April 2000
Volume 64 |
Number 4
|
| |
|
| A Methodology
for the Calculation of Anesthesia Relative Value Units |
Virginia N. Jablonski, M.S.A.
Wayne K. Marshall, M.D.
That medical practices are facing demands for higher quality
of care and decreasing reimbursement is a given today. This environment
has necessitated that physicians implement new business strategies
to decrease practice expenses and increase operational efficiencies.
Many of these strategies require new business tools.
One of the new tools being used to help practices increase their
operational efficiency is the Health Care Financing Administration's
(HCFA's) Resource-Based Relative Value Scale (RBRVS), a listing
of physician services by Current Procedural Terminology™
(CPT) code. The unit value is the relative value unit (RVU), which
is composed of three factors: 1) physician work, 2) practice expense
and 3) malpractice. HCFA's purpose in developing this scale was
to establish proportional weights for all physician services that
could then be converted into reimbursement levels; each CPT code
is reimbursed the same regardless of the medical specialty. RBRVS-reimbursed
procedures are paid at a predetermined fee calculated from the
RVU.
Not only does Medicare reimburse by this method, but a large
portion of other insurance carriers use this method for establishing
their payment levels. This scale has gone far beyond its original
purpose of setting reimbursement levels. Many medical practices
are using the RVU concept for many business applications. For
example, RVUs are used to determine practice expenses, physician
productivity, reimbursement trends analysis of managed care contracts,
budgetary projections and distribution of capitated funds. Because
RVUs are the same for all like physician services, practices can
compare themselves against established regional and/or national
practices' benchmarks.
Unfortunately, anesthesiology has been unable to use many of
these RVU business applications because anesthesia services (0XXXX
CPT codes) do not have established RVUs under RBRVS. This lack
of RVUs is especially frustrating in the current environment where
members of multispecialty group practices are commonly compared
with each other against RVU outcomes. In the faculty practice
plan studied, expense and revenue data for each department within
the practice are analyzed and compared using various RVU calculations.
For example, RVU reports show each department's expense, revenue
and production results in equal units allowing objective comparisons.
These data are used by the administration to make financial and
staffing decisions for the departments. In addition to these internal
studies, internal RVU data are also compared against various published
benchmarks for analysis with peers.
However, not having established RVUs for the anesthesia department
prevents comparison among all departments. It then follows that
the practice plan also can never be studied as a single entity.
Due to this situation, we wanted to find/develop a methodology
for the calculation of anesthesia RVUs that would be comparable
to existing RVUs for other physician services and that could be
used in our business applications. The methodology needed to be
adaptable to other anesthesia practices. It was not intended to
replace the current payment policies of any payer.
We expanded upon the Imputed Work RVUs already calculated by
HCFA, as published in the Federal Register, so that totally
new calculations would not have to be developed and proven. The
methodology first calculates total RVUs by using a portion of
HCFA's Imputed Work RVU equation and then calculates the individual
work, practice expense and malpractice factors by multiplying
the total RVUs by the corresponding anesthesia specialty share
weights. The main variable in the HCFA model is mean time per
procedure derived from Medicare claims data. We used our own internal
mean time data per procedure to determine RVUs.
After calculating these RVUs, we then used two other mean time
data sources for comparison; they included the HCFA data in the
Federal Register and claims data from a West Coast billing
company. RVUs for all three data sets were then compared to find
out if there was any significant variance. We not only wanted
to determine the variance between the RVU outcomes, but we also
wanted to see how close Medicare payments for these RVUs would
be to anesthesia payments based on units. If the RVU payments
were close to unit payments, it meant the RVUs correctly reflected
the anesthesia procedures. In addition, if the computed RVUs were
similar between the three data sets, it would mean that regardless
of type of practice, the mean time for all anesthesia procedures
is comparable.
Methods
The equation we used to calculate anesthesia RVUs for all three
RBRVS factors is an extension of HCFA's development of Imputed
Work RVU calculations (December 8, 1994 Federal Register).
