Author + information
- Received November 24, 2003
- Revision received January 5, 2004
- Accepted January 13, 2004
- Published online June 2, 2004.
- Javed Butler, MD, MPH, FACC*,‡,§∥,* (, )
- Patrick G. Arbogast, PhD†,
- James Daugherty, MS†,
- Manoj K. Jain, MD, MPH∥,
- Wayne A. Ray, PhD†,‡ and
- Marie R. Griffin, MD, MPH†,‡,§
- ↵*Reprint requests and correspondence:
Dr. Javed Butler, Cardiology Division, 383-PRB, Vanderbilt University Medical Center, Nashville, Tennessee 37232, USA.
Objectives This study was designed to assess outpatient angiotensin-converting enzyme inhibitor (ACEI) use after heart failure (HF) hospitalization.
Background Assuring therapy with ACEIs at discharge after HF hospitalization is a key Medicare quality measure. The benefits of such quality improvement attempts will be limited if therapy is not continued long-term.
Methods To assess the factors associated with filling an ACEI prescription in the 30 days postdischarge and the proportion of patients who filled such prescriptions subsequently up to 365 days postdischarge, we studied 219 patients with depressed ejection fraction (EF) specifically and 960 HF patients in general.
Results Sixty-seven percent of patients with depressed EF and 55% of the total cohort were discharged with ACEIs. Overall 81.2%/77.1% (depressed EF/total cohort) of survivors discharged with ACEIs had filled a prescription by 30 days postdischarge; only 66.3%/63.3% were current users at 365 days. In contrast, for patients with no discharge order for ACEIs, only 12.7%/12.0% (depressed EF/total cohort) had filled such a prescription by 30 days and 12.5%/18.8% were current users at 365 days postdischarge. Patients with a discharge order for ACEIs were more likely to fill a prescription within 30 days postdischarge (hazard ratio 10.93, 95% confidence interval 5.28, 22.61, for patients with depressed EF).
Conclusions For patients with HF who are discharged while taking ACEIs, there is a significant decline in use after discharge. Patients not discharged with ACEIs are unlikely to be started as outpatients. Quality improvement efforts therefore need to be focused on both discharge planning and outpatient care.
Clinical trials have convincingly demonstrated the benefits of angiotensin-converting enzyme inhibitor (ACEI) therapy for heart failure (HF) patients with depressed ejection fraction (EF) (1). Other data suggest potential benefit with renin-angiotensin system modulation in patients with HF and preserved EF also (2). The role of ACEI use in prevention of HF in high-risk patients continues to evolve (3). Finally, many patients with HF also have comorbidities such as hypertension and diabetes that may warrant ACEI use (4,5). Therefore, some experts suggest that ACEIs should be used for all HF patients irrespective of EF (6).
The American College of Cardiology and American Heart Association guidelines for HF management recommend routine lifelong ACEI therapy, unless contraindicated, for HF patients with depressed EF (7). Clinical use of these drugs, however, remains suboptimal (8). Both the Joint Commission for Accreditation of Healthcare Organizations (JCAHO) and Centers for Medicare and Medicaid Services (CMS) have selected ACEI use in HF patients with depressed EF as a key quality measure for hospitalized HF and myocardial infarction patients (9,10). This has led to development of many quality improvement interventions to improve compliance with these quality measures (11,12). Interest in implementing quality improvement interventions has become more important recently because JCAHO now mandates the reporting of data on select quality measures and CMS is preparing to report quality-of-care data to the public (13).
Medication prescription at discharge, however, does not guarantee long-term use. Patients may fail to fill prescriptions for discharge medications. Many patients who initially use these medications may later discontinue their use because of intolerance or poor adherence to recommended care (7” ok?>). Thus, quality improvement efforts restricted to hospitals may not achieve their promised effectiveness.
Although predictors of ACEI use have been described, long-term adherence once therapy is initiated is not well known (14–17). We sought to assess the outpatient utilization of ACEI therapy during the first year postdischarge after a hospitalization for HF. We also assessed the new use of ACEI therapy for patients who were not discharged while taking them. We identified a population-based cohort of Medicare beneficiaries discharged alive after being admitted with decompensated HF. The study cohort was that subset of patients who were also enrolled in the Tennessee Medicaid Program at discharge and thus eligible to receive pharmacy benefits, which enabled us to determine their use of ACEIs in the year following discharge.
