Skilled Nursing Facilities: Providers Have Responded to Medicare 
Payment System By Changing Practices (23-AUG-02, GAO-02-841).	 
                                                                 
In 1998, the Health Care Financing Administration implemented a  
prospective payment system (PPS) for skilled nursing facility	 
(SNF) services provided to Medicare beneficiaries. PPS is	 
intended to control the growth in Medicare spending for skilled  
nursing and rehabilitative services that SNFs provide. Two years 
after the implementation of PPS, the mix of patients across the  
categories of payment groups has shifted, as determined by the	 
patients' initial minimum data set assessments. Although the	 
overall share of patients classified into rehabilitation payment 
group categories based on their initial assessments remained	 
about the same, more patients were classified into the high and  
medium rehabilitation payment group categories, and fewer were	 
initially classified into the most intensive (highest paying) and
least intensive (lowest paying) rehabilitation payment group	 
categories. Two years after PPS was implemented the majority of  
patients in rehabilitation payment groups received less therapy  
than was provided in 199. This was true even for patients within 
the same rehabilitation payment group categories. Across all	 
rehabilitation payment group categories, fewer patients received 
the highest amounts of therapy associated with each payment	 
group.								 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-841 					        
    ACCNO:   A04664						        
  TITLE:     Skilled Nursing Facilities: Providers Have Responded to  
Medicare Payment System By Changing Practices			 
     DATE:   08/23/2002 
  SUBJECT:   Beneficiaries					 
	     Health care cost control				 
	     Health care planning				 
	     Health care programs				 
	     Health care services				 
	     Payments						 
	     Skilled nursing facilities 			 
	     Medicare Program					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-02-841

   Report to Congressional Requesters

   United States General Accounting Office

   GAO

   August 2002 SKILLED NURSING FACILITIES

   Providers Have Responded to Medicare Payment System By Changing Practices

   GAO- 02- 841

   Page i GAO- 02- 841 SNF Responses to Payment System Letter 1 Results in
   Brief 3 Background 4 Distribution of Patients Across Payment Categories
   Has Changed 11 Since PPS, SNFs Provide Fewer Minutes of Therapy 15
   Concluding Observations 16 Agency Comments 17 Appendix I Scope and
   Methodology 18

   Appendix II Therapy Minutes, Activities of Daily Living, and Medicare
   Payment Rates to SNFs 20

   Appendix III Comments from the Centers for Medicare & Medicaid Services 22

   Tables

   Table 1: Share of Medicare Patients at Initial Assessment By Category and
   Percent Change, First Quarters 1999, 2000, and 2001 13 Table 2: Share of
   Medicare Patients at Initial Assessment in

   Rehabilitation RUGs and Percent Change, Before and After BBRA Payment
   Increases 14 Table 3: Median Therapy Minutes Per Week Provided on or
   Before

   the Initial Medicare Assessment By Rehabilitation Category, 1999 and 2001
   16 Table 4: Therapy Minutes, Activities of Daily Living, and Medicare

   Payment Rates to SNFs in Fiscal Year 2002 a 20 Figure

   Figure 1: Resource Utilization Group (RUG) Classification Scheme 6
   Contents

   Page ii GAO- 02- 841 SNF Responses to Payment System Abbreviations

   ADL activity of daily living CMS Centers for Medicare & Medicaid Services
   BBA Balanced Budget Act of 1997 BBRA Medicare, Medicaid, and SCHIP
   Balanced Budget

   Refinement Act of 1999 BIPA Medicare, Medicaid, and SCHIP Benefits
   Improvement

   and Protection Act of 2000 HCFA Health Care Financing Administration MDS
   minimum data set PPS prospective payment system RUG resource utilization
   group SNF skilled nursing facility

   Page 1 GAO- 02- 841 SNF Responses to Payment System August 23, 2002 The
   Honorable Charles E. Grassley

   Ranking Minority Member Committee on Finance United States Senate

   The Honorable Larry Craig Ranking Minority Member Special Committee on
   Aging United States Senate

   In 1998, the Health Care Financing Administration (HCFA) implemented a
   prospective payment system (PPS) for skilled nursing facility (SNF)
   services provided to Medicare beneficiaries. 1 Mandated in the Balanced
   Budget Act of 1997 (BBA), PPS is intended to control the growth in
   Medicare spending for skilled nursing and rehabilitative services that
   SNFs provide. 2 Medicare pays SNFs a daily rate to cover most services
   provided to a patient during each day of a covered SNF stay. Because each
   patient differs in the amount of care required, the rate is adjusted for
   the patient*s

   expected care needs and therapy based on information gathered by SNFs
   using a uniform clinical assessment instrument known as the nursing home
   minimum data set (MDS). Information from the MDS* such as a patient*s
   diagnoses, the amount of rehabilitative therapy, 3 and ability to perform
   activities of daily living (ADL)* is used by SNFs to assign each patient
   to a Medicare payment group.

   Patients assigned to the same Medicare payment group exhibit similar care
   needs, so Medicare*s daily payment rate is the same for each patient
   within a group. The payment rate is based on the national average cost of

   1 On July 1, 2001, the Secretary of Health and Human Services changed the
   name of the Health Care Financing Administration (HCFA) to the Centers for
   Medicare & Medicaid Services (CMS). In this report, we will continue to
   refer to HCFA where our findings apply to the organizational structure and
   operations associated with that name.

   2 Pub. L. No. 105- 33, S: 4432( a), 111 Stat. 251, 414 (codified as
   amended at 42 U. S. C. S: 1395yy( e) (2000)). 3 This report uses therapy
   to refer to rehabilitation therapy, which includes physical, speech, and
   occupational therapies.

