Medicare Program; Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 2000 Rates

SUMMARY: We are revising the Medicare hospital inpatient prospective

payment systems for operating costs and capital-related costs to

implement changes arising from our continuing experience with the

systems. In addition, in the addendum to this final rule, we describe

changes in the amounts and factors necessary to determine rates for

Medicare hospital inpatient services for operating costs and capital-

related costs. These changes are applicable to discharges occurring on

or after October 1, 1999. We also set forth rate-of-increase limits as

well as policy changes for hospitals and hospital units excluded from

the prospective payment systems. Finally, we are revising certain

policies governing payment to hospitals for the direct costs of

graduate medical education.

DATES: The provisions of this final rule are effective October 1, 1999.

This rule is a major rule as defined in Title 5, United States Code,

section 804(2). Pursuant to 5 U.S.C. section 801(a)(1)(A), we are

submitting a report to Congress on this rule on July 30, 1999.

FOR FURTHER INFORMATION CONTACT:

Steve Phillips, (410) 786-4531, Operating Prospective Payment,

Diagnosis-Related Group (DRG), and Wage Index Issues.

Tzvi Hefter, (410) 786-4487, Capital Prospective Payment, Excluded

Hospitals, and Graduate Medical Education Issues.

SUPPLEMENTARY INFORMATION:

Availability of Copies and Electronic Access

To order copies of the Federal Register containing this document,

send your request to: New Orders, Superintendent of Documents, P.O. Box

371954, Pittsburgh, PA 15250-7954. Specify the date of the issue

requested and enclose a check or money order payable to the

Superintendent of Documents, or enclose your Visa or Master Card number

and expiration date. Credit card orders can also be placed by calling

the order desk at (202) 512-1800 or by faxing to (202) 512-2250. The

cost for each copy is $8.00. As an alternative, you can view and

photocopy the Federal Register document at most libraries designated as

Federal Depository Libraries and at many other public and academic

libraries throughout the country that receive the Federal Register.

This Federal Register document is also available from the Federal

Register online database through GPO Access, a service of the U.S.

Government Printing Office. Free public access is available on a Wide

Area Information Server (WAIS) through the Internet and via

asynchronous dial-in. Internet users can access the database by using

the World Wide Web; the Superintendent of Documents home page address

is

http://www.access.gpo.gov/nara__docs/, by using local WAIS client

software, or by telnet to swais.access.gpo.gov, then login as guest (no

password required). Dial-in users should use communications software

and modem to call (202) 512-1661; type swais, then login as guest (no

password required).

I. Background

A. Summary

Section 1886(d) of the Social Security Act (the Act) sets forth a

system of payment for the operating costs of acute care hospital

inpatient stays under Medicare Part A (Hospital Insurance) based on

prospectively set rates. Section 1886(g) of the Act requires the

Secretary to pay for the capital-related costs of hospital inpatient

stays under a prospective payment system. Under these prospective

payment systems, Medicare payment for hospital inpatient operating and

capital-related costs is made at predetermined, specific rates for each

hospital discharge. Discharges are classified according to a list of

diagnosis-related groups (DRGs).

Certain specialty hospitals are excluded from the prospective

payment systems. Under section 1886(d)(1)(B) of the Act, the following

hospitals and hospital units are excluded from the prospective payment

systems: psychiatric hospitals or units, rehabilitation hospitals or

units, children's hospitals, long-term care hospitals, and cancer

hospitals. For these hospitals and units, Medicare payment for

operating costs is based on reasonable costs subject to a hospital-

specific annual limit.

Under section 1886(a)(4) of the Act, costs incurred directly by a

hospital in connection with approved graduate medical education (GME)

programs are excluded from the operating costs of inpatient hospital

services. Hospitals with approved GME programs are paid for the direct

costs of GME in accordance with section 1886(h) of the Act; the amount

of payment for direct GME costs for a cost reporting period is based on

the hospital's number of residents in that period and the hospital's

costs per resident in a base year.

The regulations governing the hospital inpatient prospective

payment systems are located in 42 CFR part 412. The regulations

governing excluded hospitals and hospital units are located in parts

412 and 413, and the GME regulations are located in part 413.

B. Summary of the Provisions of the May 7, 1999 Proposed Rule

On May 7, 1999, we published a proposed rule in the Federal

Register (64 FR 24716) that set forth proposed changes to the Medicare

hospital inpatient prospective payment systems for both operating costs

and capital-related costs that would be effective for discharges

occurring on or after October 1, 1999. We also proposed changes

concerning GME costs and excluded hospitals and units, as well as

critical access hospitals (CAHs). On June 15, 1999, we issued a

correction notice (64 FR 31995) for the May 7, 1999 proposed rule. That

notice corrected Table 3C of the Addendum (which lists each hospital's

case-mix index and adjusted average hourly wage based on data on file

at HCFA as of February 22, 1999) and made several other technical

corrections.

In the proposed rule, we noted that the efforts that we were

undertaking to make the Medicare computer systems compliant on January

1, 2000, would not delay our ability to make timely and updated

payments to hospitals under the FY 2000 prospective payment systems

final rule. This statement still applies and the changes and updated

rates set forth in this final rule will be implemented on October 1,

1999.

The following is a summary of the contents of the proposed rule:

In order to avoid compromising our ability to process and

pay hospital claims during the period leading up to and immediately

following January 1, 2000, we did not propose to implement any

revisions to the International Classification of Diseases, Ninth

Revision, Clinical Modification (ICD-9-CM) coding system. We did

propose to make some limited changes to certain DRG classifications for

FY 2000 and described other proposed decisions

[[Page 41491]]

concerning DRGs. We also recalibrated the DRG relative weights based on

the proposed DRG changes and updated Medicare claims data.

We proposed an FY 2000 hospital wage index update, using

FY 1996 wage data, and revisions to the wage index based on hospital

redesignations. In addition, we proposed to begin excluding from the

wage index Part A physician wage costs that are teaching-related, as

well as resident and Part A certified registered nurse anesthetist

(CRNA) costs.

We proposed several policy changes in the regulations in

42 CFR parts 412 and 413 and proposed to continue existing policy

concerning classifications of sole community hospitals; the indirect

medical education adjustment; and Medicare Geographic Classification

Review Board (MGCRB) decisions. In addition, we updated the qualifying

criteria for rural referral centers and proposed several changes to the

regulations governing payments for the direct costs of GME programs.

We discussed the special exceptions process for certain

eligible hospitals to receive additional payments for major

construction or renovation projects that began soon after the start of

the capital prospective payment system and proposals that we had

received to change the eligibility criteria for these payments.

We discussed a number of proposals concerning Medicare

payments to excluded hospitals and hospital units and CAHs. These

proposed changes related to limits on and adjustments to the proposed

target amounts for FY 2000; changes in bed size or status of excluded

hospitals or hospital units; payment for Medicare services furnished at

satellite hospital locations; responsibility for care of patients in

hospitals-within-hospitals; the allowable emergency response time for

CAHs located in frontier or other specifically defined remote areas;

and compliance with minimum data set requirements by CAHs with swing

bed approval.

In the addendum to the proposed rule, we set forth

proposed changes to the amounts and factors for determining the FY 2000

prospective payment rates for operating costs and capital-related

costs. We also addressed update factors for determining the rate-of-

increase limits for cost reporting periods beginning in FY 2000 for

hospitals and hospital units excluded from the prospective payment

system.

In Appendix A of the proposed rule, we set forth an

analysis of the impact that the proposed changes would have on affected

entities.

In Appendix B of the proposed rule, we set forth the

technical appendix on the proposed FY 2000 capital cost model.

In Appendix C of the proposed rule, as required by section

1886(e)(3)(B) of the Act, we set forth our report to Congress on our

initial estimate of a recommended update factor for FY 2000 for both

hospitals included in and hospitals excluded from the prospective

payment systems.

In Appendix D of the proposed rule, as required by

sections 1886(e)(4) and (e)(5) of the Act, we included our

recommendation of the appropriate percentage change for FY 2000 for--

--Large urban area and other area average standardized amounts (and

hospital-specific rates applicable to sole community hospitals and

Medicare-dependent, small rural hospitals) for hospital inpatient

services paid for under the prospective payment system for operating

costs; and

--Target rate-of-increase limits to the allowable operating costs of

hospital inpatient services furnished by hospitals and hospital units

excluded from the prospective payment system.

In the proposed rule, we discussed the recommendations

concerning hospital inpatient payment policies made by the Medicare

Payment Advisory Commission (MedPAC) and presented our responses to

those recommendations. Under section 1805(b) of the Act, MedPAC is

required to submit a report to Congress, not later than March 1 of each

year, that reviews and makes recommendations on Medicare payment

policies.

C. Public Comments Received in Response to the Proposed Rule

We received a total of 82 timely items of correspondence containing

multiple comments on the proposed rule. The main areas of concern

addressed by the commenters were removal of teaching-related and CRNA

costs from the wage index, payments for services furnished at satellite

hospital locations, and limits on the transfer of patients in

hospitals-within-hospitals. We also received a number of comments

relating to the eligibility criteria for hospitals to qualify for

capital exceptions payments.

Summaries of the public comments received and our responses to

those comments are set forth below under the appropriate section.

II. Changes to DRG Reclassifications and Recalibrations of Relative

Weights

A. Background

Under the prospective payment system, we pay for inpatient hospital

services on the basis of a rate per discharge that varies by the DRG to

which a beneficiary's stay is assigned. The formula used to calculate

payment for a specific case takes an individual hospital's payment rate

per case and multiplies it by the weight of the DRG to which the case

is assigned. Each DRG weight represents the average resources required

to care for cases in that particular DRG relative to the average

resources used to treat cases in all DRGs.

Congress recognized that it would be necessary to recalculate the

DRG relative weights periodically to account for changes in resource

consumption. Accordingly, section 1886(d)(4)(C) of the Act requires

that the Secretary adjust the DRG classifications and relative weights

at least annually. These adjustments are made to reflect changes in

treatment patterns, technology, and any other factors that may change

the relative use of hospital resources.

As discussed in more detail in section II.B.8 of this preamble, we

are not implementing any revisions to the ICD-9-CM codes. We have

undertaken, and continue to undertake, major efforts to ensure that all

of the Medicare computer systems are ready to function on January 1,

2000. If we were to implement changes to the ICD-9-CM codes on October

1, 1999, we would endanger the functioning of the Medicare computer

systems, and, specifically, we might compromise our ability to process

hospital bills. We can, however, reclassify existing codes into

different DRGs, if appropriate.

The changes to the DRG classification system, and the recalibration

of the DRG weights for discharges occurring on or after October 1,

1999, are discussed below.

B. DRG Reclassification

1. General

Cases are classified into DRGs for payment under the prospective

payment system based on the principal diagnosis, up to eight additional

diagnoses, and up to six procedures performed during the stay, as well

as age, sex, and discharge status of the patient. The diagnosis and

procedure information is reported by the hospital using ICD-9-CM codes.

The Medicare fiscal intermediary enters the information into its claims

processing system and subjects it to a series of automated screens

called the

[[Page 41492]]

Medicare Code Editor (MCE). These screens are designed to identify

cases that require further review before classification into a DRG can

be accomplished.

After screening through the MCE and any further development of the

claims, cases are classified by the GROUPER software program into the

appropriate DRG. The GROUPER program was developed as a means of

classifying each case into a DRG on the basis of the diagnosis and

procedure codes and demographic information (that is, sex, age, and

discharge status). It is used both to classify past cases in order to

measure relative hospital resource consumption to establish the DRG

weights and to classify current cases for purposes of determining

payment. The records for all Medicare hospital inpatient discharges are

maintained in the Medicare Provider Analysis and Review (MedPAR) file.

The data in this file are used to evaluate possible DRG classification

changes and to recalibrate the DRG weights.

Currently, cases are assigned to one of 499 DRGs in 25 major

diagnostic categories (MDCs). Most MDCs are based on a particular organ

system of the body (for example, MDC 6, Diseases and Disorders of the

Digestive System); however, some MDCs are not constructed on this basis

since they involve multiple organ systems (for example, MDC 22, Burns).

In general, cases are assigned to an MDC based on the principal

diagnosis before assignment to a DRG. However, there are five DRGs to

which cases are directly assigned on the basis of procedure codes.

These are the DRGs for liver, bone marrow, and lung transplants (DRGs

480, 481, and 495, respectively) and the two DRGs for tracheostomies

(DRGs 482 and 483). Cases are assigned to these DRGs before

classification to an MDC.

Within most MDCs, cases are then divided into surgical DRGs (based

on a surgical hierarchy that orders individual procedures or groups of

procedures by resource intensity) and medical DRGs. Medical DRGs

generally are differentiated on the basis of diagnosis and age. Some

surgical and medical DRGs are further differentiated based on the

presence or absence of complications or comorbidities (CC).

Generally, GROUPER does not consider other procedures; that is,

nonsurgical procedures or minor surgical procedures generally not

performed in an operating room are not listed as operating room (OR)

procedures in the GROUPER decision tables. However, there are a few

non-OR procedures that do affect DRG assignment for certain principal

diagnoses, such as extracorporeal shock wave lithotripsy for patients

with a principal diagnosis of urinary stones.

We proposed several changes to the DRG classification system for FY

2000 and other decisions concerning DRGs. The proposed changes, the

comments we received concerning them, and the final DRG changes are set

forth below. Unless otherwise noted, our DRG analysis is based on the

full (100 percent) FY 1998 MedPAR file, which contains data from bills

received through March 31, 1999.

2. MDC 15 (Newborns and Other Neonates with Conditions Originating in

the Perinatal Period)

In the May 7, 1999 proposed rule, we noted that the following codes

in the newborn observation series are included in the allowable

secondary diagnoses under DRG 391 (Normal Newborn):

V29.0, Observation for suspected infectious disease

V29.1, Observation for suspected neurological condition

V29.8, Observation for other specified suspected condition

V29.9, Observation for unspecified suspected condition

There are two related codes, however, that currently are not included

as allowable secondary diagnoses under DRG 391: V29.2 (Observation for

suspected respiratory condition) and V29.3 (Observation for suspected

genetic or metabolic condition). (In the proposed rule, we incorrectly

stated that V29.3 was titled ``Observation for other genetic

problem.'') Diagnosis codes V29.2 and V29.3 (as well as the other V29.x

codes noted above) are used to indicate that the newborn was suspected

of having an abnormal condition resulting from exposure from the mother

or the birth process, but is without signs or symptoms and, after

examination and observation, no abnormal condition is found to exist.

Currently, when either V29.2 or V29.3 is the only secondary diagnosis

for an otherwise healthy newborn, the case is assigned to DRG 390

(Neonate with Other Significant Problems). Based on a belief that the

presence of diagnosis code V29.2 or V29.3 should not exclude a newborn

from being classified as normal, we proposed to include diagnosis codes

V29.2 and V29.3 in the list of allowable secondary diagnoses under DRG

391 (Normal Newborn).

We received one comment on this proposal.

Comment: The commenter questioned whether any of the codes in the

V29 series should be assigned to DRG 391. The commenter believes that

the infants assigned to diagnosis code in the V29 series do not belong

in the same clinical group as ``normal newborn.'' The commenter

recommended that, before moving codes V29.2 and V29.3 to DRG 391, we

should examine data such as the average length of stay for DRGs 390 and

391 and those cases coded with V29.x. Citing one hospital's experience,

the commenter noted that 2.7 percent of the cases in DRG 391 were

assigned a secondary diagnosis of V29.0 (Observation for suspected

infectious disease). In addition, cases with secondary diagnosis codes

V29.1, V29.8, and V29.9 represented less than 1 percent each of all

cases in DRG 391. The commenter also reported that, for DRG 390, less

than 1 percent of cases were assigned a secondary diagnosis code of

V29.2 or V29.3. The commenter believes that the length of stay and

resource consumption for these cases should be compared to other cases

assigned to DRG 390 and DRG 391 to determine whether a separate DRG

should be created to adequately categorize these infants.

Response: The experience of the hospital reported by the commenter

indicates that newborn cases with a secondary diagnosis of V29.2 or

V29.3 represent a small percentage of newborn cases. Medicare data do

not contain enough data on newborns to verify this.

