Improving Evidence-Based Care for Heart Failure in Outpatient Cardiology PracticesCLINICAL PERSPECTIVE
Primary Results of the Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF)
Background— A treatment gap exists between heart failure (HF) guidelines and the clinical care of patients. The Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF) prospectively tested a multidimensional practice-specific performance improvement intervention on the use of guideline-recommended therapies for HF in outpatient cardiology practices.
Methods and Results— Performance data were collected in a random sample of HF patients from 167 US outpatient cardiology practices at baseline, longitudinally after intervention at 12 and 24 months, and in single-point-in-time patient cohorts at 6 and 18 months. Participants included 34 810 patients with reduced left ventricular ejection fraction (≤35%) and chronic HF or previous myocardial infarction. To quantify guideline adherence, 7 quality measures were assessed. Interventions included clinical decision support tools, structured improvement strategies, and chart audits with feedback. The performance improvement intervention resulted in significant improvements in 5 of 7 quality measures at the 24-month assessment compared with baseline: β-blocker (92.2% versus 86.0%, +6.2%), aldosterone antagonist (60.3% versus 34.5%, +25.1%), cardiac resynchronization therapy (66.3% versus 37.2%, +29.9%), implantable cardioverter-defibrillator (77.5% versus 50.1%, +27.4%), and HF education (72.1% versus 59.5%, +12.6%) (each P<0.001). There were no statistically significant improvements in angiotensin-converting enzyme inhibitor/angiotensin receptor blocker use or anticoagulation for atrial fibrillation. Sensitivity analyses at the patient level and limited to patients with both baseline and 24-month quality measure data yielded similar results. Improvements in the single-point-in-time cohorts were smaller, and there were no concurrent control practices.
Conclusions— The Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting, a defined and scalable practice-specific performance improvement intervention, was associated with substantial improvements in the use of guideline-recommended therapies in eligible patients with HF in outpatient cardiology practices.
Clinical Trial Registration— URL: http://www.clinicaltrials.gov. Unique identifier: NCT00303979.
Received December 23, 2009; accepted June 1, 2010.
Heart failure (HF) is a chronic progressive disease that results in substantial morbidity, mortality, and expenditure of healthcare resources.1,2 Clinical trials have established that a number of therapies may improve clinical outcomes for patients with HF and reduced left ventricular ejection fraction (LVEF).2 Despite extensive clinical trial evidence and recommendations in national guidelines, prior studies have demonstrated that treatment guidelines are adopted slowly, are applied inconsistently, and thus often fail to lead to improvements in patient care quality and outcomes.3–9 For hospitalized HF patients and outpatients, gaps, variations, and disparities between evidence-based guideline recommendations and actual treatments provided have been documented.3–9 New, more effective approaches to improve the use of evidence-based, guideline-recommended therapies for HF are needed for each healthcare setting in which these patients are encountered.
Editorial see p 561
Clinical Perspective on p 596
Various strategies have been recommended to facilitate implementation of practice guidelines, including the use of performance measures based on practice guidelines, chart audits with feedback of results, reminder systems (including pathways and tools), and educational outreach visits.2 Although certain hospital-based performance improvement programs deploying these strategies have been successful in improving the quality of care,10,11 improvement in the delivery of evidence-based care in the outpatient practice setting through the use of a targeted and scalable performance improvement initiative has not been tested.12 Building on features of previous hospital-based programs, the aim of the Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF) was to facilitate more consistent delivery of evidenced-based, guideline-recommended care to eligible patients with HF and reduced LVEF or previous myocardial infarction (MI) and reduced LVEF in the outpatient cardiology practice setting by providing performance data feedback and a practice-specific performance improvement intervention.12
IMPROVE HF is a prospective study designed to evaluate the effectiveness of a practice-specific performance improvement intervention on the use of guideline-recommended therapies for patients with diagnosed HF and reduced LVEF or prior MI and reduced LVEF in outpatient cardiology practices. The methods and overall study objectives have been described in detail elsewhere.12 Community and academic cardiology or multispecialty outpatient practices from all regions of the country were invited to participate as previously described.12 Patients eligible for enrollment in the IMPROVE HF cohort included those with a clinical diagnosis of HF or prior MI documented by a cardiologist on at least 2 separate visits. Left ventricular systolic dysfunction was required to be demonstrated quantitatively by a LVEF ≤35% or qualitatively by findings of moderate to severe left ventricular systolic dysfunction on the most recent echocardiogram, nuclear multiple gated acquisition scan, contrast ventriculogram, or magnetic resonance imaging scan. Patients with a noncardiovascular medical condition associated with an estimated survival of <1 year and those who had undergone cardiac transplantation were excluded.
A baseline assessment of utilization rates for guideline-recommended therapies was completed at all participating practices. After completion of the baseline data collection, a practice-specific performance improvement intervention was implemented, and care was reassessed at prespecified time points during a 24-month period. Baseline and follow-up data were collected by medical chart review. Patient data included demographic and clinical characteristics, medical history, previous treatments, New York Heart Association (NYHA) functional status, QRS duration, laboratory results, diagnostic tests and results, treatments, and patient education. Contraindications and other reasons for not administering evidence-based therapies were also collected from patient records if documented. Other reasons included patient noncompliance, patient refusal, and medical, economic, social, and religious reasons for withholding recommended therapies.
