Effect of an American Heart Association Get With the Guidelines Program-Based Clinical Pathway on Referral and Enrollment Into Cardiac Rehabilitation After Acute Myocardial Infarction
Article Outline
Cardiac rehabilitation (CR)/secondary prevention programs are an important part of patient care after acute myocardial infarction (AMI). However, only 10% to 15% of eligible patients enroll in such programs. The purpose of this study was to evaluate the effect of an American Heart Association Get With the Guidelines (GWTG)-based clinical pathway on referral and enrollment into CR after AMI. Patients (n = 780) admitted to a single center during an 18-month period with AMI and discharged to home were evaluated retrospectively for referral and enrollment into CR programs. A total of 714 patients (92%) were on the GWTG pathway; 392 (55%) were referred and 135 (19%) were enrolled into CR. Higher referral was associated with pathway use (odds ratio [OR] 2.3, 95% confidence interval [CI] 1.1 to 4.9, p = 0.03), percutaneous coronary intervention (OR 3.1, 95% CI 1.9 to 5.2, p <0.0001), and in-patient physical therapy consultation (OR 13, 95% CI 8.2 to 20.5, p <0.0001). Ethnicity did not affect referral, but was the only variable associated with lower enrollment. Hispanic and black patients had 92% (OR 0.08, 95% CI 0.01 to 0.55, p = 0.02) and 57% (OR 0.43, 95% CI 0.19 to 1.05, p = 0.06) lower odds to enroll compared with white patients, respectively. In conclusion, use of the American Heart Association GWTG pathway showed a significantly higher referral rate to CR after AMI than previously reported in the literature. Nonetheless, most referred patients did not enroll. Strategies to bridge the gap between referral and enrollment in CR should be incorporated into AMI clinical pathways, with special emphasis on increasing enrollment in ethnic minorities.
Despite the well-founded benefits, limited risk, established cost-effectiveness, and numerous published guideline recommendations, referral rates for cardiac rehabilitation (CR) programs after acute myocardial infarction (AMI) are low. In the managed care setting, reported referral rates were as low as 24% after hospitalization for acute coronary syndrome.1 The baseline referral rate noted in the Get With The Guidelines (GWTG) Pilot study was 34%, but this included referral to CR or exercise counseling without further specification and was without follow-up to confirm enrollment and/or completion of CR.2 Programs outside the United States reported referral rates as high as 41%.3 In the recently published largest and most comprehensive evaluation to date regarding use of CR, Suaya et al4 assessed 270,000 Medicare beneficiaries after hospitalization for myocardial infarction (MI) or coronary artery bypass surgery across the United States. They report that only 14% of patients with AMI ultimately enrolled in CR. Pilot study results from the American Heart Association GWTG collaborative model for quality improvement showed increased smoking cessation counseling, lipid treatment, lipid measurement, and referral to CR or exercise counseling.2 The purpose of this study was to evaluate the effectiveness of the GWTG-based clinical pathway on referral and enrollment into CR after AMI, evaluate factors related to referral and enrollment, and determine reasons for lack of enrollment despite patient referral.
Methods
Data for all (n = 945) consecutive patients during an 18-month period from 2002 to 2003 with the primary International Classification of Diseases, Ninth Revision diagnosis code for AMI5 (410) at an urban university hospital, Boston Medical Center, Boston, Massachusetts, were retrospectively reviewed. Data were extracted from hospital-specific data previously entered by 1 of the investigators (DW) into the American Heart Association GWTG6 database. Data were excluded for patients who either died before hospital discharge or were discharged to a facility other than home.
Information derived from the database included demographics (age, gender, and race/ethnicity), type of MI (ST-elevation or non–ST-elevation MI), cardiovascular risk factors (history of hypertension, hyperlipidemia, diabetes, and smoking), co-morbidities (history of peripheral arterial disease, dialysis-dependent chronic kidney disease, stroke, and chronic obstructive pulmonary disease), presence of atrial fibrillation, measured left ventricular ejection fraction using echocardiogram or left ventriculography, revascularization during hospitalization (percutaneous coronary intervention or coronary artery bypass surgery), and discharge medications.
Boston Medical Center uses a clinical pathway and standardized order set for all patients admitted with the diagnosis of AMI. The clinical pathway and order set were developed from the GWTG Program,6 and all resident physicians are trained to use it. All orders are written exclusively by the resident physician team. Use of the standardized order set, although strongly encouraged, is at their discretion at the time of admission. The order set includes a physical therapy consult to evaluate each patient before discharge. The pathway calls for each patient to be provided with a written referral (program name, contact person, and telephone number) to a specific outpatient CR program at the time of discharge by the physical therapist, nurse, or physician team assigned to the patient. CR programs are selected based on their proximity to the patient’s home or patient preference.