In the FY 1998 MedPAR file, there are only nine cases assigned to

DRG 390 and none to DRG 391. In fact, in FY 1998, there were only 18

cases assigned to all of MDC 15. Because of the lack of data on

newborns in the Medicare claims file, the relative weights and lengths

of stay for the DRGs in MDC 15 are based on non-Medicare data collected

from 19 States. (See the September 1, 1995 final rule (60 FR 45781) for

a detailed discussion of this policy.) Therefore, we rely closely on

experts outside of HCFA when we make any changes in MDC 15. We had

received information before publication of the proposed rule suggesting

that V29.2 and V29.3 should be included with the other V29.x codes in

DRG 391. After verifying with our medical consultants that this

information was clinically accurate, we proposed to make this DRG

classification change. We do note that the average lengths of stay for

DRG 390 and 391 do not differ dramatically (3.4 and 3.1 days,

respectively). However, the relative weight for DRG 390 is

significantly higher than that for DRG 391 (1.5908 and 0.1516,

respectively). Thus, we believe the amount of resource use devoted to

newborns in DRG 390 is not

[[Page 41493]]

connected to the amount of time spent in the hospital.

The commenter did not provide any length of stay or resource use

data nor did the commenter provide any reason that codes V29.2 or V29.3

should be treated differently than the other codes in category V29.x.

We believe that DRG 390, as its title indicates, should be used to

classify newborns with significant problems. Newborns who exhibit no

signs or symptoms and are merely evaluated or observed for a suspected

condition that is ruled out should not be classified with newborns who

have significant problems that require treatment.

We note that DRG 391 includes newborns who have minor problems or

conditions that require treatment. For example, some newborns with

jaundice, newborns with scalp injuries or mild birth asphyxia, and

newborns with minor skin infections are all classified to DRG 391.

Thus, that DRG does contain newborn cases for which some medical

treatment must be provided. We believe that including newborns observed

for suspected respiratory, genetic, or metabolic conditions in DRG 391

is clinically appropriate. Therefore, as proposed, we will include

V29.2 and V29.3 as allowable secondary diagnoses under DRG 391, as are

the rest of the codes in that category.

3. MDC 19 (Mental Diseases and Disorders)

We proposed to revise the title of DRG 425, ``Acute Adjustment

Reaction and Disturbances of Psychosocial Dysfunction'' under MDC 19 to

read ``Acute Adjustment Reaction and Psychosocial Dysfunction.''

Correspondents had stated that the terms ``disturbances'' and

``dysfunction'' were redundant since the terms have similar meanings.

We received one comment in support of this revision. Therefore, we

are adopting this proposed revision as final.

4. MDC 22 (Burns)

In the July 31, 1998 final rule (63 FR 40957), we implemented an

extensive redesign of the DRGs for burns to more appropriately capture

the variation in resource use associated with different classes of burn

patients. After these DRGs went into effect on October 1, 1998, we were

contacted by several hospitals about our inclusion of the fifth digit

``0'' on codes 948.10 through 948.90 to capture cases of full-thickness

burns. These hospitals stated that codes in category 948 with a fifth

digit of ``0'' should not be assigned to DRGs 506 through 509 as full-

thickness burns since not all of these cases will have a full-thickness

(third degree) burn. The fifth digit ``0'' can capture cases in which

there actually is no third degree burn. The hospitals requested that we

consider removing from the full-thickness burn DRGs 506 through 509 all

codes in the 948 category with a fifth digit of ``0'' as follows:

948.00 Body burn involving less than 10 percent of body surface, third

degree less than 10 percent or unspecified

948.10 Body burn involving 10 to 19 percent of body surface, third

degree less than 10 percent or unspecified

948.20 Body burn involving 20 to 29 percent of body surface, third

degree less than 10 percent or unspecified

948.30 Body burn involving 30 to 39 percent of body surface, third

degree less than 10 percent or unspecified

948.40 Body burn involving 40 to 49 percent of body surface, third

degree less than 10 percent or unspecified

948.50 Body burn involving 50 to 59 percent of body surface, third

degree less than 10 percent or unspecified

948.60 Body burn involving 60 to 69 percent of body surface, third

degree less than 10 percent or unspecified

948.70 Body burn involving 70 to 79 percent of body surface, third

degree less than 10 percent or unspecified

948.80 Body burn involving 80 to 89 percent of body surface, third

degree less than 10 percent or unspecified

948.90 Body burn involving 90 percent or more of body surface, third

degree less than 10 percent or unspecified.

We agreed with the hospitals and proposed that the codes listed

above be removed from DRGs 506 through 509 and added to DRG 510

(Nonextensive Burns with CC or Significant Trauma) and DRG 511

(Nonextensive Burns without CC or Significant Trauma). Hospitals have

been instructed in Coding Clinic for ICD-9-CM, Fourth Quarter, 1994

(pages 22 through 28) to code the site of the burn first (codes 940

through 947), when known. Codes from category 948 may be used as a

principal diagnosis only when the site of the burn is not specified.

Category 948 is used as an additional code to provide information on

the percentage of total body that is burned or to show the percentage

of burn that was third degree. When hospitals report codes properly,

full-thickness burns would be assigned to a code for burn of the

specific site (940 through 947). This site code also shows the degree

of the burn. Furthermore, for those rare cases in which the site is not

provided, but it is known that 10 percent or more of the body has a

third degree burn, hospitals may report this information through the

use of category 948 with a fifth digit of ``1'' through ``9.'' All of

these cases would continue to be classified as full-thickness burns in

DRGs 506 through 509. Therefore, the proposed removal of codes 948.1

through 948.9 with a fifth digit of ``0'' would not prevent cases from

being assigned to one of the full-thickness DRGs when there is a third

degree burn and the case is correctly coded.

Comment: One commenter stated that while it is true that codes in

category 948 with a fifth digit of ``0'' may be assigned when there is

no third degree burn, fifth digit ``0'' is also used to report cases

that have a body surface of 1 to 9 percent involved in third degree

burns. The commenter suggested that consideration be given to these

cases as the presence of a third degree burn represents additional risk

to the patient.

Response: We agree with the commenter that the presence of third

degree burns represents additional risk to the patient and may result

in a higher resource use. More accurately capturing this fact was one

of the primary purposes in revising the burn DRGs in FY 1999. However,

as the commenter noted, in category 948, the fifth digit of ``0''

includes cases with no third degree burns as well as third degree burns

involving 1 to 9 percent of the body surface. It is precisely because

many of the cases coded in 948 with a ``0'' fifth digit have no third

degree burns that we believe it is not appropriate to include these

codes in DRGs 506 through 509. As stated above, hospitals have been

instructed to code the site of the burn first (codes 940 through 947),

when known. These codes capture information on the site of the burn as

well as whether the burn is a third degree burn. Therefore, by using

the more precise codes in the 940 through 947 series, hospitals will be

appropriately assigning cases with minor third degree burns to DRGs 506

through 509.

We are adopting as final our proposal to remove codes in the 948

category with a fifth digit of ``0'' from the list of full-thickness

burns.

5. Surgical Hierarchies

Some inpatient stays entail multiple surgical procedures, each one

of which, occurring by itself, could result in

[[Page 41494]]

assignment of the case to a different DRG within the MDC to which the

principal diagnosis is assigned. It is, therefore, necessary to have a

decision rule by which these cases are assigned to a single DRG. The

surgical hierarchy, an ordering of surgical classes from most to least

resource intensive, performs that function. Its application ensures

that cases involving multiple surgical procedures are assigned to the

DRG associated with the most resource-intensive surgical class.

Because the relative resource intensity of surgical classes can

shift as a function of DRG reclassification and recalibration, we

reviewed the surgical hierarchy of each MDC, as we have for previous

reclassifications, to determine if the ordering of classes coincided

with the intensity of resource utilization, as measured by the same

billing data used to compute the DRG relative weights.

A surgical class can be composed of one or more DRGs. For example,

in MDC 5, the surgical class ``heart transplant'' consists of a single

DRG (DRG 103), and the class ``major cardiovascular procedures''

consists of two DRGs (DRGs 110 and 111). Consequently, in many cases,

the surgical hierarchy has an impact on more than one DRG. The

methodology for determining the most resource-intensive surgical class

involves weighting each DRG for frequency to determine the average

resources for each surgical class. For example, assume surgical class A

includes DRGs 1 and 2 and surgical class B includes DRGs 3, 4, and 5.

Assume also that the average charge of DRG 1 is higher than that of DRG

3, but the average charges of DRGs 4 and 5 are higher than the average

charge of DRG 2. To determine whether surgical class A should be higher

or lower than surgical class B in the surgical hierarchy, we would

weight the average charge of each DRG by frequency (that is, by the

number of cases in the DRG) to determine average resource consumption

for the surgical class. The surgical classes would then be ordered from

the class with the highest average resource utilization to that with

the lowest, with the exception of ``other OR procedures'' as discussed

below.

This methodology may occasionally result in a case involving

multiple procedures being assigned to the lower-weighted DRG (in the

highest, most resource-intensive surgical class) of the available

alternatives. However, given that the logic underlying the surgical

hierarchy provides that the GROUPER searches for the procedure in the

most resource-intensive surgical class, this result is unavoidable.

We note that, notwithstanding the foregoing discussion, there are a

few instances when a surgical class with a lower average relative

weight is ordered above a surgical class with a higher average relative

weight. For example, the ``other OR procedures'' surgical class is

uniformly ordered last in the surgical hierarchy of each MDC in which

it occurs, regardless of the fact that the relative weight for the DRG

or DRGs in that surgical class may be higher than that for other

surgical classes in the MDC. The ``other OR procedures'' class is a

group of procedures that are least likely to be related to the

diagnoses in the MDC but are occasionally performed on patients with

these diagnoses. Therefore, these procedures should be considered only

if no other procedure more closely related to the diagnoses in the MDC

has been performed.

A second example occurs when the difference between the average

weights for two surgical classes is very small. We have found that

small differences generally do not warrant reordering of the hierarchy

since, by virtue of the hierarchy change, the relative weights are

likely to shift such that the higher-ordered surgical class has a lower

average weight than the class ordered below it.

Based on the preliminary recalibration of the DRGs, we proposed to

modify the surgical hierarchy as set forth below. However, in

developing the proposed rule, we were unable to test the effects of

proposed revisions to the surgical hierarchy and to reflect these

changes in the proposed relative weights due to the unavailability of

revised GROUPER software at the time the proposed rule was prepared.

Rather, we simulated most major classification changes to approximate

the placement of cases under the proposed reclassification and then

determined the average charge for each DRG. These average charges then

serve as our best estimate of relative resource use for each surgical

class. We tested the proposed surgical hierarchy changes after the

revised GROUPER was received. The final changes in the DRG relative

weights are reflected in this final rule.

We proposed to revise the surgical hierarchy for the Pre-MDC DRGs

and MDC 3 (Diseases and Disorders of the Ear, Nose, Mouth and Throat)

as follows:

In the Pre-MDC DRGs, we proposed to reorder Lung

Transplant (DRG 495) above Bone Marrow Transplant (DRG 481).

In MDC 3, we proposed to reorder Tonsil and Adenoid

Procedure Except Tonsillectomy and/or Adenoidectomy Only (DRGs 57 and

58) above Cleft Lip and Palate Repair (DRG 52).

We received two comments in support of the two surgical hierarchy

proposals. In addition, based on a test of the proposed revisions using

the most recent MedPAR file and the revised GROUPER software, we have

found that the revisions are still supported by the data and no

additional changes are indicated. Therefore, we are incorporating the

proposed revisions and reorders in this final rule.

6. Refinement of Complications and Comorbidities (CC) List

There is a standard list of diagnoses that are considered CCs. We

developed this list using physician panels to include those diagnoses

that, when present as a secondary condition, would be considered a

substantial complication or comorbidity. In previous years, we have

made changes to the standard list of CCs, either by adding new CCs or

by deleting CCs already on the list. In the May 7, 1999 proposed rule,

we did not propose to delete any of the diagnosis codes on the CC list.

In the September 1, 1987 final notice concerning changes to the DRG

classification system (52 FR 33143), we modified the GROUPER logic so

that certain diagnoses included on the standard list of CCs would not

be considered a valid CC in combination with a particular principal

diagnosis. Thus, we created the CC Exclusions List. We made these

changes to preclude coding of CCs for closely related conditions, to

preclude duplicative coding or inconsistent coding from being treated

as CCs, and to ensure that cases are appropriately classified between

the complicated and uncomplicated DRGs in a pair.

In the May 19, 1987 proposed notice concerning changes to the DRG

classification system (52 FR 18877), we explained that the excluded

secondary diagnoses were established using the following five

principles:

Chronic and acute manifestations of the same condition

should not be considered CCs for one another (as subsequently corrected

in the September 1, 1987 final notice (52 FR 33154)).

Specific and nonspecific (that is, not otherwise specified

(NOS)) diagnosis codes for a condition should not be considered CCs for

one another.

Conditions that may not co-exist, such as partial/total,

unilateral/bilateral, obstructed/unobstructed, and benign/malignant,

should not be considered CCs for one another.

[[Page 41495]]

The same condition in anatomically proximal sites should

not be considered CCs for one another.

Closely related conditions should not be considered CCs

for one another.

The creation of the CC Exclusions List was a major project

involving hundreds of codes. The FY 1988 revisions were intended to be

only a first step toward refinement of the CC list in that the criteria

used for eliminating certain diagnoses from consideration as CCs were

intended to identify only the most obvious diagnoses that should not be

considered complications or comorbidities of another diagnosis. For

that reason, and in light of comments and questions on the CC list, we

have continued to review the remaining CCs to identify additional

exclusions and to remove diagnoses from the master list that have been

shown not to meet the definition of a CC. (See the September 30, 1988

final rule for the revision made for the discharges occurring in FY

1989 (53 FR 38485); the September 1, 1989 final rule for the FY 1990

revision (54 FR 36552); the September 4, 1990 final rule for the FY

1991 revision (55 FR 36126); the August 30, 1991 final rule for the FY

1992 revision (56 FR 43209); the September 1, 1992 final rule for the

FY 1993 revision (57 FR 39753); the September 1, 1993 final rule for

the FY 1994 revisions (58 FR 46278); the September 1, 1994 final rule

for the FY 1995 revisions (59 FR 45334); the September 1, 1995 final

rule for the FY 1996 revisions (60 FR 45782); the August 30, 1996 final

rule for the FY 1997 revisions (61 FR 46171); the August 29, 1997 final

rule for the FY 1998 revisions (62 FR 45966); and the July 31, 1998

final rule for the FY 1999 revisions (63 FR 40954).) In the May 7, 1999

proposed rule, we did not propose to add or delete any codes from the

CC list.

In addition, because we are not making changes to the ICD-9-CM

codes for FY 2000, we are not modifying the current list for new or

deleted codes. Therefore, there are no revisions to the CC Exclusions

List for FY 2000.

7. Review of Procedure Codes in DRGs 468, 476, and 477

Each year, we review cases assigned to DRG 468 (Extensive OR

Procedure Unrelated to Principal Diagnosis), DRG 476 (Prostatic OR

Procedure Unrelated to Principal Diagnosis), and DRG 477 (Nonextensive

OR Procedure Unrelated to Principal Diagnosis) in order to determine

whether it would be appropriate to change the procedures assigned among

these DRGs.

DRGs 468, 476, and 477 are reserved for those cases in which none

of the OR procedures performed is related to the principal diagnosis.

