This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

H E A L T H  T R A C K I N G :
M A R K E T W A T C H
 
9 August 2005  
  

When The Price Isn’t Right: 
How Inadvertent Payment Incentives 
Drive Medical Care

If payment rates are not made more accurate,
another powerful driver of health cost trends could be created.


by Paul B. Ginsburg and Joy M. Grossman


ABSTRACT:

Unintended overpayment of some services, in combination with other market factors, is driving increased use of expensive care, which in turn could be an important driver of health care cost trends. Reimbursement systems are highly dependent on provider charge data that rarely provide accurate and up-to-date indicators of relative costs. As a result, newer services, in which productivity is increasing over time, tend to be more lucrative. As the largest payer, and one whose reimbursement policies are followed by private insurers and Medicaid programs, Medicare can address this issue by taking steps to make its prospective payment rates reflect relative costs more accurately.

Public and private decisionmakers continue to search for effective ways to control health care costs. The spotlight has focused most recently on efforts to increase consumers’ sensitivity to the full costs of care—for example, through consumer-directed health plans—and on providers’ financial incentives to improve the quality and cost-effectiveness of care—for example, through insurer- and purchaser-sponsored pay-for-performance initiatives. In this paper we address an avenue that has received less attention: the potential for cost savings from refining the systems used by public and private payers to reimburse providers. Unintended relative overpayment of some services and the relative underpayment of other services, in combination with other market factors, is driving increased use of expensive care, which in turn could become an important driver of health care cost trends.

Under current payment mechanisms, the provision of some medical services is much more profitable than others for physicians, hospitals, and other health care providers. For example, recent analysis of inpatient costs by the Medicare Payment Advisory Commission (MedPAC) shows that patients admitted for a surgical procedure, such as a heart bypass, generally are much more profitable than patients admitted for a medical condition, such as pneumonia.¹ Differences in the profitability of individual services appear to be contributing to important changes in care delivery, as both physicians and hospitals expand their capacity to compete to deliver the most profitable services.

The differences in profitability that are driving increased use of some services appear to be inadvertent but systematic—a shortcoming of third-party systems to pay providers that do not accurately reflect the relative costs of providing different types of medical services. Although payment systems vary by payer and by care setting, one common characteristic is that billed charges play a key role, even when payment methods attempt to reflect relative costs. And relative charges systematically differ from relative unit costs.

This paper was inspired by market developments that we have observed since the 2000–2001 round of site visits under the Community Tracking Study (CTS). We have supplemented information from the site visits with interviews of Medicare officials and outside advisers on reimbursement, MedPAC staff, executives at private insurance companies who have responsibility for provider reimbursement, and industry consultants. Most of our assertions about the profitability of certain procedures are based not on analysis of financial data, since these generally are not publicly available, but on consensus among our sources. There is notable consistency in the responses of representatives of numerous hospital and physician organizations and health plans interviewed in twelve markets over three rounds of site visits.

Financial Pressures Lead To Emphasis On Profitable Services

Although variation in profitability across medical services is unlikely to be a new phenomenon, market pressures appear to be driving providers to respond more directly to incentives in the payment system. In most CTS communities, site-visit respondents have reported that reduced margins for physicians’ professional services have been a stimulus to entrepreneurship by physicians.² Hospitals also appear to have been changed by their experiences of being squeezed by both managed care plans and Medicare during the 1990s and are focused on expanding those services that are most profitable.³ Although some services may be highly profitable because of shortcomings in provider payment systems, these shortcomings also may drive very low, or even negative, margins for other core services. Hospitals report that a strong motivation for expanding more profitable services is to cross-subsidize these less-well-paid services as well as to provide sufficient margins to fund uncompensated care. In some cases, the strategies are designed to preempt or respond to competition from physician-owned facilities. Advances in technology have facilitated this increased competition by making it feasible to offer more services in settings outside of hospitals, such as in freestanding centers and in physicians’ practices.

The most frequently mentioned services that hospitals and physicians are competing over include cardiovascular procedures, orthopedics, selected neurosurgical procedures, and high-end imaging. Hospitals are using multiple strategies to increase volume in one or more of these areas from hiring service-line managers and attracting well-known specialists to focusing capital expenditures on new facilities and equipment to draw physicians and patients.

