Preferred provider organizations and physician fees.

Preferred provider organizations (PPOs) represent a form of managed care in which providers agree to accept discounted fees in exchange for the expectation that their patient volume will increase or at least be maintained. Managed care plans that rely on discounted fee-for-service (FFS) payments have increased from about 10 plans in 1981 to over 700 plans in 1994. In this study, we document levels of discounts achieved by two large national insurers and discuss how the size of the discount varies by type of service and how the discounted rates relate to Medicare fees. Our results show that, despite achieving large discounts (approximately 10 20 percent) relative to their indemnity plans, the two nationwide PPOs studied here pay at rates substantially above Medicare levels.

Introduction

One of the most significant changes in the health care delivery system in the past decade has been the movement away from traditional FFS insurance to managed care arrangements. This has been largely driven by payers' demands for cost control and a growing provider surplus. Payers are becoming more discriminating consumers, searching for low-cost providers and using market power to negotiate discounted prices. Spurred by a need to maintain patient volume, physicians are increasingly accepting discounted payments.

Among the broad range of managed care arrangements that are evolving, PPOs represent one of the fastest growing alternatives to FFS or traditional health insurance. In PPOs, an insurer or other third-party payer contracts with selected physicians, hospitals, and other providers to deliver services at discounted rates. Since financial incentives are provided for enrollees to use these preferred providers, providers are willing to accept discounted fees in exchange for the expectation that their patient volume will increase or, at least, be maintained. PPOs are an attractive type of managed care arrangement because they offer something for everyone. The payer hopes to control costs through discounted fees and utilization management controls. Enrollees save money by choosing a contracted provider, although they are free to go to other providers. Providers can potentially increase their patient loads by agreeing to discounted fees and forms of utilization review, while often avoiding capitated payments.

Contracting with a panel of providers for health care services, which occurs in PPOs, is not entirely new. In fact, the Blue Cross and Blue Shield insurance plans' participating provider program rests on a variant of PPO contracting. However, only recently has the notion of selective contracting been used extensively as a cost-containment device. The recent growth in PPOs has been remarkable. Over a decade ago, PPOs played almost no role in the health insurance market. According to one source, in 1981 fewer than 10 PPOs had contracts to serve enrollees (Barger, 1985). By 1987, that number increased to over 100 plans (American Association of Preferred Provider Organizations, 1995). In 1994, there were over 700 plans. Similarly, the number of people enrolled in PPOs has also increased from an estimated 10.4 percent of individuals with private insurance in 1988 to 43.0 percent in 1993 (Employee Benefit Research Institute, 1995). Physician participation in PPOs mirrors these enrollment trends—increasing from 45 percent in 1988 to 64 percent in 1993 (American Medical Association, 1994). Overall, physicians receive about one-fifth of their revenue from PPO arrangements.

According to a recent study of 30 PPO plans, the predominant payment method for providers was FFS (Gold et al., 1995). In fact, none of the PPOs surveyed used capitation as a basic form of physician reimbursement. Although this study investigated provider payment methods, it did not address the level of discount payers are able to obtain from panels of preferred providers relative to what they pay for indemnity claims. The larger the size of the PPO discount the payer achieves, the greater the payer's ability to control health plan costs.

PPO discounting is not merely a private-sector issue. It may also play an important role in payment decisions among public payers. Historically, policymakers have been concerned that relatively low public fees in comparison with private payers could create access problems for beneficiaries. Their concerns, for example, have led to annual tracking of the relationship between Medicare and private fees by the Physician Payment Review Commission (PPRC). Their findings suggest that Medicare fees have been, on average, between 30 and 40 percent below private fees for most of the 1990s (Physician Payment Review Commission, 1995). However, it has been argued that, by relying on very little data from managed care plans, PPRC's results are misleading and “in many instances, Medicare has started paying more to doctors than private payers do” (Miller, 1995). This conclusion, though, appears to be based on fee data from a limited number of services and geographic areas.

If Medicare fees were, in fact, becoming more generous relative to private payers, then this could allow public payers to save money by lowering their fees without a great risk of impeding access. Alternatively, if generous private fees had been providing cross-subsidies for public (and uninsured) patients, discounting could pressure public payers to raise their fees under certain circumstances. For example, lower private payments brought about by market forces could threaten some essential providers' financial viability, causing public payers to respond by providing additional sources of revenue (e.g., a higher Medicare bonus in personnel shortage areas). Although it is not possible to predict how private discounting will affect public payers' decisions, it is important to understand how the growth in privately-discounted fees may be changing the relationship between public and private payments.