These DRGs are intended to capture atypical cases, that is, those cases

that do not occur with sufficient frequency to represent a distinct,

recognizable clinical group. DRG 476 is assigned to those discharges in

which one or more of the following prostatic procedures are performed

and are unrelated to the principal diagnosis:

60.0 Incision of prostate

60.12 Open biopsy of prostate

60.15 Biopsy of periprostatic tissue

60.18 Other diagnostic procedures on prostate and periprostatic tissue

60.21 Transurethral prostatectomy

60.29 Other transurethral prostatectomy

60.61 Local excision of lesion of prostate

60.69 Prostatectomy NEC

60.81 Incision of periprostatic tissue

60.82 Excision of periprostatic tissue

60.93 Repair of prostate

60.94 Control of (postoperative) hemorrhage of prostate

60.95 Transurethral balloon dilation of the prostatic urethra

60.99 Other operations on prostate

All remaining OR procedures are assigned to DRGs 468 and 477, with

DRG 477 assigned to those discharges in which the only procedures

performed are nonextensive procedures that are unrelated to the

principal diagnosis. The original list of the ICD-9-CM procedure codes

for the procedures we consider nonextensive procedures, if performed

with an unrelated principal diagnosis, was published in Table 6C in

section IV of the Addendum to the September 30, 1988 final rule (53 FR

38591). As part of the final rules published on September 4, 1990,

August 30, 1991, September 1, 1992, September 1, 1993, September 1,

1994, September 1, 1995, August 30, 1996, and August 29, 1997, we moved

several other procedures from DRG 468 to 477, and some procedures from

DRG 477 to 468. (See 55 FR 36135, 56 FR 43212, 57 FR 23625, 58 FR

46279, 59 FR 45336, 60 FR 45783, 61 FR 46173, and 62 FR 45981,

respectively.) No procedures were moved in FY 1999, as noted in the

July 31, 1998 final rule (63 FR 40962).

a. Adding Procedure Codes to MDCs

We annually conduct a review of procedures producing DRG 468 or 477

assignments on the basis of volume of cases in these DRGs with each

procedure. Our medical consultants then identify those procedures

occurring in conjunction with certain principal diagnoses with

sufficient frequency to justify adding them to one of the surgical DRGs

for the MDC in which the diagnosis falls. Based on this year's review,

we identified several procedures that we proposed to move to surgical

DRGs for additional MDCs so that they are not assigned to DRG 468. We

did not identify any necessary changes in procedures under DRG 477 and,

therefore, did not propose to move any procedures from DRG 477 to one

of the surgical DRGs.

First, we proposed to move three codes from DRG 468 to MDC 1

(Diseases and Disorders of the Nervous System), all of which would be

assigned to DRGs 7 and 8 (Peripheral and Cranial Nerve and Other

Nervous System Procedure).\1\ Procedure code 38.7 (Interruption of the

vena cava) is sometimes performed in conjunction with treatment for the

principal diagnosis 434.11 (Cerebral embolism with infarction), which

is assigned to MDC 1. Our medical advisors believe that procedure code

38.7 is appropriately performed for some neurological conditions such

as a cerebral embolism with infarction. Because the current DRG

configuration does not allow this assignment, we proposed to add

procedure code 38.7 to DRGs 7 and 8.

---------------------------------------------------------------------------

\1\ A single title combined with two DRG numbers is used to

signify pairs. Generally, the first DRG is for cases with CC and the

second DRG is for cases without CC. If a third number is included,

it represents cases with patients who are age 0-17. Occasionally, a

pair of DRGs is split between age >17 and age 0-17.

---------------------------------------------------------------------------

Second, we proposed that procedure codes 83.92 (Insertion or

replacement of skeletal muscle stimulator) and 83.93 (Removal of

skeletal muscle stimulator) both be categorized with other procedures

on the nervous system. These procedures can be performed on patients

with a principal diagnosis in MDC 1, such as 344.00 (Quadriplegia

unspecified) or 344.31 (Monoplegia of lower limb, affecting dominant

side). Therefore, these two codes would also be assigned to DRGs 7 and

8.

Third, procedure code 39.50 (Angioplasty or atherectomy of

noncoronary vessel) is not currently assigned to MDC 4 (Diseases and

Disorders of the Respiratory System). This procedure is performed for

patients who develop pulmonary embolism. The principal diagnosis for

pulmonary embolism is in MDC 4, and, to increase clinical coherence, we

proposed to add procedure code 39.50 to that MDC in DRGs 76 and 77

(Other Respiratory System OR Procedures).

Fourth, insertion of totally implantable infusion pump (procedure

code 86.06) is not assigned to MDC 5 (Diseases and Disorders of the

Circulatory System) in the current DRG configuration. Infusion pumps

should

[[Page 41496]]

be assigned to all MDCs in which subcutaneous insertion of the pump is

appropriate. Procedure code 86.06 may be performed on patients with a

principal diagnosis in MDC 5 such as 451.83 (Phlebitis and

thrombophlebitis of the deep veins of other extremities). Therefore, we

proposed to add procedure code 86.06 to DRG 120 (Other Circulatory

System OR Procedures) in MDC 5.

We received two comments on these MDC and DRG assignments, both of

which concurred with our proposed changes. Therefore, we are adopting

them as final.

b. Reassignment of Procedures Among DRGs 468, 476, and 477

We also reviewed the list of procedures that produce assignments to

DRGs 468, 476, and 477 to ascertain if any of those procedures should

be moved from one of these DRGs to another based on average charges and

length of stay. Generally, we move only those procedures for which we

have an adequate number of discharges to analyze the data. Based on our

review this year, we did not propose to move any procedures from DRG

468 to DRGs 476 or 477, from DRG 476 to DRGs 468 or 477, or from DRG

477 to DRGS 468 or 476.

8. Changes to the ICD-9-CM Coding System

As described in section II.B.1 of this preamble, the ICD-9-CM is a

coding system that is used for the reporting of diagnoses and

procedures performed on a patient. In September 1985, the ICD-9-CM

Coordination and Maintenance Committee was formed. This is a Federal

interdepartmental committee, co-chaired by the National Center for

Health Statistics (NCHS) and HCFA, that is charged with the mission of

maintaining and updating the ICD-9-CM system. That mission includes

approving coding changes, and developing errata, addenda, and other

modifications to the ICD-9-CM to reflect newly developed procedures and

technologies and newly identified diseases. The Committee is also

responsible for promoting the use of Federal and non-Federal

educational programs and other communication techniques with a view

toward standardizing coding applications and upgrading the quality of

the classification system.

The NCHS has lead responsibility for the ICD-9-CM diagnosis codes

included in the Tabular List and Alphabetic Index for Diseases, while

HCFA has lead responsibility for the ICD-9-CM procedure codes included

in the Tabular List and Alphabetic Index for Procedures.

The Committee encourages participation in the above process by

health-related organizations. In this regard, the Committee holds

public meetings for discussion of educational issues and proposed

coding changes. These meetings provide an opportunity for

representatives of recognized organizations in the coding field, such

as the American Health Information Management Association (AHIMA)

(formerly American Medical Record Association (AMRA)), the American

Hospital Association (AHA), and various physician specialty groups as

well as physicians, medical record administrators, health information

management professionals, and other members of the public, to

contribute ideas on coding matters. After considering the opinions

expressed at the public meetings and in writing, the Committee

formulates recommendations, which then must be approved by the

agencies.

The Committee presented proposals for coding changes for FY 2000 at

public meetings held on June 4 and November 2, 1998. Even though the

Committee conducted public meetings and considered approval of coding

changes for FY 2000 implementation, we are not implementing any changes

to ICD-9-CM codes for FE 2000. We have undertaken, and continue to

undertake, major efforts to ensure that all of the Medicare computer

systems are ready to function on January 1, 2000. If we were to make

system changes to capture additions, deletions, and modifications to

ICD-9-CM codes for FY 2000, we would endanger the functioning of the

Medicare computer systems, and, specifically, we might compromise our

ability to process hospital bills. Therefore, the code proposals

presented at the public meetings held on June 4 and November 2, 1998,

that (if approved) ordinarily would have been included as new codes for

October 1, 1999, are not included in this final rule. These code

changes to ICD-9-CM will be considered for inclusion in the annual

update for FY 2001. The initial meeting for consideration of coding

changes for implementation in FY 2001 was held on May 13, 1999.

Copies of the minutes of the 1998 meetings and the May 13, 1999

meeting can be obtained from the HCFA Home Page at http://www.hcfa.gov/

medicare/icd9cm.htm or from http://www.hcfa.gov/events, click on

``meetings and workshops'' link, and then click on ``reports of the

ICD-9-CM coordination and maintenance committee'' link. Paper copies of

these minutes are no longer available and the mailing list has been

discontinued. We encourage commenters to address suggestions on coding

issues involving diagnosis codes to: Donna Pickett, Co-Chairperson;

ICD-9-CM Coordination and Maintenance Committee; NCHS; Room 1100; 6525

Belcrest Road; Hyattsville, Maryland 20782. Comments may be sent by E-

mail to [email protected].

Questions and comments concerning the procedure codes should be

addressed to: Patricia E. Brooks, Co-Chairperson; ICD-9-CM Coordination

and Maintenance Committee; HCFA, Center for Health Plans and Providers,

Plan and Provider Purchasing Policy Group, Division of Acute Care; C4-

07-07; 7500 Security Boulevard; Baltimore, Maryland 21244-1850.

Comments may be sent by E-mail to [email protected].

We received one comment in support of our decision not to update

ICD-9-CM codes given the magnitude of system changes needed during the

period leading up to the year 2000.

9. Other Issues

a. Implantation of Muscle Stimulator

In the July 31, 1998 final rule, we responded to a comment on the

DRG assignment for implantation of a muscle stimulator (63 FR 40964).

In that document, we stated that we would readdress this issue after

reviewing the FY 1998 MedPAR file.

There is concern in the manufacturing industry that the current DRG

assignment for the implantation of a muscle stimulator and the

associated tendon transfer for quadriplegics is inappropriate. When the

procedures are performed during two separate admissions, the tendon

transfer (procedure code 82.56 (Other hand tendon transfer or

transplantation)) is assigned to DRGs 7 and 8, and the insertion of the

muscle stimulator (procedure code 83.92 (Insertion or replacement of

skeletal muscle stimulator)) is assigned to DRG 468. However, when both

procedures are performed in the same admission, the case is assigned to

DRGs 7 and 8.

As discussed in section II.B.7.a of this preamble, in the May 7,

1999 proposed rule, we proposed to assign code 83.92 to DRGs 7 and 8 in

MDC 1. Therefore, if a case involves either procedure code 82.56 or

83.92, or both procedure codes, the case would be assigned to DRGs 7

and 8.

A presentation on one type of muscle stimulator was made by a

device manufacturer before the ICD-9-CM Coordination and Maintenance

Committee on November 2, 1998. The manufacturer strongly suggested that

a

[[Page 41497]]

new code assignment be made for the procedure for insertion of this

stimulator and that it be placed in category 04.9 (Other operations on

cranial and peripheral nerves). However, based on comments received by

the Committee, there was an overwhelming response from the coding

community that a new code should not be created. The commenters believe

that these codes (82.56 and 83.92) adequately described the procedures

since the patient receives a tendon transfer in addition to the

skeletal muscle stimulator insertion. This is done so that the

quadriplegic patient can achieve some hand grasping ability where there

was none before. Some quadriplegic patients receive the tendon transfer

on one admission and the stimulator insertion on a subsequent

admission. Others have both procedures performed on the same admission.

Since the tendon transfer and stimulator insertion are being performed

on quadriplegic patients, a condition found in MDC 1, we proposed to

add procedure codes 82.56 and 83.92 to DRGs 7 and 8. We did not receive

any comments on this proposal. Therefore, we are adopting it as final.

b. Pancreas Transplant

Through a Medicare Coverage Issues Manual revision (Transmittal No.

115, April 1999), HCFA announced that, effective July 1, 1999, Medicare

covers whole organ pancreas transplantation (procedure codes 52.80 or

52.83) if it is performed simultaneous with or after a kidney

transplant.

Pancreas transplantation is generally limited to those patients

with severe secondary complications of diabetes, including kidney

failure. However, pancreas transplantation is sometimes performed on

patients with labile diabetes and hypoglycemic unawareness.

Pancreas transplantation for diabetic patients who have not

experienced end-stage renal failure secondary to diabetes continue to

be excluded from coverage. Medicare also excludes coverage of

transplantation of partial pancreatic tissue or islet cells. Claims

processing instructions to intermediaries were contained in Program

Memorandum Transmittal No. A-99-16 (April 1999).

We received one comment regarding the coverage and claims

processing instructions for pancreas transplants.

Comment: The commenter requested clarification on the date of

coverage for services related to pancreas transplantation services

furnished on or after July 1, 1999. Specifically, the commenter asked

whether coverage is effective for admissions, discharges, or actual

transplant surgery on or after that date. In addition, the commenter

believes that if the resource use for a pancreas-kidney transplant is

significantly greater than for a kidney transplant alone, then a new

DRG should be created for the dual transplant. Finally, the commenter

was unsure how hospitals should report the organ acquisition costs

attributable to pancreas. Specifically, the commenter wanted to know if

the costs should be included, on the hospital cost report with the

kidney costs or whether a separate organ acquisition cost center will

be established for pancreas acquisition costs.

Response: As stated in Transmittal No. 115, coverage is effective

for dates of service on or after July 1, 1999. Therefore, any pancreas

transplant performed on or after July 1, 1999 is covered by Medicare if

all other qualifying criteria are met.

Under the current DRG classification, if a kidney transplant and a

pancreas transplant are performed simultaneously on a patient with

chronic renal failure secondary to diabetes with renal manifestations

(diagnosis codes 250.40 through 250.43), the case is assigned to DRG

302 (Kidney Transplant) in MDC 11 (Disease and Disorders of the Kidney

and Urinary Tract. If a pancreas transplant is performed following a

kidney transplant (that is, in a different hospital admission) on a

patient with chronic renal failure secondary to diabetes with renal

manifestations, the case is assigned to DRG 468 (Major OR Procedure

Unrelated to Principal Diagnosis) because pancreas transplant is not

assigned to MDC 11, the MDC to which a principal diagnosis of chronic

renal failure secondary to diabetes is assigned.

If a kidney and pancreas transplant are performed simultaneously or

if a pancreas transplant is performed following a kidney transplant, on

a patient with chronic renal failure secondary to diabetes with

ketoacidosis (diagnosis codes 250.10 through 250.13), diabetes with

hyperosmolarity (diagnosis codes 250.20 through 250.23), diabetes with

other coma (diagnosis codes 250.30 through 250.33), diabetes with other

specified manifestations (diagnosis codes 250.80 through 250.83), or

diabetes with unspecified complication (diagnosis codes 250.90 through

250.93), the case would be assigned to DRG 292 or 293 (Other Endocrine,

Nutritional and Metabolic OR Procedures) in MDC 10 (Endocrine,

Nutritional, and Metabolic Diseases and Disorders). As the commenter

notes, it is possible that the resource use for a pancreas-kidney

transplant or a pancreas-only transplant might be significantly

different from a kidney-only transplant. We intend to review the

Medicare data in our FY 1999 MedPAR file in order to analyze whether we

should either reassign these transplants to a different DRG or create a

new DRG. We will announce any proposals on that issue in the FY 2001

proposed rule, which will be published in the Spring of 2000.

A separate organ acquisition cost center has been established for

pancreas transplantation. The Medicare cost report will include a

separate line to account for pancreas transplantation costs. In

addition, in this final rule, we are making a conforming change to '

412.2(e)(4) to include pancreas in the list of organ acquisition costs

that are paid on a reasonable cost basis.

c. Immunotherapy

Effective October 1, 1994, procedure code 99.28 (Injection or

infusion of biological response modifier [BRM] as an antineoplastic

agent) was created. This procedure is also known as BRM therapy or

immunotherapy. At that time, we designated the code as a [email protected] code

that does not affect DRG assignment.

Comment: One commenter, a manufacturer of a biologic response

modifier, requested that we create a new DRG for BRM therapy or assign

cases in which BRM therapy is performed to an existing DRG with a high

relative weight. The commenter suggested that DRG 403 (Lymphoma and

Non-Acute Leukemia with CC) would be an appropriate DRG. The

manufacturer=s particular drug is used in the treatment of metastatic

renal cell carcinoma and metastatic melanoma.

Response: Using the 100 percent FY 1998 MedPAR file that contains

bills through December 31, 1998, we performed an analysis of the cases

for which procedure code 99.28 was reported. Based on the commenter's

request, for purposes of this analysis we examined cases only for

hospitals that use the particular drug manufactured by the commenter.

We identified 121 cases in 19 DRGs in 9 MDCs. No more than 31 cases

were assigned to any one particular DRG. Of the 121 cases identified,

31 cases were assigned to DRG 318 (Kidney and Urinary Tract Neoplasms

with CC) and 30 of the cases were assigned to DRG 82 (Respiratory

Neoplasms). There was a wide range of charges (between approximately

$1,300 and $125,000 per case) associated with this therapy. The average

length of stay was approximately 5 days. Due to the limited number of

cases that were

[[Page 41498]]

distributed throughout 19 DRGs and the variation of charges, we

concluded that it would be inappropriate to classify these cases into a

single DRG. Because of the numerous principal diagnoses reported with

BRM therapy, a single DRG for procedure code 99.28 would need to be

placed in the pre-MDC DRG category. Similarly, it would be impossible

to classify these cases into DRG 403 because only a few cases were

coded with a principal diagnosis assigned to MDC 17 (Myeloproliferative

Diseases and Disorders, and Poorly Differentiated Neoplasms), the MDC

that includes DRG 403. Finally, the variation in charges reflected in

the 121 cases do not persuade us that there is an analytic basis for

combining these cases into one DRG. Using the FY 1999 MedPAR, we intend

to do a full analysis of these cases, which we will discuss in the FY

2001 proposed rule.

