Medicine is experiencing an unprecedented focus on quantifying and improving health care quality. The American College of Cardiology (ACC) and the American Heart Association (AHA) have developed a multifaceted strategy to facilitate the process of improving clinical care. The initial phase of this effort was to create clinical practice guidelines that carefully review and synthesize available evidence to better guide patient care. Such guidelines are written in a spirit of suggesting diagnostic or therapeutic interventions for patients in most circumstances. Accordingly, significant judgment by clinicians is required to adapt these guidelines to the care of individual patients, and these guidelines can be generated with varying degrees of confidence based upon available evidence. Occasionally, the evidence supporting a particular structural aspect or process of care is so strong that failure to perform such actions reduces the likelihood that optimal patient outcomes will occur. Creating a mechanism for quantifying these opportunities to improve the outcomes of care is an important and pressing challenge.

In the next phase of its quality improvement efforts, the ACC and the AHA created the ACC/AHA Task Force on Performance Measures in February 2000 to spearhead the development of performance measures that allow the quality of cardiovascular care to be assessed and improved. Three nominees from each organization were charged with the task of assembling teams of clinical and methodological experts, both from within the sponsoring organizations and from other organizations dedicated to the care of patients covered by the performance measurement set. These writing committees were given careful guidance with respect to the necessary attributes of good performance measures and the process of identifying, constructing, and refining these measures so that they can accurately achieve their desired goals (1).

The role of the performance measurement writing committee is not to perform primary evaluation of the medical literature. This is undertaken by ACC/AHA guidelines committees. However, performance measurement writing committees work collaboratively with guidelines committees so that the guideline recommendations are written with a degree of specificity that supports performance measurement and so that new knowledge can be rapidly incorporated into performance measurement. Development of ACC/AHA guidelines includes a detailed review and ranking of the evidence available for the diagnosis and treatment of specific disease areas. Published guideline recommendations employ the ACC/AHA Classification of Recommendations I, IIa, IIb, and III (Figure 1).

So as not to duplicate performance measure development efforts, writing committees are also instructed to evaluate existing nationally recognized performance measures using the ACC/AHA “attributes of good performance measures.” The measure specifications are adopted for those performance measures that meet these criteria. Such measures have established validity, reliability, and feasibility and will form the foundation of the ACC/AHA measurement sets. Furthermore, writing committees are encouraged to identify additional performance measures that correspond to those key areas of quality proven to improve patient outcomes.

ACC/AHA Performance Measurement Sets are to be applied in either the inpatient and/or outpatient setting depending on the topic. Inpatient measures are usually, but not always, captured by retrospective data collection; outpatient reviews lend themselves to retrospective or prospective collection. The latter is itself a continuous quality improvement process. The performance measures quantify explicit actions performed in carefully specified patients for whom adherence should be advocated in all but the most unusual circumstances. In addition, the measures are constructed with the intent to facilitate both retrospective and prospective data collection using explicit administrative and/or easily documented clinical criteria, respectively. Furthermore, the data elements required to construct the performance measures are identified and linked to existing ACC/AHA Clinical Data Standards to encourage the standardization of cardiovascular measurement.

While the focus of the performance measures writing committee is to develop measures for internal quality improvement, it is appreciated that other organizations may use these measures for external reporting of provider performance. Therefore, it is within the scope of the writing committee’s task to comment on the strengths and limitations of externally reporting potential performance measures. Specifically, this was done in the inpatient measurement set, where a “Challenges to Implementation” section was included subsequent to the specification, when appropriate (Appendix A).

All the measures contained in this set have limitations and challenges to implementation that might result in unintended consequences when used for accountability purposes. The implementation of these measures for purposes other than quality improvement (QI) require field testing to address issues related to, but not limited to, sample size, reasonable frequency of use for an intervention, comparability, and audit requirements. The way in which these issues are addressed will be highly dependent on the type of accountability system developed, including data collection method, assignment of patients to physicians for measurement purposes, baseline measure setting, incentive system, and public reporting method, among others. The ACC/AHA encourages those interested in working on implementation of these measures for purposes beyond QI to work with the ACC/AHA to understand these complex issues in pilot testing projects that can measure the impact of any limitations and provide guidance on possible refinements of the measures that would make them more suitable for additional purposes.