A representative sample of medical records from patients with a clinical diagnosis of HF and LVEF ≤35% or prior MI and LVEF ≤35% at each cardiology practice was screened and selected at random to yield an average of 90 eligible patients for each practice at each assessment period (baseline, 6 months, and 18 months) using the methodology described in the trial design (Figure 1).12 The primary end point and secondary analyses were prespecified in the study protocol and statistical analysis plan. The study design included rigorous measures to ensure data quality and accuracy. These included the use of 34 trained, centralized chart review specialists who received ongoing training and testing to maintain the accuracy of data abstraction. Oversight of the data abstraction training was provided by members of the IMPROVE HF steering committee. Average interrater reliability between chart reviewers was 0.82 (κ statistic). Additional efforts to ensure the completeness and accuracy of collected data included monthly data quality reports and inclusion of an average of 1.7 automated data quality checks for each data field to ensure that values met prespecified ranges, formats, and units. An audit of all patient data compared with source documentation was conducted for 20% of the entire patient sample for a 10% random sample of participating sites. Mean data concordance rate was 94.5% (range, 92.3% to 96.3%). Practices were required to obtain institutional review board approval or waivers to participate in IMPROVE HF. Outcome Sciences Inc (Cambridge, Mass) served as the registry coordinating center for the trial.
Guideline-Recommended Quality Measures
Seven quality measures were prospectively selected by the IMPROVE HF steering committee12: use of (1) angiotensin-converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB), (2) β-blocker, (3) aldosterone antagonist, (4) anticoagulant therapy for atrial fibrillation or flutter, (5) cardiac resynchronization therapy (CRT) (CRT with a defibrillator/CRT with a pacemaker), (6) implantable cardioverter-defibrillator (ICD) (ICD or CRT with a defibrillator), and (7) HF education for eligible patients. The rationale for the selection of each metric was based on its potential to improve patient outcomes, the precision of its definition, its construct and content validity, and its feasibility.12 It is important to note that although 4 of the 7 quality measures selected for IMPROVE HF (use of ACEI or ARB, β-blocker, HF education, and anticoagulation for atrial fibrillation or flutter) are American College of Cardiology/American Heart Association outpatient HF performance measures and are endorsed by the National Quality Forum, 3 (use of aldosterone antagonist, ICD, and CRT) are not.13 Patients eligible for inclusion in an individual quality measure calculation were those who met the criteria for each specific therapy and for whom there were no contraindications, intolerance, or other documented reasons to explain why the indicated therapy should not be provided.12 The detailed quality measure specifications have previously been published,12 and eligibility, inclusion, and specific exclusion criteria for each measure are detailed in Table I in the online-only Data Supplement. HF education was based on medical record documentation that written and/or verbal HF education had been provided. Documentation of NYHA functional class is required to be considered eligible for an ICD, CRT, or aldosterone antagonist, and analyses of these quality measures included only patients with quantitative or qualitative documentation of NYHA functional class at a level consistent with the measure specifications.12 Two summary (composite) measures were also calculated: a total composite score (defined as the percentage of the total indicated quality measures provided) and an all-or-none care measure (defined as the proportion of patients who received each quality measure for which they met eligibility criteria).12,14
Performance Improvement Intervention
The IMPROVE HF performance improvement intervention focused on helping cardiology practices improve systems for treating patients with HF and has been previously described in detail.12 The intervention included a guideline-based clinical decision support tool kit, educational materials, practice-specific data reports, benchmarked quality-of-care reports, and structured educational and collaborative opportunities.12 As part of an enhanced treatment plan, IMPROVE HF provided evidence-based best-practices algorithms, clinical pathways, standardized encounter forms, checklists, pocket cards, chart stickers, and patient education and other materials to facilitate improved management of outpatients with HF. Examples of IMPROVE HF tools are provided in the study design 12 and are included in the online-only Data Supplement (A through L). These materials can also be downloaded from www.improvehf.com. The IMPROVE HF Steering Committee followed a structured, rigorous, evidenced-based, guideline-driven process to develop these pathways and tools. Use of the tools at participating practices was encouraged but not mandatory, and practices could adopt or modify tools at their discretion. A Web-based system provided quality-of-care reports for each practice that included benchmark comparisons with regional and national cardiology practices. The investigators and practice support personnel attended a 1-day workshop after completion of baseline data collection. During the workshop, study goals, guideline recommendations, the IMPROVE HF tool kit, performance improvement methodology, incentives to promote change in clinical practice, tips on how to serve as champions and engage colleagues in the performance intervention, and strategies to use the collected data effectively to provide practice-specific performance feedback were described and discussed. Clinicians at participating practices were encouraged to participate in bimonthly educational and collaborative Web-based seminars and to continually evaluate, refine, and reassess care delivery throughout the intervention phase of the study.