As a final note, any DRG classification change for procedure code

99.28 must be appropriate for all cases that receive BRM therapy, not

just those that use the commenter's drug. Even if we might consider

such an assignment appropriate, we have no way to distinguish between

different drug therapies assigned to the same procedure code. The FY

1998 MedPAR file we analyzed contained 930 cases with procedure code

99.28. These 930 cases were assigned to 18 MDCs.

d. Heart Assist Devices

Effective May 5, 1997, we revised Medicare coverage of heart assist

devices to allow coverage of a ventricular assist device used for

support of blood circulation postcardiotomy if certain conditions were

met. In the August 29, 1997 final rule (62 FR 45973), we moved

procedure code 37.66 (Implant of an implantable pulsatile heart assist

device) from DRGs 110 and 111 (Major Cardiovascular Procedures) to DRG

108 (Other Cardiothoracic Procedures) to improve payment for these

procedures. In the July 31, 1998 final rule (63 FR 40956), in a further

effort to improve payment for these cases, we moved procedure code

37.66 to DRGs 104 and 105 (Cardiac Valve and Other Major Cardiothoracic

Procedures).

We received one comment regarding the DRG classification of

procedure code 37.66.

Comment: The commenter recommended that we either reclassify heart

assist device cases to DRG 103 (Heart Transplant) or create a new DRG

specifically for this device and technology. The commenter cited a

discrepancy between the cost of the device implantation and payment for

DRGs 104 and 105 as the basis for these recommendations.

Response: We refer the reader to our response to a similar comment

in the August 29, 1997 final rule (62 FR 45967). We note that the FY

1998 MedPAR file has 22 cases coded with procedure code 37.66. Of these

22 cases, 8 cases were assigned to DRG 103 (Heart Transplant) and 4

cases to DRG 483 (Tracheostomy Except for Face, Mouth, and Neck

Diagnoses). The remaining 10 cases would have been assigned to DRGs 104

and 105 under the current classification.

C. Recalibration of DRG Weights

We proposed to use the same basic methodology for the FY 2000

recalibration as we did for FY 1999. (See the July 31, 1998 final rule

(63 FR 40965).) That is, we recalibrated the weights based on charge

data for Medicare discharges. However, we used the most current charge

information available, the FY 1998 MedPAR file. (For the FY 1999

recalibration, we used the FY 1997 MedPAR file.) The MedPAR file is

based on fully coded diagnostic and surgical procedure data for all

Medicare inpatient hospital bills.

The final recalibrated DRG relative weights are constructed from FY

1998 MedPAR data, based on bills received by HCFA through March 1999,

from all hospitals subject to the prospective payment system and short-

term acute care hospitals in waiver States. The FY 1998 MedPAR file

includes data for approximately 11.3 million Medicare discharges.

The methodology used to calculate the DRG relative weights from the

FY 1998 MedPAR file is as follows:

All the claims were regrouped using the DRG classification

revisions discussed above in section II.B of this preamble.

Charges were standardized to remove the effects of

differences in area wage levels, indirect medical education (IME) and

disproportionate share hospital (DSH) payments, and, for hospitals in

Alaska and Hawaii, the applicable cost-of-living adjustment.

The average standardized charge per DRG was calculated by

summing the standardized charges for all cases in the DRG and dividing

that amount by the number of cases classified in the DRG.

We then eliminated statistical outliers, using the same

criteria as were used in computing the current weights--that is, all

cases that are outside of 3.0 standard deviations from the mean of the

log distribution of both the charges per case and the charges per day

for each DRG.

The average charge for each DRG was then recomputed

(excluding the statistical outliers) and divided by the national

average standardized charge per case to determine the relative weight.

A transfer case is counted as a fraction of a case based on the ratio

of its length of stay to the geometric mean length of stay of the cases

assigned to the DRG. That is, a 5-day length of stay transfer case

assigned to a DRG with a geometric mean length of stay of 10 days is

counted as 0.5 of a total case.

We established the relative weight for heart and heart-

lung, liver, and lung transplants (DRGs 103, 480, and 495) in a manner

consistent with the methodology for all other DRGs except that the

transplant cases that were used to establish the weights were limited

to those Medicare-approved heart, heart-lung, liver, and lung

transplant centers that have cases in the FY 1998 MedPAR file.

(Medicare coverage for heart, heart-lung, liver, and lung transplants

is limited to those facilities that have received approval from HCFA as

transplant centers.)

Acquisition costs for kidney, heart, heart-lung, liver,

and lung transplants continue to be paid on a reasonable cost basis.

Unlike other excluded costs, the acquisition costs are concentrated in

specific DRGs (DRG 302 (Kidney Transplant); DRG 103 (Heart Transplant

for Heart and Heart-Lung Transplants); DRG 480 (Liver Transplant); and

DRG 495 (Lung Transplant)). Because these costs are paid separately

from the prospective payment rate, it is necessary to make an

adjustment to prevent the relative weights for these DRGs from

including the effect of the acquisition costs. Therefore, we subtracted

the acquisition charges from the total charges on each transplant bill

that showed acquisition charges before computing the average charge for

the DRG and before eliminating statistical outliers.

When we recalibrated the DRG weights for previous years, we set a

threshold of 10 cases as the minimum number of cases required to

compute a reasonable weight. We used that same case threshold in

recalibrating the DRG weights for FY 2000. Using the FY 1998 MedPAR

data set, there are 40 DRGs that contain fewer than 10 cases. We

computed the weights for the 40 low-volume DRGs by adjusting the FY

1999 weights of these DRGs by the percentage change in the average

weight of the cases in the other DRGs.

The weights developed according to the methodology described above,

using the final DRG classification changes, result in an average case

weight that is different from the average case weight

[[Page 41499]]

before recalibration. Therefore, the new weights are normalized by an

adjustment factor, so that the average case weight after recalibration

is equal to the average case weight before recalibration. This

adjustment is intended to ensure that recalibration by itself neither

increases nor decreases total payments under the prospective payment

system.

Section 1886(d)(4)(C)(iii) of the Act requires that, beginning with

FY 1991, reclassification and recalibration changes be made in a manner

that ensures that the aggregate payments are neither greater than nor

less than the aggregate payments that would have been made without the

changes. Although normalization is intended to achieve this effect,

equating the average case weight after recalibration to the average

case weight before recalibration does not necessarily achieve budget

neutrality with respect to aggregate payments to hospitals because

payment to hospitals is affected by factors other than average case

weight. Therefore, as we have done in past years and as discussed in

section II.A.4.b of the Addendum to this final rule, we make a budget

neutrality adjustment to ensure that the requirement of section

1886(d)(4)(C)(iii) of the Act is met.

D. Use of Non-MedPAR Data for Reclassification and Recalibration of the

DRGs

1. Introduction

As in past years, in the DRG reclassification and recalibration

process for the FY 2000 final rule, we used the MedPAR file, which

consists of data for approximately 11.3 million Medicare discharges. In

the FY 1999 final rulemaking process, we used the FY 1997 MedPAR file

to recalibrate DRGs and evaluate possible changes to DRG

classifications; for this FY 2000 final rule, we used the FY 1998

MedPAR file. The Conference Report that accompanied the Balanced Budget

Act of 1997 stated that ``in order to ensure that Medicare

beneficiaries have access to innovative new drug therapies, the

conferees believe that HCFA should consider, to the extent feasible,

reliable, validated data other than Medicare Provider Analysis and

Review (MedPAR) data in annually recalibrating and reclassifying the

DRGs'' (H.R. Conf. Rep. No. 105-217 at 734 (1997)).

Consistent with that language, we considered non-MedPAR data in the

rulemaking process for FY 1999 and in developing the May 7, 1999

proposed rule for FY 2000. We received non-MedPAR data from entities on

behalf of the manufacturer of a specific drug, platelet inhibitors. The

manufacturer was seeking to obtain a new DRG assignment for cases

involving platelet inhibitors. The non-MedPAR data purported to show

cases involving platelet inhibitors. As discussed in the proposed rule,

we concluded it was not feasible to use the non-MedPAR data submitted

to us because, among other things, we did not have information to

verify that the cases actually involved the drug, nor did we have

information to verify that the cases reflected a representative sample

(and did not simply reflect high cost cases).

Effective October 1, 1998, we implemented a code for platelet

inhibitors, but until we receive bills for Medicare discharges

occurring during FY 1999, the MedPAR data do not enable us to

distinguish between cases with platelet inhibitors and cases without

platelet inhibitors (63 FR 40963). Representatives of the

pharmaceutical company first presented us with non-MedPAR data during

the rulemaking process for FY 1999. The data were compiled by a health

information company, and purported to show, for cases from a sample of

hospitals, the average standardized charges (as calculated by the

health information company) for different classes of patients.

In the FY 1999 final rule, we stated a number of reasons for

rejecting the non-MedPAR data we had received. Basically, the data were

unreliable and the data's use was not feasible--the data could not be

validated or verified.

After publication of the July 31, 1998 final rule, we met and

corresponded on several occasions with the manufacturers, vendors, and

legal representatives of the pharmaceutical company in an effort to

resolve data issues. We reiterated that, among other things, we needed

to know for each case the hospital that furnished the services. Before

the publication of the proposed rule, we had not received information

necessary to validate the data or the data's representativeness.

We remain open to considering non-MedPAR data in the DRG

reclassification and recalibration process, but, consistent with the

Conference Report, as well as our longstanding policies, the data must

be ``reliable'' and ``validated.'' The July 31, 1998 final rule

reflected the major factors that we consider in evaluating whether data

are feasible, reliable, and validated; however, because we believed it

might be useful, we discussed these issues in much greater detail in

the May 7, 1999 proposed rule.

2. The DRG Reclassification and Recalibration Process

In order to understand whether it is feasible to use non-MedPAR

data, and whether the data are reliable and validated, it is critical

to understand the DRG recalibration and reclassification process. As

described earlier, one of the first steps in the annual DRG

recalibration is that the Medicare hospital inpatient claims (in the

MedPAR file) from the preceding Federal fiscal year are classified

using the DRG classification system (proposed or final) for the

upcoming year. Cases are classified into DRGs based on the principal

diagnosis, up to eight additional diagnoses, and up to six procedures

performed during the stay, as well as age, sex, and discharge status of

the patient. Each case is classified into one and only one DRG.

As the term suggests, the relative weight for each DRG reflects

relative resource use. The recalibration process requires data that

enable us to compare resource use across DRGs. As explained earlier, as

part of the recalibration process, we standardize the charges reflected

on each Medicare claim to remove the effects of area wage differences,

the IME adjustment, and the DSH adjustment; in order to standardize

charges, we need to know which hospital furnished the service. For each

DRG, we calculate the average of the standardized charges for the cases

classified to the DRG. To calculate DRG relative weights, we compare

average standardized charges across DRGs.

In evaluating whether it is appropriate to reclassify cases from

one DRG to another, we examine the average standardized charges for

those cases. The recalibration process and the reclassification process

are integrally related; to evaluate whether cases involving a certain

procedure should be reclassified, we need to have information that (1)

enables us to identify cases that involve the procedure and cases that

do not involve the procedure, and (2) enables us to determine

appropriate DRG relative weights if certain cases are reclassified.

3. Feasible, Reliable, Validated Data

As indicated above, the Conference Report reflected the conferees'

belief that, ``to the extent feasible,'' HCFA should consider

``reliable, validated data'' in recalibrating and reclassifying DRGs.

The concepts of reliability and validation are closely related. In

order for us to use non-MedPAR data, the non-MedPAR data must be

independently validated. When an entity submits non-MedPAR data, we

[[Page 41500]]

must be able to independently review the medical records and verify

that a particular procedure was performed for each of the cases that

purportedly involved the procedure. This verification requires the

identification of a particular Medicare beneficiary and the hospital

where the beneficiary was treated, as well as the dates involved.

Although it is unlikely that we would review 100 percent of thousands

of cases submitted for review, at a minimum, we must be able to

validate data through a random sampling methodology. We must also be

able to verify the charges that are reflected in the data.

Independent validation is particularly critical in part because the

non-MedPAR data might be submitted by (or on behalf of) entities that

have a financial interest in obtaining a new DRG assignment and in

obtaining the highest possible DRG relative weight. If we receive non-

MedPAR data that purport to reflect cases involving a certain procedure

and a certain level of charges, we must have some way to verify the

data.

Even if non-MedPAR data are reliable and verifiable, that does not

mean it is necessarily ``feasible'' to use the data for purposes of

recalibration and reclassification. In order to be feasible for these

purposes, the non-MedPAR data must enable us to appropriately measure

relative resource use across DRGs. It is critical that cases are

classified into one and only one DRG in the recalibration process, and

that we have information that enables us to standardize charges for

each case and determine appropriate DRG relative weights. Moreover, the

data must reflect a complete set of cases or, at a minimum, a

representative sample of hospitals and claims.

If cases are classified into more than one DRG (or into the

incorrect DRG) in the recalibration process, or if the non-MedPAR data

reflect an unrepresentative sample of cases, the measure of relative

resources would be distorted. For example, cases of percutaneous

transluminal coronary angioplasty (PTCA) treated with GPIIb/IIIa

platelet inhibitors (procedure code 99.20) are currently classified to

DRG 112. Prior to the publication of the proposed rule, the same drug

manufacturer discussed above provided us with information on the

average charges for a sample of cases that purportedly involve PTCA,

for the purpose of evaluating whether these cases should be moved to

the higher-weighted DRG 116. However, without adequate identification

of the cases to allow us to specifically identify all of the cases

treated with platelet inhibitors, the relative weight for DRG 112 would

reflect the costs of platelet inhibitor cases. This distortion would

result in excessive payments under DRG 112, and thus undermine the

integrity of the recalibration process.

Therefore, in order for the use of non-MedPAR data to be feasible,

generally we must be able to accurately and completely identify all of

the cases to be reclassified from one DRG to another. At a minimum, we

must have some mechanism for ensuring that DRG weights are not

inappropriately inflated (or deflated) to the extent that a DRG weight

reflects cases that would be reclassified to a different DRG.

In short, then, for use of non-MedPAR data to be feasible for

purposes of DRG recalibration and reclassification, the data must,

among other things (1) be independently verifiable, (2) reflect a

complete set of cases (or a representative sample of cases), and (3)

enable us to calculate appropriate DRG relative weights and ensure that

cases are classified to the ``correct'' DRG, and to one DRG only, in

the recalibration process.

4. Submission of Data

Finally, in order for use of non-MEDPAR data to be feasible, we

must have sufficient time to evaluate and test the data. The time

necessary to do so depends upon the nature and quality of the data

submitted. Generally, however, a significant sample of the data should

be submitted by August 1, approximately 8 months prior to the

publication of the proposed rule, so that we can test the data and make

a preliminary assessment as to the feasibility of the data's use.

Subsequently, a complete database should be submitted no later than

December 1 for consideration in conjunction with the next year's

proposed rule.

5. How the Prospective Payment System Ensures Access to New

Technologies

As noted at the outset of this discussion, the Conference Report

that accompanied the BBA indicated that we should consider non-MEDPAR

data, to the extent feasible, ``in order to ensure that Medicare

beneficiaries have access to innovative new drug therapies'' (H.R.

Conf. Rep. No. 105-217 at 734 (1997)). There seems to be a concern

that, if a new technology is introduced, and if the new technology is

costly, then Medicare would not make adequate payment if the new

technology is not immediately placed in a new DRG. This concern is

unfounded. As explained below, the Medicare hospital inpatient

prospective payment does ensure access to new drug therapies, and to

new technologies in general.

First, to the extent a case involving a new technology is extremely

costly relative to the cases reflected in the DRG relative weight, the

hospital might qualify for outlier payments, that is, additional

payments over and above the standard prospective payment rate.

