Racial differences in the use of invasive cardiac procedures: A continuous quality improvement approach☆☆☆★★★♢

See related article on page 507.

Several large studies have demonstrated that, in a variety of clinical settings, black patients are significantly less likely than white patients to undergo cardiac catheterization, coronary angioplasty, and coronary artery bypass surgery.1, 2, 3, 4, 5 Given the potential clinical implications of these disparities in patient management, it is essential to search for underlying reasons for these differences. This effort should attempt to determine whether the disparities are clinically justified. If not, then the issue of black/white differences in invasive cardiac procedure use merits public policy initiatives to equalize access to care. To date, however, the search for underlying causes has been unrewarding. It has proven difficult to identify specific clinical, socioeconomic, or psychological factors that explain the black/white disparities in the use of these procedures. Thus studies of previously unexplored factors, such as the investigation by Gregory et al6 in this issue of the Journal, are sorely needed.

These authors attempt to determine whether black/white differences in the use of invasive cardiac procedures in patients with acute myocardial infarction can be explained by the availability of these services at the first admission hospital, or whether they are related to any of several other factors, including comorbidity, infarct location, or socioeconomic issues. They found that, after controlling for these factors, black patients remained significantly less likely than white patients to undergo cardiac catheterization, angioplasty, or bypass surgery. However, for those younger than 65 years, in the model predicting angioplasty or bypass surgery there was a significant interaction between race and the availability of these procedures at the first admission hospital, suggesting that procedure availability does contribute to this black/white difference. In addition, the racial disparities in cardiac catheterization, angioplasty, and bypass surgery were greatest among individuals admitted to hospitals that did not provide these services.

One suggestion from these results is that, after myocardial infarction, blacks are less likely than whites to be transferred to a second hospital to undergo an invasive cardiac procedure. The investigators also demonstrate that after the initial admission for myocardial infarction, blacks are less likely than whites to be readmitted to a hospital over the next year. Although there are several possible explanations for these findings, they intimate that there may be a black/white difference in the intensity of treatment provided after initial management of myocardial infarction.

Taken together, the primary and secondary findings of this study suggest that procedure availability at the index admission hospital, and the related issues of transfer and follow-up, do contribute to racial differences, but that they do not fully explain them. This result is consistent with data from the National Hospital Discharge Survey and the Veterans Affairs hospitals, in which controlling for admission hospital also failed to explain black/white disparities in the use of these procedures.7

Consequently, Gregory et al6 were able to provide only limited insight into the potential reasons for the disparities in procedure use that they observed. This limitation is related both to problems inherent in administrative data sets and to the lack of direct assessment of patients’ clinical, psychological, and socioeconomic characteristics. For example, it is possible that the nature and severity of coronary artery disease varies between blacks and whites, and thus a higher proportion of whites has an appropriate indication for catheterization and/or revascularization. Alternatively, blacks may have a higher frequency of contraindications to these procedures, such as renal failure. Yet another possibility is that white patients are receiving unnecessary procedures. Analyses of cardiac procedures performed in New York state have demonstrated that 24% of coronary angiograms, 42% of coronary angioplasties, and 9% of bypass operations had either uncertain or inappropriate indications. These studies included very few black patients, and possible associations between race and appropriateness were not addressed.8, 9, 10 These forms of clinical data are not available in administrative claims records, and thus Gregory et al6 could not address these issues.

Although it is necessary to address more thoroughly the possibility that differences in procedure use are based on clinical differences between blacks and whites, most likely the results of this investigation will also be unrevealing. In a retrospective analysis of the patients undergoing catheterization at Duke University Medical Center, after controlling for the number of significantly narrowed coronary arteries, a significant black/white difference in the rate of bypass surgery persisted. In fact, in the subset of patients with the strongest indication for surgery, the black/white difference was the greatest. Moreover, controlling for multiple measures of comorbidity, such as ejection fraction, peripheral vascular disease, and coronary risk factors, also failed to explain this difference.

Another set of factors that may contribute substantially to racial differences in invasive cardiac procedure use is socioeconomic in nature. It has been suggested that race may actually serve as a proxy for distinct cultural, economic, and demographic factors that may influence the delivery of health care.11 The primary patient level data on socioeconomic factors in New Jersey’s administrative data set is information on health insurance status, and here there are large differences in insurance status by race. For those individuals less than 65 years of age, only 39.5% of blacks, compared with 67.5% of whites, had indemnity insurance. Also, the impact of potential differences in supplemental insurance coverage for those older than 65 years was not assessed in this study.

A final set of factors that may affect black/white differences is unfortunately not found in administrative or clinical data sets. These factors relate to interactions between patients and physicians and between patients and health systems. Gregory et al6 did not address these issues, but they have been studied in prior prospective investigations and the results provide insight into the impact of these interactions. For example, Schecter et al12 demonstrated that blacks were more likely than whites to disagree with a physician’s recommendation to undergo catheterization. In the Coronary Artery Surgery Study (CASS), blacks were significantly more likely than whites to refuse bypass surgery.13 The specific reasons for these disagreements are unclear, but they may relate to a lack of trust in the health system, or in specific physicians, particularly if they are of a different race.14 They may also reflect a relatively greater fear or dislike of invasive procedures, which may be grounded in cultural or religious beliefs.15 Alternatively, the disagreements may emanate from poor communication between physicians and patients. A patient’s familiarity with a cardiovascular procedure has been demonstrated to be strongly associated with willingness to ultimately undergo that procedure.

