Search Frequenty Asked Questions

Executive Summary

Background

Methodology

Exhibits

Executive Summary

E.1. Change in Methodology Between the Resource-Based and Charge-Based MP RVUs

KPMG LLP ("KPMG") has contracted with HCFA to develop resource based malpractice relative value units ("RVU") subject to guidelines specified by HCFA in the initial request for proposal. These guidelines have been augmented by ongoing discussions with HCFA, including their review of an earlier preliminary draft report. This final report reflects the changes to that document, both in approach and presentation, requested by HCFA. The objective in developing a resource-based malpractice RVU ("RBMRVU") is to allocate actual malpractice insurance premium costs to specific procedures. The methodologies that have been selected to calculate the RBMRVUs allocate the malpractice costs of each physician specialty across the procedures that the specialty performs, according to the frequency the specialty performs it. In other words, each procedure is assigned a RBMRVU in proportion to the average costs of all the physician specialties that perform the procedure.

The previous methodology was based on the product of "base allowed charges and service specific malpractice expense percentages. The new resource based methodology for malpractice costs effectively ends the resource based relative value scale's ("RBRVS") dependence on charge based data. In lieu of allocating malpractice cost compensation across allowed charges, we have supplied two alternative allocation methods. One method allocates RVUs by averaging the malpractice costs of the specialties that perform the procedure. The other allocates the malpractice RVUs by using the averages developed in the first method multiplied by the procedure's work RVUs, accounting for the procedure specific "Risk of Service" ("ROS"). Certain adjustments to the methodology have been applied based on specific knowledge of malpractice insurer rating practices. For example, obstetrical procedures and spinal surgeries have been increased. Obstetrical procedures have been increased to a level appropriate for Obstetricians rather than the lower costs of a broader OB/GYN grouping. Spinal surgeries have been increased because it is not uncommon for Orthopedic Surgeons to be charged a higher premium if they perform these procedures. Since malpractice insurers typically rate physicians that perform surgeries higher than physicians that focus on evaluation and maintenance, we have applied a different set of expense allowances for surgical procedures than we have for non-surgical procedures. In spite of these changes to the calculation of malpractice RVUs, the new resource based methodology has a few similarities with the previous methodology. Specifically,

• Much of the statutory formula used to calculate the percentage of total charges has been maintained in the resource-based formula. Service specific malpractice expense RVUs are still based on a weighted average of the malpractice expense of the specialties performing the service. (But note the previously mentioned exceptions).
• The new RBMRVUs are calculated so that the same level of aggregate malpractice RVUs will be generated as were under the previous methodology.
• Codes, which had malpractice RVUs under the previous methodology, but which have no Work RVUs have been left unchanged in the ROS adjusted methodology. The majority of these codes have a technical modifier.

E.2. Impact Analysis

Impact on Physician Specialty

The exhibit on pages 6-7 illustrates the impact that the proposed resource-based methodology will have on various physician specialties. All figures are expressed as a percentage of total RVUs including the impact percentages. Please note that the impact to Anesthesiologists as presented in this exhibit shows the change in the portion of their income attributable to RVUs; the majority of their income will be earned through anesthesia codes, which are not being changed.

The Risk Factor methodology produces the largest change in physician specialty income. The biggest gains are for Allergist/Immunologists (+7.4%), Occupational Therapists (+6.2%), and Physical Therapists (+5.9%). The common feature of these specialties is that they perform many low intensity services. The biggest losses are for Cardiac Surgeons (-6.0%), Thoracic Surgeons (-5.5%), and Neurosurgeons (-4.8%). These surgical specialties perform fewer procedures but higher intensity services. This bias in the standard Risk Factor methodology is adjusted for in the Risk of Service adjusted methodology, which multiplies the risk factors by the procedure's work RVUs.

The Risk of Service Adjusted methodology produces more modest changes in physician specialty income. The biggest gains are for Emergency Medicine (+2.4%) and Nephrologists (+1.2%). The increase in Emergency Medicine may be due to the inclusion of Emergency Medicine malpractice data for the first time. The biggest losses are for Allergists and Immunologists (-1.1%).

Impact by Type of Procedure

The exhibit on page 10 shows the impact that the proposed resource-based methodology will have on different types of procedures. All figures are again expressed as a percentage of total RVUs including the impact percentages.