[((base + time) * anesthesia CF)/surgical
CF] * specialty share weight = Imputed workRVUs
Where:
base = base units per anesthesia CPT code. Obtained from
the American Society of Anesthesiologists (ASA) 1988 Relative
Value Guide.
time = time units based on 15-minute increments. These
are the 1993 mean time units from national Medicare claims data
for anesthesia services personally performed by the physician.
anesthesia CF = 1994 national anesthesia conversion factor
(CF) of $15.32.
surgical CF = 1994 national surgical conversion factor
of $39.45.
specialty share weight = anesthesia specialty share weight
for work in 1994 or 0.695. (Share weights are the proportion of
the total RVUs attributable to each of the physician work, practice
expense and malpractice factors.
Base plus mean time units (base + time) for each anesthesia
CPT code are multiplied by the anesthesia CF and then divided
by the surgical CF. These results are then multiplied by the work
specialty share weight to arrive at imputed work RVUs.
Our method uses the HCFA equation, but instead of calculating
only work RVUs we used all three anesthesia share weights to calculate
work, practice, malpractice and total anesthesia RVUs for each
CPT code. The equations for the RBRVS factors and total RVUs are:
[((base + time) x anesthesia CF)/surgical CF] x work share weight
= work (W) RVUs
[((base + time) x anesthesia CF)/surgical CF] x practice expense
share weight = practice expense (PE) RVUs
[((base + time) x anesthesia CF)/surgical CF] x each share weight
= Malpractice(MP) RVUs
W RVUs + PE RVUs + MP RVUs = total RVUs for each CPT code
Anesthesia RVUs were calculated using these equations to compare
three different data sets that contained mean anesthesia time
units: Medicare's 1993-94 claims data, those from a commercial
billing service for surgical CPT codes using both Medicare and
non-Medicare claims data from several states, and those from an
academic medical center's own information systems. The RVU calculated
payments that are closest to base + time payments best reflect
the level of resources utilized in performing anesthesia procedures.
1. HCFA
We used HCFA's 1994 data and updated it to 1997 national conversion
factors and specialty share weights. A sample calculation of anesthesia
RVUs using the formula described above is:
CPT code - 00100
((base units + time units) * anesthesia CF)/surgical CF = total
RVUs
((5+7.26) * 17.76)/40.96 = 5.31
total RVUs * specialty share weight = RVUs per factor
5.31 * .782 = 4.15 work RVUs
5.31 * .166 = 0.88 practice RVUs
5.31 * .052 = 0.28 malpractice RVUs
2. Billing Service
This is a database of more than 86,000 claims submitted by the
clients of a nationwide billing service in 1996. These data were
obtained for analysis by ASA and were collected by surgical code,
not by anesthesia code. They include mean surgical times for Medicare
and non-Medicare patients from academic and nonacademic practices.
Data came from six states: California, Hawaii, Idaho, Illinois,
Washington and Wisconsin. Anesthesia times matched the anesthesia
record times. Next, mean time units were recalculated using 15-minute
increments. These surgical codes (N=3,968) were then matched with
the appropriate Relative Value Guide's anesthesia codes and base
units. The same calculations as above were applied to this data
to arrive at anesthesia RVUs.
3. Virginia Commonwealth University (VCU) Data
Permission was obtained from the VCU Department of Anesthesiology
for the collection of patient charge data from both Medicare and
non-Medicare patients during June 1996 through May 1997. Anesthesia
times for all surgical codes during this period were collected
from billing software. All codes were used in our calculations.
Data from minimally performed procedures were included as valid
because they accurately represent the normal occurrences of these
procedures. All like codes were combined and mean unit times were
calculated for the resulting set (N=2,076). Anesthesia RVU calculations
were performed.
