Author + information
- Received May 30, 2006
- Revision received February 2, 2007
- Accepted February 5, 2007
- Published online June 12, 2007.
- Alan S. Pearlman, MD, FACC⁎,⁎ (, )
- Thomas Ryan, MD, FACC†,
- Michael H. Picard, MD, FACC‡ and
- Pamela S. Douglas, MD, MACC†,a
- ↵⁎Reprint requests and correspondence:
Dr. Alan S. Pearlman, Division of Cardiology, Box 356422, University of Washington School of Medicine, 1959 NE Pacific Street, Seattle, Washington 98195.
Objectives We studied the use of echocardiography among Medicare beneficiaries between 1999 and 2004 to determine the rate of growth in these services and evaluate the drivers of growth.
Background Concerned about increasing health care costs, federal and private payers have highlighted growth in diagnostic imaging studies and begun to develop approaches to curb this growth.
Methods Medicare Part B Physician/Supplier Procedure Summary Master File and enrollment data from 1999 to 2004 were reviewed. Total Medicare carrier-paid spending for echocardiography services was determined from procedure volumes and allowed charges. The 5% standard analytic file of physician claims was used to evaluate geographic variations in use and to document the specialties of physicians who request and those who interpret echocardiograms.
Results Between 1999 and 2004, echocardiography services grew at a rate similar to that for all medical services subject to Medicare’s sustainable growth rate (SGR) calculation. Increasing provision of echocardiograms in physicians’ offices contributed to increased spending under the SGR. Accounting for this shift, actual annualized per capita growth in echocardiography was 7.7%. Variations in the prevalence of heart disease contributed to geographic variations in use. Although cardiologists were the most common providers of echocardiographic services, primary care physicians ordered the majority of these diagnostic procedures.
Conclusions Growth in the use of echocardiography is in keeping with the general growth in medical services. Nonetheless, physicians who order echocardiograms and those who provide them must work together to ensure that in the future these diagnostic services are used appropriately and not excessively.
The growing use of diagnostic medical imaging procedures has attracted a good deal of attention. In its March 2005 report (1), the Medicare Payment Advisory Commission (MedPAC) noted that during the period 1999 to 2002 the average annual rate of growth of all services furnished to Medicare beneficiaries was 5.2% while imaging services grew by 10.1%/year. According to the MedPAC report, the average annual rate of growth for echocardiography was 11.8%. For certain techniques such as advanced magnetic resonance imaging, average annual growth rate was as high as 19.5%. Some experts have suggested that the bulk of growth in health spending is caused not by increasing numbers of Medicare beneficiaries, but rather by “medical innovation” (1) and “technological change” (2). Concerned about the growth in Medicare spending in general and the rapid increase in use of diagnostic imaging services in particular, MedPAC has proposed that strategies to manage the use of imaging procedures be considered.
In the case of echocardiography (cardiovascular ultrasound), a number of factors could contribute to an increase in use. For example, growth in the Medicare population and a general aging of Medicare beneficiaries might lead to an increase in total use even if practice patterns remained unchanged. Alternatively, advances in technology or the replacement over time of older diagnostic approaches by more powerful noninvasive modalities, such as echocardiography, might result in an increased use. In addition, some new treatments (such as cardiac resynchronization and other therapies for advanced heart failure [HF]) require more precise quantitation of cardiac function and follow-up evaluation of treatment efficacy. A shift in Medicare fee for service care from the hospital setting to physician’s offices would also tend to inflate the apparent numbers of services performed annually, because estimates of growth are based on Medicare’s Sustainable Growth Rate (SGR) calculation, which includes all physician services paid by Part B carriers but does not include the hospital outpatient costs that are paid by Medicare “fiscal intermediaries.”