Sources of data
The Medicare data for this study came from two different sources. First, CMS collected data on 756 random patients from Tennessee from April 1, 1998, to September 30, 1998, as part of baseline data for the Sixth Scope of Work initiative. Later, Tennessee's Medicare Quality Improvement Organization collected data on 90 random charts/month on hospitalized patients with HF across Tennessee from June 1, 1999, to September 30, 2001, in order to provide ongoing feedback of data to providers over time. Data collection methodology was similar for both initiatives.
The study was conducted among enrollees of the Tennessee Medicaid (TennCare) program. Medicaid computerized files permitted linkage to the Quality Improvement Organization data and identification of prescriptions filled. The Medicaid pharmacy file contains reimbursed prescriptions for outpatients and nursing home residents. This file identifies the date the prescription was filled, which specific drug and dose was dispensed, the number of pills dispensed, and the number of days of drug supply (maximum 30 days in Medicaid).
A total of 3,174 cases of hospitalized HF patients (756 CMS baseline data for Sixth Scope of Work and 2,420 Quality Improvement Organization data) were identified. A total of 1,617 of these patients (51%) were also enrolled in Medicaid at the time of hospital discharge and had prescription drug benefits through the program. We excluded 54 patients who were transferred to another acute care facility and 63 patients who died during hospitalization. To avoid including patients who may get retrospective activation of their Medicaid privileges and would not have been eligible for medication benefits at discharge, we required enrollment for at least 60 days before hospitalization, leaving 1,059 patients. Although all enrollees in Medicaid have access to prescription drugs through the program, some may choose to use other sources of prescription medications (e.g., VA hospitals). To assure that all study subjects were able to fill a prescription through the Medicaid program, the analysis was restricted to 1,041 of the remaining patients who filled at least one prescription in the year before discharge or in the first 30 days postdischarge. Finally, 81 of the 1,041 patients in the study cohort had documented allergy or intolerance to ACEIs, leaving the final study cohort of 960 patients. Of these 960 patients, 871, 686, and 550 subjects were alive and enrolled in Medicaid at 30, 180, and 365 days post-HF discharge, respectively.
This study was approved by Vanderbilt University and the State of Tennessee Institutional Review Boards, and reviewed and approved by the CMS and by the Tennessee Medicaid Program.
Details of CMS data collection and definitions have been described previously (8). Data collection by Quality Improvement Organization was similar to CMS methodology. For assessing whether ACEI therapy was prescribed at discharge or not, the abstractors reviewed the chart comprehensively, including clinical notes (both physician and nursing), physician orders, pharmacy sheets, and discharge summaries. The abstractors were provided with a list of both generic and trade names of all available ACEIs.
Patients with hypertension and diabetes were identified by either the specific documentation of these comorbidities in the data collection tool or coding of specific International Classification of Diseases-9 codes for these two conditions in their Medicaid records. Finally, patients who filled prescriptions for either insulin or other hypoglycemic medications were classified as having diabetes.
Depressed EF was defined either as a quantitative estimate of ≤40% or a qualitative estimate of at least moderate systolic dysfunction. The qualitative EF information was collected under the following six categories: 1) normal, preserved, or satisfactory; 2) mild dysfunction; 3) moderate or moderate to severe dysfunction, or low function; 4) severe or very severe dysfunction, or very low or poor function; 5) systolic dysfunction, severity not specified; and 6) description not classified in the other categories. For the purpose of this analysis, only patients in categories 3 and 4 were classified as having systolic dysfunction with EF ≤40%. Angiotensin-converting enzyme inhibitor intolerance was defined as documentation by a physician of moderate to severe aortic stenosis or bilateral renal artery stenosis; angioedema, rash, hives, “allergy,” or “adverse reaction” associated with ACEI; or “ACEI intolerance” or other reason for not prescribing an ACEI.
Clinical data from the medical records were collected by formally trained project data specialists. The data abstraction tool was piloted on randomly selected HF charts that were available from a previous study. The clinical consultant, project data specialists, and analysts used the results from the pilot test to refine the abstraction tool. Reliability was measured using the Project Data Specialist Team standard procedures. These include the testing and refinement of data collection tools, educational sessions, and an assessment of the consistency of results among the data specialists. Inter-rater reliability was assessed with a requirement of ≥95% agreement on ACEI use. An internal quality control process was also conducted using a 5% random sample. A separate project data collection specialist reabstracted these charts and the results were compared to ensure reliability using similar standards.
The outcomes studied were the proportion of patients who filled a prescription for an ACEI within 30 days after discharge and the proportion that had a current prescription (filled in the prior 30 days) at 180 and 365 days postdischarge. Prescriptions in Medicaid usually may not exceed 30 days. We also studied the predictors of filling an ACEI prescription by 30 days for all patients and predictors of outpatient adherence at one year postdischarge among those who were discharged while taking ACEIs and filled a prescription by 30 days. We performed an additional analysis that assessed the use of ACEIs or angiotensin receptor blocker (ARB) therapy, because these latter drugs are often used for patients who become intolerant of ACEIs.
