|
TRENDSU.S. Trends In CABG Hospital Volume: The Effect Of Adding Cardiac Surgery Programs
Hospital coronary artery bypass graft (CABG) volume is inversely related to mortality—with low-volume hospitals having the highest mortality. Medicare data (1992–2003) show that the number of CABG procedures increased from 158,000 in 1992 to a peak of 190,000 in 1996 and then fell to 152,000 in 2003, while the number of hospitals performing CABG increased steadily. Predictably, the proportion of CABG procedures performed at low-volume hospitals increased, and the proportion in high-volume hospitals declined. An unintended consequence of starting new cardiac surgery programs is declining CABG hospital volume—a side effect that might increase mortality.
CORONARY ARTERY bypass graft (CABG) surgery is among the most frequently performed surgical procedures in the United States. Because this surgery is both profitable and prestigious, many community hospitals have introduced cardiac surgery programs.1 Recently, hospitals devoted entirely to cardiac care, so-called specialty cardiac hospitals, have appeared, adding to the number of new cardiac surgery programs.2 The rapid proliferation of independent specialty hospitals precipitated a congressional moratorium on new facilities, to allow an assessment of their impact.3 Although the growth in the number of hospitals performing CABG has likely improved the availability of surgical revascularization, new programs might have the unintended consequence of causing a decline in the volume of CABG operations performed at other hospitals within that region. Previous studies have shown that hospitals with lower volume have higher operative mortality.4 Others have shown that states with no regulations governing the opening of cardiac surgery programs have lower hospital volumes, and consequently higher operative mortality.5 In this paper we look at trends in CABG surgery since 1992, using Medicare data. We found a substantial increase in the number of hospitals performing CABG and a decline in the average volume of procedures performed—especially in those regions that added facilities. We discuss the implications of these findings for both clinical care and public policy.
Study population. We used Medicare claims for 1992–2003 to identify all fee-for-service (FFS) beneficiaries older than age sixty-five who underwent CABG as determined using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure codes.6 The research project was approved by the Committee for the Protection of Human Subjects, Dartmouth Medical School (Approval no. 15475). Measures. CABG hospitals. CABG hospitals were defined as hospitals submitting at least five claims for CABG during the analysis period. Hospitals that submitted sporadic claims or had low volume (but reported five or more CABG procedures) were checked via the Internet to confirm the presence of a cardiac surgery program, and those that merged during a given year were treated as a single hospital to avoid double counting. Volume categories. Each CABG hospital was assigned to one of five volume categories previously shown to be associated with differences in operative mortality among Medicare patients.7 Hospitals were categorized using the absolute number of Medicare CABG procedures performed during the year (very low volume: fewer than 130; low volume: 130–199; medium volume: 200–314; high volume: 315–484; very high volume: more than 484). Volume was recalculated annually, allowing hospitals to change volume categories from year to year. If the hospital was open for only part of the year, its volume was annualized (for example, a hospital that opened in July and did 100 CABGs in six months would be annualized to 200 CABGs for that year). Hospital Referral Regions (HRRs). Based on their ZIP code, hospitals were assigned to one of the 306 HRRs defined in the Dartmouth Atlas of Health Care.8 Analysis. National trends. For each year in the analysis period (1992–2003), we calculated the total number of Medicare CABG procedures, the total number of CABG hospitals, and the number and proportion of CABGs performed in each of the five volume categories. Regional effects. We grouped HRRs into three categories based on the net change in the number of CABG hospitals that had occurred within the region between 1992 and 2003: (1) those that had added CABG hospitals, (2) those that had lost CABG hospitals, and (3) those that had no net change. For each of these net-change categories, we calculated the proportion of CABGs performed in each volume category.
National trends. The number of CABG procedures increased from approximately 158,000 in 1992 to a peak of 190,000 in 1996 and then fell to 152,000 in 2003 (Exhibit 1  ). The number of CABG hospitals increased steadily, from 891 in 1992 to 1,069 in 2003.