In this article, we measure PPO discounts achieved by two large private payers in 1993. For the two payers, we consider these discounts from a national perspective and explore how discounts vary across types of physician services. In addition, we contrast the discounted PPO rates with Medicare fees during the study year. Although these private payers may not be representative of all private payers, they are large and national in scope and, as such, provide a reasonable basis for making comparisons with Medicare. In the section that follows, we describe the data available to this study and outline our methods. We then present our findings and conclude with a discussion of the potential implications of these results.

Data and Methods

Data Sources

This analysis relies on data from Medicare's National Claims History System and two private-sector third-party payers. These private-sector payers include two large national insurers, one covering 4.5 million lives. Both payers operate PPO and indemnity plans in all 50 States and the District of Columbia and provided data from all geographic areas. Due to data use agreements, neither payer providing data can be identified by name. Therefore, the data sources are referred to as Payer 1 and Payer 2. We recognize that data from two payers are not generalizable to all private payers. However, these payers were willing to cooperate and provide data that is rarely available publicly; we contacted other payers who were not receptive to the idea. Moreover, the insurer covering 4.5 million lives may also apply the same payment rules to many other health plans across the country. Therefore, the payments derived from this payer are likely to reflect a larger share of the private market than is represented in the claims available for us to analyze. It would require arbitrary assumptions, however, to estimate the market share of either payer in this study.

Table 1 provides a summary of the private payer data used in the analysis. The claims experience represented in Payer 1 reflects 3 months of data from 1993 while Payer 2 represents the entire year's data. Because we are primarily interested in estimating average payment rates and discounts, there is a sufficient amount of data in 3 months of claims experience to produce reliable estimates. Although Payer 1 data represents a much shorter time period than Payer 2, the number of indemnity services reported in these data bases are comparable. PPO claims in each data set are defined by whether the service was rendered by a PPO provider and not by the health plan in which the patient is enrolled. Claims for patients enrolled in PPO plans who receive services by non-PPO providers are considered indemnity services and classified accordingly. There are twice as many PPO services reported in the Payer 2 data. In total, there are over $950 million in payments for Payer 1 and $1,100 million in payments for Payer 2.

Table 1

Characteristics of Private Payer Data Sources: 1993

Payer		Months of Data	Total Physician Services(Millions)	Total Payments(Millions of Dollars)
Payer 1	PPO	3	3.8	$195.8
	Indemnity	3	11.3	771.2
Payer 2	PPO	12	13.6	812.3
	Indemnity	12	5.8	377.5

NOTE: PPO is preferred provider organization.

SOURCE: Urban Institute analysis of 1993 claims from two large private payers.

We reviewed and edited each source of data in order to develop analytic files which consisted of claims for physician services and clinical laboratory services only. This editing involved omitting claims with invalid Current Procedural Terminology (CPT) codes, claims for medical supplies, durable medical equipment, and ambulance services, and claims for oncology, dialysis, and anesthesia services. In each data set, about 10-15 percent of the total charges were dropped as a result of this process. It was possible to identify claims for surgical assistants and the professional component of radiology and other services. This is important for expressing payment rates in terms of relative value units (RVUs) (discussed later).

In addition, claims with apparently erroneous payment data in the private and Medicare sources are screened out. The objective was to remove those claims that seemed to have very high or very low average payments and, specifically, eliminate claims for partial payments (i.e., payments for surgical assistants) that are not adequately identified in the data. The approach we used eliminates all claims that were more than three times or less than one-third of the mean payment for the service. In order to avoid screening out disproportionate numbers of claims from high- or low-cost areas, we first adjust for differences in payments using HCFA's Geographic Practice Cost Index (GPCI). Overall, outlier claims accounted for only about 2 percent of payments.

Also, we identified some Medicare services with very low or very high payments per RVU relative to the 1993 fee schedule conversion factors. These services tended to be unusual services within the Medicare program (e.g., newborn care, obstetrical care, and antigen therapy). To address this issue, all national average Medicare payments per RVU that were less than one-third or greater than three times the fee schedule conversion factors were excluded from the computations.