   United States General Accounting Office Washington, DC 20548

   Page 2 GAO- 02- 841 SNF Responses to Payment System providing care to
   patients in each group, not the actual costs for any given patient. As a
   result, a SNF profits when its costs are below the payment rate and loses
   if its costs are above it. A SNF can control its costs by

   treating less expensive patients within a group, providing care more
   efficiently, or providing fewer or a less expensive mix of services
   compared to the national average. A SNF can maximize its payments by
   admitting patients who are likely to be classified into particular payment
   groups or by modifying its patient assessment and documentation practices
   to support higher payment group assignments.

   The implementation of PPS raised providers* concerns about whether
   payments would cover their costs. 4 Our work and the work of others
   indicated that the payment groups may not adequately identify high- cost
   patients and distribute payments accordingly. The Congress addressed these
   concerns by increasing Medicare payments across all payment groups. At the
   same time, the Congress mandated additional increases for selected payment
   groups in response to concerns that payments for some types of patients
   were too low. Payments were increased for some of the payment groups for
   patients receiving primarily rehabilitation therapy and all payment groups
   for patients requiring extensive or special care and for

   clinically complex patients. Given possible provider responses to PPS, you
   asked us to analyze (1) shifts in the mix of Medicare beneficiaries across
   payment groups and (2) the amount of therapy services provided to patients
   within payment groups. To do so, we examined MDS data to determine the mix
   of patients treated and the services that Medicare beneficiaries received
   across and

   within payment groups for three points in time* early in PPS (January
   through March 1999), 1 year later and 2 years later. (For more details,
   see app. I, Scope and Methodology). We also interviewed staff from the
   Centers for Medicare & Medicaid Services (CMS) responsible for SNF payment
   policy and we reviewed regulations, literature, and other documents
   relating to SNF PPS and MDS. We performed our work from March 2000 through
   June 2002 in accordance with generally accepted

   government auditing standards. 4 U. S. General Accounting Office, Nursing
   Homes: Aggregate Medicare Payments are Adequate Despite Bankruptcies, GAO/
   T- HEHS- 00- 192 (Washington, DC: Sept. 5, 2000), U. S. General Accounting
   Office, Skilled Nursing Facilities: Medicare Payment Changes Require
   Provider Adjustments But Maintain Access, GAO/ HEHS- 00- 23 (Washington,
   DC: December 14, 1999), and Medicare Payment Advisory Commission, Report
   to Congress: Medicare Payment Policy (Washington, DC: March 2001).

   Page 3 GAO- 02- 841 SNF Responses to Payment System Two years after the
   implementation of PPS, the mix of patients across the categories of
   payment groups has shifted, as determined by the patients*

   initial MDS assessments. Although the overall share of patients classified
   into rehabilitation payment group categories based on their initial
   assessments remained about the same, more patients were classified into
   the high and medium rehabilitation payment group categories, and fewer

   were initially classified into the most intensive (highest paying) and
   least intensive (lowest paying) rehabilitation payment group categories.
   This shift is consistent with industry assertions that payments in
   relation to the cost of caring for patients in the high and medium
   rehabilitation payment group categories were more favorable than those for
   other categories of payment groups. Further, the share of patients
   initially assigned to the

   selected rehabilitation payment groups for which the Congress gave
   additional payment increases grew, while the share of patients assigned to
   almost all of the other rehabilitation payment groups remained the same or
   declined. Among patients who were not in rehabilitation payment groups,
   the share that was initially classified into categories requiring more
   extensive services increased almost 12 percent. SNFs changed two patient
   assessment practices that could have contributed to the shift in patients,
   at admission, across payment groups. First, SNFs increasingly used
   estimates of therapy needed, rather than actual therapy delivered, to
   assign patients to payment group categories. Second, SNFs conducted

   their initial patient assessments later in the stays of patients needing
   therapy, expanding the period of time over which they could receive
   therapy and increasing the likelihood that they would be classified into

   categories with the highest possible payments. Two years after PPS was
   implemented the majority of patients in rehabilitation payment groups
   received less therapy than was provided in 1999. This was true even for
   patients within the same rehabilitation payment group categories. The
   patients categorized into the two most common (high and medium)
   rehabilitation payment group categories typically received 30 minutes less
   therapy during their first week of care, a 22 percent decline. Across all
   rehabilitation payment group categories, fewer patients received the
   highest amounts of therapy associated with each payment group.

   In its written comments on a draft of this report, CMS agreed with our
   findings and noted that they were generally consistent with its analyses
   of provider responses to the PPS. Results in Brief

   Page 4 GAO- 02- 841 SNF Responses to Payment System Generally, Medicare
   covers SNF stays for patients needing skilled nursing and therapy for
   conditions related to a hospital stay of at least 3 consecutive calendar
   days, if the hospital discharge occurred no more

   than 30 days prior to admission to the SNF. For qualified beneficiaries,
   Medicare will pay for medically necessary services, including room and
   board, nursing care, and ancillary services such as drugs, laboratory
   tests, and physical therapy, for up to 100 days per spell of illness. 5
   For more than a decade beginning in 1986, Medicare SNF spending rose

   dramatically* averaging 30 percent annually. During this period, Medicare
   payments to each SNF were based on the costs incurred by the SNF in
   serving its Medicare patients. There was minimal program oversight,
   providing few checks on spending growth. Although Medicare imposed payment
   limits for routine services, such as room and board, it did not limit
   payments for capital and ancillary services, such as therapy. Cost
   increases for ancillary services averaged 19 percent per year from 1992
   through 1995, compared to a 6 percent average increase for routine service
   costs.

