the Shapiro-Wilk test and nonparametric tests were used when appropriate. Participant characteristics and referral rates were compared using unpaired t tests or [chi]2 tests when appropriate. Associations between categorical patient characteristics and CR referral as well as graduation of >=25 CR sessions were assessed using OR, 95% CI, and P values from univariate logistic regression. In all cases, two-tailed P < .05 was considered statistically significant.
RESULTS
The Table shows the patient characteristics for the 2009-2012 and 2014-2019 cohorts. In the 2014-2019 cohort, 85 patients (33%) were >=75 yr and 105 (50%) exhibited an LVEF <25%.
For the 2009-2012 cohort, 15 (28%) of the 54 patients were referred to CR of which 10 enrolled in CR. Two of the enrolled patients completed >25 sessions. For the 2014-2019 cohort, 59 (28%) of the 211 patients were referred to CR. Among the 59 patients referred to CR, 41 (69%) enrolled in CR and attended >=1 session with an average of 21 +/- 15 CR sessions completed. Further, among the 41 patients enrolled in CR, 17 (41%) patients graduated from CR.
Univariate predictors of CR referral included age (ie, >=75 yr; OR = 0.53: 95% CI, 0.27-0.99), ischemic etiology (OR = 2.30: 95% CI, 1.25-4.28), and chronic obstructive pulmonary disease diagnosis (OR = 0.18: 95% CI, 0.04-0.54). Age (ie, >=75 yr) was the only univariate predictor of completing >=25 CR sessions in this cohort (OR = 27.00: 95% CI, 4.17-539.74). Sex, care center, LVEF, systolic blood pressure, diabetes, hypertension, as well as [beta]-blockers, angiotensin-converting enzyme inhibitor/angiotensin receptor blockers, and diuretic use were not significant predictors of CR referral or completing >=25 CR sessions (all, P >= .10).
DISCUSSION
We expected the CR referral rate to be substantially higher in the present study following introduction of coverage by the CMS. In contrast, we found that the referral rates were not different between our 2009-2012 and 2014-2019 cohorts (28% vs 28%), albeit higher than previously reported (ie, 10.4%).1 Patient factors associated with low likelihood of CR referral in the 2014-2019 cohort included older age (ie, >=75 yr) and chronic obstructive pulmonary disease diagnosis, consistent with previous studies in HF and coronary artery disease patients.1,7,8 An additional factor contributing to the low referral rates may be the CMS referral restriction of 6 wk for the HFrEF patients to become stable on medication therapy.
A recent study investigating CR participation rates following CMS approval found that CR participation rates were 41% for HFrEF patients.3 In contrast, the enrollment rate (relative to referral rate) in the 2014-2019 cohort presented was substantially greater (ie, 69%). This finding suggests that referral is a major barrier limiting HFrEF patients from enrolling (and graduating from) CR. An additional important finding of the present study was that the CR graduation rate was 41% in the 2014-2019 cohort. In contrast to coronary artery disease,7 older age (ie, >=75 yr) was the only significant predictor of completing >=25 CR sessions in the 2014-2019 cohort. In conclusion, we found that CR was underutilized and that participation rates did not significantly improve following the introduction of CMS coverage, suggesting the presence of other barriers to CR participation, other than lack of insurance coverage.
Rongjing Ding, MD
Department of Cardiovascular Medicine,
Mayo Clinic, Rochester, Minnesota
Department of Cardiovascular Medicine, Peiking
University People's Hospital, Beijing, China
Joshua R. Smith, PhD
Jose R. Medina-Inojosa, MD
Sisi Zhang, MD
Department of Cardiovascular Medicine,
Mayo Clinic, Rochester, Minnesota
Marta Supervia, MD
Department of Cardiovascular Medicine,
Mayo Clinic, Rochester, Minnesota
Department of Physical Medicine and Rehabilitation, Gregorio Maranon General University Hospital, Gregorio Maranon Health Research Institute, Madrid, Spain
Karen M. Fischer, MPH
Health Sciences Research, Mayo Clinic,
Rochester, Minnesota
Ray W. Squires, PhD
Thomas P. Olson, PhD
Shawn E. Leth, MEd
Wenliang Zhang, MD
Grace Lin, MD
Randal J. Thomas, MD
Department of Cardiovascular Medicine,
Mayo Clinic, Rochester, Minnesota
Acknowledgments
This work was supported by Mayo Clinic, Division of Preventive Cardiology, National Institutes of Health (T32HL007111 to J.R.S.) and American Heart Association (18POST3990251).
REFERENCES