The factors involved in physician decision-making, and in physicians’ interactions with their patients, also merit investigation. In CASS, although black patients on average had the same number of diseased coronary arteries as white patients, and although blacks had more severe angina, physicians were less likely to recommend bypass surgery to the black patients. More recently, in a study of a multimedia computer survey instrument to assess physician management of patients with chest pain, black women were significantly less likely than others to be referred for cardiac catheterization.16 A concerning potential culprit here is that physicians may harbor subconscious prejudices.17 This factor may contribute not only to black/white differences in invasive cardiac procedures, but also to disparities that have been demonstrated for several other medical interventions.18, 19

The use of administrative data sets to document racial differences in access to care, and to generate hypotheses regarding underlying causes, is a valuable component of health services research. However, these endeavors are only the beginning of a process that is intended to ultimately improve public health. The focus of future investigations must therefore shift to the confirmation of underlying reasons for these differences in treatment. Unfortunately, the currently available data sets do not contain sufficient information to achieve this goal. The study by Gregory et al6 highlights the lack of data on why racial differences in access to care occur in the health care system. This information is essential because it may identify targets for programs to remedy this problem.

The challenge now is to modify data collection efforts, such as New Jersey’s Myocardial Infarction Data Acquisition System (MIDAS), to gather information on several of the more complex issues that may underlie racial differences in the use of invasive cardiac procedures. As part of this effort, New Jersey should collect additional clinical and socioeconomic data on patients admitted with myocardial infarction. For example, to address the issue of intensity of post–myocardial infarction care, the state could specifically track the number of patients transferred to a second institution for a procedure as well as the frequency of posthospitalization clinic visits. Attending physicians could be asked to complete care maps that explain postinfarction procedure use. These plans might include questions relating to indications and contraindications to catheterization and revascularization, as well as further inquiry into infarct severity, including Killip class. Plans are already in place in New Jersey for physicians to more thoroughly document post–myocardial infarction catheterization. Key pieces of information to collect include the number of diseased vessels and the morphologic features of the atherosclerotic lesions, the caliber of target vessels, the ejection fraction, and the presence of coexistent valvular disease. Documentation of the results of adjunctive stress tests and assessments of myocardial viability may also be helpful.

Additional pieces of socioeconomic information that could assist this effort might include more detailed insurance information, such as data on the use of supplemental insurance policies for those aged 65 or older, and, for those younger than 65 years, assessments of benefits available through indemnity insurance. Also, detailed patient-level data on income, education, and employment status may provide valuable insight into the impact of patient characteristics on procedure use.

Given the suggestive evidence that physician-patient interactions are a key determinant of the ultimate use of invasive procedures, the state should also gather information relating to these relationships. Patients should be queried regarding whether they like their physicians, whether they trust them, and whether the explanations of procedures were sufficient to help them understand their options. They also could be asked about their personal beliefs regarding medical procedures.

The continued use of the MIDAS system also has the potential to better define the impact of differential access to care on outcome. Gregory et al6 note that the overall adjusted 1-year mortality rate did not significantly differ between blacks and whites. This finding may simply be related to the size of the study population. In fact, there was an overall trend toward increased unadjusted 1-year mortality in the black patients. This trend was caused by an increase in unadjusted mortality rate in blacks younger than age 65 compared with whites younger than 65 of nearly 25%. The power of this study to assess the statistical significance of this finding was only 0.43. Thus, if more patients had been included, this result could have achieved significance. This finding is consistent with prior studies demonstrating that mortality rate after myocardial infarction is greater in blacks than in whites.20, 21

If these efforts ultimately identify factors that explain black/white differences in the use of invasive cardiac procedures, and if they also confirm that in New Jersey there are racial differences in outcomes after myocardial infarction, then the state will have the proper framework in which to begin to solve this problem. The underlying factors could serve as targets for health care policy initiatives designed to improve the care of black patients. Given that the explanations are most likely multifactorial, the solutions will almost certainly need to be multifaceted as well.

The MIDAS system could then be used to assess the impact of these programs. These assessments, and subsequent modification of policy efforts, could serve as a continuous quality improvement program for postinfarction care. As such, the New Jersey effort could serve as a model for other states.

In summary, the study by Gregory et al6 reaffirms that there is a significant black/white difference in the use of invasive cardiac procedures. This disparity cannot be explained simply by the standard clinical factors contained in administrative data sets. Instead, the explanations are most likely multifactorial, and may include other clinical variables as well as complex socioeconomic and psychological issues, many of which relate to the doctor-patient relationship. The authors’ findings extend the current literature, which has documented racial disparities in access to invasive cardiac procedures for the past 3 decades.22 Additional studies of systematic strategies to address these issues should be promptly initiated. The results of these analyses may point to potential means to reduce racial disparities in evaluation and treatment.

This strategy holds the potential to improve the quality and length of life of a significant subset of patients after myocardial infarction.