   To curb the rise in Medicare SNF spending, BBA required a change in
   Medicare*s payment method. HCFA began phasing in the SNF PPS on July 1,
   1998. 6 Under PPS, SNFs are paid a prospectively determined rate intended
   to cover most services provided to a patient during each day of a
   Medicare- covered SNF stay. 7 The SNF payment rate is based on the 1995
   national average cost per day, updated for inflation. Because the costs of
   treating patients vary with their clinical conditions and treatments,
   daily payments for each patient are adjusted for the patient*s expected
   care needs depending on the patient*s assignment into one of 44 different
   payment groups, also called resource utilization groups (RUG). A RUG

   5 A spell of illness is a period that begins when a Medicare beneficiary
   is admitted to a hospital or a SNF and ends when a beneficiary has not
   been an inpatient of a hospital or a SNF for 60 consecutive days. A
   beneficiary may have more than one spell of illness per year and maintain
   Medicare coverage.

   6 SNFs came under PPS beginning with their new fiscal year. Over 90
   percent of SNFs came under PPS before or during January 1999. The
   remainder came under PPS later in 1999. 7 Payments are adjusted for the
   local variation in wages. Certain high- cost, infrequently provided
   services, such as cardiac catheterizations and radiation therapy, are paid
   for separately outside the daily SNF rate. See U. S. General Accounting
   Office, Skilled Nursing Facilities: Services Excluded From Medicare*s
   Daily Rate Need to be Reevaluated,

   GAO- 01- 816 (Washington, DC: Aug. 22, 2001). Background SNF PPS

   Page 5 GAO- 02- 841 SNF Responses to Payment System describes patients
   with similar therapy, nursing, and special care needs and has a
   corresponding payment rate.

   The RUG classification system is hierarchical. The first distinction made
   is whether the patient has received (or is expected to receive) at least
   45 minutes a week of therapy (see fig. 1). For these rehabilitation
   patients, further divisions* into ultra high, very high, high, medium, and
   low therapy categories* are made based on the total minutes and type of
   physical, occupational, and speech therapy provided over 7 days. Each of
   these categories is defined by a range of therapy minutes and the type of
   therapy provided. For example, patients in the very high category receive
   between 500 and 719 minutes of therapy over 7 days. Each category is
   further subdivided into RUGs, based on a patient*s dependency in
   performing ADLs, such as eating, transferring from a bed to a chair, or
   using the toilet. There are 14 rehabilitation RUGs, which account for
   threefourths of Medicare- covered stays.

   Page 6 GAO- 02- 841 SNF Responses to Payment System Figure 1: Resource
   Utilization Group (RUG) Classification Scheme

   a For patients classified based on estimates of care needs, patients are
   expected to receive at least 45 minutes of therapy a week. b Care
   generally not paid by Medicare because patient does not require skilled
   nursing care.

   Source: Medicare Program: Prospective Payment System and Consolidated
   Billing for Skilled Nursing Facilities, 64 Federal Register 41,644 (July
   30, 1999) and Department of Health and Human Services, Health Care
   Financing Administration, Payment Safeguard Review of Skilled Nursing
   Facility Prospective Payment Bills, Program Memorandum Transmittal A- 99-
   20 (Baltimore, MD: May 1999).

   Page 7 GAO- 02- 841 SNF Responses to Payment System Among patients who
   have not received (or are not expected to receive) 45 minutes a week of
   therapy, the system distinguishes between patients

   requiring extensive or special care or who are clinically complex (12
   RUGs) and those receiving custodial care (18 RUGs). 8 The classification
   system uses specific medical conditions (such as having multiple sclerosis
   or being comatose) and special care needs (such as

   requiring tracheostomy care or ventilator support) within the past 14 days
   to group patients into extensive services, special care, and clinically
   complex categories. Patient characteristics such as the ability to perform
   ADLs, signs of depression, and conditions requiring more technical
   clinical knowledge and skills are used to assign patients into RUGs within
   these categories.

   Since 1991, SNFs have carried out a requirement to periodically assess and
   plan for residents* care using the MDS, 9 which documents 17 aspects of a
   patient*s clinical condition, including the amount of therapy provided or
   planned, diagnoses, certain care needs, and the ability to perform ADLs at
   the patient*s most dependent state. In addition to determining Medicare
   payments, these data are used to measure patient needs, develop a plan of
   care, and monitor the quality of care.

   To gather the MDS, an in- house interdisciplinary team assesses each
   patient*s clinical condition at established intervals throughout the
   patient*s stay. The Medicare assessment schedule requires that the initial
   assessment be performed during days 1 through 5 of a patient*s stay, but
   may be performed as late as days 6 through 8, termed *grace days,* which
   give staff additional flexibility in conducting the assessments. The
   initial assessment is used to assign patients to a RUG that establishes
   payments

   for the first 14 days of care. For patients staying longer than 14 days, a
   second assessment must be conducted during days 11 through 14 that
   determines the RUG assignment and payment rate for days 15 through

   30 of the patient*s stay. An additional assessment is performed prior to
   the 30th day of care and every 30 days thereafter; each of these
   assessments establishes the payment for the next 30 days up to the 100th
   day.

   8 Patients in the custodial RUGs are divided into three broad categories*
   impaired cognition, behavior only, and physical function reduced* based on
   the need for nursing services and the patient*s ability to perform ADLs.
   Patients classified into a custodial care

   RUG typically do not meet the skilled nursing care requirements for
   Medicare coverage. 9 42 U. S. C. S:S: 1395i- 3( b)( 3)( A) and 1396r( b)(
   3)( A) (2000). MDS Patient Assessments

   Page 8 GAO- 02- 841 SNF Responses to Payment System SNFs can classify
   patients primarily needing therapy into the high, medium, or low
   rehabilitation payment group categories for the initial

   assessment using either actual minutes of therapy provided or an estimate
   of the amount that will be provided over the 2 weeks covered by the
   initial assessment. If a patient is classified into one of these
   rehabilitation

   categories using an estimate, but actually receives less than the amount
   of therapy to qualify into that category, payments to the SNF for the
   initial assessment period are not reduced. To classify patients into the
   very high or ultra high payment group categories on the initial
   assessment, SNFs

   must have already provided the minimum amount of therapy that defines
   these categories when the assessment is done. 10 The accuracy and
   completeness of the patient assessment information are

   critical to ensure appropriate categorization of patients into payment
   groups. For example, to distinguish between different levels of assistance
   required in performing ADLs, a SNF needs to document how often and how
   much assistance was provided to a patient during the past 7 days. For a
   patient receiving over 720 minutes of therapy a week (the ultra high
   rehabilitation category), the difference between assessing a patient as
   needing *extensive* versus *limited* assistance in performing one ADL,
   such as eating, may result in an additional payment of up to $48 per day
   to the SNF. (See app. II for a comparison of ADLs and payment rates for
   each

   RUG.) Thus, a SNF might respond to the PPS by increasing the resources
   devoted to completing the MDS.