In the process of facilitating the measurement of cardiovascular health care quality, ACC/AHA Performance Measurement Sets may serve as a vehicle for more rapidly translating the strongest clinical evidence into practice. These documents are intended to provide practitioners with “tools” for measuring the quality of care and for identifying opportunities to improve. Because the target audience and unit of analysis for these measures is the practitioner, they were constructed from the provider’s perspective and were not intended to characterize “good” or “bad” practice, but to be part of a system with which to assess and improve health care quality. It is our hope that an application of these performance measures within a QI system will provide a mechanism through which the quality of medical care can be measured and improved.

Robert O. Bonow, MD, FACC, FAHA

Chair, ACC/AHA Task Force on Performance Measures

The ACC/AHA Heart Failure Performance Measures Writing Committee (hereafter, Writing Committee) was charged with the development of performance measures concerning the diagnosis and treatment of heart failure (HF). These performance measures refer to both hospitalized adult patients (age 18 years or older) with a principal discharge diagnosis of HF and to adults with HF evaluated in the outpatient setting. The Writing Committee independently developed the inpatient performance measures for HF and collaborated with the Physician Consortium for Performance Improvement (Physician Consortium) to develop the outpatient measures. In a concurrent process, the ACC/AHA Task Force on Clinical Data Standards launched the development of the ACC/AHA Key Data Elements and Definitions for Measuring the Clinical Management and Outcomes of Patients With Chronic Heart Failure (2) intended to provide a standardized informational platform for clinical trials, clinical registries, and quality performance measurement for the condition of HF.

Heart failure is a major and growing public health problem in the U.S. It affects 4.9 million people, and 550,000 new cases are diagnosed each year (3). Heart failure is primarily a disease of the elderly (4). Thus, the aging of the population and the prolongation of the lives of cardiac patients by modern therapeutic innovations have led to an increasing incidence of HF. The incidence of HF approaches 10 in 1,000 people over the age of 65, and hospital discharges for HF increased by 157% between 1979 and 2002 (3). In the U.S., the disorder is the underlying reason for 12 to 15 million office visits and 6.5 million hospital days each year (5).

Despite improvements in therapy, the mortality rate in patients with HF has remained high, making implementation of therapies demonstrated to slow the progression of HF imperative. In 2001, 52,800 people died from HF as a primary cause (3). The variability in care documented in the literature demonstrates the opportunity for improvement, which makes performance measurement in cardiovascular disease important.

The costs associated with HF are also large. Heart failure is the most common Medicare diagnosis-related group (DRG), and more Medicare dollars are spent for diagnosis and treatment of HF than for any other diagnosis (6). Medicare reported paying $3.6 billion to beneficiaries for care of HF in 1999, equating to $5,456 per discharge (3). It has been estimated that in 2005, the total direct and indirect cost of HF in the U.S. will be equal to $27.9 billion (3).

Members of the ACC/AHA Heart Failure Performance Measures Writing Committee included a senior clinician, a content expert on HF performance measures, a statistician, various representatives from HF subspecialties, and representatives from the ACC/AHA Heart Failure Guideline Update Writing Committee and ACC/AHA Heart Failure Clinical Data Standards Writing Committee to ensure consistency across the documents. The Writing Committee also included members of the Heart Failure Society of America (HFSA), the American Academy of Family Practitioners (AAFP), the American College of Physicians (ACP), and a nurse scientist from the Nursing Council of the American Heart Association.

The work of the Writing Committee was supported exclusively by the ACC and the AHA. Writing Committee members volunteered their time, and there was no commercial support. Meetings of the Writing Committee were confidential and attended only by committee members and staff. All Writing Committee members with relationships with industry relevant to this topic declared these in writing according to standard ACC and AHA reporting requirements; additionally, members verbally acknowledged these relationships to the Writing Committee. Please see Appendix D for relevant Writing Committee relationships with industry. The collaboration with the Physician Consortium to develop the outpatient measures received additional volunteer support from Physician Consortium members and staff support from the American Medical Association. Please see Appendix E for ACC/AHA/Physician Consortium Writing Group relevant relationships with industry. In addition, Appendix F includes relevant relationships with industry information for all peer reviewers of this document.

The previously published outpatient ACC/AHA/Physician Consortium Heart Failure Clinical Performance Measures (7) underwent a period of public comment from January 22, 2003 through February 12, 2003, peer review and approval from the respective boards of the ACC and the AHA, and approval by the Physician Consortium. These outpatient measures are being validated through pilot testing as part of the Doctors’ Office Quality (DOQ)–Information Technology Project of the Centers for Medicare and Medicaid Services (CMS). Some clarifications, modifications, and additions to the published ACC/AHA/Physician Consortium HF outpatient measures have been made in this document and have been incorporated in their updated ACC/AHA/Physician Consortium Heart Failure Clinical Physician Performance Measurement Set.