Methods

The size of the PPO discount for each payer in this study can be measured as the ratio of the average payment for services paid through the payer's PPO plans to the average payment for services paid through the indemnity plan. Analogously, the payment differential between Medicare and the PPOs equals the ratio of the average Medicare payment to the average PPO payment. To compute these ratios nationally, two issues must be addressed. First, how can payments for different services (e.g., office visits and surgeries) be expressed in terms of consistently-defined units of volume so that they can be combined into summary measures of PPO discounts or Medicare-to-PPO differentials? Second, if our goal is to focus solely on the price discount, how do we control for differences in the mix of services between PPO, indemnity, and Medicare claims when computing the price ratios?

The first issue can be addressed directly by expressing payments for individual services in terms of the number of RVUs contained in the service. RVUs, as defined by the Medicare Relative Value Scale, can be thought of as basic units of service volume that have the same meaning in all physician services. For example, if an office visit contains one RVU and is paid at a rate of $45, its payment per RVU would be $45. If an arthroscopy, on the other hand, contains 25 RVUs and is paid at a rate of $1,500, its payment per RVU would be $60. By expressing these payments in terms of RVUs, we can combine the payment rates for different services into a weighted average payment rate, using the distribution of RVUs across services as the weights. These weighted average payments per RVU can be computed separately for all PPO and indemnity claims, as well as for groups of services within these payment categories, allowing for fairly easy computation of PPO discounts. The average payment per RVU can be computed similarly for Medicare.

The issue of service mix differences between PPO, indemnity, and Medicare claims can also be dealt with in a straightforward fashion. The solution relates to the weights used in computing the average payment per RVU when aggregating services. Our basic choice is to use the PPO shares of RVUs, the indemnity shares of RVUs, or the Medicare shares in all instances. Since our goal is to measure the size of the discount the PPO receives relative to the indemnity part of the payer's business, it makes sense to weight all prices by the distribution of indemnity service RVUs. In this way, the discount we compute will tell us how much lower indemnity payments might have been if their prices paid were at PPO levels. Therefore, PPO and indemnity average payments per RVU and the resulting discounts are computed as if the indemnity service mix prevailed within the PPO claims.

Once the indemnity service mix is selected as the basis for computing the PPO discount, it makes sense to be consistent and use it in deriving the Medicare-to-PPO differential. This is done separately for each PPO, since the indemnity service mix varies by payer. When indemnity weights are applied to Medicare payments per RVU, the resulting average payment per RVU can be thought of as the level of payment indemnity plans would have offered if they paid at Medicare rates. The differential then reflects the potential additional savings (beyond the PPO discount) that might have accrued to indemnity plans had they adopted Medicare's rates.

The GPCI adjustment used in the data screening (described previously) was performed in order to remove potential distortions resulting from a disproportionate number of claims coming from either high- or low-cost areas. For similar reasons, the GPCI-adjusted payment rates were used in computing national average payments per RVU for all payers.

Results

Table 2 presents weighted average payment rates per RVU for each payer's PPO and indemnity plan for All Services and four broad type of service categories. Recall that the weights used in both the PPO and indemnity calculations represent the service mix of indemnity claims. First, the results show sizeable differences in indemnity payment rates across the two payers. For Payer 1, the average payment per RVU for all indemnity claims is $58.66, 17.2 percent higher than the average indemnity payment per RVU of $50.05 for Payer 2. In all services categories indemnity payments for Payer 1 exceed those for Payer 2 (although, for imaging the difference is negligible). Second, as a result of the discounts each payer is able to achieve, PPO payment levels for Payer 1 and Payer 2 are fairly close to each other, with average weighted payments per RVU for All Services of about $46 and $44, respectively. For these two payers, this new form of managed care seems to result in a reduction in rate differentials across payers. Data from a broader set of payers would be required before more general conclusions can be reached.

Table 2

Average PPO Payments per RVU for Broad Categories of Physician Services: 1993

Type of Service Categories	Payer 1			Payer 2
	PPO	Indemnity	Discount(Percent)	PPO	Indemnity	Discount(Percent)
All Physician Services	$46.48	$58.66	20.8	$44.29	$50.05	11.5
Evaluation and Management	33.75	43.34	22.1	37.24	40.49	8.0
Procedures	57.96	73.23	20.9	55.81	65.09	14.3
Imaging	55.35	66.96	17.3	57.53	66.86	14.0
Tests	48.19	59.81	19.4	46.25	55.08	16.0

NOTE: PPO is preferred provider organization.