Although Congress has focused its attention on threats to general hospitals from the growth of physician-owned specialty hospitals, the phenomenon of competition between hospitals and physicians, leading to a proliferation of capacity to provide the more profitable services, is much broader. For example, findings from recent CTS site visits are consistent with MedPAC reports and news articles indicating that more-sophisticated services such as magnetic resonance imaging (MRI) and other high-end imaging, radiation therapy and other outpatient cancer care, and many cardiac services—which traditionally were provided in hospital outpatient departments or freestanding facilities by referral—have been brought into physician practices to an increasing degree.⁴ Physician groups report that a major motivation for the many mergers among single-specialty medical groups has been to achieve the scale needed to invest in equipment to offer more ancillary services.⁵

Ramifications For Cost, Access, And Quality

The inadvertent variation in profitability of medical services does more than create winners and losers among providers. It has important implications for the cost and quality of health care and access to that care. For example, if specialty facilities and physician practices draw the most profitable services and patients away from general hospitals, these institutions will find it more difficult to subsidize money-losing activities, such as charity care, and services with large standby costs, such as trauma and burn centers. Moreover, many believe that increased capacity leads to higher rates of service use, or “supplier-induced demand.”⁶ Service use also is likely to increase because more care is being provided in settings where physician self-referral incentives come into play. Research shows that physician referrals are much higher when physicians have an ownership interest in the facility.⁷

These trends also could lead to excess capacity—with additional implications for spending. Excess capacity could spur price reductions, as it generally does outside of health care. But with third-party payment having removed much sensitivity to price, excess capacity in health care has the potential of leading to higher prices, as providers pass on higher unit costs to payers. Excess capacity also can lead to fewer providers’ having the volume of services needed to provide high-quality care.⁸

Many of these capacity expansions have positive implications as well. Physician ownership of facilities may lead to better management by aligning the incentives of the practicing physicians and the owners. Hospitals seeking to align the financial incentives of the medical staff with those of the institution have long struggled with prohibitions against gain sharing that preclude many ways of rewarding physicians for practicing more efficiently. Facilities that are new and specialize in a narrower range of services could lead to innovation in delivery. Greater convenience for patients and higher productivity for physicians are potential benefits. Capacity expansions that are driven by these factors—or to accommodate growth in clinical needs for specialty services—have a better chance at being a positive for society than those stimulated by inadvertently high payment rates.

Why Payment Rates Do Not Reflect Provider Costs

In most of the economy, competitive markets are seen as the force that leads to a structure of prices that reflects the structure of costs. But since medical care is financed largely by third parties, this mechanism does not necessarily function. Third-party payers traditionally paid providers on the basis of their billed charges, screening them for reasonableness by comparing them with charges of others in the community, or else they reimbursed providers for their incurred costs. In the 1980s public and private payers began to set payment rates prospectively, based on estimates of average costs across a category of providers. Consumers were not exposed to these payment rates and played little role in price setting.

Medicare has led this movement by payers toward prospective payment systems (PPSs). It now has separate PPSs for inpatient hospital care, outpatient hospital care, ambulatory surgical centers (ASCs), outpatient laboratories, physician services, and other services. Medicaid programs also use prospective payment extensively, in many cases following Medicare methodologies, as do many other public programs such as some states’ worker’s compensation. Most private insurers have moved to PPSs for at least a subset of services, sometimes adopting Medicare’s approach—most notably for physician payment.⁹ When plans do base payment methods on Medicare, they benefit from both the technical work of the Centers for Medicare and Medicaid Services (CMS) as well as providers’ familiarity with and acceptance of Medicare’s methodologies.