Descriptive analyses were performed and the data were tabulated as means or frequencies. Proportional hazards regression was used to estimate adjusted hazard ratios (HR) and 95% confidence intervals (CI) assessing patient characteristics related to the time to filling an ACEI prescription within the first 30 days postdischarge. The following characteristics were studied as potential predictors: age (<65, 65 to 74, >75), gender, race, admission and discharge ACEI use, cardiology involvement in care, EF (preserved, depressed, or missing), and presence of hypertension or diabetes. We studied the proportion use of ACEI and ACEI or ARB at 30, 180, and 365 days postdischarge. However, considering the relatively low use of ARB, we restricted the analysis of predictors of adherence to ACEI therapy alone.
To assess the predictors of long-term adherence, we defined a cohort of patients who were discharged with ACEIs and who filled a prescription for them by 30 days after discharge. Similar variables used to assess 30-day predictors were also studied for 365-day adherence. Logistic regression was used to calculate adjusted odds ratios and 95% CIs for the relationship of ACEI status at 365 days to patient characteristics. Odds ratios were converted to relative risks. Regression diagnostics were performed to assess the appropriateness of the proportional hazards and logistic regression models.
Because data on cardiology involvement in care and EF were missing in a large proportion of patients, we did two sets of analysis to assess predictors of adherence for both 30 and 365 days post-discharge. In the first set, data on patients with all predictor variables present were assessed and a second analysis was performed excluding provider and EF as variables. There were no significant differences noted between the two analyses, so data on only the latter analysis is presented. Because recommendation for ACEI therapy for HF patients is more conclusive for patients with depressed EF, we did two sets of analyses, primarily on a on a subgroup of HF patients with depressed EF, and second on the overall cohort. All analyses were performed using SAS Version 8.2 (SAS Institute, Cary, North Carolina).
Overall we studied 960 patients who were discharged alive after HF hospitalization and had no documented intolerance or contraindication for an ACEIs (Table 1). Ejection fraction was depressed in 219 (23%) of the 960 patients, was preserved in 265 (28%), and missing for the rest of the cohort. Hypertension or diabetes each was present in almost half of the patients, with a higher proportion in patients with depressed EF. Sixty-seven percent of those with depressed EF and 55% of the overall cohort were discharged while taking an ACEI. A discharge prescription for an ACEI was the strongest predictor of filling a prescription within 30 days for patients with depressed EF (HR 10.93, 95% CI 5.28, 22.61) and also for the overall cohort (HR 14.09, 95% CI 8.28, 23.97), after adjusting for patient characteristics listed in Table 1. None of the other characteristics, including demographics, ACEI use on admission, cardiology involvement, and presence of hypertension or diabetes, were independent predictors of ACEI use after discharge.
Overall ACEI use
A total of 447 of the 960 patients (47%) were taking ACEIs at admission, including 118 of 219 (54%) of those with depressed EF. Admission with ACEIs was highly associated with discharge prescription of ACEIs. Among patients with depressed EF, 87% of patients taking ACEIs at admission were discharged with ACEIs, and 44% of patients not taking ACEIs at admission were discharged with them (p < 0.001). For the overall cohort, among patients taking ACEIs at admission 89% were discharged on ACEI, whereas only 29% of patients not taking ACEIs at admission were discharged with them (p < 0.001).
Daily outpatient adherence to ACEI therapy after discharge among patients with depressed EF is shown in Figure 1. Among patients with a documented discharge prescription for ACEIs, approximately 81% were using them during the first 30 days as measured by prescriptions filled. There was a sharp decline around day 30, which represents gaps in prescriptions refilled, followed by an increase and then a gradual decrease that remained at around 60% to 65% from approximately 90 through 365 days. Among patients without a discharge prescription for ACEIs, approximately 10% to 15% were using ACEIs during the first 30 days; use plateaued at around 15% to 20% by day 180. Figure 2shows the use of ACEIs for the entire cohort, which shows results similar to those with depressed EF only specifically.
ACEI versus ARB use
Only 4% of the 960 study patients were discharged with ARB versus 55% with ACEIs. For patients discharged with ACEIs, 77% were taking ACEIs at 30 days and an additional 1% was taking ARB, whereas at one year, 63% were still taking ACEIs and 3% were taking ARB. Of those with no documented discharge order for ACEIs, ACEI use increased modestly over the following year from 12% to 19% overall while ARB use remained at approximately 8% to 10% (Table 2).