Echocardiography is a robust family of non-invasive diagnostic imaging tests for which published evidence documents utility in virtually every type of acquired and congenital heart disease and in the assessment of patients with a range of cardiovascular symptoms or signs (3). To understand the rate of growth in echocardiography procedures, to compare this with other types of heart imaging services, and to investigate some of the potential drivers of increasing use, we studied trends in the use of echocardiography services by Medicare beneficiaries between 1999 and 2004. We hypothesized that: 1) the actual rate of growth in use and Medicare spending for echocardiography services has not exceeded growth of other diagnostic services but rather is consistent with growth in health services in general; 2) neither the number of echocardiography services/beneficiary nor the proportion of patients undergoing repeat studies has changed appreciably; 3) regional variations in use are not random and might reflect the prevalence of underlying cardiovascular disease; and 4) “self-referral” (4) is not the primary reason for growth in echocardiography services.
Medicare carrier-paid echocardiography services from 1999 to 2004 were reviewed using the national Centers for Medicare and Medicaid Services (CMS) Part B Physician/Supplier Procedure Summary Master File and enrollment data for each of those years. These files summarize all Medicare fee-for-service carrier-paid claims for each calendar year, by Healthcare Common Procedure Coding System (HCPCS) code, modifiers, carrier and locality, provider specialty, and type and place of service. The Part B Physician/Supplier Procedure Summary Master File does not include payments (made through Medicare “fiscal intermediaries”) to hospital outpatient facilities or services provided in Medicare Advantage (managed care) plans. With the CMS Berenson-Eggers Type Of Service (BETOS) category for echocardiography and other heart imaging procedures, the volume of services provided to beneficiaries in Medicare fee-for-service plans was evaluated. Notably, stress echocardiography services were not analyzed in the present study, because stress echocardiography is considered (by CMS) as a “test” rather than an “imaging” procedure and therefore not included in the BETOS category for heart imaging procedures. On the basis of HCPCS modifiers, claims were assigned to “professional component,” “technical component,” or “global” (professional component + technical component) categories. The quantity of services was determined in 2 ways. First, the number of echocardiography procedures was counted to determine how often Medicare was billed for these services. Second, the number of 2005 relative value units (RVUs) for each procedure was tallied, to determine the quantity of services provided on the basis of a constant RVU scale. Because RVU values can change from year to year, the change in spending was decomposed into parts by calculating the impact of changes in RVUs from 1999 values to 2005 values.
In addition, actual Medicare spending under this program was calculated from Medicare allowed charges, which include program payments made by Medicare itself and also beneficiary coinsurance/deductible amounts. The Medicare allowed charge represents the dollar figure that enters into the SGR calculation. Medicare carrier-paid claims for echocardiography and other diagnostic heart imaging procedures were tabulated both by numbers of services provided and by total Medicare allowed charges for those services. Cross-sectional differences in use were evaluated as a function of geography and patient demographics. To assess possible effects of disease prevalence on the use of echocardiography services, the prevalence of congestive HF (obtained from 2004 claims in the 5% sample standard analytic file) was matched with the rates of echocardiography use on a state-by-state basis. Because geographic variations in volumes of services provided to Medicare beneficiaries have been ascribed to variations in physician supply (5,6), the number of cardiologists obtained from the 2003 Area Resource File also was included in this analysis.
To determine whether increased use of echocardiography reflects provision of more services/person or is due to more persons being evaluated, the 5% sample standard analytic file of physician claims from 2001 to 2004 was used to evaluate the location, diagnoses, and demographics of echocardiography use. These files reflect all claims and enrollment data for roughly 1.8 million Medicare beneficiaries. To avoid “double counting,” only professional component and global claims for echocardiographic imaging services were considered; “add on” services and technical-component-only claims were not included.
To evaluate trends in the ordering and provision of echocardiograms, uniform provider identification numbers were used to assess the self-reported specialties of both physicians requesting echocardiographic studies and physicians providing the diagnostic services.