Assessment of Guideline-Recommended Care
Five chart reviews using representative samples of patients were conducted at each practice. These occurred at baseline (before implementation of any practice improvement intervention activity) and at 6, 12, 18, and 24 months after initiation of the performance improvement intervention (Figure 1). The medical records of patients enrolled in the longitudinal cohort were reviewed at baseline and subsequently at 12 and 24 months after the start of the practice performance intervention. These records provided data on temporal changes in patient-specific quality measures. Two additional single-point-in-time chart reviews were conducted at 6 and 18 months. These chart reviews involved separate and unique patient cohorts from the baseline cohort, providing additional data to examine practice-specific changes in care. Patients entered in the 6- and 18-month single-point-in-time cohorts included patients not selected for the longitudinal cohort at baseline and those more recently referred to the practice. All 167 practices completed the baseline assessment, and 154, 155, 151, and 155 practices completed the 6-, 12-, 18-, and 24-month follow-up assessments, respectively.
The prespecified primary objective of IMPROVE HF was to achieve a relative 20% or greater improvement in at least 2 of the 7 quality measures at 24 months compared with baseline for the aggregate of IMPROVE HF practice sites.12 Secondary predefined objectives included changes in the 2 summary care measures, changes in care evaluated at the individual practice level for each of the 7 quality measures and the 2 summary measures, changes in care confined to patients with both baseline and 24-month quality measure data (excluding patients who were lost to follow-up, died, or did not have a follow-up visit in the specified time frame), and evaluation of changes in quality measures in the single-time-point cohorts at 6 and 18 months compared with baseline.
All statistical analyses were performed by independent biostatisticians contracted by Outcome Sciences. Descriptive statistics for baseline patient and practice characteristics were calculated and reported. Large-sample tests on proportions and large-sample tests for differences in means were used to evaluate statistical associations for ordinal and continuous data, respectively. Each quality measure was evaluated separately. For each quality measure, the proportion of patients receiving therapy among those eligible for that therapy was calculated for each practice and across all practices. The composite score and all-or-none care measure were also calculated for each practice. To determine the total composite score, the numerators of all individual quality measures were summed to produce a composite numerator, and the denominators of all individual quality measures were summed to produce a composite denominator. The final composite score was produced by dividing the composite numerator by the composite denominator. The all-or-none care measure was calculated as the percentage of patients who received all eligible measures (ie, patients who received all the care for which they were documented to be eligible). The rates of care for each individual practice over time (practice-level analyses) at 24 months were compared with baseline rates. The mean values for each measure across all practices and care limited to patients with both baseline and 24-month quality measure data were also compared. In addition, utilization rates for each therapy in the single-point-in-time evaluation cohorts at baseline, 6 months, and 18 months were compared to evaluate changes in quality measures for unique patients over time. Finally, general estimating equation logistic regression models were developed to quantify how each quality measure changed after performance improvement intervention, after adjustment for patient clinical characteristics and practice characteristics and taking into account the correlation of data within patients and practices. The logistic regression models included all practice and patient characteristics shown in Tables 1 and 2⇓⇓ as possible predictors in the model if P<0.10 in univariate analyses. Analyses were completed with SAS statistical software, version 9.1 (SAS Institute, Cary, NC). All statistical inference testing was 2 sided with results considered statistically significant at P<0.05. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Clinical and Practice Characteristics
The medical records of 34 810 patients at 167 outpatient cardiology practices in the United States were included in this analysis (Figure 1). There were 15 177 patients in the longitudinal cohort evaluated at baseline, 12 months, and 24 months. A median of 90 patients (25th to 75th percentile, 58 to 107 patients) per practice were entered. There were 9992 unique patients in the 6-month and 9641 unique patients in the 18-month single-time-point cohorts. At 24 months, 7605 (50.1%) of the 15 177 patients in the longitudinal cohort at baseline were living and had a 24-month follow-up visit, 1446 (9.5%) were living but did not have a 24-month follow-up visit, 2569 (16.9%) were dead, 2508 (16.5%) were lost to follow-up, and 1049 (6.9%) were from practices that did not complete a 24-month visit.
Baseline patient characteristics are shown in Table 1⇑. The mean and median patient ages were 68.7 and 70.0 years, and 71.1% of the patients were male. Mean LVEF was 25.4%, and an ischemic origin of HF was identified in 65.4% of patients. A history of hypertension (62.2%) and diabetes mellitus (34.1%) was common, and other important conditions, including atrial fibrillation (30.7%) and chronic obstructive pulmonary disease (16.7%), were frequently identified. The median initial blood pressure of registry patients was 120/70 mm Hg. The median creatinine level was 1.2 mg/dL. Among IMPROVE HF outpatient cardiology practices, NYHA functional class was quantitatively documented in 34.2% and qualitatively documented by symptoms and functional limitations in an additional 60.3% of medical records (94.5% total). Characteristics of participating practices at baseline are shown in Table 2. Most of the practices were not associated with an academic or university center, and fewer than half reported having a dedicated HF clinic. Of participating practices, 96.0% reported on survey adoption of ≥1 performance improvement strategies, 85.2% reported use of the benchmarked quality reports, and 60.4% reported use of ≥1 IMPROVE HF tools.