   This possible SNF response to the new payment system may be similar to how
   hospitals responded to the inpatient hospital PPS. Under the inpatient
   PPS, hospitals are paid a prospectively determined rate per patient stay,
   which is adjusted for expected resource needs based on factors such as
   patient diagnoses and treatment. After the implementation of the inpatient
   PPS in 1983, hospitals expanded the number of diagnoses they reported to
   describe patients. These changes in documentation resulted in some
   patients being classified into higher payment categories, which increased
   hospital payments.

   10 For the second and all subsequent assessments, a SNF must have provided
   the minimum amount of therapy in the range to classify a patient into any
   of the therapy categories. This categorization establishes payment for the
   next period. Possible SNF Responses to PPS Incentives

   Page 9 GAO- 02- 841 SNF Responses to Payment System A SNF also has an
   incentive to change the amount of care provided to minimize its costs and
   maximize its payments. Because the amount of

   therapy provided is key to classifying the majority of patients into RUGs,
   a SNF benefits when it provides an amount of therapy on the low end of the
   range of therapy minutes associated with that RUG. For example, furnishing
   1 additional minute of therapy a week could move a patient from the very
   high to the ultra high category. The SNF would receive an additional $63
   or $99 more per day, depending on the patient*s ADL needs, but there may
   not have been a proportionate increase in costs.

   To ensure that its patients are grouped into the highest possible payment
   groups, a SNF may adjust the timing of its initial patient assessments.
   Grace days are intended to give SNFs the flexibility to delay care until
   patients are ready to receive therapy, while ensuring that payments
   reflect the treatment levels that are provided to the patient. SNFs may
   opt to use grace days when conducting the initial assessment of patients
   who may be grouped into the payment group categories that require actual
   minutes of therapy (ultra and very high rehabilitation). Otherwise, if
   initial assessments are done before the grace days, patients may not have
   received enough therapy to reach the weekly threshold for placement into
   one of these categories.

   Since the implementation of the SNF PPS, some nursing home chains have
   claimed that payments are inadequate and that this has caused their
   financial condition to erode. We have reported that total SNF PPS payments
   are likely to be adequate and may be excessive given that the payment
   rates include the costs of inefficient delivery, unnecessary care,

   and improper billings. 11 But the Medicare Payment Advisory Commission and
   we have raised concerns that the payment rates for certain types of
   patients may be inadequate because the patient classification system may
   not appropriately reflect the differing needs of patients who require
   multiple kinds of health care services, such as extensive or special care,
   rehabilitative therapy, and ancillary services. 12 We have also expressed

   11 GAO/ T- HEHS- 00- 192 and GAO/ HEHS- 00- 23. 12 See Medicare Payment
   Advisory Commission, Report to Congress: Medicare Payment Policy
   (Washington, DC: March 2001) and U. S. General Accounting Office, Skilled
   Nursing Facilities: Medicare Payments Need to Better Account for
   Nontherapy Ancillary Cost Variation, GAO/ HEHS- 99- 185 (Washington, DC:
   Sept. 30, 1999). Refinements to the SNF

   PPS

   Page 10 GAO- 02- 841 SNF Responses to Payment System concern that the use
   of therapy minutes provided to patients as a way to classify patients
   might encourage the provision of unnecessary services. 13 In response to
   concerns about the overall adequacy of Medicare payments

   and their distribution across different types of patients, the Congress
   has raised payments twice since the PPS implementation. These actions
   increased payments across- the- board for all RUGs and, in addition, for
   certain RUGs. The Medicare, Medicaid, and SCHIP Balanced Budget Refinement
   Act of 1999 (BBRA) temporarily increased Medicare*s payments for all RUGs
   by 4 percent, beginning in fiscal year 2001 through the end of fiscal year
   2002. 14 In addition, BBRA increased payments for

   15 RUGs (3 rehabilitation RUGs and all extensive services, special care,
   and clinically complex RUGs) by 20 percent beginning in April 2000. 15 The
   Congress intended this increase to be temporary* until refinements to the
   RUGs patient classification system were implemented. However, refinements
   have not been implemented and the Congress again revised the payment
   rates. 16 The Medicare, Medicaid, and SCHIP Benefits Improvement and
   Protection Act of 2000 (BIPA) temporarily increased the portion of the
   payment related to nursing costs by 16.66 percent for all payment groups,
   17 which raised the overall payment rates from 4 to 12 percent, depending
   on the RUG, beginning April 1, 2001, through September 30, 2002. In
   addition, BIPA replaced the 20 percent BBRA

   increase that applied to 3 out of the 14 rehabilitation RUGs with a 6.7
   percent increase for all rehabilitation RUGs. 18 CMS has also responded to
   concerns about PPS. In July 2001, CMS

   awarded a contract to determine the feasibility of refinements to PPS,
   including alternatives to the RUGs patient classification system. To date,
   this contract has not resulted in proposed refinements to the RUGs system
   and the contractor*s preliminary report is not due until fall 2004. CMS
   has

   13 U. S. General Accounting Office, Medicare Post- Acute Care: Better
   Information Needed Before Modifying BBA Reforms, GAO/ T- HEHS- 99- 192
   (Washington, DC: Sept. 15, 1999). 14 Pub. L. No. 106- 113, App. F, S: 101(
   d), 113 Stat. 1501, 1501A- 325. 15 BBRA S: 101( a) and (b). This 20
   percent increase is calculated separately from the 4 percent increase. 16
   65 Fed. Reg. 46, 770 (July 31, 2000).