Second, Medicare promotes access to new technologies by making

payments under the prospective payment system that are designed to

ensure that Medicare payments for a hospital's cases as a whole are

adequate. We establish DRGs based on factors such as clinical coherence

and resource utilization. Each diagnosis-related group encompasses a

variety of cases, reflecting a range of services and a range of

resources. Generally, then, each DRG reflects some higher cost cases

and some lower cost cases.

For some cases, the hospital's costs might be higher than the

payment under the prospective payment system; this does not mean that

the DRG classifications are ``inappropriate.'' For other cases, the

hospital's costs will be lower than the payment under the prospective

payment system. We believe that Medicare makes appropriate payments for

a hospital's cases as a whole.

Each year we examine the best data available to assess whether DRG

changes are appropriate and to recalibrate DRG relative weights. As we

have indicated on numerous occasions, it usually takes 2 years from the

time a procedure is assigned a code to collect the appropriate MedPAR

data and then make an assessment as to whether a DRG change is

appropriate. This timetable applies to reclassifications that would

lead to decreased payment as well as those that would increase payment.

In fact, the introduction of new technologies itself might lead to

either higher than average costs or lower costs.

Our ability to evaluate and implement potential DRG changes depends

on the availability of validated, representative data. We believe that

our policies ensure access to new technologies and are critical to the

integrity of the recalibration process. We still remain open to using

non-MedPAR data if the data are reliable and validated and enable us to

appropriately measure relative resource use.

We received a number of comments regarding this issue, including

comments from MedPAC, pharmaceutical manufacturers (including two

manufacturers of platelet inhibitor drugs), an industry manufacturers'

association, and several

[[Page 41501]]

cardiologists. We received only one comment from a State hospital

association; otherwise, hospital associations were silent on this

issue.

Comment: MedPAC stated that HCFA's general criteria provide a valid

basis for assessing the feasibility and appropriateness of using

outside data to establish DRG assignments and relative weights for

specific technologies. MedPAC believes that it would be helpful to

entities that desire to submit useful data if HCFA would establish and

publish explicit data standards to guide their efforts. MedPAC

suggested the criteria might include the format and content of the

patient care records; the minimum sample size; required documentation

of sampling procedures; acceptable methods for ensuring that the

sampled providers were representative of the relevant provider

universe; and any other information that HCFA considered essential to

establish the validity and reliability of the submitted data. MedPAC

believes that the criteria would help to prevent misunderstandings and

ensure HCFA's ability to assess whether the submitted data were

adequate to serve as a basis for DRG assignment before actual MedPAR

claims become available.

Response: We appreciate the Commission's support of our general

criteria. We would prefer to gain further experience working with non-

MedPAR data before we develop any specific criteria regarding sample

sizes or methodologies. This will enable us to establish criteria that

realistically reflect the availability of such data and the general

suitability of the data for use in the DRG reclassification and

recalibration process. Our intent at this time is to address some

fundamental criteria that must be taken into consideration by outside

parties interested in submitting non-MedPAR data.

We note that the timetable we set forth in the proposed rule is

intended to provide adequate opportunity to permit outside parties to

conform their data to our needs through testing and resubmission. This

is the primary reason we believe it is generally necessary to have a

sample of the data 8 months prior to the publication of the proposed

rule. We are willing to meet with outside parties interested in

submitting non-MedPAR data for consideration, and would suggest that

those interested in submitting such data in the future should contact

us to discuss the specific data they wish to submit and whether the

data may be adequate.

Comment: One commenter, while supporting the idea that the data

must be reliable and verifiable, indicated that HCFA should consider

other means by which to accomplish this purpose. The commenter stated

that many of the sources for data are restricted from releasing

identifying elements of the data they collect. The commenter claimed,

for example, that they could validate the method by which the data were

assembled, thereby alleviating our concern that the cases may not

represent Medicare beneficiaries or that the reported charges are

inaccurate.

Response: We are open to considering any feasible method for

validating non-MedPAR data, and that is why at this time we are not

specifying explicit criteria for the types of data we will or will not

consider. Instead, we have outlined general guidelines and fundamental

objectives that must be met. One of those fundamental objectives is

that we must be able to validate the data and to accurately identify

cases to be reclassified during DRG recalibration.

In order to preserve the integrity of the DRG reclassification and

recalibration process, we generally believe it is imperative that we

are able to independently validate the data submitted. As noted

previously, if we receive non-MedPAR data that purport to reflect cases

involving a certain procedure and a certain level of charges, we must

have some way to verify that data. In addition, it is not enough to

simply decide that a particular diagnosis or procedure code should now

be classified to a higher-weighted DRG. Cases in the MedPAR data used

for recalibration with that diagnosis or procedure code should be

reclassified accordingly. Otherwise, these cases will affect the

calculation of the relative weights of other DRGs. Therefore, in order

to allow us to ensure the accuracy of DRG recalibration, we must have

some mechanism for ensuring that DRG weights are not inappropriately

inflated.

Comment: Some commenters stated that the criteria regarding the

feasibility of using the data are inconsistent with the intent of the

Conference Report language. The commenters contend that there is no

need to identify each case involving a new technology. Rather, the

agency can extrapolate the findings from a representative sample of

cases and estimate which cases must be moved from one DRG to another.

Two of the commenters stated that this approach was used in

reclassifying lithotripsy to an appropriate DRG, and that extrapolation

is used to some degree in setting the physician fee schedule and was

used in the proposed outpatient prospective payment system. One

commenter wanted us to clarify that we would accept a representative,

statistically valid sample of both non-HCFA and HCFA data that reflect

cases for a period of less than a full year, as well as requesting that

we specify the sources (for example, private payers, manufacturers of

medical technologies, or suppliers) from which we are willing to accept

such data.

Response: We did not rule out the use of extrapolation based on

non-MedPAR data in the proposed rule. In fact, we stated that the data

must reflect either a complete set of cases, or, at a minimum, a

representative sample of hospitals and claims. However, as stated

previously, the process of recalibrating the DRG weights requires that

cases be moved consistent with the reclassification of diagnosis or

procedure codes from one DRG to another. Failure to do so could lead to

inflated or deflated relative weights, which, in turn, result in over

or underpayments for cases in the affected DRGs.

We are attempting to accommodate the realities faced by outside

parties as they attempt to collect and present non-MedPAR data for

consideration. In addition, we will continue to explore our processes

for ways to incorporate such data while preserving the empirical and

clinical integrity of the recalibration process.

As noted by two commenters, in the September 3, 1986 final rule (51

FR 31486), we did, based on analysis by the Prospective Payment

Assessment Commission (ProPAC), assign all cases involving a principal

diagnosis of urinary stones treated by extracorporeal shock wave

lithotripsy (ESWL) to DRG 323 (Urinary stones, age >69 and/or CC).

Prior to this DRG change, ESWL cases were assigned to either DRG 323 or

DRG 324, depending on the presence of a CC or based on the patients age

(over 69). The Commission, an independent advisory body established by

Congress (and MedPAC's predecessor organization), obtained information

on ESWL procedure costs and other routine and ancillary hospital

service charges from the American Heart Association (AHA), the American

Urological Association, and seven hospitals that furnished ESWL. In

addition, ProPAC obtained a preliminary summary of a study conducted by

the Institute for Health Policy Analysis at Georgetown University

Medical Center. This study included cost data from 16 hospitals that

furnished lithotripsy. At the time of these studies, approximately 50

hospitals were furnishing ESWL. Because the ProPAC data were obtained

directly from hospitals and were verified by the Commission at the

[[Page 41502]]

hospital level, we believed the data were reliable and used the data as

a basis for reassigning ESWL cases to DRG 343 only. A full explanation

of the study and ProPAC's analysis and recommendations can be found in

the Technical Appendixes that accompanied ProPAC's April 1, 1986 Report

to Congress.

We have not precluded using either external or internal data that

represent less than a full year's worth of cases. For example, we could

examine a partial year's worth of cases from the current Federal fiscal

year rather than the preceding year's complete MedPAR. Once again,

however, a feasible approach must be developed to enable the

appropriate classification and recalibration of the DRG weights.

Finally, we do not believe it is necessary, or appropriate, to

identify in advance the sources from which we are willing to accept

data. At this time, we remain open to considering any data source that

is reliable, verifiable, and feasible. We would note, however, that

involving hospitals in any data collection would probably aid HCFA in

any validation effort. Generally, if we receive non-MedPAR data, we

will be contacting the hospitals that furnished the sources to verify

some or all of the data.

Comment: Two commenters stated the timeframe for submission of the

non-MedPAR data is unreasonable. They suggested that the submission of

data 7 months before the updated DRGs take effect (March 1) in the case

of internal HCFA data, and 8 months (February 1) in the case of

external data, would more appropriately ensure beneficiary access.

Response: The length of time necessary to validate non-MedPAR data

depends on the nature and quality of the data. In the proposed rule, we

stated that a significant sample of the data should be submitted by

August 1, approximately 8 months prior to the publication of the

proposed rule, so that we can verify and test the data and make a

preliminary assessment as to the feasibility of the data's use.

Subsequently, a complete database should be submitted no later than

December 1, approximately 4 months prior to the publication of the

proposed rule.

We do not believe that this timeframe is unreasonable. If we were

to adopt the commenter's suggestion, we would receive non-MedPAR data

only 2 months before the proposed rule is scheduled to be published

(April 1). This might not allow us sufficient time to ensure that the

data are reliable or valid prior to their use in preparing the proposed

rule.

We believe the timeframe we set forth is necessary to enable us to

independently validate any non-MedPAR data submitted. In order to

verify the data's reliability and validity, we believe we need to

review a sufficient number of the medical records associated with the

data. Expecting us to be able to accomplish this in a matter of weeks

after receiving the data (which is all the time that would be available

for data received in February due to the requirement to begin the

process of reclassifying and recalibrating the proposed DRGs by the end

of February in order for the proposed rule to be published by April 1)

is unrealistic.

Comment: Many of the commenters, including the manufacturer of the

platelet inhibitor drug, national associations representing device and

drug manufacturers, and individual cardiologists, argued that our

current process has inhibited the development of new medical

technologies, and that the criteria for the use of non-MedPAR data are

unworkable and would further slow the development of new technologies.

Several commenters asserted that certain new technologies (including

platelet inhibitors) are denied to Medicare beneficiaries due to

insufficient payment.

Response: After 15 years of administering the prospective payment

system, we do not have any independent evidence that Medicare

beneficiaries are being denied access to new technologies by hospitals

or physicians. Although we have always acknowledged that there is a

time-lag between the time new technologies are introduced and the point

at which we can begin to accurately identify their associated costs, we

believe this has not hampered Medicare beneficiaries' access to these

new technologies. The fact that under the prospective payment system a

hospital might lose money on some cases but will gain money on other

cases is well understood by hospitals. We received no comments from

hospitals or beneficiary advocates complaining about access to new

technologies in general or drug therapies in particular, and only a

brief comment from a State hospital association that indicated that the

use of non-MedPAR data should extend beyond drug therapies.

Furthermore, as provided in Sec. 489.53(a)(2), HCFA may terminate its

participation agreement with any hospital if HCFA finds that the

hospital places restrictions on the persons it will accept for

treatment and it fails either to exempt Medicare beneficiaries from

those restrictions or to apply them to Medicare beneficiaries the same

as to all people seeking care.

Comment: Several commenters, including the manufacturer of a

platelet inhibitor drug and individual cardiologists, specifically

commented on our discussion in the proposed rule of the attempts by the

manufacturer of the drug to introduce its data into the process, with

the objective that cases in which platelet inhibitor therapy is

administered should be reclassified from DRG 112 (Permanent

Cardiovascular Procedures) to DRG 116 (Other Permanent Cardiac

Pacemaker Implant or PTCA with Coronary Artery Stent Implant) for FY

2000. The commenters stated that HCFA has been unwilling to consider

the data. One commenter stated that HCFA refused to accept these data

when they were offered in December 1998.

Response: As discussed in great detail above, and also in the FY

1999 final rule, our review of the previous data submitted by the drug

manufacturer found the data to be insufficient. Despite our

consultation with the manufacturer's representatives in advance of

their submission of data during the rulemaking process for FY 1999

(that is during the first half of calendar year 1998), in which we

advised them that we must be able to identify individual hospitals and

patients in order to utilize the data, this information was not

included on over 90 percent of the cases submitted in May 1998. As

noted in the May 7, 1999 proposed rule, we continued to meet and

correspond with the manufacturers, contractors, and legal

representatives of the pharmaceutical company in an effort to resolve

data issues. At no time have we refused to consider any data offered by

the company or its agents.

However, our discussions with these parties led us to the

conclusion that it might be helpful to identify general criteria for

submission of non-MedPAR data in the proposed rule. In particular, we

were concerned that outside parties wishing to submit non-MedPAR data

were unfamiliar with our current process and the importance of

accurately reclassifying and recalibrating the DRGs. The DRG relative

weights are the principle factor in adjusting the prospective payments

for each of approximately 11 million Medicare discharges each year. In

addition to the potential financial implications to the Medicare Trust

Fund and to hospitals themselves if these weights are inaccurate,

inappropriately assigning cases to higher-weighted DRGs may create

incentives that are not in the best interest of Medicare beneficiaries.

We are hopeful that, by explaining the general criteria for

submitting non-

[[Page 41503]]

MedPAR data and receiving public comments on those criteria, we can

help to ensure that in the future those interested in submitting non-

MedPAR data will be better informed regarding how the process can work.

In particular, we believe the timeframe we set will enable us to work

effectively with those interested in submitting non-MedPAR data to help

them provide data that can be used.

Comment: A manufacturer of a platelet inhibitor drug expressed

concern that HCFA may assign a special DRG classification for patients

who receive coronary intervention with an angioplasty and treatment

with platelet inhibitor therapy, but not for acute coronary syndrome

patients who receive the same drugs without coronary intervention.

These latter cases are assigned to DRG 124 (Circulatory Disorders

Except Acute Myocardial Infarction, with Cardiac Catheterization and

Complex Diagnoses) or DRG 140 (Angina Pectoris). The commenter stated

that if we were to modify payment for one use and not the other, it

would potentially create a financial incentive for expensive, risky,

and invasive treatment. Making payment provisions for both indications

at the same time, on the other hand, will give neither use an advantage

over the other. We were asked by the commenter to evaluate platelet

inhibitor therapy cases assigned to DRG 124 or DRG 140.

Response: Because this is the first comment we have received

regarding the noncoronary intervention use of the therapy, an extensive

study of DRGs 124 and 140 before publication of this final rule was not

feasible. We will evaluate this issue as part of our annual update for

FY 2001, when we will have MedPAR data capturing injection or infusion

of platelet inhibitor (ICD-9-CM procedure code 99.20). This commenter's

concern that increasing payment for one application of platelet

inhibitors but not for others could actually create an inappropriate

incentive in favor of a more invasive treatment, illustrates the

importance of proceeding cautiously in the process of DRG

reclassification and recalibration. We have a responsibility not to

inadvertently create financial incentives that adversely affect

clinical decisionmaking.

Comment: During the comment period, we received a revised set of

data from the manufacturer seeking to have platelet inhibitor therapy

cases receiving angioplasty reclassified from DRG 112 to DRG 116. The

data contain 27,673 cases from 164 hospitals in which Medicare patients

underwent an angioplasty. The commenter describes the data as Athe

public MedPAR file with an additional field that identifies the MedPAR

case as involving an angioplasty with or without platelet inhibitor

therapy. Thus, HCFA can identify the patient and the hospital from

these data such that they are reliable and verifiable. It also is a

representative sample of claims and, therefore, it is feasible for the

agency (HCFA) to use the data set. In light of the significant number

of angioplasty cases contained in the data, HCFA should be able to

utilize accepted statistical methods to extrapolate the results of

these data and recalibrate the DRG [email protected] The manufacturer indicated

that HCFA should reclassify angioplasty cases with platelet inhibitor

therapy on the basis of these data.

Included with the comment are tables summarizing the results of the

commenter's analysis of the data, showing that angioplasty cases

receiving platelet inhibitor therapy are more expensive than those not

receiving platelet inhibitors. According to the commenter, the

approximate average standardized charges for the different classes of

patients are as follows:

No drug, no stent: $19,877.

No drug, with stent: $22,968.

Drug, no stent: $26,389.

Drug, stent: $30,139.

