American College of Cardiology

The modern cardiac catheterization laboratory is an amalgamation of complex, highly sophisticated medical and radiological instrumentation used in the diagnosis and management of patients with not only chronic stable disease, but also acute life-threatening illnesses. In any complex, procedure-oriented area, it is necessary to have a well-organized program of quality assurance that focuses on individual and laboratory outcomes. In addition, a continuous program of quality improvement should be implemented to provide ongoing feedback and structure for change. The following discussion summarizes the key components of a QA program for both diagnostic and interventional cardiac catheterization laboratories. These components are (1) clinical proficiency, (2) equipment maintenance and management, and (3) a QI process. A fourth component, radiation safety, is discussed separately later in this document.

A. Clinical Proficiency

The assessment of clinical proficiency in the catheterization laboratory is based on a composite of cognitive skills, procedural conduct, and clinical judgment. A deficiency in any one element is enough to worsen clinical outcomes; thus, all elements must be considered. Unfortunately, there is no unique source that details “how to do things correctly.” Although clinical experience is the sine qua non of proficiency, the myriad of techniques and technology preclude rigid delineation of a singular “right way.” There is, however, one incontrovertible bottom line—patient outcomes.

1. Patient Outcomes in the Diagnostic Cardiac Catheterization Laboratory

a. Rates of “Normal” Cardiac Catheterizations

The frequency of normal hemodynamic and angiographic findings at diagnostic catheterization is a function of the pretest likelihood of disease and the physician’s clinical acumen. For purposes of definition, “normal” coronaries are defined as those with no or physiologically insignificant diameter stenosis by visual inspection in patients studied specifically to assess coronary anatomy. Few contemporaneous sources of data give an acceptable percentage of normal cases. Administrative databases generally lack the requisite clinical information, whereas most clinical databases fail to include such preprocedural data as preopertive (procedure) diagnosis and appropriate ancillary diagnostic data. In an exhaustive, although now dated, review of coronary arteriography, the RAND Corporation study group found rates of normal diagnostic cardiac catheterization studies ranging from 9% to 36% (average, 21%) ⁽¹⁴⁾. These data must be viewed circumspectly, given the relatively unsophisticated x-ray imaging systems in use at that time, the variable criteria for “normal,” and the differing pretest likelihood of finding significant disease. In the Coronary Artery Surgery Study (CASS) Registry ⁽¹⁵⁾, the rate of normal arteriograms was 19%, although the appropriateness of extrapolating these data to the present is also questionable. More recent data from the Society for Cardiac Angiography and Interventions indicate that the frequency of normal angiograms is 20% to 27%, which appears to vary little over a reporting period of several years ^(16,17).

It is recognized that many cardiac catheterization studies include patients with insignificant disease (<50% coronary diameter narrowing by visual estimate). Among those considered “normal” it is evident that many patients may have significant coronary plaque burden before the coronary lumen is obviously reduced. Clearly many acute coronary syndromes occur in patients without significant luminal narrowing. In addition, certain clinical syndromes may relate to coronary endothelial dysfunction. Some laboratories may also have a high prevalence of patients studied for non-coronary issues, such as pulmonary hypertension, cardiomyopathy, valvular disease, or adult congenital heart disease. These issues should be taken into account when assessing the rate of “normal” cardiac catheterization procedures performed by any facility.

The vast majority of reported data refer to coronary artery disease. It is of interest to note that no data are reported for evaluation of hemodynamic problems. This may reflect the proliferation of noninvasive modalities as an integral part of the cardiologic evaluation of patients with suspected valvular or myocardial disease. Nevertheless, there are occasions when cardiac catheterization is recommended to clarify uncertainties related to valvular disease or ventricular function that remain after noninvasive assessment.

b. Complication Rates During Diagnostic Catheterization

There is extensive literature on the major complications of diagnostic cardiac catheterization ⁽¹⁸⁾. Fortunately, the (composite) rate of major complications is “acceptably” low at 1% to 2% (Table 4). As expected, the likelihood of major complications increases significantly with the severity of the underlying cardiac and noncardiac disease ⁽¹⁹⁾. Patients with both valvular and coronary artery disease are slightly more likely to sustain a complication than patients with isolated coronary artery disease ⁽²⁰⁾. Although complications encountered in patients with valvular or myocardial disease are more likely to reflect the patient's underlying clinical status, specific complication rates for transseptal catheterization ⁽²¹⁾ and endomyocardial biopsy ⁽²²⁾ have been reported and fall within the range referenced above. Because of patient selection, the likelihood of complications during outpatient studies is less than that found during inpatient examinations ⁽¹⁹⁾, although the constantly changing definition of “outpatient” may blur this distinction. It must be acknowledged that at present, dynamic changes are occurring in the choice of access site for procedures, the caliber of diagnostic catheters, and the means of achieving access site hemostasis. How these variables will change complication rates is unknown, although it is unlikely that any alternative access sites or vascular occlusion devices will significantly affect the already low major complication rate.