   17 Pub. L. No. 106- 554, App. F, S: 312( a), 114 Stat. 2763, 2763A- 498.
   18 The remaining 12 RUGs* extensive services, special care, and clinically
   complex* retained the 20 percent increase. BIPA S: 314.

   Page 11 GAO- 02- 841 SNF Responses to Payment System also supported work
   to assess and verify the MDS data that underlie PPS. However, we recently
   reported that CMS*s proposed on- site and off- site

   review of MDS assessments may not be sufficient to ensure the accuracy of
   MDS assessments in most nursing homes or to systematically evaluate the
   performance of state efforts to do so. 19 In September 2001, CMS awarded a
   contract to determine if there are differences between the documentation
   of patient care needs and actual patient care needs and to detect
   irregularities in MDS assessments. The contractor began these data
   monitoring activities in the spring of 2002, which include checking that
   the RUGs reported on the Medicare claims match those on the MDS
   assessments and examining the distribution of patients across the payment
   groups.

   Among patients primarily receiving rehabilitation care, more were
   classified at their initial assessment into moderate rehabilitation
   payment group categories and fewer into the intensive and low
   rehabilitation categories since the implementation of PPS. Providers
   reported that the payments for the moderate rehabilitation payment groups
   were more

   favorable, relative to their costs, than other payment groups. Further,
   the share of patients initially classified into the rehabilitation RUGs
   whose payments were increased by BBRA provisions grew, while the share of
   patients initially classified into most of the other payment groups
   declined or stayed the same. Across patients initially assigned to the
   extensive,

   special care, or clinically complex categories, more were classified as
   requiring extensive services* the highest paying category* and fewer into
   the special care or clinically complex categories. SNFs changed two
   patient assessment practices that could have contributed to these shifts
   in patients* initial payment group assignments. First, SNFs increased
   their use of estimated* rather than actual* therapy minutes to assign
   patients to rehabilitation categories. Second, SNFs assessed patients
   later in their stays, making it more likely that they received more
   therapy and therefore would be classified into categories with higher
   payments.

   19 For more information on CMS and state efforts to monitor the accuracy
   of the MDS data see U. S. General Accounting Office, Nursing Homes:
   Federal Efforts to Monitor Resident Assessment Data Should Complement
   State Activities, GAO- 02- 279 (Washington, DC: Feb. 15, 2002).
   Distribution of

   Patients Across Payment Categories Has Changed

   Page 12 GAO- 02- 841 SNF Responses to Payment System Although the
   proportion of SNF Medicare patients initially classified into
   rehabilitation payment group categories remained the same overall, the

   distribution of patients within these categories changed considerably from
   first quarter 1999 to first quarter 2001 (see table 1). 20 By 2001, more
   Medicare patients receiving therapy were initially classified into the two
   moderate rehabilitation categories* medium (16 percent more) and high (17
   percent more), which made up about two- thirds of Medicare SNF admissions.
   21 The share of patients initially classified into ultra high* the most
   intensive rehabilitation category* decreased to comprise just 3 percent of
   all Medicare SNF patients at their initial assessment in 2001. This shift
   is consistent with the industry*s assertions that the high and medium
   categories have more favorable payments, relative to their costs, than
   other categories. We do not know if this shift reflects a change in the
   care needs of patients from 1999 to 2001.

   20 Over three- quarters of Medicare SNF patients were classified into
   rehabilitation categories, while a little less than one quarter were
   classified into extensive services, special care, or clinically complex
   categories.

   21 Our findings about the share of rehabilitation versus extensive,
   special care, and clinically complex patients and the distribution of
   patients across rehabilitation categories are consistent with other
   analyses of claims data. See Department of Health and Human Services,
   Office of the Inspector General, Trends in the Assignment of Resource
   Utilization Groups by Skilled Nursing Facilities, OEI- 02- 01- 00280
   (Washington, DC: HHS,

   July 2001). More Patients Initially

   Categorized into Payment Groups with Payment Increases

   Page 13 GAO- 02- 841 SNF Responses to Payment System Table 1: Share of
   Medicare Patients at Initial Assessment By Category and Percent Change,
   First Quarters 1999, 2000, and 2001 Predominant

   type of care Category JanuaryMarch

   1999 (percent)

   JanuaryMarch 2000 (percent)

   JanuaryMarch 2001 (percent)

   Percent change 1999- 2001

   Ultra high 6.6 3.8 3. 2 -51.5 Very high 15.6 14.1 11.8 -24.4 High 37.1
   42.1 43.5 17.3 Medium 15.9 16.9 18.5 16.4 Rehabilitation therapy

   Low 0.5 0.3 0. 2 -60.0 Extensive services 11.9 13.1 13.3 11.8 Special care
   5.8 5.1 4. 9 -15.5 Extensive,

   special care, and clinically complex Clinically complex 4.1 3.1 2. 9 -29.3

   Custodial care Other 2.6 1.6 1. 5 -42.2

   Total 100 100 100 *

   Note: Percentages do not add to 100 due to rounding. Source: GAO analysis
   of the nursing home MDS, Medicare initial assessments, first quarters of
   1999, 2000, and 2001. Some of the shifts in the distribution across
   individual rehabilitation RUGs

   paralleled changes in payment rates made by the Congress. Within the high
   and medium rehabilitation payment group categories, the shares of patients
   initially classified into RUGs that received congressionally mandated
   payment increases in 2000 grew substantially more than the shares of
   patients classified into rehabilitation RUGs that did not (see

   table 2). For 8 of the 11 rehabilitation RUGs without this special
   increase, the shares of patients at their initial assessment declined and
   only one experienced an increase.