However, Medicare and other payers have found it difficult to devise payment rates that closely follow relative costs for different services. Most PPSs, with the notable exception of payment for physician services, have relied in various ways on charge data as a proxy for relative costs. For example, Medicare prospective payment for inpatient care, which is based on diagnosis-related groups (DRGs), determines the relative payment for each DRG on the basis of average charges for patients in that DRG across all hospitals. To the degree that the pattern of charges systematically diverges from the pattern of unit costs, then relative payments for different DRGs will not reflect relative costs; as a result, patients in some DRGs are more profitable than others. For example, a recent MedPAC analysis estimated that DRG 107, coronary artery bypass graft with catheterization, is 9 percent more profitable than the average DRG, while DRG 127, heart failure and shock, is 6 percent less profitable than the average.¹⁰

Few Medicare systems are based on actual cost data, and most of the data are very old—from the 1980s or 1990s (Exhibit 1). Although the program relies on more recent data to help in updating payments, experts suggest that the lag between collecting the data and implementing the payment updates means that the information is often out of date.¹¹ An important exception to reliance on charge data is Medicare’s physician fee schedule. It is based on surveys of the relative costs of different inputs that make up physician services, including physicians’ time and effort and practice expenses. In contrast to hospital relative payments, those for physician services are based directly on estimates of relative costs. But even if the starting relative values were highly accurate, shortcomings related to updating the relative values have likely introduced important divergences from relative costs.

Private insurers’ PPSs have similar weaknesses. Many private payers pay negotiated per diem rates for a large portion of inpatient care, and these per diems are typically adjusted for high-cost or “outlier” cases based on charges for those admissions. In addition, some inpatient care and much outpatient care other than for physician services are paid for on the basis of negotiated discounts from provider charges. Plans are mostly likely to use prospective payment methods besides negotiated per diems or discounted charges for hospital-based surgical procedures, emergency department visits, and laboratory and imaging services provided outside the hospital. In developing these payment systems, private plans also typically rely on charge data to set relative values.

Like Medicare, private payers face challenges in updating payments. Respondents stated that they focus primarily on controlling upward trends in aggregate payments and rarely gather market information to identify and adjust payments for specific services that are paid too much or too little. When payers notice a rapid increase in the volume of a service, they are more likely to use tools other than payment rates to address the issue.¹² A recent example was Highmark Blue Cross Blue Shield’s limiting payment for MRIs and other advanced imaging services to facilities that provide a comprehensive range of imaging services—including basic, less profitable services—and that have a radiologist on site.¹³

In some cases, payers have adjusted relative payment rates to more accurately reflect relative costs. Medicare has occasionally reduced payment rates for certain services on the basis of evidence that relative payments were too high—most recently, payments for physician-administered drugs.¹⁴ Private plans report that in the case of some technological advances that drive up the costs of devices, such as drug-eluting stents, they may use cost data from invoices to update payment rates—often in response to providers’ concerns that they are being underpaid.

Why Charges Do Not Reflect Relative Costs

With most payment systems highly dependent on the accuracy of providers’ charges, the degree to which payers achieve the goal of all services’ being similarly profitable depends on the degree to which charges for different services reflect the relative costs of those services. Although some of the divergences in the relationship between charges and costs reflect strategic pricing behavior by providers, such as smaller markups for services with the most visible prices, many come from the cost of collecting accurate unit cost information together with the limited potential for higher revenues. Assessing the unit costs of each of a large number of services requires major effort to gather information on direct inputs, such as labor, space, supplies, and equipment, and sophisticated formulas to allocate indirect costs, such as general management. Although estimating unit costs can be relatively easy for services in which a purchased input constitutes a large share of the cost, such as drugs or devices, estimating the costs of most services is far more difficult because the most important inputs produce numerous distinct services. Few hospitals use sophisticated cost accounting systems to set charges; although advances in information technology (IT) are making it easier for hospitals to better understand what different services cost, we see little evidence of their being put to use to have charges more accurately reflect relative costs or to update those charges to account for the effects of rapid technological change.

Providers have limited incentives to devote the resources to set charges that accurately reflect relative costs because a large part of their revenue is not directly influenced by their own charge structure. Any one provider’s charge schedule has negligible impact on calculation of prospective payment rates by insurers. Even when private payers negotiate discounts from charges for each hospital, little is gained from a structure that more accurately reflects costs.

Recent research conducted for MedPAC suggests that factors other than costs play important roles in determining charges and that once charges are set for a new service, the vast majority are not revisited but are updated by uniform percentage increases.¹⁵ The exceptions among new services are those that involve expensive but easy-to-cost devices or other inputs, but few of these services get serious attention in future years, even if input prices have changed greatly. This mirrors what health plan executives told us about their approach to updating payments.