The unadjusted Risk Factor methodology produces a larger change in income than the Risk of Service Adjusted methodology. The biggest gains are for Musculoskeletal Surgery (+3.3%) and HCPCS/Other (+2.6%). These codes contain many individual codes that are high frequency but low intensity. The biggest losses are for Cardiac Surgeries (-6.5%) and Laparoscopy (-5.7%). These surgical procedures are often high intensity. This methodology's bias which generally increases the RVUs for low intensity procedures and decreases RVUs for high intensity procedures relative to the historical malpractice RVUs is even more apparent in this exhibit than the previous exhibit showing Impact by Specialty.

The Risk of Service Adjusted methodology produces more modest changes in income compared to the unadjusted Risk Factor methodology. The largest increases are for maternity (+2.3%) and Musculoskeletal surgeries (+1.5%). Special adjustments have been made to maternity codes to account for their higher risk. The largest reductions are for gynecological procedures ("Surgical, female genital") (-2.3%) and surgeries to the Arms & Legs (-1.5%). Gynecological procedures were subject to a decrease due to the use of a Gynecology malpractice risk factor as opposed to a combined Obstetrical/Gynecological malpractice percentage.

Background

B.1. Background of the Malpractice RVU

Until 1992, under Medicare Part B, the reasonable charge for a physician’s service was the lowest of (1) the physician’s actual charge, (2) the physician’s customary charge, or (3) the prevailing charge in the locality for similar services. The customary charge is defined as the median physician charge for the service during the preceding July through June data collection period. These charges are arrayed in ascending order and the median, or midpoint, of the charge data is selected as the customary charge. The prevailing charge for a particular service in a locality is traditionally defined as the 75^th percentile of physician charges in the locality for that service. Since 1975, however, changes in prevailing charge limits from year to year have been constrained by statute to the amount of inflation in medical costs as measured by the Medical Economic Index ("MEI").

A major change in the Medicare physician payment rules was enacted as part of the Omnibus Budget Reconciliation Act (OBRA) of 1989 (Public Law 101-239) on December 19, 1989. Section 6102 of Public Law 101-239 amended title XVII of OBRA by adding a new section 1848, "Payment for Physicians’ Services". The new section contained three major elements: (1) establishment of volume performance standard rates of increase for physician services expenditures; (2) replacement of the reasonable charge payment mechanism with a fee schedule for physician services; and (3) replacement of the maximum actual allowable charge ("MAAC"), which constrains the total amounts that nonparticipating physicians can charge Medicare beneficiaries for covered services, with a new limiting charge.

Section 1848 requires that the fee schedule include national uniform relative values for all physicians’ practice expenses net of malpractice expenses and the cost of professional liability insurance (malpractice insurance). Nationally uniform relative values must be adjusted for each locality by a geographic adjustment factor ("GAF"). The new fee schedule was phased in over 4 years, beginning in 1992, with the new rules fully effective in 1996. From 1992 through 1995, transition provisions generally blended the old payment amounts with the new.

The relative value unit is primarily based on the resources needed to furnish the service. It is for this reason that the fee schedule is often called a resource based relative value scale, or RBRVS. Each physician fee schedule comprises three components: (1) work RVUs, (2) practice expense RVUs, and (3) malpractice RVUs. The work component reflects the relative value of the doctor’s work for a particular procedure, in terms of time and intensity of effort, and compares it to the value of work required for other procedures. The practice expense component reflects the doctor’s overhead. The malpractice insurance component attributes a portion of a doctor’s malpractice insurance premium to each procedure.

Work RVUs have been resource-based since the inception of the fee schedule in 1992. Resource based practice expense RVUs are currently being phased in over a four-year period. The malpractice RVU is the last of the three components to be converted to a resource based relative value unit.

In most cases, the current malpractice expense RVUs are calculated based on a statutory formula. They are derived from the product of "base allowed charges" and service specific malpractice expense percentages. The base allowed charge is the national allowed charge for the service furnished during 1991. The service specific malpractice expense is a weighted average of the malpractice expense percentages of the specialties performing the service.