Adherence to ACEI use
For patients discharged with an ACEI who filled such a prescription in the first 30 days, there were no significant associations between the baseline characteristics and one-year adherence with ACEI for either the patients with depressed EF specifically or the overall cohort (Table 3).
Despite the benefits of ACEI therapy in patients with HF, our study once again documents underutilization of ACEIs in this relatively contemporaneous cohort of HF patients. Although underutilization of ACEI therapy and its predictors has been described before, we also demonstrated that the use of these agents diminishes further over time after therapy is initiated, leaving a minority of patients with long-term treatment (14–17). In our cohort, the only independent predictor of filling an ACEI prescription as an outpatient was prescription at the time of discharge. New use among those not discharged with ACEIs was low, underscoring the importance of careful discharge planning.
Suboptimal cardiovascular pharmacotherapy has been well described (8). Based on these results, both CMS and JCAHO have designated a combined set of quality measures for both myocardial infarction and HF, including ACEI use (9,10). Moreover, CMS is now assessing the possibility of public disclosure of hospital health care quality data (13). Such assessments may be used to rate hospitals, usually resulting in a significant attention by the popular press (18). In response, both professional organizations and individual hospitals are rapidly trying to implement new systems to improve compliance with such quality measures (11,12). Although such interventions are commendable, if the use of these medications is not continued in outpatients, the potential benefits from these costly interventions are likely to be significantly lessened. In this study of patients with no documented intolerance to ACEIs, not only were almost half of the overall cohort and one-third of those with depressed EF not prescribed ACEIs, almost one-third of the patients who were discharged with ACEIs had stopped taking these medications by one year postdischarge. Considering that almost 50% of HF patients are readmitted within six months of discharge, underutilization of ACEIs poses a significant risk to the patients (19). Moreover, if outpatient care specifically after discharge correlates with overall general outpatient care, then the majority of the HF patients who have not required hospitalization yet are likely to be receiving suboptimal care also.
Discharge prescription of ACEIs was the single most important predictor of filling an ACEI prescription as an outpatient, underscoring the importance of careful discharge planning. Previous investigations have shown similar correlation with discharge prescription and long-term medication use (20). We found a strong correlation between ACEI use at admission and at discharge. Thus, most patients discharged with ACEIs may represent a select group of patients known to be tolerant to these medications. However, during a HF hospitalization, both the patient and the family members are likely to be more responsive to education regarding medication use and may be more motivated to comply with the recommendations given during the acute phase. Similarly, nurses and doctors may be able to give more detailed information in the hospital setting than in the busy outpatient clinic where provider-patient contact is usually very limited. Thus, failure to prescribe an ACEI at discharge may represent a missed opportunity. Systems to support appropriate prescribing in hospital setting may help address this problem. However, even if patients are discharged on appropriate therapy, our and other data suggest that adherence continues to deteriorate in the outpatient setting (21). Therefore, interventions restricted to hospital discharge may be beneficial but are likely to be inadequate. The fluctuation in use seems to stabilize by six months postdischarge. Thus, the first three to six months postdischarge may be an appropriate target for additional interventions. Besides the discharge prescription, there were no other significant predictors of either 30- or 365-day adherence with ACEI therapy. Data on some key variables such as provider specialty and EF were missing in a substantial number of patients in our study, however, reducing the power of detecting difference. However, lack of other predictors also suggests that the use, or lack thereof, of ACEIs is not explainable on the basis of clinical characteristics alone.
Beyond allergies, tolerability with ACEI therapy is a significant problem and may restrict the use of these drugs. Side effects such as hyperkalemia, worsening renal function, and hypotension may require careful monitoring (22). Dose adjustment and/or changes in other medications may be more preferable to discontinuation, given that lower doses also significantly improve survival (23). It is difficult to determine the causal relationship of cough to ACEI in individual patients; however, a switch to an ARB is a reasonable option for patients with a chronic cough and no other clear etiology. Angiotensin receptor blocker therapy has also been shown to significantly improve outcomes among HF patients (24,25). Among our cohort of patients discharged with ACEIs, few were started on ARB therapy when ACEIs were discontinued. Therefore, side effects and intolerance are unlikely explanations for most of the drug discontinuations. A higher proportion of patients not discharged on ACEIs filled ARB prescriptions after discharge. This may represent either provider preference or undocumented/unknown ACEI intolerances during the index hospitalization. Irrespective of the debate whether ACEI and ARB have comparable benefit or not, the overall use of either of the two drugs was still observed in <30% of patients not discharged on ACEI.