Growth in use of echocardiography
Figure 1summarizes changes in numbers of services and Medicare enrollees, professional service RVUs, and Medicare allowed charges for all services subject to the SGR system for the years 1999 to 2004. Over this time period, the number of echocardiographic services rose by an average of 10.6%/year, a rate lower than that of other heart imaging services (14.2%/year) but higher than the rate of growth of non-heart imaging services (5.9%/year). The number of Medicare beneficiaries grew by an average of 2.3%/year between 1999 and 2004, contributing to the growth in echocardiography (and other medical services). Expressed in terms of professional service RVUs (which more closely reflect numbers of services than do total RVUs, because the latter are affected by shifts in site of service), the average annual rate of growth of echocardiography services was 9.9%, similar to the rate of growth of imaging services in general (9.5%). Total Medicare allowed charges for echocardiography grew by 10.9%/year, a rate similar to the average annualized rate of growth (11.2%) for all services covered under the SGR but not as high as the rates of growth of all heart imaging (14.9%) and non-heart imaging (14.5%) services.
Growth in Medicare spending for echocardiography
Changes in payment and RVUs for the 4 most commonly performed procedures (which together account for 99% of all echocardiographic services provided) are summarized in Table 1.In 1999, echocardiography accounted for 57% of the Medicare dollars spent for cardiac imaging services; in 2004, this percentage fell to 48%. Medicare carrier-paid spending for echocardiography increased by 68% between 1999 and 2004; this is the net result of several factors. Increases in the conversion factor accounted for approximately 8% of the increase in Medicare payments for echocardiography services, whereas decreases in RVUs for the various echocardiography codes tended to reduce Medicare expenditures for all echocardiographic services by approximately 16%. If the decline in RVUs/echocardiographic service is accounted for by holding RVUs constant at 2005 levels, then total RVUs for echocardiography services grew by 84% between 1999 and 2004.
Because technical component costs for echocardiography services performed in the hospital outpatient department are paid through Medicare “fiscal intermediaries,” these costs are not counted in the SGR. If the same services were provided in a physician’s office, however, then the technical component costs would be paid by the Medicare carrier and counted in the SGR. Technical component costs are larger than physician work components for all echocardiography services, so that site of service shifts from hospital facilities to physician offices would tend to inflate Medicare carrier-paid spending for echocardiography out of proportion to the numbers of services provided. Table 2demonstrates that between 1999 and 2004, base imaging services (primarily transthoracic and transesophageal echocardiography) grew by an average of 9%/year. “Add on” Doppler services grew at a slightly higher rate of 11%/year. Rates of growth were higher in the “non-facility” (physician office or independent diagnostic testing facility) setting than in either hospital inpatient or outpatient settings. Figure 2demonstrates that between 1999 and 2004, the percentage of transthoracic echocardiography imaging and Doppler procedures done in the office setting increased, whereas the percentage done in hospital inpatient and outpatient settings declined. Because Medicare carriers pay for the technical component of office-based echocardiography services but not for hospital-based services, this shift in reimbursement from professional component bills (for services performed in hospital settings) to total component bills (for services provided in physician’s offices) resulted in a net 12.5% increase in carrier-paid RVUs (Table 3).
Roughly one-half of the growth in the volume of echocardiography service RVUs can be attributed to the combined effects of shifting site of service and increasing numbers of Medicare fee-for-service beneficiaries. If site of service is held constant and growth in the Medicare population is considered, growth in echocardiography services between 1999 and 2004 increased by 45%, as demonstrated in Table 4.This indicates an average annual growth rate of 7.7%.
Growth in repeat studies
Using the 5% standard analytic file of physician claims for 2001 to 2004 (the only data available to us for this beneficiary-level analysis), growth in echocardiography services was due almost entirely to increases in the numbers of persons served, with a small contribution from the higher use of spectral and color Doppler “add on” codes. Table 5demonstrates that the mean number of echocardiograms/individual having an echocardiogram was relatively constant at 1.3/year. Of patients having at least 1 echocardiogram in a given year, 80% of them had only 1 scan/year, and this figure remained constant from 2001 to 2004. If analysis is restricted to patients undergoing evaluation in the office setting only, only 10% of subjects having an echocardiogram in a given year underwent a second echocardiography evaluation during that same calendar year. The percentages of patients having multiple scans/year were also constant over the time period investigated. Thus, the growth in total echocardiography volume cannot be explained by more frequent repetition of echocardiographic services in individual patients.