c. Diagnostic Accuracy and Adequacy

An important, although generally ignored area, is that of the completeness and diagnostic accuracy of catheterization procedures. Incomplete or aborted procedures, technically inadequate procedures that fail to obtain the critical information for diagnostic purposes, and erroneous interpretation of the acquired information are markers of quality no less important than the previous 2 areas. Failure to engage coronary arteries selectively often results in insufficient opacification of the artery to accurately assess stenosis. Failure to identify or engage all bypass grafts selectively is frequently another reason that angiograms are incomplete and need to be repeated. Inability to recognize the presence of coronary arteries with anomalous origins also contributes to this problem. Understandably, there is an absence of literature on this subject. The implications of inadequate or incomplete studies are significant and range from the need to perform repeat procedures to obtain the key information to performance of unnecessary and more invasive procedures. In the PCI era, the need for high-quality angiography is great. Inadequate attention to the details of accurate hemodynamic recording in patients with valvular heart disease and the failure to accurately demonstrate coronary anatomy must be viewed as important measures of outcome. It seems clear that inadequate diagnostic procedures as defined above should occur in far fewer than 1% of cases.

d. The Special Case of the “Ad Hoc” PCI

The performance of a coronary interventional procedure at the conclusion of the diagnostic session presents several important issues for assessment of quality. Complications engendered during diagnostic catheterization and angiography, e.g., coronary dissection or abrupt occlusion, may well be treated with prompt intervention. Does the success of the intervention mitigate the inciting event? Although the composite procedure was “successful,” how is the original complication recorded? The majority of such “ad hoc” procedures are currently performed as the result of efforts to improve cost-efficiency as well as patient convenience and satisfaction. The ad hoc procedure also facilitates the management of patients with both stable and unstable coronary syndromes. In these cases, complications encountered during the interventional portion of the procedure should be attributed to the interventional procedure and not to the antecedent diagnostic study. Given the increasing use of the hybrid approach, it will be important to carefully define its indications, clinical outcomes, and overall cost-effectiveness.

2. Patient Outcomes in the Interventional Cardiac Catheterization Laboratory

Although patient outcomes are clearly the most important indicators of proficiency and competency in interventional cardiology ⁽²⁾, they are arguably the most difficult to quantify accurately. The importance of risk-adjustment of crude event frequencies cannot be overstated ⁽³⁵⁾. Therefore, it is essential that careful and complete preprocedural and intraprocedural information be reliably collected, sorted, and analyzed. Given that operator and institutional outcomes depend on many demographic, clinical, anatomic, and administrative variables, an adequate information system within the laboratory is mandatory. Without a complete recording of such variables, meaningful analysis of event rates is impossible. It is very difficult to risk-adjust variables for low-volume operators based on the wide confidence intervals for outcomes in this situation.

Given this caveat, the emphasis on individual and institutional outcomes is appropriate ⁽²⁾. Operators must be responsible for their actions and resulting consequences. The ability to estimate the likelihood of significant complication ^(36,37), choose devices and conduct procedures appropriately ⁽³⁸⁾, promptly recognize and treat ischemic complications ⁽³⁹⁾, select cases appropriately, and be able to say “no” are hallmarks of an experienced, competent operator. It is the responsibility of the director of the cardiac catheterization laboratory to establish a method of QA to track major events, (e.g., death and serious hemodynamic and/or arrhythmic events). In addition, periodic review of less severe complications (e.g., hematoma or pseudoaneurysm rates) should be part of any ongoing QI program. Admittedly, many outcomes are hard to measure, but there is little ambiguity when outcomes for PCI are either consistently superior (e.g., <2% major complication rate) or consistently suboptimal (e.g., >5% major complication rate). At present, with overall in-hospital mortality averaging 2% and rates of emergent CABG averaging <1%, a major complication rate 3% (95% CI 1.9%, 4.1%) is to be expected.

Table 5 summarizes in-hospital outcomes from recently published data on this subject. Each series includes patients undergoing PCI for a variety of indications, e.g., stable angina, post-infarct angina, and acute MI. The definitions of “elective,” “urgent,” and “emergent” vary among studies. Complication rates (especially bleeding and access site complications) in the GP IIb/IIIa inhibitor era not only vary according to the definition applied, but almost universally reflect the clinical trial literature. Complication rates in community-based practice must await the development of an appropriate data collection instrument. The use of 30-day event rates to benchmark operator performance has been advocated by some ⁽⁴⁰⁾.