The proportion of CABGs being performed in very-low-volume hospitals increased from 19 percent to 29 percent during this period, while the proportion being performed in very-high-volume hospitals declined from 20 percent to 14 percent (Exhibit 2 ). The proportion of CABGs done in the middle three volume categories also changed over time to a lesser degree (data not shown): There was a modest increase in low-volume hospitals (17 percent to 20 percent) and a decrease in middle- and high-volume hospitals (25 percent to 22 percent and 20 percent to 15 percent, respectively).
Regional effects. From 1992 to 2003, 123 regions added CABG hospitals, 18 regions lost CABG hospitals, and 165 regions had the same number of CABG hospitals (Exhibit 3 ). Regions that added CABG hospitals experienced a larger increase in the proportion of CABGs done in very-low-volume hospitals (from 13 percent to 29 percent) than regions that were stable or lost CABG hospitals (Exhibit 4). Furthermore, regions that added CABG hospitals were the only ones to experience a decrease in the proportion of CABGs done in very-high-volume hospitals (from 28 percent to 15 percent). On the other hand, regions that lost CABG hospitals experienced a substantial increase in the proportion of CABGs done in very-high-volume hospitals (from 0 percent to 17 percent).
Our analysis shows that the distribution of Medicare CABG procedures has been shifting to lower-volume hospitals, a trend that is most pronounced in regions that have added CABG hospitals. Because of this shift, we estimate that more than 15,000 CABGs were performed in very-low-volume hospitals in 2003 that would have been performed in higher-volume hospitals had the distribution observed in 1992 persisted. Assuming that the reported volume/mortality relationship for Medicare CABG persists, 190 deaths occurred in 2003 that were attributable to the shift toward lower-volume hospitals. This shift is essentially the result of two phenomena: (1) the declining amount of CABG surgery, and (2) the increasing number of CABG hospitals. Improved preventive care (such as smoking cessation) and medical care (such as beta-blockers and lipid-lowering agents) has reduced the incidence of coronary artery disease and the need for revascularization.9 And for those who need revascularization, the introduction and growth of angioplasty and stenting has further reduced the demand for CABG.10 Because cardiac surgery both is highly profitable under current reimbursement formulas and can serve as a high-profile program for marketing more generally, hospitals have been motivated to offer cardiac surgical services.11 The drive to establish interventional cardiology (which is also highly profitable) in community hospitals has also led to an increasing number of cardiac surgery programs, as hospitals with interventional cardiology programs often recruit cardiac surgeons to "back up" the cardiologists.12 These community hospitals are then able to market themselves as having comprehensive cardiac care. Role of specialty cardiac hospitals. One portion of the growth in new cardiac surgery programs is due to a relatively new phenomenon: the specialty cardiac hospital. These hospitals provide cardiac services almost exclusively and are usually located in urban areas near major general hospitals. Although specialty hospitals often become high-volume hospitals quickly, they adversely affect the volume and case-mix of nearby general hospitals.13 The proliferation of specialty cardiac hospitals has been most intense in recent years: Fourteen of the seventeen U.S. specialty cardiac hospitals opened between 1998 and 2003, prompting a congressional moratorium on specialty hospitals.14 Study limitations. Our analysis has several limitations. First, we only examined CABGs performed on the Medicare population. However, our analysis is almost certainly relevant to the non-Medicare population as well. Medicare pays for more than half of all CABGs performed. This proportion has been stable throughout the analysis period, and population-based analyses of geographic variations in utilization—the focus of the current study—reveal a strong correlation between findings in the elderly and the general population.15 Second, our analysis did not assess the impact on mortality of the changes in operative volume. Although it would be feasible to report changes in mortality, we believe that the multiple competing factors at play would make any attempt to draw convincing causal inferences virtually impossible. Because of the rapid growth of angioplasty and stents, the case-mix of patients undergoing bypass surgery has changed dramatically (toward more severely ill patients and the use of CABG as an attempt to "rescue" those with failed angioplasty). Evolving surgical techniques have improved operative outcomes at many centers. Also, a recent study of trends in mortality following myocardial infarction (MI) documents strong regional variation in these trends related to the quality of local delivery systems.16 The proper analysis would therefore need to assess the outcomes for all patients with coronary artery disease. In contrast, the well-established association between volume and outcome, recently confirmed by Peter Cram and colleagues in the era of specialty hospitals allows us to be confident that operative mortality rates are worse than they might have been if adequate operative volumes had been maintained during this period.17 Fur ther considerations. The