Rate of echocardiography use versus prevalence of heart disease
When analyzed on a state-by-state basis, rates of use of echocardiography services grew in every state but at different rates. Use rates varied widely, with more than a 3-fold difference from lowest (0.23 per capita in Wyoming, in 2004) to highest (0.75 per capita in Florida and New Jersey). The geographic variation in rates of echocardiography services seemed to be related to differences in the prevalence of cardiovascular disease. Figure 3plots by state the number of echocardiographic services per capita against the prevalence of congestive HF. By least squares regression, more than one-half of the state-to-state variation in use is explained by differences in the prevalence of HF (r2= 0.59). When the number of cardiologists/1,000 Medicare beneficiaries was also considered and the rate of echocardiography use compared with both the supply of cardiologists and the prevalence of HF, the correlation with echocardiography use rates did not change appreciably (r2= 0.60). With a 2-step regression analysis, both the supply of cardiologists and the prevalence of HF were statistically significant predictors of echocardiography use (p < 0.001). Although preliminary, this observation suggests that in this population multiple factors might influence echocardiography use.
Trends in the provision of echocardiography
The specialty distribution of physicians providing echocardiography services is summarized in Table 6.In 2004, cardiologists submitted claims for 75% of the Medicare allowed charges for echocardiography services, with internists making up the next most common specialty providing these services (15%). The specialty distribution was similar in 1999.
Table 7summarizes the self-reported specialty of physicians requesting echocardiograms in Medicare beneficiaries in 2004. Notably, cardiologists ordered only 29% of the studies. Generalists and non-cardiologist specialists requested 71% of all echocardiograms, with primary care physicians (internists and general practitioners) requesting the majority of these studies (56%).
This study of echocardiography use in Medicare beneficiaries demonstrates that the rate of growth in echocardiography services during the period 1999 to 2004 was comparable to the rate of growth of medical services in general and not disproportionately higher. After accounting for increasing numbers of Medicare beneficiaries, changes in the conversion factor, adjustments in Medicare RVUs, and shifts in service from hospitals to physician’s offices, the volume of echocardiography services rose by 45% over the study period, representing an annualized growth rate of 7.7%. In addition, neither the percentage of patients having repeat studies in the same year nor the number of repeated studies/person having an echocardiogram has increased. Large state-by-state variations in use rates are evident, but these seem related at least in part to variations in the prevalence of cardiovascular disease. Although cardiologists provide most echocardiography services, primary care physicians and non-cardiology specialists order the large majority of these services.
Growth in services and allowed charges
The methodology used in this study differs from some prior analyses. Other investigators (7–9) have examined use of diagnostic imaging from the standpoint of total numbers of tests. However, different tests have different costs. Thus, if a more expensive test were substituted for a less costly alternative, the true impact on Medicare spending would not be appreciated if only numbers of tests were considered. In addition, growth in the numbers of diagnostic services raises several important questions: “Are these services appropriate?”, “What number of services would be correct?”, and “How did these services affect clinical outcomes?” Answering these questions would require consideration of the clinical circumstances that prompted ordering of the tests, the other “downstream” services that might have been used or avoided appropriately, and how clinical outcomes might have been influenced by the diagnostic services. These questions cannot be answered from a study of an administrative database.
Because health care spending (rather than just numbers of services) has been the focus of considerable recent attention (1), we not only analyzed growth in numbers of echocardiographic services, but also examined total expenditure of Medicare dollars. To accomplish this, allowed charges were calculated and corrected for changes in RVUs over time by using the 2005 RVUs as a constant factor. This approach allowed us to determine the magnitude of changes in Medicare spending for echocardiography services and to investigate the reasons for these changes. Although the results of this study do not tell whether the rate of growth in echocardiography services is too high, too low, or just right, they do indicate that this growth rate is similar to that of other services under the Medicare program and not disproportionately higher.