The frequency of use of guideline-recommended quality measures at baseline in eligible patients is shown in Tables 3 and 4⇓ and Figure 2. For aggregate practices at baseline, an ACEI/ARB was prescribed for 11 165 (79.8%) of 13 987 eligible patients; a β-blocker, for 11 868 (86.2%) of 13 772 eligible patients; and an aldosterone antagonist, for 987 (34.4%) of 2870 eligible patients. Anticoagulant therapy was prescribed for 2910 (68.6%) of 4244 eligible patients with permanent, persistent, or paroxysmal atrial fibrillation or flutter. Of patients eligible for CRT, 580 (37.7%) of 1540 patients received therapy; ICD therapy was provided to 4799 (48.8%) of 9830 eligible patients. The baseline assessment showed that 9373 (61.8%) of 15 177 patients received documented HF education.
Primary Analyses of Quality Measures
A significant improvement was demonstrated in 5 of the 7 quality measures at the practice level at 24 months after implementation of the performance improvement intervention (P<0.001; Table 3). By 24 months, adherence to evidence-based, guideline-recommended care among practices increased in absolute terms by +25.1% for aldosterone antagonist use, +29.9% for CRT use, and +27.4% for ICD use compared with baseline results (P<0.001; Table 3). Use of β-blocker (+6.2%) and HF education (+12.6%) also showed significant improvement (P<0.001; Table 3). In contrast, use of anticoagulation in eligible patients with permanent, persistent, or paroxysmal atrial fibrillation or flutter did not improve over time (−0.1%). Use of ACEI/ARB at the practice level increased (+6.8%), but this was not a statistically significant improvement at the practice level (P=0.063). Measures with the lowest level of baseline utilization showed the greatest relative and absolute rates of improvement. A modified intention-to-treat analysis of the relative improvements at 12 and 24 months at the practice level showed statistically significant improvements for 6 of 7 quality measures for the longitudinal cohort alive and with complete follow-up (P<0.001; Table II in the online-only Data Supplement). There were 123 individual practices (79.4%) with ≥20% relative improvement in ≥2 measures from baseline to 24 months. Analysis of quality measure improvement in practices with and without electronic health record systems is shown in Table III in the online-only Data Supplement.
When the quality measures were analyzed at the aggregate patient level, there were significant improvements in 6 of 7 measures (Figure 2 and Table 4). The absolute magnitude of improvement at the patient level in these 6 measures ranged from +6.7% (ACEI/ARB) to +30.9% (CRT), and the relative improvement ranged from 8.4% to 81.9% (Table 4). Because 3 of the 7 quality measures improved by ≥20% at 24 months relative to baseline, the prespecified primary end point for the study was met. Significant improvements in the 6 quality measures that increased from baseline were observed by 12 months with further improvement by 24 months (Figure 2). General estimating equation univariate and multivariate analyses, which adjusted for patient and practice characteristics, as well as the correlation of data within practices showed significant improvement at 24 months compared with baseline for 5 of 7 measures in the longitudinal cohort (Table IV in the online-only Data Supplement). Use of guideline-recommended therapies at baseline in patients alive with 24-month follow-up and patients who died or were lost to follow-up at the patient level is provided in Table V in the online-only Data Supplement. When the quality measure results were stratified by patient age, improvements were seen for each age group for 6 of 7 measures without evidence of heterogeneity except for the ICD measure (Tables VI and VII in the online-only Data Supplement).
A modified intention-to-treat patient-level analysis confined to only patients alive and with complete follow-up at 24 months showed that these patients had similar relative and absolute improvement in the 6 quality measures at 12 and 24 months (Table 5). This paired patient cohort had baseline characteristics similar to those of the primary analysis population (Table 1⇑). The portion of patients with newly documented contraindications or intolerance at the 24-month assessment that were not documented at baseline and the portion of patients newly treated are shown in Table 6. The improvements in the ICD, CRT, and HF education measures were predominantly the result of more eligible patients being treated. For the ACEI/ARB, β-blocker, and aldosterone antagonist measures, both improved treatment and better documentation of contraindications or intolerance were operative.
Composite Score and All-or-None Care Measures
At baseline, the mean total composite score was 68.4%, and the all-or-none care measure was met in 24.3% of patients (95% confidence interval [CI], 23.6–25.0). At 24 months after intervention, the mean total composite score increased to 80.1% (+11.6% absolute improvement and 17.0% relative improvement; P<0.001). For the all-or-none care measure, the percentage increased to 43.9% (95% CI, 42.8–45.0; +19.6% absolute improvement, 80.6% relative improvement; P<0.001).
The 6- and 18-month single-point-in-time cohorts of unique patients had characteristics similar to those of the baseline cohort (Table 1⇑). Improvement at the patient level was demonstrated at 6 and 18 months for each quality measure except anticoagulation for atrial fibrillation or flutter. The magnitude of absolute and relative improvement in use of guideline-recommended therapies was smaller (+1.5% to +14.1% and 1.9% to 22.8%, respectively) in the single-point-in-time cohorts (Table 7) compared with the longitudinal cohort. When analyzed at the practice level, 5 of 7 measures showed modest but statistically significant improvement in the 18-month cohort compared with the baseline cohort (Table VIII in the online-only Data Supplement). General estimating equation multivariate analysis showed improvement in 4 of 7 measures in the 18-month cohort compared with the baseline cohort (Table IX in the online-only Data Supplement).
