Payment for Physician Services 1 Kathy Bryant, Esq. David Hilgers, Esq. Sidney Welch, JD, MPH

Transcription

1 Payment for Physician Services 1 Kathy Bryant, Esq. David Hilgers, Esq. Sidney Welch, JD, MPH Medicare s system for paying physicians has evolved in a manner somewhat similar to the hospital payment system. Initially, Medicare generally paid a physician s reasonable charges with few limitations. The payment system gradually accumulated restrictions resulting from cost containment efforts and then was converted to a fee schedule. In 2014, Medicare pays providers the lesser of the actual charge for the service or the allowed amount under the Medicare physician fee schedule. However, this system also is evolving as Medicare is experimenting with new payment methodologies in order to reduce costs and improve quality. This chapter describes Medicare s current system for paying physicians, as well as the new payment models to the degree that they can be determined as of this edition. 4-1 The Physician Fee Schedule 4-1(a) History Prior to 1992, Medicare paid for most physician services based on reasonable charges, as limited since 1973 by the rate of inflation. This charge-based system resulted in payment levels that some viewed as overcompensating procedural services, such as surgery, and underpaying for cognitive services, such as primary care. In addition, the system was seen as underpaying physicians in certain areas, particularly rural areas. In response to these concerns, statutory amendments enacted in 1989 required CMS to develop and implement a fee schedule derived from a resource-based relative value scale (RBRVS) and adjusted geographically only for cost differences. 2 The same legislation also contained Medicare s first attempt to control aggregate spending for physician services and protections for beneficiaries through new restrictions on physician charges. Effective January 1992, the basis of the physician s payment was and is the lower of actual charge or the fee schedule amount. 1 This paper was originally developed for AHLA with an initial set of authors and has been expanded, revised, and updated over the years. This year s presenters gratefully acknowledge the assistance of Samuel Shapiro, a third year dual degree JD/MSHA candidate at Georgia State University. 2 SSA 1848; 42 U.S.C. 1395w-4; Medicare Claims Processing Manual, CMS Pub , Ch

2 The new PFS system used today (described in detail below) altered Medicare s historical payment policy in several fundamental ways. First, it moved Medicare from a market-based approach of basing payment amounts on physician charges to an ostensibly objective method based on the resources used. Although the previous payment system was not closely linked to current charges because of the long-standing limits on permitted annual inflation in payment rates, and it is questionable whether charges for physician services are actually based on traditional market forces, the current payment method is conceptually far different from its predecessor. Second, the fee schedule converted what had been a relatively decentralized payment system into a uniform national structure. Previously, policies of local Medicare carriers frequently controlled how particular services were paid for. an approach that was consistent with a system based on local charging practices. Because : the new fee schedule is nationally uniform, however, the interpretations and implementing policies must also be uniform. This change required creating numerous new policies at the national level to replace local rules. Finally, the fee schedule attempted to control aggregate Medicare spending for physician services by reducing payment rates for services if overall spending rises too fast. Basing annual increases in payment rates in part on the rate of increase in aggregate spending contrasts with the prior method of looking largely to the rate of inflation. 4-1(b) Current System Overview Under the fee schedule, CMS has established a relative value for each type of physician service identified in the coding system used by Medicare. For each of the three components of a relative value physician work (approximately 52%), practice expense (approximately 44%), and malpractice costs (approximately 4%) CMS has also developed indexes to reflect geographic variations in costs. In general, the payment amount for a particular service and physician is figured by multiplying each component of the relative value for the service by the geographic index for the area in which the physician service is furnished. The result is then converted to dollars by multiplying it by a conversion factor. The formula for calculating 2014 physician fee schedule payment amount is as follows: 2014 Non-Facility Pricing Amount = [(Work RVU * Work GPCI) + (Transitioned Non-Facility PE RVU`* PE GPCI) + 2

3 (MP RVU * MP GPCI)) * Conversion Factor (CF) 2014 Facility Pricing Amount = [(Work RVU * Work GPCI) + (Transitioned Facility PE RVU * PE GPCI) + (MP RVU * MP GPCI)] * CF As an example, consider how Medicare determines the payment amount for a three-graft venous artery bypass operation (CPT code 33519) performed in Chicago, Illinois (locality 95216). The number of relative value units (RVUs) for each of the three components for this code is multiplied by the geographic index for Chicago: physician work: RVUs x = practice expense: 3.58 RVUs x = malpractice cost: 2.52 RVUs x = The sum of these calculations ( ) is multiplied by the conversion factor ( ) to determine the payment amount ($673.99). Medicare pays 80% of the fee schedule amount (or of the physician s actual charge if it is lower), and the beneficiary is responsible for paying the remaining 20% (the patient s copay). The conversion factor published in the final rule for 2014 was $ , which would have been a reduction of 20.1%. 3 However, in December, 2013 Congress passed legislation providing for a temporary 0.5% increase in the conversion factor. 4 Accordingly, the 2014 conversion factor through March 31, 2014 is $ (c) Applicability of the Fee Schedule The fee schedule applies to services furnished by physicians. By virtue of Medicare s definition of physician, 6 the fee schedule is applicable to physicians, dentists, podiatrists, optometrists, and chiropractors to the extent that their services are covered by Medicare. The fee schedule also applies to services furnished by a variety of other practitioners specified in the 3 The reduction would have been even larger under the SGR system discussed on page 21, but due to some changes in PE values an adjustment of approximately 5 percent was made. 4 Pathway to SGR Reform Act of 2013, P.L (Dec. 26, 2013); see also, American Academy of Sleep Medicine (AASM), SGR cuts delayed Obama signs pathway for SGR Reform Act of 2013, Dec. 30, 2013, 5 Id. 6 SSA 1861(r); 42 U.S.C. 1395x(r). 3