A common criticism of the current malpractice RVU system is that for many services the RVUs, which are based on charges under the reasonable charge system, are not based directly on the resources involved in furnishing the service. Rather the charge-based nature of the current fee schedule malpractice expense retains historical charge patterns that existed before the implementation of the physician fee schedule on January 1, 1992. For example, those charge patterns favor procedures and tests performed in hospitals rather than evaluation and management services and other office based services, whereas the current trend in reimbursement philosophy appears to encourage performing services outside of a hospital setting.

HCFA has asked KPMG to address this criticism by allocating malpractice premium RVUs by existing total RVUs by CPT code. These malpractice premiums will be weighted by provider specialty within CPT code. This methodology will result in a malpractice RVU, which is more resource based than the current malpractice RVU formula.

B.2. Malpractice Insurance Premium Rate Survey

The fundamental resource driving the malpractice RVU is the actual malpractice premium charged to the physician. Allocating these amounts to physician procedures would finally align malpractice RVU revenue with the costs that are intended to be offset. To put in place any such allocation system, it is necessary to collect information regarding malpractice premiums. HCFA has previously contracted with Allied Technology Group Inc. ("Allied") to collect malpractice premium data for the updating of the GPCI and has requested that KPMG apply this available information.

KPMG relied on the 1995 premium data collected through Allied’s survey. Allied used an OMB-approved survey to collect physician specialty sub-state region specific malpractice data. The questionnaire solicits information on the following:

• availability of physician liability policy premiums;
• specific premiums;
• limits of liability;
• mature liability premium rates;
• geographic adjustment to premiums;
• risk classifications by specialty;
• alternative insurance company contacts;
• the existence of a patient compensation fund or joint underwriting association; and
• insurance company market share.

Overall Allied collected physician malpractice insurance premium data on a total of 55 companies across all 50 states, Puerto Rico, and Washington, D.C. Insurance companies that operated in more than one state are counted once in this total. Market share information on the top state insurance companies (i.e. >50% of the state’s physicians) was solicited as part of the survey process. In 36 states Allied collected data from at least two companies. In the remainder of the states, Allied received information from one company, but in each of these cases the insurers represented 60% or greater of the market share in that state. On average the data collected from insurance companies represented approximately 77% of the state’s insured physicians.

Although insurance companies normally set premiums according to risk class, the physician specialties included within a particular risk class are not consistent across insurance companies, nor are the numbers of risk classes consistent. Thus risk classes are defined differently across states and do not always contain the same group of physician specialties. Given this inconsistency in the data it is not possible to compare insurance premium data across companies by risk class. To eliminate this problem and to ensure data comparability and compatibility across time (i.e., with past physician malpractice insurance premium data collection efforts), Allied collected premium data based on the 20 physician specialty groups used in previous data collection efforts. Malpractice premium data collected consists of the top 20 physician specialties whose ranking reflect Medicare Part B spending in 1990. In practice there are more than 90recognized specialties. This is an acknowledged limitation of the Allied data.

In many instances, due to insurance company variation in the specialties included in their risk classes, Allied chose a specialty premium among several choices within that particular specialty. The premiums for the specialty primarily varied according to whether or not the physician performed surgery and to what degree. For these cases, Allied chose the specialty (and corresponding premium) by following decision rules for categorizing a physician specialty. Allied developed a set of general rules for categorizing an insurance company’s physician specialty nomenclature into one of the 20 physician specialty groups, as well as rules for calculating premiums. This resulted in a general physician specialty grouping consistent across malpractice insurance companies and with comparable data categories.

The decision rules Allied used for categorizing a physician specialty are as follows:

1. Choosing the correct specialty if there is a discrepancy:

2. If only one specialty is listed, but is not exactly the specialty (by definition) listed in the top 20 Medicare specialties, choose it anyway.

For the majority of the data Allied collected, company specific premium rates for a specialty applied statewide. However, for fifteen states, premiums varied by sub-state regions. Each insurance company has uniquely defined sub-state regions that are used in its rating calculation. These regions were not necessarily uniform across companies within the same state.

The Allied methodology was modified to expand the list of specialty categories to include an additional 20 specialties. Including a larger number of specialties was particularly desirable as a refinement to the RBMRVU estimates due to the high degree of variation among insurance company classifications.