Despite compelling scientific evidence and readily accessible national guidelines, beneficial therapies for HF remain underused in many care settings.3–9 IMPROVE HF is the first large-scale outpatient performance improvement initiative designed to assess the effects of a practice-specific, process-of-care improvement intervention on HF patient care. By virtue of its design, combining cardiology practice–based structured data collection and feedback with guideline-based performance improvement tools, IMPROVE HF has contributed to the current description of care patterns for patients with HF with reduced LVEF or prior MI with reduced LVEF. These results provide a model to enhance clinical management of these patients in the outpatient setting. IMPROVE HF has shown that use of evidence-based, guideline-recommended HF therapies can be enhanced by the use of decision support tools, patient data collection, and performance feedback, concentrating on the specific therapies proven to improve outcomes. However, the determination of eligibility for treatment by medical record review is complex, and this study highlights many of the challenges faced in trying to assess and improve care in outpatient cardiology practices.
Gaps in the use of evidence-based therapies may reflect differences in training, familiarity with guidelines, patient case mix, and lack of systems to ensure that recommended care is provided. Successful strategies to improve the implementation of national practice guidelines include clinical decision support, provider education, patient education data collection, and performance feedback.2 A number of large, national, US hospital-based programs have sought to monitor and improve the care of patients hospitalized with HF by encouraging the use of process improvement systems and evidence-based, guideline-recommended HF therapies, including the Acute Decompensated Heart Failure National Registry (ADHERE), Organized Program to Initiate Life-Saving Treatment in Hospitalized Patients With Heart Failure (OPTIMIZE-HF), and Get With the Guidelines–Heart Failure (GWTG-HF).7,10,11 These performance improvement registries focused on HF care in the inpatient setting or the early postdischarge follow-up period. Certain HF disease management programs involving specialty care and multidisciplinary teams have been shown to be effective in improving quality of care and outcomes in the outpatient setting.15 Unfortunately, despite the potential benefits, these programs are not the norm and are often offered to only a small number of HF patients. Thus, there is a substantial need for strategies to be broadly implemented to improve the quality of outpatient HF care.
Although there are many major obstacles to overcome to improve care in the outpatient setting, cardiology practices participating in IMPROVE HF were able to improve performance on 5 of 7 quality measures. At the time of the 24-month assessment, 80.1% of the opportunities to provide evidence-based, guideline-recommended care to patients documented to be eligible were fulfilled compared with 68.4% at baseline. Importantly, all-or-none care demonstrated an 80.6% relative improvement from baseline to 24-month assessment. The relative and absolute rates of improvement in guideline-recommended use of CRT, ICD, and aldosterone antagonist therapies in eligible patients were substantial. Other studies have suggested very little change in the use of these 3 therapies for HF from 2005 to 2008, making it less likely that the improvements observed were a result of secular trends.6,7,16 Although the baseline rates of β-blocker use were reasonably high, there was still significant improvement at 24 months in this care measure. Documentation that HF education was provided also improved over the course of the program.
Those quality measures with lower baseline use showed greater absolute and relative improvement over time. These findings may reflect a greater effort to improve those measures with the best opportunities to improve, a ceiling effect on measures with higher levels of conformity, or other factors. Despite being a focus of the process improvement intervention and despite significant improvements observed in all other quality measures, there was no increase in the use of anticoagulation in eligible patients with atrial fibrillation or flutter. Other quality improvement initiatives in the hospital setting have also failed to demonstrate improvement in the use of this guideline-recommended therapy.10,17 The reasons for this are unclear but may be due in part to physician perception that warfarin is not necessary for atrial fibrillation in certain patients with HF, reluctance to prescribe warfarin when indicated because of the inconvenience associated with monitoring, undocumented concerns about risks, contraindications that were present but not documented, undocumented patient refusal, or other factors. Further study is needed to understand the underlying causes of the anticoagulation treatment gap and to develop more effective strategies to overcome these barriers.
Although performance improvements were evident for the longitudinal and single-point-in-time (6- and 18-month) cohorts, the magnitudes of both relative and absolute improvements were not equivalent. This may be due to the fact that practice- and patient-level performance reports and the clinical decision support tools were provided for the longitudinal cohort but patient-level reports were not provided for the single-point-in-time patient cohorts. The patient-level reports in the longitudinal cohort identified patients who were eligible for specific therapies but were not being treated. Physicians and nurses could use this information to recall or flag patients for initiation of the indicated therapy at a subsequent appointment. The greater magnitude of improvement over time in the longitudinal cohort suggests that receipt of patient-level performance feedback that is actionable appears to be a particularly important component of HF performance improvement efforts.
The improvements in certain quality measures in this study may reflect improvements in the documentation of care provided. For example, clinicians may have provided complete HF education at baseline but failed to document this in the medical record, and over the course of IMPROVE HF, they started to include this documentation. Medication and device therapy treatment rates also could have improved as a result of increased documentation of contraindications, intolerances, and patient and physician reasons for not providing guideline-recommended therapies rather than an increase in the number of eligible patients actually treated. However, the sensitivity analyses comparing patients who qualified for treatment at both baseline and 24 months showed similar improvements in the use of evidence-based therapies, which suggests that this finding is not solely attributed to changes in documentation. Furthermore, there were only modest changes in the number of patients with documented contraindications or intolerance at 24 months compared with baseline for most measures except aldosterone antagonists.