4 statute including, physician assistants, nurse practitioner, physical therapists, certified nurse midwives and others. The fee schedule also applies to certain services, regardless of whether they are furnished by physicians: outpatient physical and occupational therapy; antigens; diagnostic tests other than clinical diagnostic laboratory tests (which are subject to their own fee schedule); and radiology services. 4-2 Relative Values Historically, the relative value established for each service is supposed to reflect the resources typically used in furnishing the service. As a resource-based relative value scale, the fee schedule does not reflect factors such as the value of a service to the patient. Each relative value ordinarily consists of three components: 7 A work component that reflects physician time and intensity in providing the service, includes before and after patient contact pre- and postoperative services; A practice expense component that reflects overhead costs (excluding malpractice costs); and A malpractice component that reflects malpractice expenses associated with the service. 4-2(a) Work Component 4-2(a)(1) Initial Relative Values In assessing physician time and intensity to set the work component of relative values, CMS relied on an elaborate survey methodology for a relatively small number of codes and determined values for the remainder of the codes more informally. The primary work in sizing the work components was performed by a group at the Harvard School of Public Health under William Hsiao, pursuant to a cooperative agreement with CMS. 8 Hsiao developed vignettes 7 SSA 1848(c)(1); 42 U.S.C. 1395w-4(c)(1). 8 See 56 Fed. Reg. 59,502 (Nov. 25, 1991) (CMS s implementing regulations); WILLIAM C. HSIAO ET AL., A NATIONAL STUDY OF RESOURCE-BASED RELATIVE VALUE SCALES FOR PHYSICIAN SERVICES: PHASE II (1990); WILLIAM C. HSIAO ET AL., A NATIONAL STUDY OF RESOURCE-BASED RELATIVE VALUE SCALES FOR PHYSICIAN SERVICES: PHASE III (1992); William C. Hsiao et al., Resource-Based Relative Values: An Overview (and accompanying articles), 260 J. AM. MED. ASS N (1988). 4

5 that corresponded to some of the codes in the American Medical Association s code book for physician services, Current Procedural Terminology (CPT). The vignettes described a hypothetical patient s condition or a service furnished to a patient. For example, the vignette for CPT code was aspiration of synovial fluid from the knee, and one of the vignettes for (old) CPT code was office visit with a 65-year-old established patient with eruption on both arms from poison oak exposure. A group of vignettes covering CPT codes was developed for each of a number of physician specialties. Hsiao s surveys in the first phases of his work resulted in relative values for 460 codes, and based on this work, he determined values for a total of about 4,300 codes by extrapolation. Initially, Hsiao attempted to extrapolate by comparing physician charges for a surveyed service to charges for related services. This method produced problematic results, however, and Hsiao substituted a process in which small groups of physicians assigned relative values to unsurveyed procedures by comparing them to the surveyed procedures. In addition to the 4,300 relative values developed by Hsiao, approximately 825 codes were covered by the pre-existing radiology and anesthesiology relative value scales. CMS itself, in a process involving a group of carrier medical directors, determined relative values for about 400 other codes, generally involving low-volume or non-physician services. Also, as a result of comments received from the public, CMS, assisted by carrier medical directors, revised hundreds of relative values that Hsiao had assigned to codes. CMS determined the relative values for evaluation and management (E/M) services (e.g., visits and consultations) somewhat differently than the relative values for procedures. 9 Based on Hsiao s work, CMS estimated the relative value units per minute of intra-service (face-to-face) work involved in each type of E/M service. These estimates of RVUs per minute were then multiplied by the average number of minutes in each service, as stated for typical services in the CPT, to determine the total intra-service work for each service. A standard percentage add-on was applied to account for pre- and post-service work. CMS decided that the RVUs per minute of intra-service work should be linear with respect to time for any category of service, rejecting the arguments of some specialties that RVUs per minute were greater in short visits or, alternatively, in especially long visits. 9 See 57 Fed. Reg. 55,896, 55, (Nov. 25, 1992); 56 Fed. Reg. 59,502, 59,529 (Nov. 25, 1991). 5