Eight states have legally required Patient Compensation Funds ("PCFs") funded through a surcharge on physician’s malpractice premium. The PCF pays for any claim amount beyond the statutory amount, thereby limiting an individual physician’s liability in cases of a large suit. The PCF data reflects a physician malpractice premium that is equal to the premium plus the patient compensation fund surcharge.

In three of the eight states participation in the PCF is mandatory and in the remainder participation is voluntary. The premium rates provided by each of the insurance companies in the PCF states were adjusted by the corresponding PCF surcharge to obtain the true costs of malpractice insurance coverage in these states. PCF factors were only applied to premium rates in those states that mandate a PCF contribution.

Joint Underwriting Associations ("JUAs") were created by some legislatures during the withdrawal from the market (thereby decreasing availability) of many commercial carriers during the 1970s. JUAs are typically state government administered risk pooling insurance arrangements that provide professional liability insurance to health care providers. In some states, JUAs also maintain additional premium contingency assessments, retroactively assessing policyholders in the face of deficits. Historically, JUAs carry the bulk of the physician malpractice insurance market share in a state or the JUA potentially is the insurer of last resort. JUA premiums can be lower than other state insurance company premiums since a JUA can pass the surcharge to purchasers of other lines of insurance. However, there are two circumstances when JUA premiums can be higher than commercial insurers. First, JUAs are heavily regulated and may not have the flexibility to move with the market forces. Second, because the JUA must accept all applicants regardless of litigation history, the JUA may have to charge higher premiums.

Allied was able to obtain information from the twelve states known to have operating JUAs. None of the four states in which physicians are insured primarily through JUAs (Massachusetts, New Hampshire, Rhode Island, and South Carolina) had assessed retroactive premium adjustments. Only one JUA had assessed their policyholders a retroactive premium adjustment. This JUA, Minnesota Medical Malpractice, assessed $1.2 million in 1993. No assessment was levied in 1994 or 1995. However, the market share of this JUA in Minnesota is <1% and, therefore, is not considered a top insurer for this state. Allied did not include this assessment in their data.

The rules Allied applied for adjusting benefit levels, PCFs, JUAs, and other coding/data entry are as follows:

1. If a company/state provides premium rates as occurrence rates and provides claims-made rates conversion factors, use these factors to convert. If no factors are provided, use St. Paul’s conversion factors.

2. If a company/state provides limits of liability other than $1M/$3M, and they provide adjustment factors, use these factors to convert. If no factors are provided use St. Paul’s adjustment factors.

3. In the case of PCF surcharges, if the sum of the surcharge and the required basic limitations do not equal $1M/$3M limits of liability, then use St. Paul’s conversion factor to reach $1M/$3M limits of liability.

Methodology

M.1. Overview

The underlying task is to allocate the costs of physician malpractice insurance to a resource based relative value scale ("RBRVS"), which compensates physicians based on the procedures they perform. The difficulty with this task is that malpractice insurers do not charge physicians per procedure performed; physicians are typically charged premiums for all services performed over a given period of time, typically a year or more. These premiums are based on certain risk characteristics such as specialty, surgical involvement, and experience. The solution to this difficulty is to develop malpractice relative value units ("RVU"s) that will be based on an assessment of the relative risk of the procedure. This assessment will be made using two methods. The first method relies solely on a weighted average of the malpractice premiums of the specialties that perform the procedure. The second method uses the analysis of the first method but adds work RVUs as an additional measure of risk. Work RVUs are believed to represent a reasonable proxy for risk since they were developed to reflect the time, intensity and level of experience needed to perform the procedure.

This task has the following obstacles:

M.2. Mapping of Medicare Specialties to ISO Codes

Medicare uses specialty codes that are distinct from the ISO codes used by malpractice insurers to identify specialists. ISO codes tend to be more detailed than HCFA codes, which already identify almost 90 distinct physician specialties. In particular, ISO codes make distinctions according to the level of surgical involvement of the practitioner. In order to perform our calculations, it was necessary to map the Medicare specialty codes to their nearest ISO codes.