Individual medical records were reviewed to evaluate whether the patient was placed on the clinical pathway, whether a physical therapy consult was ordered, and the name of the CR program to which the patient was referred. Each of 36 different CR programs was contacted initially by letter and then by telephone by 1 of the investigators (GRS) to confirm whether the patient had enrolled in that program. Questionnaires were then mailed with a stamped self-addressed envelope to the 278 patients who were referred to CR, but reportedly did not enroll. The questionnaire asked the patient to respond to only the 2 specific questions of whether the patient had enrolled in a CR program, and if so, to provide the program name, and if not, to select 1 reason (from a list of 12) why the patient had not joined a program. An identical follow-up letter was mailed 2 weeks later to patients who did not respond to the first letter.
The Institutional Review Board of Boston Medical Center approved this protocol with waiver of informed consent. All authors had full access to the data and take responsibility for its integrity. All authors read and agreed to the manuscript as written. All analyses were performed using the SAS statistical system, version 9.1 (SAS Institute, Cary, North Carolina, 2006), with a 0.05 significance level unless otherwise noted. The 2 dichotomous outcome variables were referral to CR (yes/no) and enrollment upon referral (yes/no). For each, univariate associations between probability of outcome and continuous variables were assessed using either independent samples t test or Wilcoxon’s rank-sum test. Associations were similarly estimated for categorical variables using Pearson’s chi-square test or Fisher’s exact test, as appropriate.
Multiple logistic regression models were used to assess simultaneous effects of clinical variables on probability of outcome. For each outcome, a forward stepwise regression was used to select a final multivariate model, using the criterion p <0.10 for variable entry and termination of the stepwise procedure. Variables eligible for inclusion in the final model were age at admission, gender, race, diabetes, smoking, peripheral arterial disease, dialysis-dependant chronic kidney disease, history of heart failure, stroke, chronic obstructive pulmonary disease, MI type (ST-elevation vs non–ST-elevation MI), revascularization type (coronary artery bypass graft surgery vs percutaneous coronary intervention vs none), pathway ordered, physical therapy consult ordered, and discharge service (coronary care unit vs interventional service vs other).
Results
From 2002 to 2003, a total of 945 patients with the primary diagnosis of AMI were admitted to Boston Medical Center. Of these, 165 patients either died during their hospitalization or were discharged to a facility other than home. Baseline characteristics of the remaining 780 study patients, 714 of whom were on the GWTG-based clinical pathway, are listed in Table 1. There were no significant differences in age, gender, ethnicity, coronary artery disease risk factors, or co-morbidities between pathway (n = 714) and nonpathway (n = 66) patients (data not shown).
Table 1. Patient characteristics
| Variable | All patients (n = 780) | Pathway patients (n = 714) |
|---|---|---|
| Age (yrs) | 63.2 ± 12.8 | 63.1 ± 12.5 |
| Men | 513 (65.8%) | 475(66.5%) |
| Race/ethnicity | ||
| White | 637(81.8%) | 588(82.5%) |
| Black | 59(7.6%) | 51(7.2%) |
| Hispanic | 45(5.8%) | 38(5.3%) |
| Other | 32(4.1%) | 36(5.0%) |
| MI | ||
| ST-elevation MI | 155(19.9%) | 155(21.1%) |
| Non–ST-elevation MI | 625(80.1%) | 559(78.3%) |
| Baseline left ventricular ejection fraction (%) | 47.9 ± 14.9 | 47.9 ± 14.9 |
| Revascularization | 586(75.1%) | 552(77.3%) |
| Percutaneous coronary intervention | 503(85.8%) | 469(65.8%) |
| Coronary artery bypass grafting | 83(14.2%) | 83(11.6%) |
| Diabetes | 226(29.0%) | 209(29.3%) |
| Hypertension | 505(64.7%) | 459(64.3%) |
| Hyperlipidemia | 289(37.1%) | 269(37.7%) |
| Smoker | 295(40.4%) | 269(40.2%) |
| Peripheral arterial disease | 77(9.9%) | 73(10.2%) |
| Previous heart failure | 82(10.5%) | 67(9.4%) |
| Atrial fibrillation | 51(6.5%) | 46(6.4%) |
| Dialysis-dependent chronic kidney disease | 10(1.3%) | 8(1.1%) |
| Stroke | 47(6.0%) | 40(5.6%) |
| Chronic obstructive pulmonary disease | 77(9.9%) | 68(9.5%) |
A total of 417 (54%) patients were referred to CR at hospital discharge, and 392 of these were on the pathway. Therefore, the referral rate for patients on the pathway was 392 of 718 (55%; Table 2). Multivariate analysis, determined using stepwise regression, included age, revascularization, pathway use, discharging service, and physical therapy consultation. Results showed that patients had higher odds of referral if they were discharged with the clinical pathway (odds ratio [OR] 2.3, 95% confidence interval [CI] 1.1 to 4.9, p <0.03), underwent percutaneous coronary intervention during the hospitalization (OR 3.1, 95% CI 1.9 to 5.2, p <0.0001), or had an inpatient physical therapy consultation during their hospitalization (OR 13, 95% CI 8.2 to 20.5, p <0.0001). For each additional year of age, there was a 2% lower odds of referral (OR 0.98, 95% CI 0.966 to 0.997, p = 0.02; Table 3).