6 4-2(a)(2) Relative Values for New and Revised Codes Today, as additional CPT codes are established for new and redefined physician services, CMS continues to use the magnitude estimation process to assign relative values to the physician work component of the codes. The process is carried out in coordination with the American Medical Association/Specialty Society Relative Value Update Committee (AMA RUC). The AMA RUC works with CMS to identify and review potentially misvalued codes. Since 2006, the AMA RUC has identified over 1,500 potentially misvalued services through objective screening criteria and has ultimately recommended a decrease or deletion for nearly half of the identified services. 10 Typically, physician specialty groups interested in the new code conduct surveys of physicians, who estimate a work value for the new procedure by comparing the work involved in the new procedure to the relative values of a group of reference services. The median value estimated in the survey is often used. The results of the survey are considered by the American Medical Association s Relative Value Update Committee, which makes recommendations to CMS. CMS can accept or modify these recommendations. The agency then publishes new and revised relative values in the Federal Register through interim final rules. CMS s interim final rules are immediately implemented without the issuance of a proposed rulemaking or opportunity for prior public comment, usually due to the overwhelming need such immediate implementation of a particular rule. However, public comments are often requested sometime after implementation. In such cases, CMS will often revise the rules and publish subsequent finalized versions of the rule. The Affordable Care Act requires CMS to identify, review, and ultimately adjust potentially misvalued CPT codes, with an emphasis on codes that have grown the most, have experienced substantial changes in practice expenses, are recently established for new technologies or services, are frequently billed together in conjunction with furnishing a single service, or are generally determined inappropriate by CMS, amongst others. 11 For 2014, CMS 10 See Letter from American Medical Association to Marilyn B. Tavenner (CMS), Sept. 6, 2013, available at (last accessed March 25, 2014). 11 See American Academy of Family Physicians (AAFP), Summary of the 2014 Final Medicare Physician Fee Schedule, SummaryMedicareFeeSchedule pdf (last accessed March 24, 2014). 6

7 finalized the values for around 200 codes in the CY 2014 final rule. 12 The agency also assigned interim final values for approximately 200 services (b) Practice Expense Component 4-2(b)(1) The Top Down Methodology Historically, although the physician fee schedule has always been termed resource-based, that characterization was originally not accurate for the practice expense and malpractice components, which on average constitute nearly half of the total relative value units. Instead of being resource-based, the statute at first required the relative value units of these components to be determined in a mechanical way based on past allowed charges. 14 In 1994, Congress required CMS to develop resource-based practice expense components for the physician fee schedule. 15 The agency initially used a bottom up methodology, whereby direct costs would be determined by summing the costs associated with the clinical staff, equipment, and supplies typically required to provide services; however this concept was determined to be unworkable. In response, CMS changed direction and adopted a top down methodology that was based on expenses currently being incurred by physicians, rather than on estimates of what expenses ought to be incurred. This method aggregated specialty practice costs and allocated them to specific procedures. Not surprisingly, use of the current expenses of each specialty as the starting point for determining appropriate expenses reduced the degree of redistribution among specialties compared to the bottom up method. In basic outline, CMS s top down method for making the practice expense components resource-based consists of the following steps: 16 CMS s threshold estimate was the practice expenses incurred for each hour of physician work in each physician specialty. Based on the Socio-Economic Survey by the American Medical Association (AMA) of physician office expenses and work hours, CMS computed the amount of expenses per hour of physician work incurred by physicians in each specialty for various categories of expenses (e.g., clinical labor, Fed. Reg (Dec. 10, 2013). 13 Id. 14 SSA 1848(c)(2)(C)(ii); 42 U.S.C. 1395w-4(c)(2)(C)(ii). 15 SSA 1848(c)(2)(C)(ii); 42 U.S.C. 1395w-4(c)(2)(C)(ii) Fed. Reg. 30,818, 30,802-30,840, 30,885-30,902 (June 5, 1998); 63 Fed. Reg. 58,814, 58,816-58,842 (Nov. 2, 1998). 7

8 medical supplies). The survey asked physicians how much they spent in various categories and how many hours they worked each year, thus allowing a calculation that, for example, a specialty spends $7 on supplies for each hour of physician work. For specialties that are not represented in the AMA survey data, CMS used the practice expenses per hour of what it viewed as analogous specialties. Based on the Medicare claims submitted by physicians in each specialty and on Hsiao s estimates of the physician time involved in furnishing each service, CMS estimated the total practice expenses associated with providing services to Medicare patients. For example, if the AMA data showed that cardiologists spent $30 on clinical staff for every hour of physician work, and the Medicare claims data (together with Hsiao s time estimates) indicated that cardiologists spent a million hours providing services to Medicare patients, then the clinical staff costs supporting those services cost the cardiologists $30 million. The preceding steps resulted in an estimate of the total practice expenses of a specialty for the codes it used, but the expenses still had to be allocated to individual codes. CMS used one allocation method for the categories of clinical labor, medical supplies, and medical equipment (sometimes referred to as the direct costs of a service), and a different allocation method for all the other cost categories ( indirect costs ), such as nonclinical staff, rent, utilities, and other costs. In the case of the direct costs, CMS had estimates of such costs for individual codes that had been prepared by the CPEPs when CMS was pursuing the bottom up method. Using these CPEP estimates, CMS allocated the clinical labor, medical supplies, and medical equipment costs of a specialty to individual codes in proportion to the CPEP estimates of those costs. For example, assume that the cardiologists who spent $30 million on clinical labor billed only two codes; for one code, the CPEP had estimated that the clinical labor costs were $40 and for the other code $20. Because the CPEP estimates indicated that two-thirds of the expenses were due to the first code, $20 million of the $30 million total clinical labor expenses would be allocated to the first code. The $20 million total would be converted to a per-service cost by dividing that total by the number of claims for that code in the Medicare claims file. In a controversial step, CMS eliminated costs incurred by some physicians in using their own nurses in hospital 8