During the period February 3, 2005 to March 4, 2005, the complete ACC/AHA HF measurement set, incorporating both inpatient and outpatient measures, underwent a 30-day public comment period during which time ACC and AHA members, as well as other health professionals, had an opportunity to review and comment on the document in advance of its final approval and publication. Over 30 responses were received. The official peer and content review of the document was conducted simultaneously with the 30-day public comment period, with three peer reviewers nominated by the ACC and three reviewers nominated by the AHA. Additional comments were sought from clinical content experts and performance measurement experts.

The ACC/AHA Clinical Performance Measures for the Care of Adults with Chronic Heart Failure was adopted by the respective boards of the ACC and AHA in August 2005. These measures will be reviewed for currency once a year and will be updated as needed. They will be considered valid until they are updated or rescinded by the ACC/AHA Task Force on Performance Measures.

The development of performance measures involves identification of a set of measures targeted toward a particular patient population, observed over a particular care period. To achieve this goal, the ACC/AHA Task Force on Performance Measures has outlined and published a methodology of sequential tasks that performance measures writing committees are required to complete (1). The following sections delineate how the Writing Committee applied this methodology to the topic of HF.

The ACC/AHA 2005 Guideline Update for the Diagnosis and Management of Chronic Heart Failure in the Adult (ACC/AHA 2005 HF Guideline Update) (8) classified HF into four stages (Table 1). For the purpose of this document, only the latter two stages, which qualify for the traditional diagnosis of HF (Stages C and D), were considered for inclusion in the measure population. Thus, the inpatient and outpatient performance measurement sets do not apply to patients for whom established risk factors and structural disorders occur without left ventricular systolic dysfunction or symptoms associated with HF (Stages A and B). In addition, specific diagnosis codes, based on ICD-9-CM (Table 2) should be used to screen and select the inpatient target patient population. These codes correspond to the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) and Centers for Medicare and Medicaid Services (CMS) HF cohort selection codes (i.e., the ICD-9-CM codes previously defined by JCAHO and CMS to screen and select cohorts for HF performance measures).

Given the multiple domains of providing treatment that can be measured, the Writing Committee identified and explicitly articulated the relevant dimensions of care that should be evaluated. As part of the methodology, each potential performance measure was categorized into its relevant dimension of care. Classification into dimensions of care facilitated identification of areas where evidence was lacking, as well as prevented duplication of measures within the set. Diagnostics, Patient Education (including prognosis and etiology), Treatment, and Self-Management were selected as the relevant dimensions of care for HF performance measures in both the inpatient and outpatient settings. A fifth dimension, Monitoring of Disease Status, has been addressed for the outpatient setting but will be evaluated in the future for the inpatient setting and might include items such as documentation of follow-up appointments. The Writing Committee exclusively focused on processes and did not consider outcomes since the purpose of the measures is to assist physicians in improving specific clinical care.

The Writing Committee reviewed the ACC/AHA 2005 HF Guideline Update (8) as the primary source for deriving the measures. In addition, the Writing Committee reviewed current national performance measures and other relevant literature from organizations developing clinical guidelines and quality standards, including:

Explicit criteria exist for the development of performance measures so that they accurately reflect the quality of care, including quantification of the numerators and denominators and clearly evaluating the interpretability, applicability, and feasibility of the proposed measures. To determine which measures will be selected for inclusion in the performance measurement set, the committee identified the Class I and Class III recommendations from the ACC/AHA 2005 HF Guideline Update (8) and specific relevant statements from the Team Management of Patients With Heart Failure: A Statement for Healthcare Professionals From the Cardiovascular Nursing Council of the American Heart Association (12).

Once these recommendations were identified, the Writing Committee rated their potential for use as performance measures utilizing the Rating Form and Guide (Appendix C). Writing Committee members rated 27 potential measures on 13 dimensions using 5-point Likert scales (1 = lowest rating; 5 = highest rating) against the ACC/AHA attributes for satisfactory performance measures (Table 3) (15).

The rating results of the final question on the rating form, “Overall Assessment,” were used to make the final determination for inclusion of a potential measure in the HF measurement set. Any measure that received a full committee consensus rating of 3 or above in this area (“Overall Assessment”) was advanced to the final measure set. Based on the results, 7 inpatient measures and 12 outpatient measures were advanced. These measure sets were then evaluated in light of the ACC/AHA 2005 HF Guideline Update (8) to determine where measures should be revised and clarified. The Discussion section and measurement set specifications (Appendix A and B) detail the rationale for modifications based on the ACC/AHA 2005 HF Guideline Update (8).