Our study has several limitations. Because of the retrospective nature of the study, we do not have the data on multiple clinical and laboratory characteristics at discharge and tolerability of these medications by patients after discharge. We cannot ascertain the degree to which the lack of adherence is related to the physician- or patient-related factors in our study. Our patient population was eligible for both Medicare and Medicaid benefits. This may represent a higher risk population and the results may not be generalizable. Patients dually eligible for Medicare and Medicaid may represent a more adverse group from a socioeconomic perspective. Moreover, these patients are likely to be older, less well educated, may more often belong to a racial or ethnic minority, and may have poor access to care. However, patients enrolled in Medicaid were a significant proportion of all Medicare patients with HF hospitalizations (>50%) in our state, and thus are likely to be not too dissimilar to the general Medicare population. In addition, these patients have prescription pharmacy benefits, which remove the financial barrier to obtaining their medications. Our overall sample size is relatively small. Finally, in our study cardiology involvement in care included any interaction with a cardiologist during hospitalization, either in consultation or as a primary service. We have no data on whether cardiology was involved in the outpatient care. Thus, the influence of physician specialty on ACEI adherence cannot be assessed for certain.
In order to maximize patient outcomes after hospitalization for HF, attempts must be made to improve upon current trends. Multiple interventions directed at patients have been shown to increase adherence with medications (26). Other investigators have described the hospital-related factors that were associated with quality improvement with pharmacotherapy at hospital discharge (27). Prospective studies are needed to assess the effectiveness, safety, and tolerability of these drugs in patients who are at a higher risk for side effects. However, because lower doses are associated with significant survival benefit, patients with relative intolerance should at least be tried on smaller doses if larger doses are not tolerated (23).
In conclusion, our study suggests that a significant proportion of patients who are discharged with ACEI therapy do not fill their prescription as outpatients. Discharge administration of ACEIs is the strongest predictor of continued use. The use of these drugs is less than optimal in all groups of patients, based on demographics and clinical characteristics. New outpatient prescriptions for those not discharged with ACEI therapy is infrequent. Quality improvement efforts should focus on improving discharge planning and preferably should continue up to six months postdischarge.
☆ Supported in part by the Agency for Healthcare Research and Quality, Centers for Education and Research in Therapeutics cooperative agreement (grant #HS 1-0384) and a cooperative agreement with the Food and Drug Administration (FD-U-001641).
The analyses upon which this publication is based were performed under contract number 500-99-TN01, titled “Quality Improvement Organization for the State of Tennessee,” sponsored by the Centers for Medicare and Medicaid Services, Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
The authors assume full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the Centers for Medicare and Medicaid Services, which has encouraged identification of quality improvement projects derived from analysis of patterns of care. Ideas and contributions to the authors concerning experience in engaging with issues presented are welcomed.
- angiotensin-converting enzyme inhibitor
- angiotensin receptor blocker
- Centers for Medicare and Medicaid Services
- ejection fraction
- heart failure
- hazard ratio
- Joint Commission for Accreditation of Healthcare Organizations
- Received November 24, 2003.
- Revision received January 5, 2004.
- Accepted January 13, 2004.
- American College of Cardiology Foundation
- Hunt S.A.,
- Baker D.W.,
- Chin M.,
- et al.
- ↵JCAHO Overview of Heart Failure Core Measure Set. Available at: http://www.jcaho.org/pms/core+measures/hf_overview.htmf. Accessed June 16, 2003.
- Quality Improvement Organizations: Statement of Work. Available at: http://cms.hhs.gov/qio/2.asp. Accessed June 16, 2003.
- ↵Get With The Guidelines. American Heart Association. Available at: http://www.americanheart.org/presenter.jhtml?identifier=1165. Accessed June 16, 2003.
- Guidelines Applied in Practice. American College of Cardiology. Available at: http://www.acc.org/gap/gap. Accessed June 16, 2003.
- ↵Hospital Quality Information Initiative. Available at: http://www.cms.hhs.gov/quality/hospital/hqii.asp. Accessed June 16, 2003.
- Philbin E.F.,
- Andreaou C.,
- Rocco T.A.,
- et al.
- ↵Best Hospitals. Available at: http://www.usnews.com/usnews/nycu/health/hosptl/ranking. Accessed June 16, 2003.
- Butler J.,
- Arbogast P.G.,
- BeLue R.,
- et al.
- Packer M.,
- Poole-Wilson P.A.,
- Armstrong P.W.,
- et al.
- Haynes R.B.,
- Montague P.,
- Oliver T.,
- McKibbon K.A.,
- Brouwers M.C.,
- Kanani R.