SOURCE: Urban Institute analysis of 1993 claims from two large private payers.

Taken together, the indemnity and PPO payment rates show that Payer 1 is able to achieve a higher and more uniform discount across the broad service categories shown than Payer 2. For example, for All Services, the discount achieved by Payer 1 is nearly twice the size of the discount achieved by Payer 2 (20.8 percent versus 11.5 percent). Further, the payment discount for Payer 1 varies only slightly by broad type of service category (17.3 percent for imaging services to 22.1 percent for evaluation and management services), while for Payer 2 the level of the discount ranges from 8.0 and 16.0 percent. It may be that, because Payer 1 is paying substantially more for services among its indemnity claims, it has more room to negotiate larger discounts. However, other factors, such as differences in the two payers market shares may also be playing an important role in determining these discounts.

Table 3 shows the size of the differentials between PPO payments and indemnity payments for each payer by more detailed type of service categories. To some extent, the size of the discount for each payer varies within each of the broad service categories. These varying discounts are most pronounced in the Evaluation and Management (E&M) service category. They range from 8.7 percent for emergency visits to 26.5 percent for office visits for Payer 1 and 2.1 percent for nursing home visits to 11.7 percent for hospital visits for Payer 2.

Table 3

Average PPO and Indemnity Payments per RVU for Detailed Categories of Physician Services: 1993

Type of Service Categories	Payer 1		Payer 2		Payer 1 Discount(Percent)	Payer 2 Discount(Percent)
	PPO	Indemnity	PPO	Indemnity
All Physician Services	$46.48	$58.66	$44.29	$50.05	20.8	11.5
Evaluation and Management (E/M)	33.75	43.34	37.24	40.49	22.1	8.0
Office Visits	31.02	42.22	37.49	41.03	26.5	8.6
Hospital Visits	38.40	47.29	39.56	44.80	18.8	11.7
Emergency Room Visits	55.75	61.08	45.31	50.51	8.7	10.3
Nursing Home Visits	30.58	37.44	32.92	33.62	18.3	2.1
Specialist E/M	30.04	35.72	35.55	37.68	15.9	5.7
Consultations	37.57	46.65	39.29	44.15	19.5	11.0
Procedures	57.96	73.23	55.81	65.09	20.9	14.3
Major Procedures-General	59.76	77.21	59.80	72.15	22.6	17.1
Major Procedures-Cardiac	66.98	81.14	61.45	68.78	17.5	10.7
Major Procedures-Orthopedic	65.95	82.71	59.39	69.56	20.3	14.6
Eye Procedures	62.54	74.18	59.12	66.40	15.7	11.0
Ambulatory Procedures	59.94	79.95	54.99	64.31	25.0	14.5
Minor Procedures	46.80	57.15	45.31	51.40	18.1	11.8
Endoscopy	73.18	92.08	67.18	79.57	20.5	15.6
Imaging Services	55.35	66.96	57.53	66.86	17.3	14.0
Standard Imaging	57.22	68.79	58.76	67.97	16.8	13.6
Advanced Imaging	54.12	62.54	53.03	63.90	13.5	17.0
Echography	50.25	64.56	55.28	63.21	22.2	12.5
Imaging/Procedure	65.65	79.39	69.50	79.26	17.3	12.3
Tests	48.19	59.81	46.25	55.08	19.4	16.0
Laboratory Tests	47.60	59.51	44.67	54.17	20.0	17.5
Other Tests	49.71	60.58	49.38	56.87	17.9	13.2

NOTE: PPO is preferred provider organization. RVU is relative value unit.

SOURCE: Urban Institute analysis of 1993 claims from two large private payers.

The E&M category also has the greatest difference in discounts between Payer 1 and Payer 2. For example, in the office visit category, Payer 1 is able to achieve a 26.5 percent discount while the discount achieved by Payer 2 is only 8.6 percent In fact, the PPO discounts for Payer 1 are so large among E&M services that, despite having higher indemnity rates than Payer 2 in all but one of the E&M categories, Payer 1 actually has lower PPO rates than Payer 2 in five of the six service groups. For example, the average indemnity payment per RVU for office visits is about $42 for Payer 1 and $41 for Payer 2, but drops to $31 and $37, respectively, among PPO claims. The only E&M service category where the opposite finding is true is emergency room visits, where discounts are comparable and average PPO rates for Payer 1 are about 23 percent higher than those for Payer 2.