9 procedures on the ground that such costs were implicitly paid under the hospital payment system. 17 Where more than one specialty uses the same CPT code, as is usually the case, the practice expenses for each code were determined by using an average of the costs determined for each specialty, weighted by the number of claims that each specialty submitted for the code. The result was an estimate of the average practice expenses incurred in furnishing each service. This step is responsible for much of the resulting reallocation of Medicare payments among specialties, because codes used by multiple specialties are paid based on an average of the costs of those specialties. As part of the preceding step of combining the data for all specialties, CMS adjusted the CPEP data for each specialty by a scaling factor. In essence, CMS s assumption was that the AMA s survey data on expenses were correct, but that the CPEPs had, to varying degrees, mostly overestimated the costs involved. Therefore, the CPEP data were reconciled with the AMA data by adjusting the CPEP data for each specialty by a scaling factor that would make the CPEP data for all specialties comparable to the AMA data. In the case of the indirect costs, for which there were no CPEP estimates for individual codes, CMS allocated the total indirect expenses of a specialty to each code in proportion to the total CPEP estimates of direct costs for each code plus the physician work value relative value units of that code converted to dollars based on the fee schedule conversion factor. This methodology resulted in indirect costs being allocated to individual codes based in significant part on the amount of physician work attributable to each code. CPT codes that had no physician work relative value units (because the procedures are performed by staff) were disadvantaged, and CMS compensated for this by treating such zero work value codes separately. (CMS created a separate pool of funds for these services, referred to as the zero work or nonphysician work pool.) The outcome was made budget neutral. The total number of practice expense relative value units that had existed under the prior system was left unchanged. This Fed. Reg. 59,380, 59, (Nov. 2, 1999). 9

10 number was reallocated to all codes in proportion to the practice expenses estimated for each code under the top down methodology. However, in its CY 2007 final rule, CMS revised the methodology for calculating direct practice expense RVUs. Beginning in 2007, CMS switched from the top-down methodology to the bottom-up methodology, providing a four-year transition period ending in CY The revised method bases the measurement of indirect costs on the magnitude of direct costs. 4-2(b)(2) Differential Payments Based on Site of Service There are many services that a physician could furnish either in the office or, alternatively, in a hospital outpatient department or other facility setting. A physician presumably incurs lower practice expense costs when a service is furnished in a facility setting because the physician does not ordinarily pay for the assistance of the hospital s nurses, supplies used, and other costs that the physician, would pay for personally if the patient were seen in the office. Based on this assumption, carriers were authorized, prior to establishment of the fee schedule, to limit payment for services that were routinely furnished in physician offices to a carrier s determined reasonable charge of the office-based payment amount when the services were furnished in hospital outpatient departments. 19 Prior to the adoption of the physician fee schedule, a service that was performed more than 50% of the time outside of a facility was subject to a payment reduction when the service furnished in a facility setting. Therefore, the practice expense component of the relative value was reduced by 50% of the facility s payment in what CMS called the site-of-service payment differential. When the practice expense components were made resource-based, the same principle was continued in somewhat different form. Practice expenses were estimated separately for services furnished in the facility and for services furnished in a non-facility. The facility rate published in the fee schedule applies to services furnished in hospitals, skilled nursing facilities (SNF), comprehensive inpatient rehabilitation facilities, inpatient psychiatric facilities, community mental health centers (CMHC), and for approved procedures in approved ambulatory surgical centers. The nonfacility rate applies in skilled nursing facilities for services furnished Fed. Reg (Dec. 1, 2006). 19 SSA 1861(v)(1)(K); 42 U.S.C. 1395x(v)(1)(K); 42 C.F.R (f). 10

11 to patients covered only by Medicare Part B, in a patient s home and in a facility or institution other than a hospital, SNF, CMHC, or ASC. 20 Non-facility fees are applicable to therapy procedures regardless of whether the services are furnished in a facility or non-facility setting (c) Malpractice Component Beginning in 2000, CMS revised the malpractice components of the fee schedule to be resource-based. 22 CMS s methodology for establishing resource-based malpractice component was as follows: 23 CMS collected data on malpractice insurance premiums throughout the country for the twenty largest physician specialties and calculated a national average for each specialty. The premium data for the twenty specialties were extrapolated to the other specialties by applying the data for the surveyed specialties in each risk class (as defined by the St. Paul Companies) to other specialties in the same risk class. In concept, the total malpractice costs for a specialty were allocated among each of the CPT codes used by the specialty in proportion to the physician work values of each code. CMS s assumption was that malpractice risk for the code is related to the amount of physician work for the code. For CPT codes that do not have a physician work value, CMS simply kept the previous non-resource-based relative value. In practice, because multiple specialties use most codes, the allocation of malpractice costs was preceded by determining a blended premium for the specialties that used each code. CMS did this by calculating a risk factor for each specialty, where the risk factor for the specialty with the lowest malpractice premiums (psychiatry) was set at 1.00, and risk factors for the other specialties were calculated based on their premium rates relative to psychiatry. These risk factors were then used, in combination with the proportion each specialty accounted for use of each code, to allocate premiums to individual codes C.F.R ; Medicare Claims Processing Manual, CMS Pub , Ch. 12, See MLN/MLNProducts/downloads/How_to_MPFS_Booklet_ICN pdf at 8 (last accessed March 24, 2014); see also Barry D. Alexander, Medicare Part B, AHLA Whitepaper, available at Programs/Materials/Documents/MM09/alexander.pdf (last accessed March 24, 2014). 22 SSA 1848(c)(2)(C)(iii); 42 U.S.C. 1395w-4(c)(2)(C)(iii) Fed. Reg. 39,608, 39, (July 22, 1999); 64 Fed. Reg. 59,380, 59, (Nov. 2, 1999). 11