The inpatient target population consists of patients age 18 years or older with a principal discharge diagnosis of HF based on ICD-9-CM (see Table 2). A set of exclusion criteria specific to each inpatient measure was developed. For all the inpatient measures, patients who were under 18 years of age, were transferred to another acute care facility, who left against medical advice, who were discharged to hospice, or who died during the index admission are excluded.

The general period of assessment is the related inpatient hospitalization. The specific time period of interest for each measure is further defined in Appendix A and ranges from within 30 minutes of the index admission to the hospital discharge.

(Table 4) lists the ACC/AHA Heart Failure Performance Measures Inpatient Measurement Set. The inpatient measurement set includes the dimensions of care referenced in the ACC/AHA Methodology for the Selection and Creation of Performance Measures article (1), with the exception of Monitoring of Disease Status (Table 5). Although no current measure exists for this dimension for the inpatient setting, future measure development efforts will examine how to address this gap in the measurement set.

Appendix A provides detailed specifications for each inpatient performance measure including numerator, denominator, period of assessment, method of reporting, sources of data, rationale, clinical recommendations, and challenges to implementation.

To aid in data compilation, a data collection instrument is recommended. A sample instrument is provided in (Table 6). This instrument was created to accompany the associated measures; however, individual institutions may need to modify this instrument or develop a different one based on their local practice patterns and standards. The flowsheet is intended for prospective data collection only. It is not designed to meet the reporting requirements of organizations, such as JCAHO or CMS.

The target population consists of patients age 18 years or older with at least one primary outpatient visit for HF, documentation of HF in the medical records as the primary reason for their visit, or with a principal diagnosis of HF. In addition, patients with documentation of symptoms consistent with Stage C or Stage D HF (Table 1) should be included. A set of exclusion criteria specific to each outpatient measure was developed to further specify the target population (Appendix B).

For the purpose of this document, the outpatient care period is defined as the care provided in an outpatient setting within the time period under evaluation (reporting year).

The outpatient performance measurement set for HF (Table 7) was developed, revised, and approved through collaborative efforts among the ACC, AHA, and the Physician Consortium. The ACC/AHA Heart Failure Performance Measures Writing Committee drafted the candidate outpatient measures that were published by the Physician Consortium in 2003 (7) and were reviewed for update in 2005. Although these outpatient measures, have been previously published (7), the outpatient measurement set included in this document reflects some changes to the outpatient measures based on the evidence review from the 2005 ACC/AHA HF Guideline Update (8).

As shown in (Table 8), the outpatient measurement set includes measures for each of the dimensions of care referenced in the ACC/AHA Methodology for the Selection and Creation of Performance Measures article (1).

Appendix B provides detailed specification for each outpatient performance measure including numerator, denominator, period of assessment, method of reporting, sources of data, rationale, clinical recommendations, and challenges to implementation. Through ongoing efforts, many of the outpatient measures are under consideration for widespread implementation/endorsement by national healthcare organizations (e.g., National Quality Forum [NQF] and the Ambulatory Care Quality Alliance [AQA]).

The outpatient HF performance measures are intended to be used prospectively to enhance the QI process. To aid in data compilation, a data collection instrument is recommended. A sample instrument is provided in (Figure 2), which was developed by the Physician Consortium and adapted to correspond to the outpatient measures included in this document. Individual institutions may need to modify the instrument or develop a different tool, based on their local practice patterns and standards.

The ACC/AHA Clinical Performance Measures for Adults with Chronic Heart Failure address many of the same processes of care as earlier measurement sets published by other organizations. The Writing Committee has been cognizant of the previous efforts of other groups and sought to enhance and clarify measures in ways that reflect the advancement of the underlying science, the complexity of care, and the challenges of accurate and complete data collection. In particular, the current document incorporates the performance measures developed jointly by CMS and JCAHO for management of patients with HF (9).

The Writing Committee recognizes that not all Class I guideline recommendations lend themselves to becoming excellent performance measures, as many do not easily fit the attributes of performance measures in terms of usefulness, accuracy, feasibility, and measurability. Thus, the Writing Committee selected only those Class I recommendations that were considered to perform well as performance measures in the inpatient or outpatient setting. However, in the case of angiotensin-converting enzyme (ACE) inhibitor/angiotensin receptor blocker (ARB) measures, a Class IIa recommendation was considered and used as the basis for clarifying the measure constructed by the committee. Although Class IIa recommendations are not considered for stand-alone measures, in some cases, such as this one, they provide additional information about valid alternative therapies that are considered by the committee for inclusion in a measure set. This change is made with recognition that although ACE inhibitors are preferred as the first option in HF patients who have left ventricular systolic dysfunction (LVSD), physicians should be given credit for prescribing or continuing ARB therapy.