Response: The submission of these data illustrates the problems of

attempting to ensure that non-MedPAR data are reliable, validated, and

feasible to use. Our greatest concern with respect to the data

submitted by the commenter is that we must validate the data to assess

whether they are reliable, and (as explained further below) this

validation process would take significant time and resources because

the data are not readily verifiable.

The data file submitted by the commenter is a MedPAR file with an

additional field. The commenter has ``marked'' certain cases in the

MedPAR file. The file contains variables named REO-FLAG and STENT-FLAG,

which purportedly indicate the case received the platelet inhibitor or

a coronary stent, respectively. However, the variables were placed in

the file by the commenter, based on information that was not made

available to HCFA; we did not receive any information to verify that

the cases flagged by the commenter involved platelet inhibitors.

Although we can use the FY 1998 MedPAR data to validate whether a case

received a coronary stent (because the FY 1998 MedPAR data include the

corresponding procedure code (36.06)), we cannot use the FY 1998 MedPAR

file by itself to validate whether a case involved platelet inhibitors

because the procedure code for the use of platelet inhibitors

(procedure code 99.20) was not effective until October 1, 1998.

Therefore, we cannot validate the data submitted to us without further

investigation.

In order to do so, we believe it is necessary to review the medical

records associated with the cases. Unless the entity submitting the

non-MedPAR data includes medical records (or other information that

would enable us to validate the data), the only method HCFA has to

review medical records is through Peer Review Organization (PRO)

review. Thus, we would need to request assistance in the PRO in each of

the States represented in the submitted data. The PROs would then

contact the hospitals involved to request copies of the medical

records. Finally, based on reviewing those records, the PROs would

notify HCFA whether the data can be validated.

Conducting a PRO independent validation would require a minimum of

2 to 3 months, and possibly much longer. Thus, there is not sufficient

time available to conduct a review of the data submitted by the drug

manufacturer. Since we cannot validate the data, it would compromise

the integrity of the DRG recalibration process to use these data in the

DRG reclassification and recalibration for FY 2000.

We note that the process used by the manufacturer to collect these

data is not specified. Based upon our prior discussions with the

manufacturer and its contractor that prepared the data, we believe the

164 hospitals represented in the sample have a contract for data

analysis and review with the consultant. Although we would not rule out

the possibility that this sample is statistically sufficient, we note

that in general, random sampling is necessary for generalization beyond

the sample itself.

The analysis submitted by the commenter is similar to that

presented in last year's final rule. As we indicated at that time, our

general process of waiting until we have identifiable MedPAR data

applies to changes that would enhance payment as well as those that

would decrease payment. Absent alternative data meeting the criteria

otherwise described in the proposed rule and in this final rule, we

cannot reclassify the administration of platelet inhibitors with

angioplasty (procedure code 99.20) from DRG 112 to DRG 116.

Comment: Some commenters believed that the proposed weights for

DRGs 112 and 116 are dramatically lower than they should be and the

result will be a disincentive to use these technologies.

[[Page 41504]]

Another commenter stated that by not reclassifying cases receiving

platelet inhibitors with angioplasty to DRG 116, we actually promote

the inaccuracy of the DRG weights, by grouping these higher-cost cases

with other lower-cost cases in DRG 112.

Response: With regard to the comment concerning the weights of DRGs

112 and 116, we refer the commenters to the discussion above in section

II.C of this preamble concerning the steps we take in recalibrating the

weights. Every year when the relative weights are recalibrated, we use

charge information from the most recent Medicare data available. That

is, we use the charges reported by hospitals for the cases under each

DRG to establish the relative weights. Each DRG weight represents the

average resources required to care for cases in that particular DRG

relative to the average resources used to treat cases in all DRGs. We

have not identified any problems or anomalies related to the cases in

DRGs 112 and 116 and are confident that the relative weights are

accurate.

With respect to the comment about our promoting the inaccuracy of

the DRG weights by failing to reclassify platelet inhibitor cases, the

commenter does not appear to understand the difference between

reclassification and recalibration. That is, the commenter argues that

the DRG relative weights are inaccurate because high-cost cases are not

reclassified to a higher-weighted DRG. However, our point regarding the

accuracy of the relative weights pertains to the necessity that, in the

process of recalibration, cases are grouped in the DRG to be used for

payment for similar cases during the upcoming year. Thus, the relative

weights are accurate in the sense that they are calculated by grouping

cases according to the DRG under which they would be paid.

Comment: One of the manufacturers of platelet inhibitor therapy

disagreed with our statement in the proposed rule that the prospective

payment system outlier policy would address the rationing of new

technology to Medicare beneficiaries. The commenter argues that cases

of platelet inhibitor therapy would not receive outlier payments

because the cost of the drug, while it is several thousand dollars over

the DRG payment, is not in excess of the fixed loss threshold ($14,575

over the DRG payment in the proposed rule for FY 2000).

Response: Section 1886(d)(5)(A) of the Act provides for payments in

addition to the basic prospective payments for outlier cases, cases

involving extraordinarily high costs. Our statement in the proposed

rule was meant to apply to all new technologies, and not specifically

to platelet inhibitor therapy. As stated previously, the prospective

payment system reflects ``averaging principles,'' which means, among

other things, that a hospital might lose money on some cases but will

gain money on other cases; sometimes new technologies lead to lower

costs and we might [email protected] hospitals for those cases. If a case does

not qualify for an outlier payment, then presumably the case falls

within the ``typical'' range of costs for cases in the DRG. We believe

that, as a whole, the prospective payment system does ensure access to

new technologies, including platelet inhibitor therapy.

III. Changes to the Hospital Wage Index

A. Background

Section 1886(d)(3)(E) of the Act requires that, as part of the

methodology for determining prospective payments to hospitals, the

Secretary must adjust the standardized amounts ``for area differences

in hospital wage levels by a factor (established by the Secretary)

reflecting the relative hospital wage level in the geographic area of

the hospital compared to the national average hospital wage level.'' In

accordance with the broad discretion conferred under the Act, we

currently define hospital labor market areas based on the definitions

of Metropolitan Statistical Areas (MSAs), Primary MSAs (PMSAs), and New

England County Metropolitan Areas (NECMAs) issued by the Office of

Management and Budget (OMB). OMB also designates Consolidated MSAs

(CMSAs). A CMSA is a metropolitan area with a population of one million

or more, comprised of two or more PMSAs (identified by their separate

economic and social character). For purposes of the hospital wage

index, we use the PMSAs rather than CMSAs since they allow a more

precise breakdown of labor costs. If a metropolitan area is not

designated as part of a PMSA, we use the applicable MSA. Rural areas

are areas outside a designated MSA, PMSA, or NECMA.

We note that effective April 1, 1990, the term Metropolitan Area

(MA) replaced the term Metropolitan Statistical Area (MSA) (which had

been used since June 30, 1983) to describe the set of metropolitan

areas comprised of MSAs, PMSAs, and CMSAs. The terminology was changed

by OMB in the March 30, 1990 Federal Register to distinguish between

the individual metropolitan areas known as MSAs and the set of all

metropolitan areas (MSAs, PMSAs, and CMSAs) (55 FR 12154). For purposes

of the prospective payment system, we will continue to refer to these

areas as MSAs.

Beginning October 1, 1993, section 1886(d)(3)(E) of the Act

requires that we update the wage index annually. Furthermore, this

section provides that the Secretary base the update on a survey of

wages and wage-related costs of short-term, acute care hospitals. The

survey should measure, to the extent feasible, the earnings and paid

hours of employment by occupational category, and must exclude the

wages and wage-related costs incurred in furnishing skilled nursing

services. As discussed below in section III.F of this preamble, we also

take into account the geographic reclassification of hospitals in

accordance with sections 1886(d)(8)(B) and 1886(d)(10) of the Act when

calculating the wage index.

B. FY 2000 Wage Index Update

The final FY 2000 wage index values in section VI of the Addendum

to this rule (effective for hospital discharges occurring on or after

October 1, 1999 and before October 1, 2000) are based on the data

collected from the Medicare cost reports submitted by hospitals for

cost reporting periods beginning in FY 1996 (the FY 1999 wage index was

based on FY 1995 wage data).

The final FY 2000 wage index includes the following categories of

data associated with costs paid under the hospital inpatient

prospective payment system (as well as outpatient costs), which were

also included in the FY 1999 wage index:

Salaries and hours from short-term, acute care hospitals.

Home office costs and hours.

Certain contract labor costs and hours.

Wage-related costs.

Consistent with the wage index methodology for FY 1999, the final

wage index for FY 2000 also continues to exclude the direct and

overhead salaries and hours for services not paid through the inpatient

prospective payment system, such as skilled nursing facility services,

home health services, or other subprovider components that are not

subject to the prospective payment system. (As discussed in section

III.C of this preamble, we are refining the methodology for calculating

the wage index for FY 2000.)

We calculate a separate Puerto Rico-specific wage index and apply

it to the Puerto Rico standardized amount. (See 62 FR 45984 and 46041.)

This wage index is based solely on Puerto Rico's data. Finally, section

4410 of the BBA provides that, for discharges on or after October 1,

1997, the area wage index

[[Page 41505]]

applicable to any hospital that is not located in a rural area may not

be less than the area wage index applicable to hospitals located in

rural areas in that State.

Comment: In a general comment on the wage index, MedPac noted that

new measures are needed to implement each new prospective payment

system as well as for Medicare+Choice plans and suggested that we

explore alternative strategies for obtaining labor prices that could be

applied to each type of provider affected. MedPAC offers to assist us

in examining this issue.

Response: We agree with MedPAC that this is an area warranting

further attention to determine whether it is appropriate to continue to

adjust payments for these other provider types based on the relative

average hourly wages of hospital employees, and whether the collection

of wage data for every type of Medicare provider is feasible or

necessary. Currently, the data used to calculate the hospital wage

index is used broadly in payment systems for other types of Medicare

providers. New prospective systems for skilled nursing facilities,

hospital outpatient services, and home health agencies will continue to

use the hospital wage index data for the foreseeable future. We have

collected data separately for skilled nursing facilities, but, pending

further development and auditing of these data, we continue to use the

hospital wage data (before reclassifications by the Medicare Geographic

Classification Review Board) for adjusting skilled nursing facility

payments at this time.

C. FY 2000 Wage Index Methodology Changes

In the July 31, 1998 final rule, we reiterated our position that,

to the greatest degree possible, the hospital wage index should reflect

the wage costs associated with the areas of the hospital included under

the hospital inpatient prospective payment system (63 FR 40970). That

final rule contained a detailed discussion concerning the costs related

to teaching physicians, residents, and CRNAs, all of which are paid by

Medicare separately from the prospective payment system. For reasons

outlined in detail in that final rule, we decided not to remove those

costs from the calculation of the FY 1999 wage index, but to review

updated data and consider removing them in developing the FY 2000 wage

index.

In response to concerns within the hospital industry related to the

removal of these costs from the wage index calculation, the American

Hospital Association (AHA) convened a workgroup to develop a consensus

recommendation. The workgroup, which consisted of representatives from

national and State hospital associations, recommended that costs

related to teaching physicians, residents, and CRNAs should be phased-

out of the wage index calculation over a 5-year period. Based upon our

analysis of hospitals' FY 1996 wage data, and consistent with the AHA

workgroup's recommendation, we proposed to phase-out these costs from

the calculation of the wage index over a 5-year period. The proposed FY

2000 wage index was based on a blend of 80 percent of an average hourly

wage including these costs, and 20 percent of an average hourly wage

excluding these costs.

Comment: Commenters unanimously supported our proposal to remove

teaching-related and CRNA costs from the wage index. Further, two

commenters recommended that we emphasize that Medicare pays its share

of teaching-related wage costs through direct graduate medical

education (GME) payments and that these costs are being removed from

the wage index only insofar as Medicare continues to pay the costs

outside of the hospital prospective payment system. Additionally,

commenters favored the proposed 5-year phase-out of these costs to

reduce significant redistributive impacts.

MedPAC, however, recommended that, rather than reducing the weights

for the old calculation and increasing the weights for the new

calculation by the proposed 20 percent each year, we should apply

smaller weights to the new wage index calculation for the first 2

years. Its rationale for this is its concern that inaccurate reporting

of teaching physician data, and our methodology for removing costs for

hospitals that fail to report these data, may inappropriately lower the

wage index values for nonteaching hospitals in the same labor market

areas.

Response: We are pleased to receive strong support for our efforts

to remove from the hospital wage index, wage costs that are associated

with areas of the hospital not included under the hospital prospective

payment system. Therefore, beginning with the FY 2000 wage index, and

over a 5-year period, we are phasing-out costs related to teaching

physicians, residents, and CRNAs. As recommended, we emphasize that our

rationale for removing these costs from the wage index calculation is

that Medicare pays for these costs separately, and these costs will be

excluded from the wage index as long as they are paid separately from

the hospital prospective payment system.

With respect to MedPAC's recommendation that the weight given to

the average hourly wage calculated after removing CRNAs, teaching

physicians, and residents, should be less than 20 percent for FY 2000,

we disagree. If we applied a percentage less than 20 percent for FY

2000 (and FY 2001), we then would have to apply a higher percentage

phase-out in a later fiscal year (or years) and thus increase the

redistributive impact for that year. We believe that applying 20

percent increments each year promotes the smoothest transition to total

exclusion of the costs.

1. Teaching Physician Costs

As discussed in the FY 1999 final rule and the FY 2000 proposed

rule, before FY 1999, we included direct physician Part A costs and

excluded contract physician Part A costs from the wage index

calculation. Since some States prohibit hospitals from directly

employing physicians, hospitals in these States were unable to include

physician Part A costs because they were incurred under contract rather

than directly. Therefore, for cost reporting periods beginning in 1995,

we began separately collecting physician Part A costs (both direct and

contract) so we could evaluate how to best handle these costs in the

wage index calculation. Based on our analysis of the 1995 wage data, we

decided to include the contract physician salaries in the wage index

beginning with FY 1999.

In the July 31, 1998 final rule, in response to comments regarding

the inclusion in physician Part A costs of teaching physician costs for

which teaching hospitals are already compensated through the Medicare

GME payment, we stated that we would collect teaching physician data

``as expeditiously as possible in order to analyze whether it is

feasible to separate teaching physician costs from other physician Part

A costs'' (63 FR 40968). Excluding teaching physician costs from the

wage index calculation is consistent with our general policy to exclude

from that calculation those costs that are paid separately from the

prospective payment system.

Because the FY 1996 cost reports did not identify teaching

physician salaries and hours separately from physician Part A costs, we

instructed our fiscal intermediaries to collect, through a survey,

teaching physician costs and hours from the teaching hospitals they

service. Specifically, we requested collection of data on the costs and

hours related to teaching physicians that were

[[Page 41506]]

included in Line 4 (salaried), Line 10 (contracted), Line 12 (home

office and related organizations), and Line 18 (wage-related costs) of

the Worksheet S-3, Part II. In our instructions accompanying the

survey, we indicated that these teaching-related costs are those

payable under the per resident amounts (Sec. 413.86) and reported on

Worksheet A, Line 23 of the hospital's cost report.

Survey data were received from approximately 59 percent of teaching

hospitals reporting physician Part A costs on their Worksheet S-3, Part

II (500 out of 845). Our fiscal intermediaries reviewed the survey data

for consistency with the Supplemental Worksheet A-8-2 of the hospitals'

cost reports. Supplemental Worksheet A-8-2 is used to apply the

reasonable compensation equivalency limits to the costs of provider-

based physicians, itemizing these costs by the corresponding line

number on Worksheet A.

Hospitals were given until March 5, 1999 to request changes to the

initial survey data. Fiscal intermediaries had until April 5, 1999 to

submit the revised data to the Health Care Provider Cost Report

Information system (HCRIS) for inclusion in the May 1999 final wage

data file. Due to the extraordinary effort needed to collect these data

and the importance of accurately removing teaching physician costs, we

allowed hospitals to request revisions to their teaching survey data up

until June 5, 1999.

The hospital industry workgroup also recommended that if the

teaching data collected by the intermediaries are not accurate or

reliable, HCFA should include only 20 percent of reported physician

Part A costs in the calculation, based on the assumption that 80

percent of total physician Part A costs are related to teaching

physicians. In developing the final FY 2000 wage index (as in the

proposed), if we had complete survey data for a hospital, that amount

was subtracted from the amount reported on the Worksheet S-3 for

physician Part A costs. These data had been verified by the fiscal

intermediary before submission to us. If we did not have survey data

for a teaching hospital as of June 5, 1999, we removed 80 percent of

the hospital's reported total physician Part A costs and hours for the

wage index.