12 A budget neutrality adjustment was applied to the final results so that the total number of malpractice relative value units remained the same. Updates to the malpractice RVUs occur every five years. In order to update malpractice RVUs, CMS must first collect malpractice (i.e., professional liability) insurance premium data by specialty. 24 Those premiums are then linked to the physician work conducted by different specialties. Then, using those premiums, malpractice RVUs are based on the contribution of different physician specialties and surgical involvement to different Medicare procedures. Because malpractice costs vary by state and by specialty, the malpractice premium information must be weighted geographically and across specialties (d) Special Relative Values 4-2(d)(1) Radiology Services The Medicare statute had required CMS to establish a fee schedule for radiology services in 1989 prior to the generally applicable physician fee schedule, and CMS developed the fee schedule based largely on a relative value scale of the American College of Radiology (ACR). Although Hsiao s surveys included radiology procedures, the statute required CMS to continue to use the ACR s relative values with appropriate modifications to ensure that the relative values for radiology services were consistent with the other relative values. 26 In effect, the statute required CMS to maintain the value relationships among radiology services reflected in the ACR relative value scale, rather than use the relationships determined by Hsiao. In carrying out this requirement, CMS broke the ACR-derived relative values into components for work value, practice expense, and malpractice costs. Only the work values were carried over to the 1992 relative value scale, and those were resealed to fit the Hsiao values by comparing the ACR and Hsiao relative values for the sixty-five radiology procedures that Hsiao had surveyed. The ACR s numbers were higher than Hsiao s, and CMS divided the ACR s relative value by Hsiao s relative value for the each of the sixty-five procedures. The unweighted average of these ratios was then used to convert all of the ACR relative values for use in the 24 See Margaret O Brien-Strain, et al., Final Report on GPCI Malpractice RVUs for the CY 2010 Medicare Physician Fee Schedule Rule, March 2010, 25 Id. 26 SSA 1848(b)(2)(A); 42 U.S.C. 1395w-4(b)(2)(A). 12

13 1992 fee schedule. These relative values for radiology and imaging have been substantially modified beginning in (d)(2) Anesthesia Services Medicare payment for anesthesia services was historically based on a combination of relative values and locally determined conversion factors. As required by statute, Medicare adopted a uniform relative value guide in 1989 to standardize the relative values that the carriers used. The fee schedule statute required CMS to retain the 1989 relative value guide, adjusted to make it consistent with relative values for the other specialties, in the same manner as for the pre-existing radiology services relative value scale. 28 The relative value guide for anesthesia services consisted of variable base units depending on the particular procedure, plus additional units based on the amount of time taken for the particular procedure. Hsiao had surveyed nineteen anesthesiology procedures but derived a total value for the work involved that was not divided into base and time-sensitive portions. Thus, CMS s pre-existing relative values could not be directly compared to Hsiao s as in the case of radiology services. To compare the two systems, CMS determined an average payment under the Medicare system for each service (by combining average times and conversion factors) and an average payment under Hsiao s relative value for the same service (by calculating dollar amounts for the practice expense and malpractice components based on current payment amounts and scaling Hsiao s work values to be consistent with those components). 29 By comparing the putative average payments under both approaches for the nineteen surveyed procedures (using the unweighted mean, as with the radiology scale), CMS determined the factor by which to reduce the pre-existing relative values to be consistent with Hsiao s relative values for, other specialties, known as a base unit. Base unit values have been assigned to each anesthesia procedure code and reflect the level of difficulty of the anesthesia services. The allowance is.based on time units multiplied by base units multiplied by an anesthesia conversion factor. 27 See Section 4.5(e), supra regarding the present status of radiology reimbursement levels. 28 SSA 1848(b)(2)(B); 42 U.S.C. 1395w-4(b)(2)(B) Fed. Reg. 59,562 (Nov. 25, 1991). 13