Table 2. Cardiac rehabilitation referral and enrollment rates in pathway patients (n = 718)
| Referral | 392/718(55%) | 95% CI 51–58 |
| Enrollment in referred patients | 135/392(34%) | 95% CI 30–39 |
| Enrollment in all pathway patients | 135/718(19%) | 95% CI 16–22 |
Table 3. Multivariate analysis of factors related to referral and enrollment in cardiac rehabilitation programs
| Variable | Odds Ratio (95% CI) | p Value |
|---|---|---|
| Referral | ||
| Discharged on clinical pathway | 2.3(1.1–4.9) | <0.03 |
| Percutaneous coronary intervention during hospitalization | 3.1(1.9–5.2) | <0.0001 |
| Inpatient physical therapy | 13.0(8.2–20.5) | <0.0001 |
| Increased age (each additional year) | 0.98(0.966–0.997) | 0.02 |
| Enrollment | ||
| Discharged on clinical pathway | 2.7(0.88–8.06) | 0.08 |
| Hispanic patients | 0.08(0.01–0.55) | <0.02 |
| Black patients | 0.43(0.19–1.05) | 0.06 |
The enrollment rate for the 780 patients discharged with AMI was 18% (n = 139). Restricting attention to the subgroup of patients referred for CR, 33% (n = 139) ultimately enrolled in a program and only 135 (34%) of the 392 pathway patients who were referred to CR actually enrolled (Table 2). Multivariate analysis identified no individual variable associated with increased rate of enrollment in CR, although there was a trend favoring enrollment for patients discharged on the clinical pathway (OR 2.66, 95% CI 0.88 to 8.06, p = 0.08). Although ethnicity did not appear to affect referral, it was the only variable associated with decreased enrollment after referral. Hispanic and black patients had 92% (OR 0.08, 95% CI 0.01 to 0.55, p = 0.02) and 57% (OR 0.43, 95% CI 0.19 to 1.05, p = 0.06) lower odds of enrollment in CR compared with white patients (Table 3).
To address potential reasons for nonenrollment, questionnaires were mailed with a stamped self-addressed envelope to the 278 patients who were referred to CR, but did not enroll in a CR program. The overall questionnaire response rate was 36% (n = 99). None of the patients stated they enrolled in another program. Reasons given for nonenrollment included return to work, 10%; not interested in CR, 12%; believed they were never referred, 26%; no reason specified, 40%, and other, 12%.
Discussion
Underreferral to CR programs after AMI poses an important problem for community hospitals, academic centers, and managed care organizations alike. Previous studies showed that factors associated with increased rate of referral included younger age, discharge diagnosis of AMI, and coronary revascularization.3 These data were consistent with our findings, which showed that patients had higher odds of referral if they underwent percutaneous coronary intervention and lower odds of referral if they were older. In addition to patient characteristics, health care provider involvement has an important role in referral. In our study, patients had higher odds of referral if they received an inpatient physical therapy evaluation or were placed on a clinical pathway. Similarly, Doolan-Noble et al7 showed that inpatient CR was associated with high rates of referral to outpatient CR. A pilot study in Ontario, Canada, showed 22% participation of eligible patients in CR with usual referral practices that improved to 43% using computer-generated referrals at 1 urban tertiary care center in the province.8 This was independent of the reason for referral (i.e., status post coronary artery bypass grafting, percutaneous coronary intervention, or AMI).8
Lack of enrollment in CR is a related, but distinct, problem. Referral to CR is obviously associated with enrollment because 1 should follow the other. Strong predictors of participation in CR after initial referral include surgical coronary revascularization or ≥2 cardiovascular events after the index event.8 Although patient-specific characteristics associated with improved odds of enrollment are not clear, older age,1, 4, 8, 9 female gender,4, 10 and minority ethnicity4, 11 were associated with a lower likelihood of enrollment in CR.