Although removing 80 percent from the amount reported on the

Worksheet S-3 for physician Part A costs allows an estimate of teaching

physician costs to be removed in the majority of cases in which survey

data are not available, there are instances in which a teaching

hospital did not report either survey data or any physician Part A

costs on its Worksheet S-3. We identified 19 of these teaching

hospitals in our final database (there were 72 of these hospitals

identified in the proposed rule). For purposes of calculating the FY

2000 wage index for these 19 hospitals, we subtracted the costs

reported on Line 23 of the Worksheet A, Column 1 (Resident and Other

Program Costs) from Line 1 of the Worksheet S-3. These costs (from Line

23, Column 1 of Worksheet A) are included in Line 1 of the Worksheet S-

3, which is the sum of Column 1, Worksheet A. They also represent costs

for which the hospital is paid through the per resident amount under

the direct GME payment.

We believe this approach is appropriate in situations in which

hospitals have failed to otherwise identify their teaching physician

costs. To determine the hours to be removed, we divided the costs

reported on Line 23 of Worksheet A, Column 1 by the national average

hourly wage for physician Part A costs based upon Line 4 of Worksheet

S-3 (the national average hourly wage is $54.48). We indicate these 19

hospitals by an asterisk in Table 3C of this final rule.

In the proposed rule, we invited comments as to whether the

proposed method to remove teaching-related costs based on the amount

included in Line 23, Column 1 of Worksheet A would be an appropriate

method for removing GME costs in the future (and perhaps other excluded

area costs as well). We were especially concerned that the earliest

cost report on which we would be able to make the necessary changes to

capture the separate reporting of teaching physician Part A costs would

be those submitted for cost reporting periods beginning during FY 1998.

Therefore, we were considering subtracting the costs in Lines 20, 22,

and 23 of Worksheet A from Line 1 of Worksheet S-3, Part II, in

calculating the FY 2001 wage index. The current Worksheet S-3 is not

designed to net out of Line 1 costs that are otherwise included in

Column 1 of Worksheet A, but it would be possible to use data from the

Worksheet A in a manner similar to that described above.

Comment: Two commenters disagreed with our decision to allow

changes to the teaching survey data but not to corresponding lines on

Worksheet S-3 during the final wage data correction period (June 5

deadline). They believed we should be willing to accept conforming wage

data corrections, even during the final correction period, to achieve

the goal of using the most accurate data available.

Response: If hospitals had miscategorized their teaching physician

costs on their cost report in such a way that accurately completing the

teaching survey would result in their teaching physician survey costs

being removed twice, we did authorize corresponding revisions to

Worksheet S-3. For example, some hospitals included teaching physician

costs in Line 6 of their Worksheet S-3 (which is intended for reporting

interns and residents' costs). Therefore, reporting these costs on

their teaching physician survey, which would be subtracted from Line 4

for the salaries of teaching physicians directly employed by the

hospital, would result in them being removed twice, once when the

teaching physician data are subtracted from Line 1 of Worksheet S-3,

and again when Line 6 of Worksheet S-3 is subtracted from Line 1.

Comment: We received several comments regarding our proposal to use

the teaching survey data for teaching hospitals that submitted surveys

but to remove 80 percent of the total physician Part A costs and hours

for nonresponsive teaching hospitals. Most commenters supported our

reliance on the teaching survey data for the FY 2000 wage index. One

commenter added that we should be assertive in insisting that teaching

survey data be reported accurately by hospitals and verified by fiscal

intermediaries, holding hospitals to a level of accountability that is

similar to the certification of a cost report at filing. Another

commenter urged us to incorporate the separate collection of teaching

physician Part A data into the cost report as soon as possible to

ensure that the data submitted by hospitals is consistent.

Although most commenters agreed that we should reduce reported

total physician Part A costs by 80 percent for teaching hospitals that

do not submit the teaching survey, some took issue with this approach.

One national and one State hospital association recommended we remove

100 percent of reported total physician Part A costs from nonresponsive

teaching hospitals' total costs as a penalty for not reporting their

data. The commenters believe that, for hospitals whose proportion of

teaching physician Part A costs relative to total physician Part A

costs is greater than 80 percent, there is no incentive to complete the

teaching survey. On the other hand, MedPAC recommended that, since

HCFA's preliminary teaching survey data indicate that teaching

physician Part A costs are 68 percent of total physician Part A costs,

we should have adjusted the hospital's data by that amount rather than

the higher 80

[[Page 41507]]

percent figure. MedPAC comments that, although using the 80 percent

figure may give hospitals the incentive to submit the requested survey

data if their ratio of teaching physician Part A costs to total

physician Part A costs is less than 80 percent, that amount could

inappropriately lower the wage index values for other hospitals located

in the same MSA as the nonresponsive teaching hospital. The comments do

acknowledge, however, the policy dilemma in terms of the incentives not

to report that may arise by setting the percentage too low.

Response: We appreciate the commenters' general support of using

the survey data, as well as the efforts of hospitals and the fiscal

intermediaries in this special data collection effort. We believe that,

although the response rate is less than we would have preferred, the

end result is a more accurate FY 2000 wage index.

Although Worksheet S-3 is being revised to provide for the separate

reporting of teaching physician Part A costs, this change will not be

incorporated until cost reporting periods beginning during FY 1998.

Therefore, we will have to conduct another teaching physician cost

survey corresponding with the FY 1997 wage data. We agree with the

commenter's suggestion that the accuracy and completeness of the survey

data should be certified by the hospital in the same manner as the

accuracy and completeness of the cost report data must be certified.

In our calculation of the FY 2000 wage index, we removed 80 percent

of physician Part A costs and hours for teaching hospitals that failed

to report their teaching physician costs. We will consider the comment

to remove 100 percent of these costs for nonresponsive hospitals in the

future, however. Although the 80 percent figure was taken from the

industry workgroup's recommendation, we believe it may be appropriate

to consider raising this percentage to address the problem of hospitals

failing to comply with Medicare instructions.

We appreciate MedPAC's concern that the estimation of teaching

physician costs for hospitals that did not report should not

disproportionately harm other hospitals in the same labor market area.

Similarly, however, these hospitals should not benefit from

noncompliance. Also, as noted previously, because the teaching

physician costs are being removed gradually, with 80 percent of the FY

2000 wage index based on an average hourly wage that includes all of

these costs, we do not believe it is necessary to reduce the 80 percent

estimate to an amount based on the percentage of teaching physician

Part A costs to all physician Part A costs for hospitals completing the

survey to protect other hospitals in the labor market area. Any impact

should be relatively minor for this first year.

Comment: Two commenters believed that hospitals that contract with

physicians for Part A services are disadvantaged because the cost

report and teaching survey instructions seem to be designed only for

hospitals that employ physicians.

Response: The cost report and teaching survey do account for the

costs of contract physicians. The first year contract physician Part A

costs were included in the wage index was FY 1999. Beginning with the

FY 1995 cost report, we revised Worksheet S-3 to allow a separate line

item for reporting these costs. To improve the reporting for all

physician-related wage costs, we made additional changes to the FY 1996

cost report. The teaching survey was patterned after the FY 1996

Worksheet S-3.

The salaries on the Worksheet S-3 for employed physicians derive

from column 1 of Worksheet A. Hospitals should report the labor costs

associated with contract physicians in column 2 of that same worksheet.

If hospitals report their costs properly according to the cost report

instructions, hospitals using contract physicians will not be

disadvantaged by the way the costs are reported. We encourage hospitals

to be diligent in working with their intermediaries if they have

questions about reporting costs on the cost report.

Comment: We received four comments regarding the use of Worksheet

A, Line 23, Column 1 as a proxy for teaching-related wage costs when a

teaching hospital did not report either survey data or any physician

Part A costs. One was favorable without qualifications. One commenter

recommended that, beginning with the FY 2001 wage index, we should

instruct hospitals to report on Worksheet S-3 the wage costs associated

with teaching physicians directly from Worksheet A, Line 23 and the

corresponding hours directly from hospitals' records. A national

hospital association recommended that if we use Worksheet A, Line 23

for teaching salaries and a national average hourly wage for physicians

to estimate the associated hours to be removed for nonreporting

hospitals, then we should apply this approach to all hospitals. If we

apply this method only to hospitals that do not respond to the teaching

survey, the commenter believed that we should penalize nonresponsive

hospitals by increasing the hourly rate by 25 percent to ensure they

are not advantaged by not reporting their costs.

Several hospitals contacted us to report that, although they were

listed as one of the 72 hospitals for whom we used Line 23 of Worksheet

A to remove teaching physician costs, these costs were actually

included in other lines of Worksheet S-3, such as Line 5, Physician

Part B services, or Line 6, Interns and Residents. Therefore, since

both of these lines are subtracted from Line 1 in our calculation,

subtracting Line 23 from Worksheet A would remove these costs twice.

In opposing the use of Line 23 as a proxy for teaching-related

costs, one commenter cautioned that, particularly for hospitals in

States that are prohibited from employing physicians, Line 23, Column 1

may not include any teaching physician costs. MedPAC also stated

concern with this approach, but did not cite any specific problems

associated with it.

Response: For FY 2000, we are removing the amount reported on

Worksheet A, Line 23, Column 1, only in the absence of teaching survey

or Worksheet S-3 data for a hospital but we will continue to explore

using this approach rather than the survey for identifying GME and CRNA

costs to be removed in the FY 2001 wage index. The approach we adopted

has the advantage of being straightforward and easy to apply. Line 1,

Column 1 of Worksheet S-3 is equal to Line 101 of Column 1 of the

Worksheet A. Line 23 of Column 1, which is for the reporting of

nonresidents' costs related to GME that are paid separately from the

prospective payment system, is included in Line 101. Therefore, one

could argue that the simplest way to remove GME costs from the wage

index calculation would be to subtract the costs from Line 1 of

Worksheet S-3 that are attributable to the GME cost centers on

Worksheet A (Lines 22 and 23).

In carving out an estimate of hours for the final 19 hospitals for

which we subtracted Line 23 of Worksheet A from total salaries on

Worksheet S-3, we removed an estimated amount of associated hours based

on the average hourly wage of all physician Part A salaries. We did not

increase this average hourly wage by 25 percent as a penalty for

hospitals that did not otherwise report teaching physician costs. We do

reserve the right to remove some or all of a hospital's wage data that

cannot be appropriately supported by the hospital's records. We also

reserve the right to pursue further action in the case of hospitals

that intentionally withhold, conceal, or otherwise attempt

[[Page 41508]]

to circumvent the cost reporting requirements of their participation

agreements.

If we were contacted timely by a hospital that reported its costs

from Line 23 of Worksheet A somewhere other than Line 4 of the

Worksheet S-3, we did accommodate the hospital's request to avoid

removing the teaching physician Part A costs twice. We note that the

majority of these situations involved hospitals that did not follow the

cost reporting instructions for these costs. Despite MedPAC's general

concerns about this approach to removing costs, we did not receive any

comments that would cause us to rule out this seemingly straightforward

approach for removing GME and CRNA costs from the FY 2001 wage index

for all teaching hospitals. The biggest difficulty seems to be related

to ensuring that the cost reporting instructions are uniformly

followed.

Comment: Two commenters suggested using Worksheet A-8-2 of the cost

report, ``Provider-Based Physicians Adjustments,'' to determine

physician Part A costs, particularly for costs associated with teaching

and contract physicians. The commenters reasoned that, because

Worksheet A-8-2 is used to determine allowable cost and hours to be

included in the Medicare cost report, HCFA should use Worksheet A-8-2

to determine physician Part A labor costs for wage index purposes. Use

of the Worksheet A-8-2 would also ensure the wage index includes only

those physician costs paid under Part A. One of the commenters

commended us for requesting intermediaries to compare the teaching

survey and Worksheet A-8-2 data, but suggested that we should also

require intermediaries to use Worksheet A-8-2 data for determining

teaching physician wage costs when the survey data are unacceptable.

Response: We agree that, if properly completed, Worksheet A-8-2

should be an acceptable source for teaching physician Part A data. In

February, we instructed intermediaries to review hospitals' teaching

survey data for consistency with Worksheet A-8-2, and when necessary,

revise the data accordingly. One minor problem with relying solely on

Worksheet A-8-2 is that it may include some wage-related costs that are

excluded from the wage index calculation; however, these should be

insignificant. We believe that Worksheet A-8-2 is an appropriate source

for physician Part A costs. However, we need to examine Worksheet A-8-2

more closely before requiring that it be used to determine physician

part A costs for future wage indexes.

Comment: We received two comments recommending that we remove

overhead costs associated with the teaching physician, resident, and

CRNA direct costs that are excluded from the wage index. The commenter

compared this action to our current policy in which we remove the

overhead costs associated with excluded providers such as skilled

nursing facilities or rehabilitation units from the wage data. One

commenter offered technical assistance to HCFA in this effort.

Response: We agree, in principle, that overhead costs associated

with teaching-related and CRNA labor costs should be removed from the

wage index calculation in the same way that we remove overhead costs

associated with excluded areas of the hospital. However, we believe

that the methodology we apply for specific patient care cost centers

excluded from the wage data may not be appropriate for removing

overhead related to CRNA and GME costs. Therefore, we are grateful for

the commenter's offer of technical assistance to develop an appropriate

methodology for allocating overhead costs related to CRNAs and GME. We

anticipate that this issue will be discussed by HCFA's wage index

workgroup later this year, and in next year's proposed rule for FY

2001.

2. Resident and CRNA Part A Costs

The wage index presently includes salaries and wage-related costs

for residents in approved medical education programs and for CRNAs

employed by hospitals under the rural pass-through provision

(Sec. 412.113(c)). Because Medicare pays for these costs outside the

prospective payment system, removing these costs from the wage index

calculation would be consistent with our general policy to exclude

costs that are not paid through the prospective payment system.

However, because these costs were not separately identifiable on

Worksheet S-3 before the FY 1995 wage data, we could not remove them.

We began collecting the resident and CRNA wage data separately on

the FY 1995 cost report. However, there were data reporting problems

associated with these costs. For example, the original FY 1995 cost

report instructions for reporting resident costs on Line 6 of Worksheet

S-3, Part III, erroneously included teaching physician salaries and

other teaching program costs. Also, the FY 1995 Worksheet S-3 did not

provide for separate reporting of CRNA wage-related costs. These

problems were corrected in the reporting instructions for the FY 1996

cost report, and, therefore, we proposed and are now implementing the

removal of CRNA and resident costs over a 5-year period, beginning with

the FY 2000 wage index.

We received no comments related to this change.

3. Transition Period

The FY 2000 wage index is based on a blend of 80 percent of

hospitals' average hourly wages without removing the costs and hours

associated with teaching physician Part A, residents, and CRNAs, and 20

percent of the average hourly wage after removing these costs and hours

from the wage index calculation. This methodology is consistent with

the recommendation of the industry workgroup for a 5-year phase-out of

these costs. The transition methodology is discussed in detail in

section III.E of this preamble.

Comment: One hospital believed that it has been disadvantaged by

HCFA's allowance of contract teaching physician Part A costs in the FY

1999 wage index, and that HCFA should disallow teaching physician costs

entirely, beginning with FY 2000. The hospital stated that it is

experiencing difficulty meeting the criteria for geographic

reclassification for purposes of the wage index to another MSA that

includes a teaching hospital that reports a large amount of contract

teaching physician Part A costs.

Response: Our reasons for including contract physician Part A costs

are discussed in detail in the July 31, 1998 Federal Register (63 FR

40967). In general, it was our belief that if contract physician Part A

costs were reliably reported by hospitals, they should be included in

the wage data along with the Part A costs of directly employed

physicians. In that final rule, we also discussed our position that, to

the greatest degree possible, the hospital wage index should reflect

the wage costs associated with the areas of the hospital included under

the hospital inpatient prospective payment system. Therefore, based on

data we have collected since that final rule was published, and as

discussed above, we are removing teaching physician costs (as well as

CRNA and resident costs) for the wage data, over a 5-year period.