Table 6 summarizes representative outcomes from the published literature on PCI for acute MI. Here, too, event rates are unadjusted, and rates of access site and bleeding complications reflect a complex mix of systemic anticoagulation, systemic lytic activity, and the adjunctive use of platelet antagonists. These issues are particularly critical in the interpretation of central nervous system complications during PCI in this setting.

Although the frequencies of adverse events are likely to change over time as the result of continuing improvements in technology, clinical competence and its assessment will remain the foundation on which a QI program rests. Table 7 summarizes current approaches to the assessment of proficiency in coronary intervention for both individuals and institutions.

B. Equipment Maintenance and Management

The modern diagnostic and interventional catheterization laboratory uses many sophisticated radiological, electronic, and computer-based systems, which require a program of rigorous maintenance and troubleshooting. The x-ray imaging system, a crucial component of every laboratory, must be carefully assessed at frequent intervals to detect early signs of deterioration in performance. Unfortunately, this aspect of quality control is the first to be sacrificed in an era of cost cutting.

A program of periodic assessment of system performance and (cine) image quality has been recommended by the Society for Cardiac Angiography and Interventions ⁽⁴¹⁾. Additional programs, which will address issues specific to digital imaging systems, are under evaluation ⁽⁴¹⁾. A representative outline of the performance characteristics needed to assess radiographic cardiac imaging systems is presented in Table 8.

Note that at present the only federally mandated parameter of image performance is the maximum table-top exposure rate (10 R/min) for conventional cardiac fluoroscopy. The concept of minimum performance standards must await universal acceptance of a suitable test instrument for cardiac fluoroscopy. There is considerable heterogeneity across laboratories in selective measurements of image quality ⁽⁴²⁾. Such heterogeneity precludes specific recommendations with respect to what is considered “acceptable” performance. Current-generation imaging systems must be capable at minimum of providing images of sufficient diagnostic quality to enable decision making with respect to intervention and provide sufficient spatial and contrast resolution for the conduct of contemporary coronary intervention.

Interventional procedures occur in environments of high information density. In the past, physiological recorders were used only for the acquisition and recording of analog signals. They are now required to serve as front ends for the increasingly complex gathering of data. These recorders have essentially been transformed into desktop personal computers capable of acquiring, storing, and transmitting data to other sites. Given the critical importance of these data for numerous purposes (e.g., billing, quality assurance, report generation), flawless and lossless transmission must take place all the time. Backup systems and low-cost storage media are essential.

The need for patient safety-related precautions is self-evident. The operational efficiency of infrequently used equipment (e.g., defibrillators) must be tested routinely and appropriate logs kept. Electrical isolation and grounding systems must be regularly assessed. The number of ancillary devices used in coronary intervention (e.g., Doppler and pressure-tipped sensor wires and ultrasound catheters) now requires that electrical safety precautions that were adequate in the past ⁽⁴³⁾ be revisited.

C. Quality-Improvement Program Development

A continuous QI program with regard to clinical proficiency must function under the broad rubric of system-level performance analyses, which should connote a more constructive (rather than punitive) context ⁽³⁸⁾. Table 9 outlines some of the essential elements of such a program.

An overall continuous QI program is only as effective as the commitment of all involved in the process of healthcare delivery. Clearly, the most conspicuous components are procedural outcome and individual operator proficiency. Thus, the emphasis and direction in the profession alluded to above, in which sub-subspecialty “boards” in interventional adult cardiology have been developed, is properly focused on proficiency, both cognitive and technical. For coronary interventional procedures, proficiency is intimately related to procedural volume, although the latter is not synonymous with the former. However, sound quantitative support now exists for these once presumed arbitrary cut points. The situation is less clear with respect to diagnostic catheterization. Given the absence of similar quantitative data for diagnostic procedures, as well as the significantly lower associated morbidity and mortality associated with diagnostic catheterization, operator proficiency may be better assessed in a larger overall context. Rates of normal studies, peer review of diagnostic quality of studies, rates of referral for intervention, and perhaps development of criteria of the appropriateness of these studies are suggested as methods of incorporating physician practice into the QI process of diagnostic procedures. It is recognized that the latter depends critically on the development of locale-specific “pathways” of care. However, “outliers” in this process may be readily identified and constructively advised. Standards of performance and QA in either a diagnostic or an interventional catheterization laboratory must of course originate with the individual. However, processes for credentialing activity and the ongoing assessment of proficiency must be developed in accord with both local governance policies, as well as professionally developed standards. In particular, the granting of privileges by healthcare systems is properly within the legal and ethical purview of these institutions. It is hoped that these systems use criteria similar to those outlined in this document to support the decision to credential physicians and monitor system performance.