A likely result of these practices is that charges for services with the most rapidly advancing technology tend over time to have the largest markups over costs. This is probably the reason that DRG 517 (percutaneous transluminal coronary angioplasty stent without acute myocardial infarction) is 16 percent more profitable than the average Medicare case.¹⁶ In general, higher markups for services that use advancing technology result from increasing productivity that comes from experience in designing and manufacturing equipment and devices and in providing services. Newer models of equipment are often more productive or have lower costs than what they replace. For example, recent advances in CT scanning equipment have reduced the time it takes to do a scan, which means that a machine and technician staff can perform more scans per day.¹⁷ Similarly, the physician time required to perform many cardiovascular surgical procedures has fallen over time as the technology has evolved.¹⁸ In addition, as new services are provided to more patients, equipment and dedicated space and personnel are used at a higher capacity, which is another factor behind productivity gains over time. For some supplies, such as prepackaged instruments and supplies that physician practices purchase for certain procedures, entry by additional manufacturers leads to lower prices. These declines in costs often mean that payment rates that are not revised frequently soon become out of date, leading to profit margins that are higher than those for services not experiencing these trends.

The result of many of these factors is shown by Medicare’s experience with its PPS for ASCs. Current relative payment rates are based on a 1986 survey. In 1998 the CMS proposed reducing payment rates for several high-volume procedures, such as cataract-related procedures and gastrointestinal endoscopies, based on a more recent survey conducted in 1994. These reductions were not implemented, which suggests that the actual payment rates—based on the 1986 survey data—exceeded costs for those procedures in 1998.¹⁹ Now, seven years later, it is likely that the inadequacies in the payment schedule are even greater.

The phenomenon of having payments for newer services become too high in relation to those for older services is a problem even when charges are not used to determine prospective payments. Take for example, Medicare’s resource-based relative value scale (RBRVS) used to set payments that reflect the work (time and effort) component of different physician services. Since the survey data on which the RBRVS is based are from the mid-1980s, the CMS updates the relative values every five years. Instead of repeating the survey itself, in partnership with the American Medical Association and many specialty societies, CMS panels review requests for changes in relative values for specific procedures; in the second five-year review, only a small proportion of approved changes resulted in decreased relative values.²⁰ This bias could be most problematic for the facility (practice expense) component for certain services, which is more relevant to physicians’ efforts to expand services offered in their practices. The CMS has asked medical directors of Medicare carriers responsible for processing claims to take on this role, but one can question whether they have access to the data needed to do this job.

Inadvertent differences in profitability can also result from other aspects of payment design. For example, when physicians perform complex procedures that involve discrete components, the work needed to perform the entire task is often less than the work to perform each component separately. Both Medicare and private payers take this into account for surgery, in which a single payment is established for all services performed within ninety days by the primary surgeon, but this approach has not been applied to major therapeutic nonsurgical procedures, such as endoscopies and catheterizations, which could help explain the proliferation of endoscopy and catheterization centers. Separate payment for different components of spine surgery may be a factor behind the recent proliferation of spine centers.²¹

Policy Options

Inadvertent variation in profitability across services and its contribution to rising health care costs could be addressed by revising relative payment rates so that they more accurately reflect relative costs. Although there have been major efforts in Medicare to apply prospective payment methods to additional settings in recent years, policymakers have not paid close attention to the need for making prospective payments more accurate.

Historically, a political constituency to drive improvements in payment mechanisms has been lacking. However, the fact that some organized interests, such as general hospitals, perceive the responses to differences in the profitability of different services to be very harmful to them may create the constituency for change. Indeed, the debate about the growth of specialty hospitals is already beginning to focus the broader policy community on efforts to make payments more accurate. Recent testimony by the CMS administrator included a commitment to doing this.²²

The range of options to revise PPSs to better reflect the actual costs of care include the following: (1) systematic collection of provider cost data on a broad range of services; (2) incremental efforts to improve the accuracy of the data or analyses based on them, such as for new technologies; and (3) approaches to dampening providers’ responses to the remaining distortions in profitability left after improvements have been made in the underlying data and models used to set payments.

Role of Medicare. These approaches can probably be best pursued through the Medicare program. Not only are Medicare patients of great importance for most providers, but private payers, despite not placing a priority on more accurate payment methods, often follow Medicare payment policies when they believe they have the leverage with providers. Other public programs have often followed Medicare methods as well, so the financial and clinical benefits from more accurate Medicare payment rates could have sizable spillover effects in the health care system as a whole.