As is generally true with changes in the wage index, hospitals that

may have once been eligible to reclassify to another MSA for purposes

of the wage index may find that they no longer qualify after changes

have been implemented. However, we believe that all our changes to the

wage index are designed to more accurately reflect the wage costs

incurred by hospitals. In the case of the teaching physician costs, we

[[Page 41509]]

believe that a 5-year phase out is appropriate to reduce significant

redistribution impacts. With regard to the accuracy of the teaching

hospital data, the intermediary verified the data and determined it is

consistent with audit findings.

D. Verification of Wage Data from Medicare Cost Reports

The data for the FY 2000 wage index were obtained from Worksheet S-

3, Parts II and III of the FY 1996 Medicare cost reports. The data file

used to construct the final wage index includes FY 1996 data submitted

to HCRIS as of early February 1999. As in past years, we performed an

intensive review of the wage data, mostly through the use of edits

designed to identify aberrant data. In the proposed rule, we discussed

our review and methodology for resolving questionable elements in the

hospital data (64 FR 24728). The revised data are reflected in this

final rule. Since the proposed rule, we deleted data for four hospitals

that reported aberrant and unverifiable wage data that would have

significantly distorted the wage index values, and added data for seven

hospitals that were not included in the proposed wage index but rather

whose data have now been corrected and verified. The final FY 2000 wage

index is calculated based on FY 1996 data for 5,038 hospitals.

Comment: One hospital association expressed concern that a number

of hospitals might have failed to comply with the new cost reporting

instructions for wage-related costs, causing an overreporting of these

costs in the FY 2000 wage index. Prior to the FY 1996 cost report, the

lines on Worksheet S-3 for core and other wage-related costs reflected

a hospital's total costs for those categories. However, beginning with

the FY 1996 cost report, core and other wage-related costs must be

reported net of costs associated with excluded areas. The commenter

stated that wage-related costs for a significant number of hospitals

increased at least 10 percent this year and it believed that the

increase is due to hospitals incorrectly reporting excluded area wage-

related costs on Line 13. The commenter recommended that we develop a

method to determine if a hospital misreports its wage-related costs,

and that we should require correction of the data.

Response: We believe the new cost reporting instructions for wage-

related costs, Lines 13 and 14 of Worksheet S-3, Part II, are clear

regarding the exclusion of costs associated with excluded areas.

Intermediaries were aware of the new cost reporting instructions and

instructed their auditors to closely examine the costs reported in

Lines 13 and 14 of Worksheet S-3, Part II for compliance. In addition,

the intermediaries' FY 1996 wage data review program included an edit

for hospitals having wage-related costs that increased 10 percent or

more between FY 1995 and FY 1996. Furthermore, we contacted

representatives of national hospital associations who agreed to alert

their members of the reporting change. We are aware of numerous

instances where intermediaries adjusted hospitals' wage-related costs

after review. As part of the FY 1997 wage data desk review program (for

the FY 2001 wage index), we will provide more specific instructions to

the intermediaries to review the data reported for core and other wage-

related costs to ensure no costs associated with excluded areas are

included.

Comment: One commenter disagreed with the approach we used in the

proposed rule to identify teaching hospitals to ensure that all of

these hospitals had reported teaching physician survey data. We based

our decision to remove either 80 percent of physician Part A costs and

hours or the amount on Line 23, Column 1 of Worksheet A, based on

whether the hospital had a resident-to-bed ratio greater than zero on

the latest Provider-Specific File. The commenter suggested it would be

more appropriate to base the identification of teaching hospitals on

whether the hospital reported residents on its cost report for the

period corresponding with the wage data.

Response: We agree with this comment. It is more appropriate to

base the identification of teaching hospitals on data from the same

year as the wage data we use. Therefore, we revised our method to

identify teaching hospitals based on whether they reported residents

during their cost reporting period beginning during FY 1996.

Comment: One State hospital association commented that the

underrepresentation of physician Part A costs for hospitals in its

State is due to the intermediary's exclusion of a majority of the costs

reported by hospitals. The commenter believes there are inconsistencies

between the two intermediaries that service hospitals in the State in

their treatment of contract physician Part A costs. The commenter

recommended that HCFA monitor intermediaries and enforce uniform

application of Medicare principles and standards, particularly with

regard to the determination of allowable physician costs on Worksheet

A-8-2.

Response: For wage index purposes, contract physician costs are to

be reported according to the instructions for Worksheet S-3 Part II,

Line 10. The physician Part A costs reported on Worksheet S-3 may

differ slightly from those reported on worksheet A-8-2 because there

are minor differences in the types of wage-related costs that are

allowed for each of the worksheets. The two forms serve different

purposes. The wage index worksheet (S-3) may include, to a reasonable

extent, the actual costs a hospital incurs. However, Worksheet A-8-2 is

used to determine allowable costs for Medicare cost report purposes and

includes cost limits. The commenter did not indicate exactly what

inconsistencies it had found. If there are inconsistencies, we would

like to address them as soon as possible for the FY 2001 wage index.

We note that, intermediaries have informed us that hours associated

with contract physicians are often difficult to verify because

hospitals have not developed reporting systems that accurately account

for contract physician hours. Consistent with Medicare policy,

intermediaries must exclude costs and other data that are

insufficiently supported by a hospital's documentation.

Comment: One commenter noted several errors in the proposed rule

and final wage data public use file. The commenter stated that Table 3C

of the proposed rule included some hospitals with extremely low average

hourly wages, and that the average hourly wages reported for some

hospitals marked with an asterisk do not seem to incorporate the

Worksheet A, Line 23 data as described in the footnote. Additionally,

the commenter stated that the final wage data on the Internet includes

two different date formats for fiscal year begin and end dates, an

eight digit format and a seven digit format. The commenter asked that

HCFA make the appropriate corrections in the final wage index

calculation.

Response: We were informed shortly after publication of the

proposed rule that there were several errors in Table 3C, including

those noted by the commenter. As a result, we issued a revised Table 3C

in a correction notice published in the Federal Register on June 15,

1999 (64 FR 31995). Although the extremely low average hourly wages

still appear in Table 3C of the correction notice just as they were

reported by the hospitals, the aberrant data were either corrected or

deleted in the final wage index calculation. All other errors

identified in Table 3C were corrected through the June 15 notice. Also,

fiscal year beginning and ending dates that appear in a 7-digit date

format in the final wage data public use file were

[[Page 41510]]

corrected to an 8-digit date format in the final calculation.

E. Computation of the Wage Index

The method used to compute the FY 2000 wage index is as follows:

Step 1--As noted above, we based the FY 2000 wage index on wage

data reported on the FY 1996 Medicare cost reports. We gathered data

from each of the non-Federal, short-term, acute care hospitals for

which data were reported on the Worksheet S-3, Parts II and III of the

Medicare cost report for the hospital's cost reporting period beginning

on or after October 1, 1995 and before October 1, 1996. In addition, we

included data from a few hospitals that had cost reporting periods

beginning in September 1995 and reported a cost reporting period

exceeding 52 weeks. These data were included because no other data from

these hospitals would be available for the cost reporting period

described above, and because particular labor market areas might be

affected due to the omission of these hospitals. However, we generally

describe these wage data as FY 1996 data.

Step 2--Salaries--The method used to compute a hospital's average

hourly wage is a blend of 80 percent of the hospital's average hourly

wage including all teaching physician Part A, resident, and CRNA costs,

and 20 percent of the hospital's average hourly wage after eliminating

all teaching physician, resident, and CRNA costs.

In calculating a hospital's average salaries plus wage-related

costs, including all teaching physician Part A, resident, and CRNA

costs, we subtracted from Line 1 (total salaries) the Part B salaries

reported on Lines 3 and 5, home office salaries reported on Line 7, and

excluded salaries reported on Lines 8 and 8.01 (that is, direct

salaries attributable to skilled nursing facility services, home health

services, and other subprovider components not subject to the

prospective payment system). We also subtracted from Line 1 the

salaries for which no hours were reported on Lines 2, 4, and 6. To

determine total salaries plus wage-related costs, we added to the net

hospital salaries the costs of contract labor for direct patient care,

certain top management, and physician Part A services (Lines 9 and 10),

home office salaries and wage-related costs reported by the hospital on

Lines 11 and 12, and nonexcluded area wage-related costs (Lines 13, 14,

16, 18, and 20). We note that contract labor and home office salaries

for which no corresponding hours are reported were not included.

We then calculated a hospital's salaries plus wage-related costs by

subtracting from total salaries the salaries plus wage-related costs

for teaching physicians (see section III.C.1 of this preamble for a

detailed discussion of this policy), Part A CRNAs (Lines 2 and 16), and

residents (Lines 6 and 20).

Step 3--Hours--With the exception of wage-related costs, for which

there are no associated hours, we computed total hours using the same

methods as described for salaries in Step 2.

Step 4--For each hospital reporting both total overhead salaries

and total overhead hours greater than zero, we then allocated overhead

costs. First, we determined the ratio of excluded area hours (sum of

Lines 8 and 8.01 of Worksheet S-3, Part II) to revised total hours

(Line 1 minus Lines 3, 5, and 7 of Worksheet S-3, Part II). We then

computed the amounts of overhead salaries and hours to be allocated to

excluded areas by multiplying the above ratio by the total overhead

salaries and hours reported on Line 13 of Worksheet S-3, Part III.

Finally, we subtracted the computed overhead salaries and hours

associated with excluded areas from the total salaries and hours

derived in Steps 2 and 3.

Step 5--For each hospital, we adjusted the total salaries plus

wage-related costs to a common period to determine total adjusted

salaries plus wage-related costs. To make the wage adjustment, we

estimated the percentage change in the employment cost index (ECI) for

compensation for each 30-day increment from October 14, 1995 through

April 15, 1997 for private industry hospital workers from the Bureau of

Labor Statistics' Compensation and Working Conditions. We use the ECI

because it reflects the price increase associated with total

compensation (salaries plus fringes) rather than just the increase in

salaries. In addition, the ECI includes managers as well as other

hospital workers. This methodology to compute the monthly update

factors uses actual quarterly ECI data and ensures that the update

factors match the actual quarterly and annual percent changes. The

factors used to adjust the hospital's data were based on the midpoint

of the cost reporting period, as indicated below.

Midpoint of Cost Reporting Period

------------------------------------------------------------------------

Adjustment

After Before factor

------------------------------------------------------------------------

10/14/95...................................... 11/15/95 1.023163

11/14/95...................................... 12/15/95 1.021153

12/14/95...................................... 01/15/96 1.019151

01/14/96...................................... 02/15/96 1.017157

02/14/96...................................... 03/15/96 1.015246

03/14/96...................................... 04/15/96 1.013489

04/14/96...................................... 05/15/96 1.011888

05/14/96...................................... 06/15/96 1.010428

06/14/96...................................... 07/15/96 1.009099

07/14/96...................................... 08/15/96 1.007900

08/14/96...................................... 09/15/96 1.006788

09/14/96...................................... 10/15/96 1.005719

10/14/96...................................... 11/15/96 1.004695

11/14/96...................................... 12/15/96 1.003653

12/14/96...................................... 01/15/97 1.002529

01/14/97...................................... 02/15/97 1.001325

02/14/97...................................... 03/15/97 1.000000

03/14/97...................................... 04/15/97 0.998514

------------------------------------------------------------------------

For example, the midpoint of a cost reporting period beginning

January 1, 1996 and ending December 31, 1996 is June 30, 1996. An

adjustment factor of 1.009099 would be applied to the wages of a

hospital with such a cost reporting period. In addition, for the data

for any cost reporting period that began in FY 1996 and covers a period

of less than 360 days or more than 370 days, we annualized the data to

reflect a 1-year cost report. Annualization is accomplished by dividing

the costs and hours by the number of days in the cost report and then

multiplying the results by 365.

Step 6--Each hospital was assigned to its appropriate urban or

rural labor market area before any reclassifications under sections

1886(d)(8)(B) or 1886(d)(10) of the Act. Within each urban or rural

labor market area, we added the total adjusted salaries plus wage-

related costs obtained in Step 5 for all hospitals in that area to

determine the total adjusted salaries plus wage-related costs for the

labor market area.

Step 7--We divided the total adjusted salaries plus wage-related

costs obtained under both methods in Step 6 by the sum of the

corresponding total hours (from Step 4) for all hospitals in each labor

market area to determine an average hourly wage for the area.

Because the FY 2000 wage index is based on a blend of average

hourly wages, we then added 80 percent of the average hourly wage

calculated without removing teaching physician Part A, residents, and

CRNA costs, and 20 percent of the average hourly wage calculated with

these costs removed.

Step 8--We added the total adjusted salaries plus wage-related

costs obtained in Step 5 for all hospitals in the nation and then

divided the sum by the national sum of total hours from Step 4 to

arrive at a national average hourly

[[Page 41511]]

wage (using the same blending methodology described in Step 7). Using

the data as described above, the national average hourly wage is

$21.1800.

Step 9--For each urban or rural labor market area, we calculated

the hospital wage index value by dividing the area average hourly wage

obtained in Step 7 by the national average hourly wage computed in Step

8. We note that on July 6, 1999, OMB announced the designations of two

new MSAs: Auburn-Opelika, Alabama, comprising Lee County, and

Corvallis, Oregon comprising Benton County.

Step 10--Following the process set forth above, we developed a

separate Puerto Rico-specific wage index for purposes of adjusting the

Puerto Rico standardized amounts. (The national Puerto Rico

standardized amount is adjusted by a wage index calculated for all

Puerto Rico labor market areas based on the national average hourly

wage as described above.) We added the total adjusted salaries plus

wage-related costs (as calculated in Step 5) for all hospitals in

Puerto Rico and divided the sum by the total hours for Puerto Rico (as

calculated in Step 4) to arrive at an overall average hourly wage of

$9.86756 for Puerto Rico. For each labor market area in Puerto Rico, we

calculated the hospital wage index value by dividing the area average

hourly wage (as calculated in Step 7) by the overall Puerto Rico

average hourly wage.

Step 11--Section 4410 of the BBA provides that, for discharges on

or after October 1, 1997, the area wage index applicable to any

hospital that is not located in a rural area may not be less than the

area wage index applicable to hospitals located in rural areas in that

State. Furthermore, this wage index floor is to be implemented in such

a manner as to ensure that aggregate prospective payment system

payments are not greater or less than those that would have been made

in the year if this section did not apply. For FY 2000, this change

affects 226 hospitals in 36 MSAs. The MSAs affected by this provision

are identified in Table 4A by a footnote.

Comment: Two commenters suggested that, given the complexity of the

FY 2000 wage index calculation, we should make our detailed calculation

procedures and edits publicly available. This would enable hospitals

and researchers to more easily replicate the wage index values. One of

the commenters recommended that the detailed calculations and methods

should be included in future proposed and final rules. In addition,

they requested that we release the actual computer program used to

calculate the wage index.

Response: We have fully explained the steps we take to calculate

each hospital's average hourly wage and the wage index. In addition, we

have worked with hospitals that contacted us after attempting to

replicate our calculations, by reviewing their results and identifying

discrepancies. In doing so, we have been able to identify certain

anomalies in some of the proposed wage index values, which have been

corrected in the final wage index. Therefore, we agree that it might be

useful to provide more information to make it easier for the public to

replicate our calculations, and we are exploring our options. However,

we do not generally provide our computer programs that are used to

perform the wage index calculations, or for that matter, the programs

we use for all other calculations we perform.

Comment: One commenter recommended that, for leap years HCFA should

use 366 days, rather than 365 days, when annualizing cost report data

(see step 5 of the wage index calculation).

Response: We agree that the commenter's recommended method of

annualization, which recognizes an additional day for leap years, is

theoretically more accurate than our simple, across-the-board approach.

However, due to the intense effort required to incorporate all of the

wage data changes processed in conjunction with hospitals' final

opportunity to request revisions, we were unable to evaluate and

incorporate this change into our computer program in time to be

reflected in the final FY 2000 wage index. Therefore, we are not

adopting this recommendation for the FY 2000 wage index calculation. We

would note that, as described in step 5 above, we annualize any cost

reporting period that covers a period of fewer than 360 days or more

than 370 days. The majority of cost reporting periods are not

annualized. In those instances where annualization is done, we would

further point out that it does not affect the hospital's average hourly

wage calculation, since both the costs and hours are annualized by 365.

The impact, therefore, of this commenter's suggestion is limited to the

calculat

This text is long and has been trimmed here. Open the source document for the complete record.