14 Medicare continues to use the structure of a base unit reflecting the type of surgical procedure and time units for each fifteen minutes of anesthesia time (d)(3) Separate Professional and Technical Components The statute sets forth certain services as benefits independent of the physician services benefit even though physicians themselves often furnish the services, including X rays, other diagnostic tests, and radiation therapy. 31 This separate listing allows entities other than physicians to furnish and bill for the services. When physicians furnish the service, however, the service may include a professional aspect, such as interpretation of the test results. The possibility that a service could be divided among two providers required CMS to establish relative values for the technical component (i.e., the nonphysician work) and for the professional component (i.e., the physician s service). The relative value of the technical component consists only of the practice expense and malpractice components. 4-2(e) Periodic Adjustment Historically, the statute required CMS, in consultation with the Medicare Payment Advisory Commission and physician organizations, to review the relative values at least every five years and adjust them based on changes in medical practice, new data, new procedures, and similar factors. 32 Relative values for selected codes have been re-evaluated in this process, with CMS relying on recommendations from the American Medical Association s Relative Value Update Committee in the same manner as for new codes. There are two methods by which practice expense relative values are updated. First, physician specialties can sponsor supplemental surveys of physician expenses. CMS uses the results from these surveys, after adjustment to make them comparable to the older surveys, instead of the results from the AMA surveys to determine the practice costs of the specialty surveyed. 33 Second, CMS periodically re-estimates the data estimates representing the expenses and supplies used in furnishing specific services. Such re-estimations have occurred on an annual basis for work and practice expense RVUs since C.F.R ; Medicare Claims Processing Manual, CMS Pub , Ch SSA 1861(s)(3), (4); 42 U.S.C. 1395x(s)(3) and (4). 32 SSA 1848(c)(2)(B); 42 U.S.C. 1395w-4(c)(2)(B). 33 SSA 1886(d); 42 U.S.C. 1395w-4(b)(1)(B); 42 C.F.R (b)(6). 14

15 4-2(1) Budget Neutrality Under the statute, any adjustments in relative values that are made for a year cannot increase or reduce aggregate spending by more than $20 million. 34 This limitation requires the reduction of payments for existing procedures when new codes are added if the reduction is necessary to maintain aggregate spending within the $20 million limit. How this reduction is allocated between existing procedures can be very controversial. Until 1996, CMS implemented this requirement by reducing the relative values across the board to offset new and increased relative values. Because a number of non-medicare payors use the Medicare relative value scale, this practice had effects beyond Medicare. As a result, CMS changed its policy and began making budget neutrality adjustments by revising the conversion factor. The budget neutrality adjustment made after the five-year review of physician work values in 1997, however, was applied only to the work RVUs, on the ground that adjusting the conversion factor would have in effect applied the adjustment across all components of the relative values instead of just the work RVUs Geographic Adjustment Factors One of the primary objectives of the fee schedule was to reduce the wide geographic variations in Medicare payments. Under the reasonable charge payment system, it was not unusual for physicians in some localities typically large cities to receive Medicare payments that were multiples of the payments in rural areas. The geographic adjustment factors in the fee schedule are intended to vary Medicare payment geographically only insofar as is justified by local cost differences. This aspect of the fee schedule has had the effect of reducing the range of variation in Medicare payments, thus sharply lowering Medicare payments in some large cities and increasing payments in rural areas. 4-3(a) Geographic Practice Cost Indices (GPCI) Payment rates are adjusted for geographic variations based on indices, i.e., numerical values that compare a particular locality with the national average where the national average is set at For example, a locality with costs that are 15% higher than the national average would have an index value of CMS is required to develop three indexes to reflect 34 SSA 1848(c)(2)(B)(ii)(II); 42 U.S.C. 1395w-4(c)(2)(B)(ii)(II). 35 See 61 Fed. Reg. 59,490, 59, (Nov. 22, 1996). 15

16 geographic variations in each of the three components of the relative values practice expense; malpractice costs, and physician work. 36 The statute uses varying names for the indices, but CMS refers to each of them as a geographic practice cost index (GPCI, pronounced gypsy ), a term that was used prior to enactment of the fee schedule law when the geographic adjusters were already under development for other purposes. The practice expenses GPCI is built from four underlying elements: (1) index of the relative costs for employee wages; (2) relative cost of office rents; (3) measure of costs (equal to 1.0) for equipment and supplies; and (4) cost share weights to combine these elements into the practice expense GPCI. 37 Under CMS s methodology, the GPCI for each of the three components of the relative values is multiplied by that component in calculating the payment amount. This method of applying the geographic adjustment is different from, although algebraically the same as, the calculation method set forth in the statute. The statute contemplates that a unique geographic adjustment factor will be generated for each service by creating an average of the three GPCIs, weighted by the number of relative value units in each component, and that this weighted average geographic adjustment factor will then be multiplied by the total relative value for a particular service. CMS is required to update the GPCIs every three years and to phase in any changes over two years. The Patient Protection and Affordable Care Act (Affordable Care Act) also included several new GPCI provisions. 38 In 2003, Congress established that for three years there would be a floor of 1.0 on the work component of practice expenses. This floor has been extended by legislation since then. The latest extension goes, through March 31, Though CMS updated the GPCIs using more recent data for CY 2014, the CY 2014 final rule make no major changes in the GPCI methods or payment locality definitions. 4-3(a)(1) Practice Expense The starting point for the practice expense GPCI is AMA survey data that provides a breakdown, by percentage of physician revenue, of the three categories of physician overhead costs (other than malpractice) employee wages, office rents, and equipment and other costs. 36 SSA 1848(e); 42 U.S.C. 1395w-4(e). 37 SSA 1848(e)(i); 42 U.S.C. 1395w-4(e). 38 Pub. L. No as amended by Pub. L. No , and collectively referred to as the Affordable Care Act (2010). 39 Pathway to SGR Reform Act of 2013, P.L (Dec. 26, 2013). 16