Hospital services. For hospital services, Medicare needs to develop an alternative to its use of hospital charges to develop its relative payment rates for inpatient and outpatient care. One approach would be to pursue a voluntary relationship with a representative sample of hospitals in which the program would provide a subsidy to defray the cost of sophisticated cost accounting systems in return for regular data on the relative costs of different services. This would provide Medicare with more accurate data on relative costs as well as trend data to use for updating schedules. Although Medicare could begin by adjusting its relative payments for selected services, ultimately this could fully replace the use of charge data.

A less ambitious option would be to focus on revising how payments are set for selected new high-cost technologies. Medicare could develop a process to identify services for which technological change is likely decreasing costs and gather data to adjust payment rates.

Physician services. For physician payment, Medicare could consider conducting a broad relative value survey periodically in conjunction with the AMA and specialty societies. Another interim approach, which would apply to the practice expense component of physician payment, would involve assuming higher-capacity use rates when setting payments for services that use expensive equipment. Medicare assumes a uniform utilization rate of 50 percent over the life span for all types of equipment used in physician practices.²³ Raising this to 75 percent would limit the number of practices that see potential in offering the services as well as reduce incentives to overprescribe.

Keeping current. Whatever success Medicare might have in gathering cost information at a moment in time, it will still have the challenge of keeping cost information current. The options described above would have to be repeated periodically, and this would add to Medicare administrative costs. Some of this burden could be reduced if results from studies of productivity trends for particular services could be extrapolated to determine annual revisions before the next assessment.

Increased funding. Many of these options would require more funding for program administration. Although the amounts are likely minuscule in relation to the potential reduction in outlays for benefits, funds for administering Medicare have always been tight. If experience prevails, resources will become available only when a major stakeholder decides that this investment is important and effectively advocates for it.

Variations. Even with major success in using more-accurate data to set payment rates, policymakers need to realize that variation in profitability across services will continue. Thus, other steps should be considered to minimize the negative fallout from inadvertent payment incentives. These can range from steps to address a specific imbalance, such as those taken by Highmark to address the use of imaging, to additional restrictions on physician self-referral. Indeed, with technology having expanded the range of opportunities for physicians to refer to their own facilities, policy discussions on this subject need to go beyond resolution of the specialty hospital issue. For example, the Final Rule published in 2001 to implement the Ethics in Patient Referral Act (Stark II) is believed to have encouraged physician expansion of ancillary services in their offices; this could be revisited, for example, to exempt from self-referral restrictions only those services in which the physician is directly involved.

The United States has probably lost a decade in addressing these payment problems because of the growth of, and subsequent backlash against, managed care. If the vision of integrated delivery systems had gained traction, and if capitation—fixed per person, per month payments for care—had become a more important payment tool, the problem of inadvertent incentives in fee-for-service payment would be much less important. Similarly, some tools being developed today, which involve broader payment units than fee-for-service, would also address these issues. For example, to the extent that payers determine which medical groups are most efficient in treating an episode of care for particular procedures or illnesses and use incentives to direct patients to them, physicians’ incentives would be changed. But these departures from fee-for-service are developing slowly, and it is unclear how widespread they will become.

As the medical care system continues to become more competitive and technology opens up increasing opportunities to exploit inadvertent payment system incentives, the need increases to make payment rates more accurate. Otherwise, another powerful driver of health cost trends may have been created. Society risks getting a lot more of the services for which payment rates are too high and perhaps not enough of the others.

The authors are grateful for financial support from the California HealthCare Foundation, based in Oakland, California. The paper benefited from interviews with many insurance executives responsible for provider payment and from comments or discussion with William Cleverly, John Cookson, Elizabeth Dichter, Zach Dyckman, Allen Dobson, Richard Gundling, Tom Gustafson, Mark Miller, William Rich, Deborah Williams, and Jill Yegian.