Payer 1 is also able to negotiate lower PPO rates than Payer 2 in the imaging category. Similar to the E&M category, Payer 1 starts with higher indemnity payments than Payer 2 in three out of the four imaging categories, but achieves larger discounts. For instance, the average indemnity payment rate for echographies is about $65 for Payer 1 and $63 for Payer 2 but drops to $50 and $55 for Payers 1 and 2, respectively, as a result of the discounting.

Despite the sizeable PPO discounts achieved by both payers, Table 4 shows that PPO rates are still well above those paid by Medicare in 1993. Across all services, on average, 1993 Medicare fees were 35.2 percent and 33.2 percent below the PPO rates of Payer 1 and Payer 2, respectively. These differentials are consistent with those reported by PPRC (1995). Not surprisingly, given that the Medicare fee schedule increased payments for E&M services relative to Procedures, the differential between Medicare and PPO rates was lowest for E&M services (15.3 and 24.2 percent) and highest for Procedures. With the exception of E&M services, the Medicare-to-PPO differential is similar for both payers across service categories. For E&M, payments made by Payer 1 are closer to Medicare than are those made by Payer 2.

Table 4

Average Medicare and PPO Payments per RVU for Broad Categories of Physician Services by Private Payer: 1993

Type of Service Categories	Payer 1			Payer 2
	Medicare	PPO	Differential(Percent)	Medicare	PPO	Differential(Percent)
All Physician Services	$30.11	$46.48	35.2	$29.58	$44.29	33.2
Evaluation and Management	28.57	33.75	15.3	28.21	37.24	24.2
Procedures	30.93	57.96	46.6	31.18	55.81	44.1
Imaging	32.78	55.35	40.8	33.58	57.53	41.6
Tests	29.79	48.19	38.2	29.94	46.25	35.3

NOTE: PPO is preferred provider organization. RVU is relative value unit.

SOURCE: Urban Institute analysis of 1993 claims from the Medicare National Claims History System and two large private payers.

The results in Table 4 provide a general sense of the differences between Medicare and PPO payments. However, because the computations are based on average payments per RVU, they do not show how payments for specific services vary across payers. While it would be impractical to include an exhaustive listing of Medicare and PPO prices, we have included national average payments per service for a selected set of services in Table 5. Our goal is to offer an alternative, and potentially more tangible, basis for comparing Medicare prices with PPO prices. Within each type of service group, the services shown were selected from among those accounting for the largest shares of spending among the indemnity side of private payers.

Table 5

Average Medicare and PPO Payments per Service for Selected Services: 1993

CPT Code and Service Description		Average Payment Per Service			Difference Between Medicare and PPO Payment Rate
		Medicare	Payer 1PPO	Payer 2PPO
					Payer 1	Payer 2
					Percent
Evaluation and Management
99213	Office or Other Outpatient Evaluation of an Established Patient, 15 Minutes	$30	$32	$39	6.3	23.1
99232	Subsequent Hospital Evaluation of a Patient, 25 Minutes	40	58	58	31.0	31.0
99244	Office Consultation with a New or Established Patient, 60 Minutes	107	131	141	18.3	24.1
Procedures
45378	Diagnostic Colonoscopy	284	583	572	51.3	50.3
49505	Inguinal Hernial Repair	376	892	827	57.8	5
67228	Destruction of Retinal Lesion by Laser Treatment(s)	671	1,042	996	35.6	32.6
92982	Percutaneous Transluminal Coronary Balloon Angioplasty, Single Vessel	1,181	2,195	2,061	46.2	42.7
Imaging1
70470	Contrast CAT Scan of the Head	293	495	454	40.8	35.5
70553	Magnetic Resonance Image of the Brain, Without Contrast Material	879	1,102	996	20.2	11.7
71020	Chest X Ray, Two Views	31	54	53	42.6	41.5
76805	Echography Exam of a Pregnant Uuterus	117	164	153	28.7	23.5
Tests2
93015	Cardiovascular Stress Test	109	198	205	44.9	46.8
94060	Evaluation of Wheezing	51	75	71	32.0	28.2
95904	Sensory Nerve Conduction Study	30	57	50	47.3	40.0

Includes performing procedure and interpretation of imaging results.