It therefore follows that system-wide changes are needed to improve both referral to and use of CR after AMI. To our knowledge, ours is the largest comprehensive study to address the effect of a clinical pathway on referral and enrollment in CR in patients after AMI. Data showed that use of the GWTG–based clinical pathway was associated with a higher referral rate (55%) compared with that generally quoted in the literature (20% to 30%).12 Although the GWTG-based pathway at Boston Medical Center calls for all patients discharged to home to be referred to a specific CR program, it is clear that this was not done. Patients were more likely to be referred if an inpatient physical therapy consult was ordered. This implies that inpatient physical therapists were more rigorous about referring patients to CR than were discharging nurses or physician teams. Patients who underwent percutaneous coronary intervention during the hospitalization were also more likely to be referred to CR. Patients (n = 194) who did not undergo revascularization after AMI were significantly older and had significantly more co-morbidities (diabetes, hypertension, renal failure, history of heart failure, and previous stroke) than patients who underwent revascularization (n = 586; data not shown). Therefore, these patients may not have been deemed candidates for referral by the discharging physical therapist, nurse, or physician team. Finally, older age was independently associated with lower referral rates, as shown in other studies.1, 7, 8, 9 Also, it is important to note that of patients who were documented to have been referred, but did not enroll in a CR program, 26% perceived they had not been referred.
In the present study, the enrollment rate in referred patients on the pathway was only 34%, and in all pathway patients, 19%. However, there were no specific orders or actions on the GWTG-based pathway that fostered enrollment. Race/ethnicity was an important factor for nonenrollment, particularly in Hispanic and black patients. Cohen et al11 previously showed that Hispanic patients were more often managed conservatively after acute coronary syndrome and non–ST-segment elevation MI. They underwent stress tests after the event more frequently than white patients and were less likely to be referred for cardiac catheterization or undergo subsequent percutaneous coronary intervention.11 In that study, Hispanic patients received similar discharge medical therapy; however, they were less frequently referred for CR.11 Gender-based studies showed that African-American women were less likely to enroll in CR than white women.10 Similarly, Suaya et al4 showed that women, older patients, and nonwhites were less likely to enroll in CR.
Although the findings of the present study have important implications, they are limited in that this was not a randomized controlled trial of the use of a clinical post-MI pathway. The number of patients who were not on the pathway was too small to allow robust statistical comparison between groups. However, it is important to note that the main purpose of this observational study was to evaluate referral and enrollment rates in pathway patients in the context of the low rates reported in the literature. Finally, the findings are specific to implementation of the GWTG-based pathway at a single hospital. Differences in methods by which the pathway is ordered and carried out could lead to different outcomes.
References
- . Delivery of outpatient cardiac rehabilitation in a managed care organization. J Cardiopulm Rehabil. 2004;24:157–164
- . Get With the Guidelines for cardiovascular secondary prevention (Pilot results). Arch Intern Med. 2004;164:203–209
- . Factors associated with referral to outpatient cardiac rehabilitation services. J Cardiopulm Rehabil. 2004;24:165–170
- . Use of cardiac rehabilitation by Medicare beneficiaries after myocardial infarction or coronary bypass surgery. Circulation. 2007;116:1653–1662
- . Hypotheses, design, and methods for the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) Trial. Am J Cardiol. 2006;97(suppl):9G–19G
- . Get With The Guidelines Overview. http://www.americanheart.org/presenter.jhtml?identifier=3045578Accessed July 18, 2007
- . Cardiac rehabilitation services in New Zealand: access and utilization. N Z Med J. 2004;117:U955
- . Increasing access to cardiac rehabilitation: automatic referral to the program nearest home. J Cardiopulm Rehabil. 2004;24:171–174
- . Attendance and graduation patterns in a group model health maintenance organization alternative cardiac rehabilitation program. J Cardiopulm Rehabil. 2004;24:150–156
- . Disparities in women’s referral to and enrollment in outpatient cardiac rehabilitation. J Gen Intern Med. 2004;19:747–753
- Clinical characteristics, process of care, and outcomes of Hispanic patients presenting with non-ST-segment elevation acute coronary syndromes: results from Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA Guidelines (CRUSADE). Am Heart J. 2006;152:110–117
- . Quality indicators in cardiovascular care. J Cardiopul Rehabil. 2005;25:249–256
PII: S0002-9149(07)02445-9
doi:10.1016/j.amjcard.2007.11.063
© 2008 Elsevier Inc. All rights reserved.