For the top twenty specialties chosen by Allied Technology, their already established mapping was applied. An additional twenty Medicare specialties were identified that could also be cleanly mapped to ISO codes. There are over 10,000 procedure codes in RBRVS and many of them are unique to particular specialties, some of which were not included in the Allied Technology report. By including the additional twenty specialties, the number of procedure codes for which malpractice RVUs could be calculated was expanded. The 40 specialties used in this mapping procedure encompass approximately 81% of the total procedures included in the physician profile dataset. In comparison to the Allied Medicare specialty mapping, the expanded Medicare specialty mapping represents a 12.8% increase in total procedures mapped to a physician specialty.

Exhibit 2 shows the mapping or cross-walk of the Medicare specialty codes to their closest ISO codes and the risk class assigned to the ISO codes by the St. Paul rate manual.

M.3. Calculating the National Average Specialist Premiums

As stated in the Background section of this report, Allied Technology collected the malpractice insurance data underlying our analysis. Since a detailed explanation of the Allied Technology study is provided in the Background section, it is not necessary to provide an elaborate description here.

The resource based malpractice relative value units ("RBMRVUs") will be uniform across the entire nation. (It is the function of the GPCIs to account for geographical cost differences.) Because of this, risk factors must be based on national average data. For example, the risk factors used for the twenty specialties tracked by Allied Technology were based on national average premiums for each physician specialty. These national averages were developed from the underlying county-level data.

It is important to understand that more than one choice of the weighting used to calculate the national average premiums was available. For example, any of the following could serve as a weight for determining the national average:

• General population statistics
• Number of physicians
• Medical care consumption (dollars)
• Senior population or Medicare covered lives
• Senior physician medical care consumption (Medicare RVUs)

In spite of this, the decision was straightforward, based on precedent. In fact, HCFA has both established a precedent and expressed a preference for using total RVUs as a basis for weighting county-level malpractice premiums.

HCFA supplied 1997 RVU data organized by zip code and Medicare specialty code in order to tabulate the national average malpractice premiums for the twenty specialties tracked in the Allied Technology study. (See section M2 for elaboration on this mapping process). Because the Allied Technology data was county level information and the HCFA RVU data was zip code level, it was necessary to map the zip codes into counties. Average premiums were further adjusted by dividing the premiums by the malpractice GPCI for the area, which effectively "normalizes" the data for geographic differences. Since it is the function of the GPCI and not the RVUs to account for geographical cost differences, this normalization was deemed necessary. After mapping the RVU data to convert Medicare specialty code to ISO code and zip code to county, the national averages were calculated. In equation form, the calculation was performed as follows:

(Σ SPEC_TRVUs,COUNTY X PREM_SPEC,COUNTY ÷ MP GPCI_COUNTY ) ÷ SPEC_TRVUs = PREM_SPEC

COUNTY = Abbeville, SC to Zieback, SD

Where,

TRVUs = Total 1997 RVUs (work, practice expense and malpractice)

Exhibit 1 shows the national average premium calculated for each specialty from 1990 through 1995.

M.4. Specialist Risk Factors

Risk Factor vs. Risk of Service Adjusted Methods

KPMG was asked to present two sets of resource-based malpractice RVUs. The first set is based solely on specialty risk factors and the second includes an adjustment based on the existing resource-based Work RVUs. This risk-of-service ("ROS") adjustment could have taken a number of forms including an analysis of malpractice insurance claims by procedure. This was decided against because inquiries to malpractice insurance experts revealed that this data would not be available in sufficient quantity and breadth to be useful. Following discussions with HCFA, a decision was made that the most reasonable available proxy for a risk-of-service adjuster was the work RVU. The work RVU had the advantage that it did not require additional data collection, it is already resource-based and one of the primary determinants of the work RVU is the relative difficulty or intensity needed for the procedure.

From a risk/reward perspective, the unadjusted Risk Factor approach will favor the physician that performs many low-intensity procedures. If the number of patients treated is the best measure of risk, this may be the desired alternative. The ROS method rewards the physician who generates the highest quantity of work RVUs. This method will be more favorable for surgeons than the Risk Factor method. The impact analysis (section E.2.) reveals that both methods deduct from surgical codes; however, the ROS method produces a smaller impact (3.1% vs 0.6% of total RVUs). Because work RVUs are closely tied to time and intensity of services, this would be the preferred approach, under the assumption that these factors are well correlated with malpractice risk.