The support for use of ARBs in patients with HF and reduced left ventricular ejection fraction has evolved significantly in response to published clinical trials that showed ARBs as an effective alternative therapy (16) and is recommended in the ACC/AHA 2005 HF Guideline Update (8) as a reasonable alternative therapy. Thus, the Writing Committee decided to revise both the inpatient and outpatient ACE inhibitor measures to include ARB therapy.

The inpatient and outpatient measures are designed to be implemented in either a retrospective chart abstraction process or used as part of a prospective quality improvement process. The data collection tool suggested for use with the inpatient measures (Table 6) permits prospective data capture, as well, and promotes the prospective identification of HF patients. For example, documentation of patient education is often difficult to obtain in a retrospective chart review, but can be easily implemented using a prospective patient management tool. These inpatient and outpatient measures will require testing in practice to determine their validity and may require modification in the future.

The Writing Committee also felt it was important to add exclusion criteria to the measures to recognize that there are justifiable medical and patient reasons for not meeting the performance measures. In the inpatient set, these reasons should be included in the “reasons documented by physician, nurse practitioner, or other healthcare provider for not… .” In the outpatient set, medical and patient reasons for not meeting the measure are listed separately. Documentation of such factors should be encouraged and will provide valuable data for future research and conducting in-depth quality improvement for situations where there seem to be outliers with respect to the number of patients with medical or patient-centered exclusions for the performance measures.

Challenges to implementation of measures are discussed, where applicable. In general, inadequate documentation is the initial challenge of any measurement effort. The fact that these challenges are discussed is not intended as an argument against measurement. Rather, they should be considered as cautionary notes that draw attention to areas where additional focus on research and improvement of the measures should be considered.

Although the committee considered a number of additional potential measures that focus on equally important aspects of care, either the evidence base or more significant challenges to measurement of these components of care across all patients undermined the benefits that might be gained. Such is the case for measurement of implantable cardioverter-defibrillator (ICD) implantation for the reduction of sudden death in patients with severe LVSD and biventricular pacing in appropriate candidates. Currently, there is not a sufficient number of qualified cardiac electrophysiologists in every community to implant ICDs and monitor follow-up in these HF patients as well as to ensure that complications are addressed as these patients progress from HF Class III to Class IV.

The Writing Committee also considered including an inpatient beta-blocker therapy measure. There is no specific guideline recommendation for implementation of beta-blockers in the inpatient setting, although it is recommended that such medications be started when patients are clinically stable. The complexities of establishing the right conditions under which stable HF patients would be included in the measure minus the exclusions would result in so small a denominator that the measure would not be meaningful at this time. The omission of an inpatient beta-blocker measure does not recommend against its initiation in hospitals for appropriate patients.

Three measures were removed from the measurement sets (inpatient and outpatient) post-peer review/public comment, i.e., the “Volume Status and Clinical Assessment on Admission” and “Initial Evaluation of Left Ventricular Systolic Function for Newly Diagnosed HF Patients” measures from the inpatient set, and the “Examination of the Heart” measure from the outpatient set. These measures, although potentially of value for improving patterns of care, were not believed to have been tested in clinical situations to assure their reliability and validity.

To be successful as quality improvement tools, these measures need to be adopted, implemented, and integrated into the patient care. These measurement sets should contribute to the evolution of reporting systems that allow physicians and other health care providers to improve treatment for a critical patient population. Quality improvement is a continuous process, and this document reflects the lessons the practicing community has learned to date in using existing measures and knowledge gained about how they might be improved. The clinical care team should collect data and review adherence to these measures on a routine basis, look for changes, and adjust practice patterns as necessary to improve performance.

American College of Cardiology Foundation

Christine W. McEntee, Chief Executive Officer

Joseph M. Allen, MA, Director, Clinical Decision Support

Tilithia McBride, Senior Specialist, Clinical Performance Measurement

Susan L. Morrisson, Associate Specialist, Clinical Performance Measurement

American Heart Association

M. Cass Wheeler, Chief Executive Officer

Gayle R. Whitman, PhD, RN, FAAN

Fernando Costa, MD, FAHA, Staff Scientist