NOTES

1. Julian Pettingill, Medicare Payment Advisory Commission, presentation to the commission, 29 October 2004.
2. H.H. Pham et al., “Financial Pressures Spur Physician Entrepreneurialism,” Health Affairs 23, no. 2 (2004): 70–81.
3. K.J. Devers, L.R. Brewster, and L.P. Casalino, “Changes in Hospital Competitive Strategy: A New Medical Arms Race?” Health Services Research 38, no. 1, Part 2 (2003): 447–469.
4. MedPAC, “Assessing Payment Adequacy and Updating Payments in Fee-for-Service Medicare: Ambulatory Surgical Center Services,” in Report to the Congress: Medicare Payment Policy (Washington: MedPAC, March 2003), 144; and D. Armstrong, “MRI and CT Centers Offer Doctors Way to Profit on Scans,” Wall Street Journal, 2 May 2005.
5. L. Casalino, H. Pham, and G. Bazzoli, “Growth of Single-Specialty Medical Groups,” Health Affairs 23, no. 2 (2004): 82–90.
6. Although most economists believe that increases in facilities or physician supply lead to higher spending, estimation of the magnitude of this effect is difficult and has been the subject of extensive controversy. See D. Dranove and P. Wehner, “Physician-Induced Demand for Childbirths,” Journal of Health Economics 13, no. 1 (1994): 61–73.
7. See L.G. Aronovitz, “Referrals to Physician-Owned Imaging Facilities Warrant HCFA’s Scrutiny,” Pub. no. GAO/HEHS-95-2 (Washington: U.S. Government Accountability Office, 1994); and B.E. Kouri, R.G. Parsons, and H.R. Alpert, “Physician Self-Referral for Diagnostic Imaging: Review of the Empiric Literature,” American Journal of Roentgenology 179, no. 4 (2002): 843–850.
8. The volume/quality relationship has been documented most extensively for open-heart surgery. See, for example, H.S. Luft et al., Hospital Volume, Physician Volume, and Patient Outcomes: Assessing the Evidence (Ann Arbor, Mich.: Health Administration Press, 1990).
9. Z. Dyckman and P. Hess, Survey of Health Plans Concerning Physician Fees and Payment Methodology, June 2003, www.medpac.gov/publications/contractor_reports/Aug03_PhysPaySurveyRpt.pdf (17 June 2005).
10. MedPAC, Report to the Congress: Physician-Owned Specialty Hospitals (Washington: MedPAC, March 2005), Table 10.
11. See MedPAC, “Payment for New Technologies in Medicare’s Prospective Payment Systems,” in Report to the Congress: Medicare Payment Policy, 179.
12. See MedPAC, “Purchasing Strategies,” in Report to the Congress: New Approaches in Medicare (Washington: MedPAC, June 2004), 96–97.
13. “Other Blues Plans Watch Closely as Highmark Starts Radiology Program,” The AIS Report on Blue Cross and Blue Shield Plans (Washington: Atlantic Information Services Inc., September 2004); and Highmark, “Privileging Application and Guidelines,” www.highmark.com/health/professionals/privileging-home-page.html (17 June 2005).
14. Revamping payment for physician-administered drugs was mandated by the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003.
15. Mark Miller, MedPAC, personal communication, November 2004.
16. MedPAC, Report to the Congress: Physician-Owned Specialty Hospitals, Table 10.
17. Pamela Kassing, American College of Radiology, personal communication, October 2004.
18. J. Cromwell, J.B. Mitchell, and W.B. Stason, “Learning by Doing in CABG Surgery,” Medical Care 28, no. 1 (1990): 6–18.
19. The CMS’s proposed revisions in 1998 would also have increased payments for several lower-volume procedures such as arthroscopic surgery and hernia repair, which suggests that the actual payments for these services were less than the costs. See MedPAC, “Assessing Payment Adequacy,” 140–141.
20. William Rich, RVS Updating Committee, personal communication, December 2004. This committee makes recommendations on updating physician payment relative values.
21. We are grateful to William Rich for these points.
22. Testimony of Mark McClellan, Centers for Medicare and Medicaid Services, before the House Committee on Energy and Commerce, 12 May 2005.
23. Steve Phillips, CMS, personal communication, January 2005.


Paul Ginsburg (pginsburg{at}hschange.org) is president of the Center for Studying Health System Change in Washington, D.C. Joy Grossman is a senior researcher there.

DOI: 10.1377/hlthaff.w5.376
©2005 Project HOPE–The People-to-People Health Foundation, Inc.