The growth in medical imaging procedures has received recent attention from regulators and lawmakers, but this issue is not new. Our data confirm that echocardiography use is growing but at a rate similar to other components of medical care. Noninvasive imaging, particularly echocardiography, plays a prominent role in the evidence-based evaluation and management of patients with a wide range of cardiovascular symptoms and diagnoses (3). Because an increasing proportion of beneficiaries carry such diagnoses as the population ages, it is not surprising that use of echocardiography, an important tool for proper diagnosis and management, is increasing. Figure 3, which demonstrates increasing echocardiography use as the prevalence of HF increases, illustrates this point.
Geographic variations in echocardiography services
Geographic variations in the use of medical services have been recognized for many years (10,11). Studies of Medicare beneficiaries have suggested that these variations are largely due to variations in the supply of physicians (particularly specialists) and availability of services and that higher use rates do not correlate with improved quality of care or better access to care (5,6). Although a number of studies have evaluated total spending in patients with a variety of medical disorders, rates of use of cardiovascular diagnostic and therapeutic procedures have also been examined. Wennberg et al. (12) reported that variations in the rates of coronary angiography in northern New England were highly associated with the prevalence of cardiac catheterization laboratories and that revascularization rates correlated closely with rates of diagnostic angiography. In fact, such “supply sensitive care” was mentioned by MedPAC as one important source of unwarranted variation in the use of health care services (1).
However, few studies have focused specifically on echocardiography. In a study of Medicare beneficiaries, Lucas et al. (13) did observe substantial variation in the rates of use of echocardiography during the year 1995. These investigators concluded that “the likelihood of Medicare beneficiaries having an echocardiogram is influenced by where they live” (13) but did not provide insights into the cause(s) of the variations they observed. Our study confirms, in a more contemporary Medicare population, wide regional variations in the rates of echocardiography use. Our preliminary results do not provide definitive proof as to the reason(s) for these variations. However, they do suggest the hypothesis that rates of use of diagnostic echocardiography might be related to the prevalence of heart disease. In our study, when the supply of cardiologists (who provide three-quarters of echocardiography services nationally) was considered in addition to disease prevalence, we found no improvement in the correlation with use rates. This observation suggests the possibility that higher rates of echocardiography use might not be due just to availability of services but also might be related to clinical factors such as the prevalence of documented or suspected heart disease. Further studies of regional variations in echocardiography use are clearly warranted, and we believe that these studies ought to consider clinical factors in addition to the supply of specialists and the availability of services.
Provision of echocardiography services
Training in echocardiography is required in all cardiovascular disease fellowships, and our study confirms that specialists in cardiovascular medicine provide about 75% of Medicare allowed echocardiography services. Multidisciplinary, peer-reviewed guidelines have been developed describing the training and experience needed for competence in echocardiography (14) and, together with board examinations and physician certification for echocardiography, represent efforts by the cardiology community to assure the quality of echocardiography. These guidelines are not specialty-specific. Widely used, evidence-based guidelines for the clinical application of echocardiography procedures also have been published (3). Although some have suggested that diagnostic imaging services provided by non-radiologists are less often of high quality than imaging services provided by radiologists (15), no credible data support that hypothesis with regard to echocardiography.
Our study also demonstrates that although cardiologists provide the large majority of echocardiography services, primary care providers or non-cardiology specialists typically request these diagnostic procedures. Because echocardiography is useful in evaluating virtually every form of heart disease and in patients in all age groups with common symptoms and signs (including but not limited to breathlessness, abnormal heart beats, and heart murmurs) that suggest cardiovascular disease (3), it is not surprising that echocardiographic studies frequently are ordered by many different kinds of health care providers. Seventy-one percent of the echocardiography procedures provided to Medicare beneficiaries in our study sample were requested by non-cardiologists (over one-half of all studies were ordered by primary care physicians), whereas only 29% were ordered by cardiologists. These data emphasize that if appropriate use of echocardiography is to be achieved, one critical component of such an effort will be education not only of cardiologists but also of a wide array of physicians and mid-level providers as to the clinical situations in which it is and is not appropriate to order echocardiographic studies.