Surgical Codes vs. Non Surgical Codes

Malpractice insurers often categorize physicians not only by their specialty but their level of surgical involvement. A given specialty may have several ISO codes depending upon whether the physician performs minor surgeries, major surgeries, special surgeries such as Orthopedists that also perform spinal surgery, or no surgeries at all. Because the level of surgical involvement can dramatically change the physician's premium, it was decided to develop separate risk factors for surgical procedures as opposed to non-surgical procedures. The specialist risk factor for surgery was calculated from the ISO code involving major surgical involvement; similarly, the ISO code for no surgical involvement was used to calculate the risk factor for E&M codes. We applied a special adjustment to spinal surgery codes (22100-22899) because some malpractice insurers will rate Orthopedic Surgeons based on whether or not they perform these procedures. These codes were assigned a risk factor that is consistent with a risk factor for an Orthopedic Surgeon who performs spinal surgeries.

Obstetrical Codes

Medicare specialty codes do not distinguish between Obstetricians and Gynecologists. However, malpractice premiums can be quite different for these specialties. Because the majority of Medicare OB/GYN services will be Gynecological, a Gynecology risk factor was used for all OB/GYN procedures, except those that are clearly Obstetrical, such as maternity codes (59000-59899). These codes were assigned a risk factor that is consistent with a risk factor for an Obstetrician who performs surgery.

HCFA Cross Walk

In addition to the 40 specialties that were mapped to an ISO code, 34 additional Medicare defined specialties for which ISO codes could not be identified, were cross-walked to a specialty for which a risk factor and ISO code had been assigned. The cross walk assumptions can be found in Exhibit 3.

Risk Factor Assignment Algorithm

Having discussed the background and issues surrounding the specialist risk factors, the actual mechanics of the algorithm used to assign risk factors will now be described. The assignment algorithm was as follows:

1. The lowest malpractice premium specialty was determined to be Psychiatry (ISO code 80249). This specialty was arbitrarily assigned a risk factor value of "1.00".
2. If the specialist, defined by ISO code, was tracked in the Allied Technology report, the 1993-95 average premium was divided by the average Psychiatrists' premium.

3. If the ISO code was not tracked in the Allied Technology report and an ISO code could be identified, then the appropriate ISO code was assigned to the specialty.
4. Risk classes were assigned to the ISO codes assigned factors in step 2 by using the St. Paul's risk classification system.
5. Average risk factors were calculated for each risk class from results of step 4.
6. The ISO codes from step 3 were assigned the average risk factor calculated in step 5. If no average was available, then a risk factor was interpolated based on the relationships found in the St. Paul's rate manual.
7. For specialties, which could not be mapped to an ISO code, we used the HCFA supplied cross walk to assign an appropriate risk factor (see Exhibit 3).

A summary of the risk factor assignments by Medicare code is presented in Exhibit 4.

M.5. Physician Profile Dataset

HCFA supplied physician specialty and procedure frequency data that will be referred to as the physician profile dataset. Since 1998 data was not available at the time of the study, 1997 data was used. The data layout, showing only the fields used, was:

The three records shown are for the purpose of illustration only. The first record shows the number of procedures coded 97035, "Ultrasound therapy", performed in 1997 by independently practicing physical therapists. The next two records show "Subsequent Hospital Care" provided by Neurologists and "Tissue Exams performed by Pathologists" performed by Pathologists. This dataset will be called the "Physician Profile Dataset" because it shows utilization profiles or relative frequency of procedures by Medicare physician specialty. It is useful for calculating the frequency-weighted averages of all the risk factors of the specialties that perform a procedure (this is explained in more detail in the following section). It is also useful for measuring the impact of the new malpractice RVUs versus the existing malpractice RVU schedule.

Cross-Walk of 1997 Codes to 1998 & 1999

New codes are introduced to the RBRVS fee schedule each year. Because some basic utilization assumptions are needed for each code, utilization was estimated for the new 1998 and 1999 codes. For each new 1998 and 1999 procedure code, a weighted average of 1997 codes was used to project utilization (this calculation was performed by HCFA). HCFA has supplied the utilization assumptions for each of the new codes. These assumptions are shown in exhibit 5.