The key elements of such a program are (1) the development of a consensus on variables that reflects quality of care, (2) the rigorous prospective collection of these variables, (3) appropriate statistical analysis of the data to identify deficiencies in the process of care, (4) the development of a multidisciplinary approach to problem solving, (5) subsequent data collection with analysis of the specific effect of the solution on the identified deficiency, and (6) benchmarking of the information against national database standards such as the ACC National Cardiovascular Data Registry ⁽⁴⁴⁾. These data are perhaps best presented to involved practitioners at regularly scheduled conferences for appropriate critique and problem solving.

Over a 10-year period, improvements in instrumentation, imaging, data recording, and procedural outcomes have proceeded rapidly. Consequently, continuing education for practitioners beyond the level of training programs has become the norm for the acquisition of many of these skills. Training programs themselves are changing from the traditional 1-year program in interventional cardiology to 2-year programs in some institutions. The development of sub-subspecialty certification boards in interventional cardiology reflects this burgeoning knowledge base. All of this translates into the need to provide continuing education to all members of the team. The implementation of new technology requires a critical evaluation of both the experience in the literature as well as experience within individual institutions. An organized program of didactics coupled with cautious early clinical experience is an ideal mechanism for the introduction of new therapies. These types of programs in conjunction with attendance at regional or national scientific meetings devoted to the unbiased presentation of new data provide a solid infrastructure for credentialing purposes. Attention to this aspect of laboratory QI is critical to maintaining both expertise and morale.

A recent review of cardiac catheterization laboratory settings has outlined certain practical lessons learned by the Laboratory Survey Committee of the Society for Cardiac Angiography and Interventions ⁽⁴⁵⁾. This committee noted that the major QA problems were usually not related to equipment but rather to inadequate laboratory space, lack of a physician medical director, lack of specific operating rules for the laboratory space, and lack of a functioning QA program. Not only must a QA program provide procedural complication information, but a feedback mechanism to modify behavior must be in place.

Benchmark data are important, and because these benchmark data are dependent on a high number of participating laboratories, the Committee strongly recommends that cardiac catheterization laboratories actively participate in the national data registries, such as the ACC-NCDR™.

D. Minimum Caseload Volumes

The use of a specific minimum number of cases to define the quality of operator performance is obviously fraught with problems. Because many laboratories may not adhere to appropriate oversight or may not have an established QA program, it has become popular to define minimum caseloads for both the operators and the laboratory in place of many of the issues described in detail above. Given the low risk for diagnostic cardiac catheterization, the Committee could not arrive at any consensus as to what would constitute a minimum workload for individuals with regard to diagnostic procedures. There have been no data to justify the prior recommendation of at least 150 cases per year ⁽⁵⁾. The minimum diagnostic caseload for the entire laboratory also varies widely from state to state, often depending on the presence of the certificate of need (CON) process or other occasionally arbitrary requirements. It falls upon the director of the laboratory to ensure that all studies in the cardiac catheterization laboratory are of the highest quality. In general, high-volume laboratories have consistently been shown to have fewer complications than low-volume facilities, although quality cannot be deciphered by observing the total laboratory volume alone ⁽²⁾.

Recommendations regarding interventional volumes are noted in Table 7. In general, the Committee thought that the minimum interventional caseload of 75 procedures per year for operators and a minimum performance of 200 cases per year by institutions, with the ideal being 400 cases per year per laboratory, both reasonable and supportable, based on current data ^(3,46). This minimum caseload for operators has also been adopted by the ABIM as a prerequisite for eligibility to take the interventional boards.

Issues of training, competency, and operator volume are important. It was estimated that 6100 physicians performed 428,000 interventional procedures in 1994. These physicians represented 40% of board-certified cardiologists in the United States ⁽²⁸⁾. Over half of the physicians performing interventional procedures in the United States at that time did not meet the current minimum suggested volume recommendations for proficiency within the catheterization laboratory. Operators performing a low volume of interventions might be tempted to expand the indications for diagnostic or interventional procedures in their clinical practice, yet a more aggressive approach to invasive therapies may or may not be in the patient’s best interest. Under these circumstances, low-volume operators may wish to consolidate practices and dedicate 1 individual to perform catheterization-related procedures instead of having multiple physicians perform such procedures.

The ACC/AHA guidelines for PCI ⁽³⁾ have reviewed this issue in depth, noting multiple studies that support a relationship between complications and procedural volume. The lowest complication rates are observed when interventional procedures are performed by higher-volume operators (75 cases per year) with advanced skills (e.g., subspecialty certification) at high-volume institutions. This concept has also been endorsed by the ACC/AHA Task Force on Practice Guidelines for Coronary Angiography and the Society for Cardiac Angiography and Interventions ^(18,45). Ideally, lower-volume operators (<75 cases per year) should only work at institutions that perform >600 procedures per year ⁽³⁾. Even in the high-volume setting, low-volume operators should develop a defining mentoring relationship with a highly experienced operator who performs >150 procedures per year ⁽³⁾.