   Page 14 GAO- 02- 841 SNF Responses to Payment System Table 2: Share of
   Medicare Patients at Initial Assessment in Rehabilitation RUGs and Percent
   Change, Before and After BBRA Payment Increases

   Rehabilitation RUG Share

   January- March 2000 (percent)

   Share JanuaryMarch 2001 (percent)

   Percent change 2000- 2001

   RUGs with payments increased 20 percent by BBRA RHC 17.4 19.3 11 RMC 5.2
   5. 6 8 RMB 8.4 9. 5 13

   RUGs with payments not increased 20 percent by BBRA RUC 0.6 0. 5 -17 RUB
   2.5 2. 2 -12 RUA 0.6 0. 5 -17 RVC 1.8 1. 3 -28 RVB 9.2 7. 8 -15 RVA 3.2 2.
   7 -16 RHB 18.2 18.1 -1 RHA 6.4 6. 2 -3 RMA 3.4 3. 5 3 RLB 0.1 0. 1 0 RLA
   0.2 0. 2 0

   Note: The payment increases mandated by BBRA applied to services furnished
   on or after April 1, 2000. Beginning April 1, 2001, BIPA replaced the 20
   percent BBRA increase with a 6.7 percent increase for all rehabilitation
   RUGs. Source: GAO analysis of the nursing home MDS, Medicare initial
   assessments, first quarters of

   2000 and 2001.

   Among the patients initially classified into the extensive and special
   care or clinically complex categories (all of which were increased 20
   percent by BBRA), the share of patients initially assessed as requiring
   the most intensive care* those in the extensive services category*
   increased to become about two- thirds of patients in these categories,
   while the share of patients in the special care and clinically complex
   categories decreased.

   Since the introduction of PPS, changes in SNF patient assessment practices
   have made it easier to classify patients into some categories with higher
   payments. When performing their initial patient assessments, SNFs have
   increasingly opted to use estimates of the amount of therapy they expect
   to provide (rather than actual therapy given during the first week of
   care) to categorize patients into the high, medium, and low therapy
   categories for the first 14 days of care. Because payments are based on
   these estimates, payments for some patients were higher than they would
   Changes in Assessment

   Practices May Contribute to Different Classifications and Higher Payments

   Page 15 GAO- 02- 841 SNF Responses to Payment System have been if the
   payments were based on actual therapy provision. Comparing the first
   quarters of 1999 and 2001, the practice of using

   estimated therapy minutes, rather than actual therapy provided, to
   classify patients into therapy categories increased more than 35 percent,
   becoming the mechanism for classifying nearly two- thirds of all patients
   in high, medium, and low rehabilitation categories. Of the patients who
   could be evaluated, 22 one quarter of the patients classified using
   estimated minutes of therapy did not receive the amount of therapy they
   were assessed as needing, while three- quarters eventually did. 23 SNFs
   increasingly performed initial patient assessments later in patient

   stays, during the grace days, for patients in the highest paying therapy
   categories* ultra high and very high. Because classification into these
   categories is based on the actual amount of care provided, conducting the
   patient assessments during the grace days allows additional time for more
   therapy services to be provided, making it likelier that patients would be

   classified into the ultra high and very high categories. To classify
   patients into these categories, the use of grace days increased more than
   40 percent from the first quarter of 1999 to the first quarter of 2001.

   In the 2 years following the implementation of PPS, SNFs provided less
   therapy to almost two- thirds of all Medicare SNF patients* those in the
   medium and high rehabilitation payment group categories. The typical
   patient in these categories received 22 percent less therapy, at least 30
   fewer minutes, per week during the initial assessment period between the
   first quarters of 1999 and 2001. Indeed, in 2001 half of the patients
   initially categorized in these two groups did not actually receive the
   amount of therapy required to be classified into those groups, due in part
   to the use of estimated therapy minutes for classification (see table 3).
   Further, during their initial assessment period, fewer patients received
   therapy near the higher end of the range that defines each category. For
   example, to be assigned to the high rehabilitation category, patients are
   assessed as needing between 325 and 499 minutes of therapy a week. In 22
   Only patients who stay long enough to have a second assessment done (where
   the actual minutes of therapy provided in the past 7 days are recorded)
   could be evaluated. For the

   largest share of patients, however, we do not know if they received the
   projected services because these patients did not stay in the facilities
   long enough for a second assessment. 23 It is possible that between the
   initial assessment and the end of the second assessment period the care
   needs of some patients changed and they no longer required the amount of
   therapy that had been originally estimated. Since PPS, SNFs

   Provide Fewer Minutes of Therapy

   Page 16 GAO- 02- 841 SNF Responses to Payment System 1999, 20 percent of
   patients in the high rehabilitation payment group category received 390
   minutes or more of therapy per week during their

   initial assessment period. Two years later, less than 13 percent received
   this much therapy. In 1999, 5 percent of patients initially assessed in
   the high rehabilitation payment group category received 480 minutes or
   more of therapy per week. Two years later, only 2 percent of patients
   received

   this level of therapy.

   Table 3: Median Therapy Minutes Per Week Provided on or Before the Initial
   Medicare Assessment By Rehabilitation Category, 1999 and 2001 Minutes per
   week provided in Rehabilitation category

   (required therapy minutes per week) 1999 2001 Percent change

   1999- 2001

   Ultra high (720 or more) 735 737 0 Very high (500- 719) 525 525 0 High
   (325- 499) 325 255 -22 Medium (150- 324) 150 117 -22 Low (45- 149) 80 77
   -4

   Source: GAO analysis of the nursing home MDS, Medicare initial
   assessments, first quarters of 1999 and 2001.