Analysis of Data
To determine the inherent accuracy of the RVU determinations
for all three data sets, the resulting data was compared to the
corresponding base + time units data. Because we were dealing
with two different systems, we computed Medicare payments (by
individual CPT code) for the RVUs and the base + time units. These
calculated payments gave us the common denominator we needed.
Payments were calculated using local geographic practice cost
indices (GPCIs). The equations follow:
Medicare unit payment
(mean time units + base units) x local anesthesia CF =
payments
Medicare RVU payment
[(RVUw xGPCIw) + (RVUpe x GPCIpe) + (RVUm x GPCIm)] x
surgical
CF = payments
Where:
CF = conversion factor
RVUw = physician work relative value units
RVUpe = practice expense relative value units
RVUm = malpractice relative value units
GPCI = geographic practice cost index; there are separate
indices for
each RBRVS factor
Once the payments for RVUs and base + time units were calculated
for each data set, percentage comparisons of the results were
computed. It was assumed that if the two payments matched within
a ±2 percent variance then the RVU calculations were reasonable.
A sample database of anesthesia procedures was set up to compare
the computed RVU payments of each data set with the base + time
payments of the sample. To create the sample, internal VCU patient
charge data for one month was entered into a database. Actual,
not mean, time units were calculated by dividing the reported
anesthesia time by 15 minutes and rounding to the nearest tenth.
Prior to calculations, those anesthesia or surgical codes that
were not common to all three data sets were deleted from the database.
All nonanesthesia procedures were also removed, e.g., visit codes.
Also, modifiers were not included in the calculations; all services
were treated as performed personally by the physician. A total
of 1,308 records were computed and compared for the three data
sets and with the sample database.
Results
In order to determine the reasonableness of the RVU data, we
computed the ratio of RVU payments to base + time unit payments
for each CPT code within each data set. The resulting RVU and
base + time payments were within 01 percent of each other
within the same data set.
When the total computed HCFA and billing service RVUs for the
set of procedures in the VCU sample database are matched with
the sample, there are much larger variances. The results of RVU
and payment calculations for all three data sets using internal
patient data are shown in Table 1. The VCU
data set (number 3) had the highest calculated RVU total of any
of the data sets. This data set's computed RVU payment was also
the closest to the computed base + time units payment. Percentage
comparisons of RVU payments for each of the three data sets to
the computed time + base units payment range from 89.1 percent
to 99.6 percent.
Discussion
The purpose of this study was to develop a methodology to calculate
anesthesiology work, practice expense and malpractice RVUs that
could be used in various business applications. It should be restated
that our intent was not to develop an alternative anesthesia payment
methodology.
We first determined the inherent reasonableness of the RVU determinations
by comparing computed RVU and base + time payments for each CPT
code within the three different data sets, and second, we determined
what effect any differences in mean times would have on RVU outcomes.
These two determinations were done by comparing calculated RVUs
and payment outcomes using Medicare payment methods.
RVU and base + time unit payments per anesthesia procedure for
the HCFA, billing service and VCU data sets matched within 0-1
percent. This would confirm that our equations accurately predict
RVUs per procedure because both payment calculations have coinciding
outcomes. This does not mean the RVUs are comparable across data
sets because of divergent mean time data. Another supporting argument
for the validity of this method is that the RVU equations are
based on HCFA's own determination of imputed work RVUs.
RVUs and RVU payments were calculated on a sample anesthesia
claims database for all three data sets. Also calculated were
actual reported base + time units payments. We assumed if the
RVUs and comparative payments calculated on the sample database
were close for all three data sets, we could assume any variances
in anesthesia practices/procedures are diluted in large populations
and mean times become similar. The results for our three data
sets show a variance of several percentage points because the
HCFA data used anesthesia CPT codes instead of surgical codes.
Anesthesia codes are not as specific as surgical codes since multiple
surgical codes are reported by a single anesthesia code. Therefore,
the use of resources to deliver anesthesia services could be over-
or understated. This is the main weakness in any data using anesthesia
codes.