M.6. Calculating Average Risk Factors by Procedure

As stated previously, KPMG was requested to calculate RBMRVUs using two methodologies. This section will be broken into two sub headings, "Risk Factor Method " and "Risk of Service Adjusted Method" corresponding to the two methodologies.

Risk Factor Method

Under this method, the RBMRVUs were calculated as the services (or procedure frequency) weighted average of specialists' risk factors for each procedure code, adjusted for 1997 budget neutrality. In equation form:

SPEC = Allergist/ Immunologist through Vascular Surgeon

( Σ SERVICES _{PROC/MOD, SPEC} X RF _SPEC )    = Raw RBMRVU_PROC/MOD

^{__________________________________________________}

SERVICES_PROC/MOD

1999 Aggregate MP RVUs = BN Adj.

( Σ SERVICES _PROC/MOD X Raw RBMRVU_PROC/MOD )

^{PROC=00000 to 99999 TC}

RBMRVU_PROC/MOD = Raw RBMRVU_PROC/MOD X BN Adj.

Where,

This is best demonstrated by example. Let's assume that three physician specialties perform a given procedure. We'll assume that budget neutrality does not enter into the equation. Each of the specialties perform an equal number of the procedure over a given time period.

_{SPEC = A through C}

( Σ SERVICES _{PROC/MOD, SPEC} X RF _SPEC )    = Raw RBMRVU_PROC/MOD

^{___________________________________________________}

SERVICES _PROC/MOD

(10 X 1.0 + 10 X 2.0 + 10 X 3.0 )    = Raw RBMRVU_PROC/MOD

30

60/30 = 2.0 = Raw RBMRVU_PROC/MOD

The specialist risk factor or procedure allocation technique does not produce a reasonable result without the implementation of a budget neutrality adjustment. Our preliminary analysis produced a budget neutrality adjustment of .05 for the specialist factor technique. Using this .05 adjustment, the result would be .05 X 2.00 = 0.10 malpractice RVUs for this method.

Risk of Service Adjusted Method

This method also uses a services weighted average of risk factors using the physician specialty utilization data ("Physician Profile Dataset"). For this methodology, codes without work RVUs that historically have had malpractice RVUs were left at their historical malpractice RVU levels.

The first method gives us a head start since the average risk factors have already been determined. An additional step is added, the average risk factors are multiplied by the Work RVUs and a separate budget neutrality adjustment is calculated. In equation form:

SPEC = Allergist/ Immunologist through Vascular Surgeon

(Σ SERVICES _{PROC/MOD, SPEC} X RF _SPEC )  X Work RVU_PROC/MOD = Raw RBMRVU_PROC/MOD

^{__________________________________________________}

SERVICES_PROC/MOD

1999 Aggregate MP RVUs = BN Adj.

(Σ SERVICES _PROC/MOD X Raw RBMRVU_PROC/MOD )

^{PROC=00000 to 99999 TC}

RBMRVU_PROC/MOD = Raw RBMRVU_PROC/MOD x BN Adj.

Where,

Expanding from the numerical example from the previous section,

Assuming 0.05 work RVUs,

_{SPEC = Allergist/ Immunologist through Vascular Surgeon}

( Σ SERVICES _{PROC/MOD, SPEC} X RF _SPEC X Work RVU_PROC/MOD = Raw RBMRVU_PROC/MOD

^{__________________________________________________}

SERVICES _PROC/MOD

(10 X 1.0 + 10 X 2.0 + 10 X 3.0 ) X 0.05 = Raw RBMRVU_PROC/MOD

30

(60/30) X 0.05 = 0.10 Malpractice RVUs.

M.7. Codes Whose RBMRVUs Were Not Assigned Using the Risk Factor Algorithms

Certain codes could not be assigned an average risk factor for various reasons. In the previous section it was mentioned that codes with zero work RVUs cannot be assigned malpractice RVUs using the ROS method. The solution was to leave these malpractice RVUs unchanged. This section of the report will discuss two other special sets of codes, codes with technical modifiers and codes for which 1997 utilization data did not exist.