17 The GPCI for total practice expenses was formed by determining the geographic variation for each of these three categories and combining them according to the relative weights indicated by the AMA data. Data to serve as a proxy for these costs were identified for each category. For employee wages, the GPCI uses a special tabulation of the most recent census data and survey data from the Bureau of Labor Statistics census data on salaries of individuals in ten occupational categories employed by physicians. For office rents, the GPCI uses a data series of residential rents produced annually by the Department of Housing and Urban Development because adequate public data on commercial rents are not available. For medical equipment, supplies, and other expenses, CMS concluded that these expenses did not vary geographically, so no adjustment is made. Because the statute establishes a fee schedule for each Medicare locality, the GPCIs vary by locality (see 4-3(c)). The data sources used to construct the GPCIs do not, however, contain data on a Medicare-locality basis. The GPCIs were therefore constructed on the basis of metropolitan statistical areas (MSAs). If more than one locality fell in the same MSA, they were all assigned the same GPCI values. If a locality crossed MSA lines, the values for each MSA were assigned to the locality in proportion to the total number of relative value units billed to Medicare in each locality in the MSA. Although the GPCI uses proxy data that is years old as well as data that does not precisely represent the expenses to be measured, that does not necessarily mean that the indices are flawed. Because the geographic adjustment factor is an indication of relative differences, old data is still valid if relative costs remain the same among geographic areas even though actual costs have increased. The Affordable Care Act limited recognition of locality differences in comparison to the national average in the rent and employee compensation components of the practice expense component of the GPCIs. In addition, the Affordable Care Act made permanent the 1.5 GPCI floor for Alaska and created a permanent 1.0 floor for practice expense GPCI in frontier states (Nevada, South Dakota, Idaho, Wyoming and Montana). The statute authorizes CMS to establish different practice expense GPCIs for classes of physician services if a single GPCI would be substantially inequitable because the mix of goods and services varies from class to class. CMS has, however, created only a single practice expense GPCI. 17

18 4-3(a)(2) Malpractice Costs The malpractice cost GPCI is determined in a similar way. The GPCI uses a three-year average of premiums in each geographic area for a claims made insurance policy affording a particular level of coverage ($1 million/$3 million). 40 Because premiums vary by physician specialty, the GPCI uses an average weighted by the share of Medicare spending accounted for by each risk class. 4-3(a)(3) Work A geographic adjustment for the work component is required but, unlike the adjustments for the other components, was controversial. Although the statute requires an adjustment for geographic differences in the relative value of physicians work effort, the adjustment is sometimes referred to as a cost-of-living factor. 41 Advocates of this adjustment argued that physicians should be compensated at the same real rate regardless of location. Under this approach, physicians practicing in areas where incomes and living costs are high should receive greater Medicare payments than those in lower cost areas so that their standards of living can be the same. Opponents of this adjustment argued that physicians living in high cost areas, typically cities, enjoy the offsetting amenities of urban living and that rural physicians should not receive lower payments simply because the lack of amenities in rural areas is reflected in lower incomes, real estate prices, and other costs. The statute compromised this issue by providing for only one-quarter of the geographic variation to be recognized. For example, if an area is determined to have costs that are 80% of the national average, the index value for the area would be 0.95 one-fourth of the difference between the indicated 0.8 and the average 1.0. CMS s methodology for determining the index is consistent with its statutory description as a measure of the relative value of physicians work efforts, rather than a direct measure of differences in the cost of living. The premise of the index is that areas where professionals generally receive higher incomes can be said to value physician work more highly than areas where professionals receive lower incomes. The index was therefore constructed using census data on the incomes of nonphysician professionals. In 2011, CMS began basing its physician 40 SSA 1848(e); 42 U.S.C. 1395w-4(e). 41 SSA 1848(e); 42 U.S.C. 1395w-4(e). 18