   Across all therapy patients, the median amount of therapy provided during
   the initial assessment period also declined from 1999 through 2001. The
   declines in therapy service use and resultant reductions in costs were not
   uniform across the rehabilitation payment group categories. Consequently,
   payments for some categories of RUGs are likely to be higher than their
   service costs, compared to other categories of RUGs. For patients in the
   more intensive rehabilitation payment group categories, where estimated
   minutes cannot be used to classify patients, median therapy minutes did
   not decline. Our work indicates that SNFs have responded to PPS in two
   ways that

   may have affected how payments compare to SNF costs. SNFs have (1) changed
   their patient assessment practices and (2) reduced the amount of therapy
   services provided to Medicare beneficiaries. The first change can increase
   Medicare*s payments and the second can reduce a SNF*s costs. CMS*s ongoing
   efforts to refine the payment system are particularly important in light
   of these provider responses to the PPS. Concluding

   Observations

   Page 17 GAO- 02- 841 SNF Responses to Payment System In its written
   comments on a draft of the report, CMS agreed that ongoing evaluations of
   PPS are important. CMS stated that our findings are

   generally consistent with its analyses and with its expectations regarding
   provider responses to the incentives of the PPS. CMS noted that it intends
   to examine whether therapy provided is consistent with payment levels and
   ADL coding accuracy through its program safeguard contractor project. CMS
   stated that reporting the percentage change of relatively small shares of
   patients across payment categories may overemphasize the

   changes and is somewhat misleading. However, the percentage changes
   reported in table 1 demonstrate that the shifts in shares of patients
   across payment categories are consistent with the industry*s assertions
   that high and medium categories have the most favorable payments, relative
   to costs. In addition, the percentage changes reported in table 2
   demonstrate that the shifts among RUGs parallel the congressionally
   mandated payment increases. CMS also provided technical comments, which we
   incorporated as appropriate. CMS*s comments are in appendix III.

   We are sending copies of this report to the Administrator of CMS,
   appropriate congressional committees, and other interested parties. We
   will also provide copies to others upon request. In addition, the report
   is available at no charge on the GAO Web site at http:// www. gao. gov.

   If you or your staff has any questions, please call me at (202) 512- 7114.
   Laura Sutton Elsberg, Leslie Gordon, and Walter Ochinko prepared this
   report under the direction of Carol Carter.

   Laura A. Dummit Director, Health Care* Medicare Payment Issues Agency
   Comments

   Appendix I: Scope and Methodology Page 18 GAO- 02- 841 SNF Responses to
   Payment System We used data from the 1998 Medicare cost reports to
   identify SNFs that began participating in PPS on or before January 1,
   1999. Facility

   ownership and other characteristics were taken from HCFA*s end- of- year
   Provider of Services file for 1999. We included in our analysis only those
   SNFs that had transitioned to PPS before or during January 1999, were
   active in 1999, and submitted Medicare MDS assessments in the three
   periods used in this study. This cohort comprised approximately 80 percent
   of all SNFs that filed a 1998 cost report and was representative of the
   universe of SNFs in terms of bed size, location (rural and urban), and
   ownership characteristics.

   For the SNFs in our sample, we analyzed data from the nursing home MDS
   national repository to compare differences in patient classification and
   therapy services across three points in time* early in PPS (January- March
   1999), 1 year later (January- March 2000), and 2 years later (JanuaryMarch
   2001). 1 Data to examine the distribution of Medicare patients after the
   implementation of BIPA- mandated changes (applied to services on or after
   April 1, 2001) were not available in time for this analysis. Our sample
   included over 350,000 MDS assessments for Medicare beneficiaries for each
   time period. To examine the differences in patient classification, we
   grouped patient assessments into 11 major categories* the 5 major
   rehabilitation categories (ultra high, very high, high, medium, and low),
   3 categories for patients requiring extensive or special care or who are
   clinically complex, and 3 categories for patients requiring custodial
   care, based on the RUG reported on the initial assessment.

   To examine the differences in the provision of therapy services, we
   aggregated the reported physical, occupational, and speech therapy minutes
   for each assessment. We calculated the number of initial assessments that
   had used estimated minutes to qualify patients into a rehabilitation
   category by counting the number of first assessments that reported actual
   therapy minutes below the minimum number of minutes required in the three
   rehabilitation categories (high, medium, and low). To determine the extent
   to which patients received the estimated therapies, we calculated, for the
   patients who had a second assessment, the percent who had received less
   than the minimum number of therapy minutes required for the RUG reported
   on the initial assessment. We also

   1 The national repository contains resident assessment information for
   every resident of a Medicare- or Medicaid- certified long- term care
   facility. Appendix I: Scope and Methodology

   Appendix I: Scope and Methodology Page 19 GAO- 02- 841 SNF Responses to
   Payment System interviewed CMS staff responsible for SNF policy and we
   reviewed regulations, literature, and other documents relating to SNF PPS
   and MDS.

   Appendix II: Therapy Minutes, Activities of Daily Living, and Medicare
   Payment Rates to SNFs

   Page 20 GAO- 02- 841 SNF Responses to Payment System Table 4: Therapy
   Minutes, Activities of Daily Living, and Medicare Payment Rates to SNFs in
   Fiscal Year 2002 a Predominant type of care Category Therapy minutes

   per week ADLs b Resource utilization group Medicare daily

   payment rate

   16- 18 RUC $441.18 9- 15 RUB $392.78 Ultra high 720+ 4- 8 RUA $369.27

   16- 18 RVC $342.67 9- 15 RVB $330.22 Very high 500- 719

   4- 8 RVA $298.41 13- 18 RHC $318.68

   8- 12 RHB $291.02 High 325- 499

   4- 7 RHA $264.74 15- 18 RMC $315.94

   8- 14 RMB $279.99 Medium 150- 324

   4- 7 RMA $262.01 14- 18 RLB $252.39 Rehabilitation therapy

   Low c 45- 149 4- 13 RLA $209.52 Extensive services d N/ A 7- 18 3 RUGs
   $234.06-$ 307. 35 Special care e N/ A 7- 18 3 RUGs $211.93-$ 228. 53
   Extensive or special care or clinically complex