Includes performing procedure and interpretation of test results.

NOTE: PPO is preferred provider organization.

SOURCE: Urban Institute analysis of 1993 claims from the Medicare National Claims History System and two large private payers.

The payments shown in Table 5 highlight large variations in the size of the Medicare-to-PPO differential across specific services for these payers. The size of the differential ranges from a low of 6.3 percent (established office visit for Payer 1 PPO) to a high of 57.8 percent (inguinal hernia repair for Payer 1 PPO). Within imaging services, the differentials for a two-view chest X ray (CPT 71020) and a CAT Scan of the Head (CPT 70470) are very close to the overall differential of approximately 40 percent shown in Table 4. However, for the other two imaging services reported in Table 5, the differentials are smaller. In fact, Medicare's payment for a Magnetic Resonance Image (MRI) of the Brain (CPT 70553) is only 11.7 percent below the average payment rate of Payer 2's PPO.

Discussion

If PPOs are going to thrive as a form of managed care that is able to control private spending, then they must be able to pay providers at rates well below those in traditional insurance plans. Our results show that the two payers for which we have data were able to establish heavily discounted rates within their PPOs. This suggests that, unless volume responses completely offset these discounts, purchasers covered by these payers' rules should see PPOs as an effective means of lowering their health care spending. However, the payer who was paying at the more generous level in its traditional insurance plans was able to negotiate larger PPO discounts than the low-rate payer. Though preliminary, these findings suggest that, for any given payer, the extent of the reduction in spending may be a function of their initial level of prices.

A potentially important implication of the variation in the size of PPO discounts is that the extent of heterogeneity in private payment rates across payers for physician services could gradually diminish. Based on the two payers in this study, we see a 17-percent differential in indemnity payment rates reduced to 5 percent in the PPO market. However, the process of PPO discounting did not seem to undue historical differences in payments across types of services that the Medicare fee schedule was designed to address. Despite some variations in the discounts, for both payers, indemnity and PPO payment rates for procedures, imaging services and tests were considerably higher than rates for E&M services. If the expectation is that through the discounting process fees for E&M services will increase relative to those for other services, as Medicare accomplished through the adoption of its Relative Value Scale (Levy and Borowitz, 1992), we find little evidence to suggest that has been accomplished by either of the two PPOs analyzed here.

These changes taking place in private sector fees can have important implications for future developments in the policies of public payers. Decisionmakers have always been concerned that reductions in public rates could make their beneficiaries less attractive to providers and create access barriers. As PPOs grow and their lower fees become more typical of the private market, public payers may have hoped that they would have had the opportunity to lower their rates without risking serious access problems for its beneficiaries. Of course, this opportunity may not have materialized if the lower private rates reduced cross-subsidies to public payers and created demands for higher public fees. However, based on the two payers studied here, there is little reason to conclude that the gap between public and private rates is disappearing.

Based on these findings, two conclusions revelant to Medicare policy follow:

First, simply enrolling Medicare beneficiaries in private PPOs paying at the rates reflected by the two payers in this study would not necessarily result in lower spending, on average, as a result of lower prices. If the argument that “the Medicare program could be rescued if only the Government would adopt some of the cost controls that employers have imposed on their workers under the banner of ‘managed care’” (Freudenheim, 1995) is true and these two payers' prices are reasonably representative, then savings would have to come from lower rates of service utilization. Whether or not utilization controls would be acceptable to beneficiaries would be an issue policymakers would have to confront.

Second, despite arguments to the contrary (Miller, 1995), there is no evidence from this study or earlier work by PPRC to suggest that national average Medicare fees are generous or are much closer to private payer fees than they have been historically. Therefore, the view that significant program savings can be achieved by reducing Medicare fees without access concerns being an issue may be overly optimistic. Although we acknowledge that these results are based on 1993 data and that the rapidly changing market may have already led to lower private fees than those observed here, a great deal of ground would have had to have been closed in order to put Medicare fees near those of the average private payer. If Medicare reduces its fees and access is not adversely affected, as has occurred at various times over the past decade, it is more likely to be due to providers' reliance on Medicare revenues than on the fact that Medicare is now the generous payer in the physician services market.