Although conformity with guideline recommendations substantially increased with the intervention, gaps in care persisted at 24 months, as did opportunities for further improvements in care. However, treatment rates that are <100% are not necessarily evidence of deficiencies of care. These data also may indicate that there is a need for better documentation of decision making and patient exclusions for guideline-recommended therapies if they are present. The decision to proceed with device therapy requires in-depth and often multiple discussions with patients, but these discussions may not always be documented in the medical record, especially when the decision is to forego use of an otherwise indicated device.9 Use of an aldosterone antagonist requires close monitoring of patients’ potassium levels and renal function.2 As a result, patients may have appeared in the medical record to be eligible for treatment, but an appropriate decision led to withholding that intervention. Conversely, because there was incomplete documentation of NYHA functional class and QRS duration, other patients who might have been eligible for treatment with improved documentation were not included in the quality measures. Quality improvement efforts should also focus on better documentation of patient eligibility for guideline-recommended therapies, patient exclusions if present, and clinical decision making.
Cardiology practices participating in IMPROVE HF in aggregate were better able to translate the strongest clinical evidence and national guideline-recommended therapies into routine clinical practice and/or to provide better documentation of the therapies given after implementing the performance improvement initiative. Unlike other initiatives that depended on self-reported data, which may introduce the potential for bias, data analyzed in this study were abstracted from patient medical records by independent operators. Representative samples of patients were generated, decreasing the likelihood of selective case ascertainment. National guidelines recommend the use of programs to identify appropriate patients for therapy (Class I), to provide practitioners with useful reminders based on the guidelines (Class IIa), and to continuously assess the success achieved in providing these therapies to patients who can benefit from them (Class IIa).2 The demonstration of substantial change in the use of guideline-recommended therapies among practices participating in IMPROVE HF provides direct evidence in support of these guideline recommendations. Because this program, except for practice-specific data collection, used existing personnel and resources within each practice, it may be more scalable and sustainable than alternative models to improve outpatient care.
Certain limitations are inherent in the design of IMPROVE HF and should be considered when interpreting these findings. Data were collected by medical chart review. Thus, the quality and validity of these data depend on the accuracy and completeness of the medical records and abstraction process. Importantly, determination of patient eligibility and utilization rates for each of the 7 quality measures was based on this documentation. It is possible that the changes in treatment rates found in this study may be attributable in part to variations or inaccuracies in the medical record or data abstraction process. A proportion of patients considered eligible for treatment but not treated at each time point may have had contraindications or other reasons that prevented treatment but were not documented. Documentation in the medical record of the provision of HF education may not be a reliable indication that the information was understood or sufficient to meet the educational needs of patients and their caregivers. As expected in the clinical practice setting, follow-up was not available for all patients. However, improvements in the quality measures were still observed when the analysis was confined to paired analyses of patients with 24-month follow-up data and the 6- and 18-month single-point-in-time cohorts. Although this study assessed improvements in use of therapies at outpatient cardiology practices exposed to a performance improvement initiative, the design does not support evaluation of the effect of the intervention on clinical outcomes. Additional studies are required to establish that the improved use of guideline-recommended therapies actually results in improved clinical outcomes for HF patients in the real-world outpatient cardiology practice setting. It is also not possible to determine whether the intervention had unintended consequences such as increasing the use of therapies not included in current guideline recommendations. Furthermore, individual intervention components that were more or less efficacious in facilitating changes in clinical practice cannot be determined. Although the amount of missing data in the longitudinal cohort is greater than that seen in clinical trials, there is no current benchmark for this kind of quality improvement data collection in the outpatient practice setting, and given the variances in patient mobility and access to care, the level of missing data seen in IMPROVE HF may be within expectations. Although aldosterone antagonists, CRT, and ICD therapy have been shown to reduce mortality and are Class I recommendations for eligible patients in the American College of Cardiology/American Heart Association guidelines, the outpatient performance measure sets do not include measures for these therapies. Some experts question whether such measures are appropriate for quality assessment and improvement.2,13 The practices participating in IMPROVE HF were self-selected; thus, these findings may not apply to practices that differ in patient case mix, baseline care patterns, motivation, and resources from IMPROVE HF outpatient cardiology practices. In addition, all patients received care at outpatient cardiology practices, whereas many HF patients in the United States are treated in primary care settings. We do not know how this performance improvement intervention would affect clinical practice in primary care settings. Patients in IMPROVE HF were drawn from a representative sample from each practice, had documented left ventricular function, and had at least 2 office visits in the last 2 years. These factors may have introduced ascertainment bias, so these patients may not be entirely representative of the outpatient population of patients with HF in cardiology practices.