   Clinically complex f N/ A 4- 18 6 RUGs $175.98-$ 227. 14 Impaired
   cognition c N/ A 4- 10 4 RUGs $145.55-$ 167. 68 Behavior only c N/ A 4- 10
   4 RUGs $138.64-$ 166. 30 Custodial g Physical function reduced c N/ A 4-
   18 10 RUGs $135.87-$ 181. 51

   a The payment rates in the table became effective October 1, 2001, for
   SNFs located in urban areas and include the 16. 66 percent increase for
   the nursing component as required by BIPA, but do not include the add- on
   payments for individual RUGs. There are separate payment rates for
   facilities located in rural areas.

   b ADL scores range from 4 (least dependent) to 18 (most dependent). c The
   low rehabilitation RUG and some of the custodial RUGs require at least two
   nursing rehabilitation activities, 6 days a week. Some examples include:
   passive or active range of motion, amputation care, and splint or brace
   assistance. d The extensive services category includes patients who have
   had the following specific medical or

   skilled nursing care needs in the past 14 days* intravenous medications,
   tracheostomy care, ventilator/ respirator support, suctioning, or
   intravenous feeding in the last 7 days. e The special care category
   includes patients who have any of the following clinical conditions:
   multiple

   sclerosis, cerebral palsy, quadriplegia with high ADL dependency, surgical
   wounds or open lesions, pressure or stasis ulcers on two or more body
   sites or have a fever in combination with dehydration, pneumonia,
   vomiting, weight loss, or tube feeding. It also includes patients who
   require specific medical and skilled nursing care, such as radiation
   therapy and respiratory therapy. f The clinically complex category
   includes patients who have any of the following clinical conditions:

   comatose, burns, systemic infection (septicemia), pneumonia, internal
   bleeding, dehydration, dialysis, or paralysis on one side (hemiplegia) in
   combination with a high ADL dependency. It also includes patients
   receiving chemotherapy, tube feeding of at least 26 percent of daily
   calorie intake and 501 milliliters of fluid, being treated for foot wounds
   or transfusions, receiving injections 7 days per week for diabetes while
   their condition is somewhat unstable, or those who have received oxygen
   therapy in the last 14 days. The group also includes patients with
   unstable conditions.

   Appendix II: Therapy Minutes, Activities of Daily Living, and Medicare
   Payment Rates to SNFs

   Appendix II: Therapy Minutes, Activities of Daily Living, and Medicare
   Payment Rates to SNFs

   Page 21 GAO- 02- 841 SNF Responses to Payment System g Patients are
   classified into the custodial categories according to their need for
   nursing services and assistance with ADLs. These patients typically do not
   meet the criteria for Medicare coverage because they generally do not
   require skilled nursing care.

   Sources: Medicare Program: Prospective Payment System and Consolidated
   Billing for Skilled Nursing Facilities, 63 Federal Register 26, 252 (May
   12, 1998), table 2C and Medicare Program: Prospective Payment System and
   Consolidated Billing for Skilled Nursing Facilities, 66 Federal Register
   39,562 (July 31, 2001), table 3.

   Appendix III: Comments from the Centers for Medicare & Medicaid Services

   Page 22 GAO- 02- 841 SNF Responses to Payment System Appendix III:
   Comments from the Centers for Medicare & Medicaid Services

   Appendix III: Comments from the Centers for Medicare & Medicaid Services

   Page 23 GAO- 02- 841 SNF Responses to Payment System (201016)

   The General Accounting Office, the investigative arm of Congress, exists
   to support Congress in meeting its constitutional responsibilities and to
   help improve the performance and accountability of the federal government
   for the American people. GAO examines the use of public funds; evaluates
   federal programs and policies; and provides analyses, recommendations, and
   other assistance to help Congress make informed oversight, policy, and
   funding decisions. GAO*s commitment to good government is reflected in its
   core values of accountability, integrity, and reliability.

   The fastest and easiest way to obtain copies of GAO documents at no cost
   is through the Internet. GAO*s Web site (www. gao. gov) contains abstracts
   and fulltext files of current reports and testimony and an expanding
   archive of older products. The Web site features a search engine to help
   you locate documents using key words and phrases. You can print these
   documents in their entirety, including charts and other graphics.

   Each day, GAO issues a list of newly released reports, testimony, and
   correspondence. GAO posts this list, known as *Today*s Reports,* on its
   Web site daily. The list contains links to the full- text document files.
   To have GAO e- mail

   this list to you every afternoon, go to www. gao. gov and select
   *Subscribe to daily E- mail alert for newly released products* under the
   GAO Reports heading.

   The first copy of each printed report is free. Additional copies are $2
   each. A check or money order should be made out to the Superintendent of
   Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
   copies mailed to a single address are discounted 25 percent. Orders should
   be sent to: U. S. General Accounting Office 441 G Street NW, Room LM
   Washington, D. C. 20548 To order by Phone: Voice: (202) 512- 6000

   TDD: (202) 512- 2537 Fax: (202) 512- 6061

   Contact: Web site: www. gao. gov/ fraudnet/ fraudnet. htm E- mail:
   fraudnet@ gao. gov Automated answering system: (800) 424- 5454 or (202)
   512- 7470 Jeff Nelligan, managing director, NelliganJ@ gao. gov (202) 512-
   4800

   U. S. General Accounting Office, 441 G Street NW, Room 7149 Washington, D.
   C. 20548 GAO*s Mission Obtaining Copies of

   GAO Reports and Testimony

   Order by Mail or Phone To Report Fraud, Waste, and Abuse in Federal
   Programs Public Affairs
*** End of document. ***