Codes with Professional and Technical Components

Codes that have unbundled technical and professional codes are treated differently. KPMG is only updating the professional component of these codes. It is assumed that the malpractice RVUs assigned to technical codes are based on the malpractice costs of the facility. Since the study is focused on professional malpractice costs, it was decided to leave these allocations to technical codes unchanged. The global malpractice RVUs for these codes are the sum of the updated professional component and the unchanged technical component.

Codes Without 1997 Utilization Data

The Risk of Service and Risk Factor used to develop the resource based malpractice RVUs methods combined with the unchanged technical modifier codes accounted for 98.2% of all the procedure codes included in the 1999 RBRVS. The 191 remaining procedure codes (out of 10,936) that were not assigned a resource based malpractice RVU were assigned a malpractice RVU based upon the following algorithm:

1. If a similar code could be identified, an average risk factor for the similar procedures was used.
2. If a similar code could not be identified, but an assumption could be made about the specialty which would normally perform the procedure, the risk factor associated with that specialty was used.
3. If a specialty could not be assigned, either the average surgical or non-surgical factor depending on the type of code was assigned.

A summary of the judgments used to map these codes is shown in Exhibit 6.

M.8. Budget Neutrality Adjustment

The budget neutrality adjustment used for the draft RBMRVUs is not the final budget neutrality adjustment that will be applied. The final budget neutrality adjustment is the responsibility of the Office of the Actuary at HCFA. KPMG has attempted to get the malpractice RVUs as close to budget neutral as possible with respect to 1999 RBRVS, so that the RVUs presented in this report are not radically different from the final malpractice RVUs to be implemented in the year 2000.

To perform this preliminary budget neutrality adjustment, KPMG used the 1997 frequency of procedures and calculated what the malpractice RVUs would have been if the 1999 malpractice RVU values were incorporated. KPMG also calculated preliminary aggregate malpractice RVUs using the resource-based malpractice RVUs derived from our two methodologies.

Our budget neutrality adjustment for each procedure is the aggregate malpractice RVUs calculated using the appropriate resource-based allocation methodology divided by the aggregate malpractice RVUs using the existing malpractice RVUs.

This adjustment is then multiplied by the preliminary malpractice RVUs for each method to get the budget-neutrality RVU allocations. Aggregate RVUs will not be perfectly balanced due to the RBRVS rounding conventions, which do not carry more than two decimal places of accuracy.

Exhibits

Exhibit 1 – National Average Malpractice Premiums

Exhibit 2 – Mapping of Medicare Specialty Code to ISO Code

Explanation of Exhibit

Medicare uses specialty codes that are distinct from the ISO codes used by malpractice insurers to identify specialists. ISO codes tend to be more specific than HCFA codes which already identify almost 90 distinct physician specialties. In particular, ISO codes make distinctions according to the level of surgery practiced by the physician. In order to perform our calculations, it was necessary to map or cross-walk the Medicare specialty codes to their nearest ISO codes. For the twenty Medicare specialty codes tracked by Allied Technology this was performed for us.

Explanation of Columns

Medicare Code -- The specific Medicare code assigned by HCFA to the specialty.

Medicare Description -- The description that Medicare uses for the specialty code.

ISO Code - The code used by malpractice insurers to identify the specialty.

Risk Class - St. Paul's risk classification for the particular ISO code. Specialties within the same risk class will have comparable premium rates.

St. Paul’s Description - The exact interpretation of the ISO Code according to St. Paul's rate manual.

Exhibit 3 – Cross Walk For Unassigned Specialties

Exhibit 4 – Medicare Specialties and Risk Factor Assignment

Explanation of Columns

Medicare Code - The code used by HCFA to identify the specialty.

Medicare Description - HCFA's definition of the specialty code.

Risk Factor - The risk factor is an assessment of the relative malpractice costs of the specialty. It is calculated as the specialist's national avg. premium divided by the lowest national avg. premium (Psychiatrists). For specialties that were not tracked in the malpractice survey, we first attempted to extrapolate a risk factor based upon the St. Paul rate manual. If this was not possible, we used a cross-walk assumption taken from the development of the Practice Expense RVUs.

Exhibit 5 – Crosswalk of 1997 Procedure Codes to 1998-99 Procedure Codes

Exhibit 6 – Crosswalk for Codes With No Utilization Data