Another issue is that the HCFA mean time data represents the
Medicare population only. RVUs calculated from this data will
not accurately represent physician practices that treat a dissimilar
population. In order to be used as an effective business tool,
the RVU data must reflect a physician's or group's own patient
population.
Data sets 2 and 3 have the highest number of RVUs. Both of these
data sets were computed by surgical code and include Medicare
and non-Medicare patients. The billing service data has academic
and nonacademic data, and VCU has only academic data. These data
bring up two important points. First, RVUs calculated by surgical
code are more specific because their payments are closest to actual
billed unit payments. Second, RVUs calculated on a physician's
own patient population are the most accurate. VCU data are the
closest to units payments; there is only a 0.4-percent payment
variance. The reason(s) for the time differences between the billing
service and VCU RVUs may be the result of patient demographics
or different practice types of academic and nonacademic. However,
the underlying factors that result in these time unit differences
for all three data groups could be many. They may be due to institutional
sites such as urban or rural. They may also be due to the types
of surgeries performed. Additional studies would need to be done
to find the answer.
In conclusion, we believe calculating RVUs from one's own internal
data yields valid information. This information can be applied
to various business applications to determine the efficiencies
of the practice. Until further studies can be completed, we do
not recommend calculating RVUs using time data from external populations.
What Does the Anesthesia-RVU Methodology Mean for Your Practice?
Calculating relative value units for anesthesia services on the
RBRVS allows you to compare yourself to other specialties for
purposes of measuring productivity, allocating practice expenses,
analyzing reimbursement trends and managed care contracts, making
budgetary projections and distributing capitated funds, as noted
in the introduction to the article. It is obviously particularly
useful in the multispecialty setting such as an academic medical
center or a multispecialty group, but it can also facilitate comparisons
between anesthesia services and the other procedures, including
pain management and visits that you provide.
To establish RVUs for your own practice, you will need to know
your average time units by procedure code, as well as the following
year 2000 Medicare Fee Schedule values:
Anesthesia Conversion Factor (CFa): $17.77
Surgical Conversion Factor (CFs): $36.61
Anesthesia Work Share: 0.7359
Anesthesia Practice Expense Share: 0.1955
Anesthesia Malpractice Expense Share: 0.0686
To calculate the RVUs for a total hip replacement (code 01214),
which has 8 base units and a hypothetical average of 12 time units,
use the formula described in the article:
([base+time] x CFa)/CFs, or ([8+12] x 17.77)/36.61 = 9.71
Total RVUs (rounded)
Work RVUs = 9.71 x .7359 = 7.14
Practice Expense RVUs = 9.71 x .1955 = 1.90
Malpractice Expense RVUs = 9.71 x .0686 = 0.67
Knowing the RVUs for each procedure that you perform, together
with frequencies, will permit determination of total practice
or physician RVUs.
Table 1
|
Comparison of
RVU and Base + Time Unit Payments
|
|
Data Set
|
Total
RVUs
|
Total Base + Time Units
|
RVU Payments
|
Total Unit Payment
|
% Comparison
RVU/Unit
Payments
|
| 1
HCFA |
8813.11 |
22654.6 |
$321,598 |
$360,888 |
89.1% |
| 2
Billing Service |
8912.31 |
22654.6 |
$325,230 |
$360,888 |
90.1% |
| 3
VCU |
9844.57 |
22654.6 |
$359,251 |
$360,888 |
99.6% |
Virginia N. Jablonski, M.S.A., is the Account
Manager, Per Se Technologies, Richmond, Virginia, and a former
senior health consultant at Medical College of Virginia, Richmond,
Virginia.
Wayne K. Marshall, M.D., is Professor of
Anesthesiology, Pennsylvania State University College of Medicine,
Hershey, Pennsylvania, and formerly on the faculty at Medical
College of Virginia, Richmond, Virginia.
return to top
|