Some gaps in the data source for this study might have caused overestimation or underestimation of relative use and costs of noninvasive cardiac imaging. For example, because it was not possible to allocate Medicare expenditures for radioisotopes to specific nuclear medicine procedures, in our dataset the total dollars spent on nuclear cardiology procedures did not include the radioisotope costs. In addition, billing codes specific to cardiac computed tomography services did not exist during the period of our study. Thus, the data summarized graphically in Figure 1might tend to overestimate echocardiography’s share of allowed charges for heart imaging services. Our state-by-state analysis of echocardiography use tabulated services by the state where the service was performed, but the site of primary residence defined the state population. Although probably not a major source of error, the lower use in cold weather states and higher use in southern states might in part reflect seasonal movement of retirees from colder to warmer locations. In addition, in our preliminary analysis, only the prevalence of HF was considered as a surrogate for the prevalence of heart disease. Although echocardiography is of well-documented value in the assessment of HF, it is also valuable in a wide range of other cardiac disorders, which we were not able to include in this initial evaluation. Despite these limitations, our data establish reasonable estimates of the relative use and cost of echocardiography services in the Medicare population.
We share the concern of policy-makers that continuing growth in Medicare expenditures at rates exceeding that of inflation is not sustainable in the long term. However, because the actual 7.7% average annual rate of growth of echocardiography services is not out of line with the general rate of growth in health services, efforts to regulate use of echocardiography will not solve the challenge of continuing growth of Medicare services at a rate that outstrips inflation. We believe that diagnostic studies must be performed properly and for accepted indications to provide information that is most useful for patient care. We agree that efforts to develop appropriateness criteria (16–18) ought to assist requesting physicians—particularly those who do not have special training in evaluating and managing patients with complex heart disease—in using diagnostic imaging, including echocardiography, wisely and only when medically appropriate. We encourage professional societies that are actively involved in echocardiography teaching, research, and advancement of patient care to take an active role in assuring that echocardiography services are provided by well trained, experienced cardiac sonographers and physicians and that sites of service comply with accreditation programs including continuing education and quality improvement processes. In addition, professionals who provide high-quality echocardiography services must take an active role in working with health care providers who order echocardiograms to be sure that the right studies are done at the right time for the right patients and that studies are performed only if the results will impact the patient’s treatment.
The authors appreciate the assistance of Christopher Hogan, PhD (Direct Research, LLC, Vienna, Virginia), who helped obtain and analyze the Medicare data upon which this investigation is based.
↵a Dr. Douglas serves as a consultant to GE Healthcare.
A grant from the American Society of Echocardiography supported some of the costs of data collection and analysis.
- Abbreviations and Acronyms
- Berenson-Eggers Type Of Service
- Centers for Medicare and Medicaid Services
- Healthcare Common Procedure Coding System
- heart failure
- relative value unit
- sustainable growth rate
- Received May 30, 2006.
- Revision received February 2, 2007.
- Accepted February 5, 2007.
- American College of Cardiology Foundation
- MedPAC. Medicare Payment Advisory Commission
- Cheitlin M.D.,
- Armstrong W.F.,
- Aurigemma G.P.,
- et al.
- Wennberg D.,
- Dickens J. Jr..,
- Soule D.,
- et al.
- Quinones M.A.,
- Douglas P.S.,
- Foster E.,
- et al.
- Brindis R.G.,
- Douglas P.S.,
- Hendel R.C.,
- et al.
- Patel M.R.,
- Spertus J.A.,
- Brindis R.G.,
- et al.
- Hendel R.C.,
- Patel M.R.,
- Kramer C.M.,
- et al.