IMPROVE HF was not a randomized clinical trial with a concurrent control group of outpatient cardiology practices not exposed to the performance improvement intervention. Therefore, it is possible that the improvements in quality measures may have been influenced by secular trends and factors other than participation in this study. However, participation in IMPROVE HF was associated with substantial improvements in the use of HF therapies that were not part of any other known concurrent local or regional performance improvement initiatives for HF, which makes it less likely that the observed improvements were due to external factors. Furthermore, quality-of-care improvement initiatives that did include concurrent control groups such as the Guidelines Applied in Practice (GAP) Project resulted in changes consistent with those found in IMPROVE HF.18
IMPROVE HF is the largest outpatient, cardiology practice–based performance improvement program dedicated to improving the use of guideline-recommended therapies in eligible patients with HF with reduced LVEF and previous MI with reduced LVEF. Practices participating in IMPROVE HF demonstrated a significant increase in the use of 5 of 7 guideline-recommended therapies in eligible patients without contraindications. The results of this study suggest a favorable impact of applying performance improvement techniques of clinical decision support, reminder systems, guideline-driven care improvement tools, educational outreach, collaborative support, performance profiling, and feedback in real-world cardiology practices. Implementation of this defined and scalable practice-specific intervention may enhance the use of guideline-recommended HF therapies previously demonstrated to improve outcomes. These findings may also help to establish a model and framework for future performance-improvement programs administered in the outpatient setting.
CommGeniX LLC provided technical copyediting support with funding from Medtronic Inc.
Source of Funding
The IMPROVE HF registry and this study are sponsored by Medtronic Inc, Minneapolis, Minn.
Dr Fonarow has received research grants from the National Institutes of Health; honoraria from Medtronic, GlaxoSmithKline, Pfizer, St. Jude Medical; and consultant frees from Medtronic, Novartis, and St. Jude Medical. Dr Albert has received consultant fees from Medtronic and speakers’ bureau fees from GlaxoSmithKline. Dr Curtis has received research grants from Medtronic and St. Jude Medical; speakers’ bureau fees from Medtronic, St. Jude Medical, Boston Scientific, Biotronik, and Sanofi-Aventis; honoraria from Medtronic and Sanofi-Aventis; consultant/advisory board fees from St. Jude Medical and Biosense Webster; and fellowship support from Medtronic. Dr Gattis Stough has received consultant/advisory board fees from GlaxoSmithKline, Gilead Sciences, Medtronic, and Scios. Dr Gheorghiade has received consultant fees from Abbott Laboratories, Astellas, Astra Zeneca, Bayer Schering Pharma AG, CorThera Inc, Cytokinetics Inc, DebioPharm SA, Errekappa Terapeutici (Milan, Italy), GlaxoSmithKline, Johnson & Johnson, Medtronic, Merck, Novartis Pharma AG, Otsuka Pharmaceuticals, Pericor Therapeutics, Protein Design Laboratories, Sanofi-Aventis, Sigma Tau, and Solvay Pharmaceuticals. Dr Heywood has received research grants from Biosite, Medtronic, and St. Jude Medical; speakers’ bureau fees/honoraria from GlaxoSmithKline, Medtronic, AstraZeneca, Novartis, Actelion, St. Jude Medical, Otsuka, and Boston Scientific; and consultant/advisory board fees from Emerge, Medtronic, and Actelion. Dr Mehra has received consultant fees from Medtronic, Orqis, Johnson & Johnson, Solvay, PeriCor, and St. Jude Medical and grants/research support from Maryland Industrial Partnerships, Maryland Tobacco Fund, National Institutes of Health, and Orqis. Dr O'Connor has received consultant fees from Forest, Medtronic, Amgen, Medpace, Impulse Dynamics, Actelion, Cytokinetics, Roche, and Trevena. Dr Reynolds has received research grants from Medtronic and Biotronik; speakers’ bureau fees from Medtronic and Sorin, and consultant fees from Medtronic. Dr Walsh has received consultant fees from ARCA, Boston Scientific, Medtronic, and United HealthCare. The other authors report no conflicts. Medtronic provided financial and material support for the IMPROVE HF registry. The sponsor had no role or input into the selection of endpoints or quality measures used in the study. A contract research organization, Outcome Sciences Inc (Cambridge, Mass), independently performed the practice site chart abstractions for IMPROVE HF and is responsible for performing data checks, storing site-specific and aggregate data, and providing benchmarked quality-of-care reports to practice sites. The contract research organization receives funding from Medtronic. Individually identifiable practice site data are not shared with either the steering committee or the sponsor. The authors had complete control and authority over the study design, the manuscript preparation, and the decision to submit this manuscript to Circulation for publication. The manuscript was submitted to Medtronic before submission for publication.
Lloyd-Jones D, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K, Ford E, Furie K, Go A, Greenlund K, Haase N, Hailpern S, Ho M, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott M, Meigs J, Mozaffarian D, Nichol G, O'Donnell C, Roger V, Rosamond W, Sacco R, Sorlie P, Stafford R, Steinberger J, Thom T, Wasserthiel-Smoller S, Wong N, Wylie-Rosett J, Hong Y. Heart disease and stroke statistics—2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2009; 119: e21–e181.
Hunt SA, Abraham WT, Chin MH, Feldman AM, Francis GS, Ganiats TG, Jessup M, Konstam MA, Mancini DM, Michl K, Oates JA, Rahko PS, Silver MA, Stevenson LW, Yancy CW, Antman EM, Smith SC Jr, Adams CD, Anderson JL, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Jacobs AK, Nishimura R, Ornato JP, Page RL, Riegel B. ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure): developed in collaboration with the American College of Chest Physicians and the International Society for Heart and Lung Transplantation: endorsed by the Heart Rhythm Society. Circulation. 2005; 112: e154–e235.