19 work GPCI updates on the median hourly earnings from the 2006 through 2008 Bureau of Labor Statistics (BLS) Occupational Employment Statistics (OES) wage data. For calendar year 2014, CMS used updated BLS OES data (2009 through 2011) as a replacement for the 2006 through 2008 data to compute work GPCIs. 42 During the period , the law set a floor of 1.0 for the physician work geographic adjustment, thus favoring rural and other low-cost areas. 43 The practice expense of 1.0 is equivalent to the national average. Beginning in 2009, Congress set a permanent 1.5 work GPCI floor for services furnished in Alaska. 44 The Act also established a permanent floor in frontier states (Montana, Wyoming, North Dakota, South Dakota, and Nevada) that results in a GPCI of less than 1.0, so these states receive higher payments. 45 Thus, for these states, the same GPCIs apply for calendar year (b) Revision of the Geographic Adjustors CMS is required to update the geographic adjustment factors at least every three years, 47 and, accordingly, CMS revised the GPCIs six times. In 2010, the Affordable Care Act required CMS to evaluate data that fairly and reliably establishes distinctions in the cost of operating a medical practice. If adjustments are less frequent than annual, the statute creates a two-year transition for any revisions by requiring that the first year s change to each index be only one-half of the adjustment actually indicated. CMS continues an examination of the costs and appropriate weight to give each category during the calendar year 2014 rulemaking process. Additionally, CMS will review the complete findings and recommendations from the Institute of Medicine s study of geographic adjustment factors for physician payment, HHS Medicare Geographic Payment Summit, and the Medicare Economic Index (MEI) technical advisory panel, and will continue to study the issues. 48 The MEI assesses the relevance and accuracy of the inputs, input weights. 4-3(c) Localities Fed. Reg (Dec. 10, 2013). 43 SSA 1848(e); 42 U.S.C. 1395w-4(e)(1)(I). 44 SSA 1848(e)(1)(G); 42 U.S.C. 1395w-4(e) )(1)(G). 45 SSA 1848(e)(I); 42 U.S.C. 1395w-4(e)(1) Fed. Reg (Dec. 10, 2013). 47 SSA 1848(e)(1)(C); 42 U.S.C. 1395w-4(e)(1)(C). 48 SSA 1848(e)(1)(H)(iv) (as added by section 3102(b) of the Affordable Care Act). 19

20 The statute provides that the required geographic variations in payments under the fee schedule are to be based on Medicare localities, which were the areas established by the Medicare carriers for the purpose of calculating prevailing charges under the predecessor reasonable charge methodology. 49 As CMS interprets this requirement, it does not prevent the agency from altering locality boundaries, and CMS has made extensive revisions in the localities since introduction of the fee schedule. There is a trade-off between, on the one hand, simplicity in the form of fewer fee schedule areas and, on the other hand, accurate matching of physician costs to physician payments. The compromise adopted by CMS is to use statewide localities except for areas within a state that has a geographic adjustment factor that is more than 5% greater than the geographic adjustment factor for the remainder of the state. 50 When this compromise was adopted for 1997, it reduced the number of localities from 210 to 89 and increased the number of statewide localities from twenty-two to thirty-four. The fee schedule areas are identified in an addendum to the annual fee schedule notice in the Federal Register. 4-4 Conversion Factor The last factor in the payment computation, in addition to the relative values and the geographic adjustments, is the conversion factor. Although the first two factors determine the distribution of Medicare payments among the various physician services, it is the conversion factor that determines the amount of money that Medicare will pay. Adjustment of the conversion factor is thus the primary instrument by which Medicare s aggregate expenditures for physician services can be controlled. 4-4(a) Initial Conversion Factor For the first year, 1992, the conversion factor was required by statute to be set at a budget neutral amount. 51 In this calculation, CMS was required to estimate the conversion factor that would have resulted in the same total payments that were made in 1991 (under the generally charge-based system) had the fee schedule been in effect in That estimated conversion factor was then increased by an update factor to become the conversion factor for SSA 1842(b); 42 U.S.C. 1395w-4(j)(2); 42 U.S.C. 1395(u)(b). 50 See 61 Fed. Reg. 34,614, 34, (July 2, 1996); 61 Fed. Reg. 59,490, 59, (Nov. 22, 1996). 51 SSA 1848(d)(1)(B); 42 U.S.C. 1395w-4(d)(1)(B). 20

21 The principal policy issue involved CMS s baseline adjustment to account for expected changes in physician behavior as a consequence of payment reductions under the fee schedule. In accordance with its historical practice in estimating Medicare physician expenditures, CMS assumed that physicians experiencing payment reductions would increase the services they perform such that they would make up 50% of their lost income. (No behavioral response was projected for physicians gaining income under the fee schedule.) To offset the effects of this projected behavior, the conversion factor was reduced by 10.5%. 4-4(b) Annual Updates to the Conversion Factor Each year the conversion factor is adjusted based on two factors (1) inflation in the costs incurred by physician offices, as represented by the Medicare Economic Index, and (2) the extent to which aggregate Medicare expenditures for physician services have or have not stayed below a target growth rate. 4-4(b)(1) Medicare Economic Index The Medicare Economic Index (MEI) is a measure of inflation in physician fees and practice costs. Created in 1975 as a way to estimate annual changes in physicians operating costs and earning levels, the MEI was intended to serve as a cap on physician fee increases. 52 However, since the creation of the physician fee schedule in 1992, the MEI has been used to determine allowable charges under the reasonable charge system. Although the statute does not describe the construction of the MEI, CMS followed comments in the legislative history and created an index with two broad components: (1) operating expenses of physicians, and (2) general earnings levels. 53 Based on an American Medical Association survey of a national sample of physicians, CMS concluded that operating expenses constitute 45.5% of a physician s gross income, while net income constitutes the other 54.5%. The inflation factor used for the net income portion is the Bureau of Labor Statistics index of hourly earnings of professional and technical workers. The office expense portion of the 52 David O. Barbe, Improving the Medicare Economic Index, Report of the Council on Medical Service, CMS Report 6-I-08, available at (last accessed March 25, 2014) C.F.R (d); construction of the MEI is described at 63 Fed. Reg. 58,814, 58, (Nov. 2, 1998) and 50 Fed. Reg. 39,941 (Sep. 30, 1985). 21