difficulty of precisely quantifying the magnitude of the effect of declining operative volumes underscores the need for ongoing monitoring and highlights the issues that must be considered, not only for CABG but for other clinical services where substantial technical expertise is required for their safe and effective use. For many services (such as emergency treatment of heart attacks or neonatal intensive care), the challenge is to find a proper balance between at least the following three competing interests: the need for timely access to locally available care, the need to maintain clinical expertise through adequate volumes, and the high costs of such programs. The optimal approach will depend upon whether the proposed programs are located in areas where access is currently limited, or, as in many of the programs recently introduced, in suburban settings where the new program is competing directly with existing programs and the impact is likely to be a decline in procedure volumes with no improvement in access. Regionalization (relying upon effective stabilization and transfer) or less-tested approaches such as having experienced, high-volume surgeons rotate through facilities could theoretically lead to better outcomes, but a second question must still be addressed: Can the optimal approach be put in place in the current political and policy environment? Strengthen regulation. One approach would be to expand and strengthen regulatory approaches such as certificate-of-need (CON) programs, as many have now recommended.18 Recent data indicate that those states that maintained CON had both higher average operative volumes and lower operative mortality.19 Also, CON offers at least the theoretical possibility that the multiple competing interests could be effectively balanced to achieve maximal benefits. Weaknesses of these approaches, however, include not only the possibility that they will be dominated by provider interests (and thus ineffective at controlling new growth), but also their inability—in their current forms—to deal with existing excess capacity (and resultant low volumes) in specific markets. Private-sector intervention. Could the private sector intervene more effectively to constrain the growth of low-volume centers? Although some initiatives are under way to steer patients toward high-volume centers, it is not clear that these efforts have been effective or whether they would have any impact on low-volume centers. The risk is that while some patients might be diverted to high-volume centers, hospitals below the thresholds would face even stronger incentives to perform more procedures in settings where patient benefits are marginal or even negative.20 Profitability. A third approach is to address more directly one of the major underlying causes for the proliferation of these services: the relative profitability of interventional cardiology and surgery compared with medical therapy. Much policy interest has focused on the need to align payment rates more closely with marginal costs, although the technical difficulties can be great and require much more frequent updates than is now the rule. More-comprehensive payment mechanisms would also mitigate the incentives for overuse. Although a return to capitation is not high on the current political agenda, pay-for-performance is receiving wide attention, and initiatives that focus on both quality and costs for defined populations could, if widely adopted, create powerful incentives to address all three issues—access, quality, and costs—simultaneously. Lack of information. The major barrier confronting each of these three approaches is the profound lack of timely, comprehensive, and reliable measures of the quality, outcomes, and costs of care across providers. Volume is an imprecise surrogate for risk-adjusted outcomes, especially in settings—such as cardiovascular disease—where sample sizes are adequate to support provider-specific comparisons. The Institute of Medicine recently called for the establishment of an effective national system of performance measurement, to accelerate improvement in health care.21 Such a system would provide the empirical basis for more effective regulation of the growth of new programs, would strengthen payers’ efforts to help their patients choose wisely, and, if properly designed and implemented, could help providers work to improve the quality of their own care. OUR STUDY PROVIDES additional evidence that further unregulated growth in the number of cardiac facilities could have unintended consequences on hospital volume. The potential impact of these changes on the accessibility, quality, and outcomes of care deserves careful scrutiny and continued policy consideration.
Chad Wilson (ctwilson{at}partners.org) is a surgical resident at Massachusetts General Hospital in Boston. Elliott Fisher is a professor of medicine and of community and family medicine at Dartmouth Medical School, Hanover, New Hampshire. Gilbert Welch is a professor of medicine and of community and family medicine, Dartmouth Medical School, and codirector of the VA Outcomes Group. Andrea Siewers is a research analyst, Center for Outcomes Research and Evaluation, Maine Medical Center, in Portland. Lee Lucas is the center’s associate director. This work was supported in part by grants from the National Institute on Aging (PO1-AG19783) and the Robert Wood Johnson Foundation.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Study Data And Methods
Study Results