Shah B, Hernandez AF, Liang L, Al-Khatib AM, Yancy CW, Fonarow GC, Peterson ED, for the Get With the Guidelines Steering Committee. Hospital variation and characteristics of implantable cardioverter-defibrillator use in patients with heart failure: data from the GWTG-HF (Get With The Guidelines–Heart Failure) registry. J Am Coll Cardiol. 2009; 53: 416–422.
Piccini JP, Hernandez AF, Dai D, Thomas KL, Lewis WR, Yancy CW, Peterson ED, Fonarow GC, for the Get With the Guidelines Steering Committee and Hospitals. Use of cardiac resynchronization therapy in patients hospitalized with heart failure. Circulation. 2008; 118: 926–933.
Fonarow GC, Yancy CW, Albert NM, Curtis AB, Gattis Stough W, Gheorghiade M, Heywood JT, McBride ML, Mehra MR, O'Connor CM, Reynolds D, Walsh MN. Heart failure care in the outpatient cardiology practice setting: findings from IMPROVE HF. Circ Heart Fail. 2008; 1: 98–106.
Fonarow GC, Abraham WT, Albert NM, Gattis Stough W, Gheorghiade M, Greenberg BH, O'Connor CM, Pieper K, Sun JL, Yancy CW, Young JB, for the OPTIMIZE-HF Investigators and Hospitals. Influence of a performance-improvement initiative on quality of care for patients hospitalized with heart failure: results of the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients With Heart Failure (OPTIMIZE-HF). Arch Intern Med. 2007; 167: 1493–1502.
Fonarow GC, Heywood JT, Heidenreich PA, Lopatin M, Yancy CW, for the ADHERE Scientific Advisory Committee and Investigators. Temporal trends in clinical characteristics, treatments, and outcomes for heart failure hospitalizations, 2002 to 2004: findings from Acute Decompensated Heart Failure National Registry (ADHERE). Am Heart J. 2007; 153: 1021–1028.
Fonarow GC, Yancy CW, Albert NM, Curtis AB, Gattis Stough W, Gheorghiade M, Heywood JT, Mehra M, O'Connor CM, Reynolds D, Walsh MN. Improving the use of evidence-based heart failure therapies in the outpatient setting: the IMPROVE HF performance improvement registry. Am Heart J. 2007; 154: 12–38.
Bonow RO, Bennett SJ, Casey DE Jr, Ganiats TG, Hlatky MA, Konstam MA, Lambrew CT, Normand SLT, Piña IL, Radford MJ, Smith AL, Stevenson LW, Burke G, Eagle KA, Krumholz HM, Linderbaum J, Masoudi FA, Ritchie JL, Rumsfeld JS, Spertus JA. ACC/AHA clinical performance measures for adults with chronic heart failure: a report of the American College of Cardiology/American Heart Association Task Force on Performance Measures (Writing Committee to Develop Heart Failure Clinical Performance Measures): endorsed by the Heart Failure Society of America. J Am Coll Cardiol. 2005; 46: 1144–1178.
Eagle KA, Montoye CK, Riba AL, DeFranco AC, Parrish R, Skorcz S, Baker PL, Faul J, Jani SM, Chen B, Roychoudhury C, Elma MA, Mitchell KR, Mehta RH, for American College of Cardiology’s Guidelines Applied in Practice (GAP) Projects in Michigan and American College of Cardiology Foundation’s Guidelines Applied in Practice Steering Committee. Guideline-based standardized care is associated with substantially lower mortality in Medicare patients with acute myocardial infarction: the American College of Cardiology’s Guidelines Applied in Practice (GAP) projects in Michigan. J Am Coll Cardiol. 2005; 46: 1242–1248.
Heart failure (HF) is a chronic progressive disease that results in substantial morbidity, mortality, and expenditure of healthcare resources. Despite compelling scientific evidence and professional society guidelines, beneficial therapies for HF remain underused and inconsistently applied in many care settings. New, more effective approaches to improve the use of guideline-recommended therapies for HF are needed. The Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF) prospectively tested a multidimensional practice-specific performance improvement intervention on the use of guideline-recommended therapies for HF in outpatient cardiology practices. Performance data were collected in a random sample of HF patients from 167 outpatient cardiology practices at baseline (pre-intervention), longitudinally following intervention at 12 and 24 months, and in unique patient cohorts at 6 and 18 months. Participants included 34 810 patients with reduced left ventricular ejection fraction and chronic HF or post–myocardial infarction. Interventions included clinical decision support tools, structured improvement strategies, and chart audits with performance feedback. The performance improvement intervention resulted in significant improvements in 5 of 7 quality measures at the 24-month assessment compared to baseline. Improvements in the unique single-point-in-time cohorts were smaller. The results of this study suggest a favorable impact of applying performance improvement techniques of clinical decision support, reminder systems, guideline-driven care improvement tools, educational outreach, collaborative support, performance profiling, and feedback in real-world cardiology practices. These findings may also help to establish a model and framework for future performance-improvement programs administered in the outpatient setting.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109.934471/DC1.