Statement of Felice Loverso, Ph.D., American Medical Rehabilitation Providers Association

The American Medical Rehabilitation Providers Association (AMRPA) is the leading national trade association representing over 450 freestanding rehabilitation hospitals, rehabilitation units of acute care general hospitals and numerous outpatient rehabilitation services providers.  Our members serve over 450,000 patients per year, and most, if not all, of our members are Medicare providers. We appreciate the Subcommittee’s focused attention on post-acute care services in Medicare.  Rehabilitation hospitals and units are a crucial part of the spectrum of post-acute care providers, and we believe it is important to examine the issues surrounding this complex area of care. 

An ongoing debate exists among policymakers, providers and various organizations about whether skilled nursing facilities (SNFs), inpatient rehabilitation facilities (IRFs) and possibly long term care hospitals (LTCHs) provide the same programs and activities with equivalent outcomes to patients needing medical rehabilitation services.  Facilities should be compared both by their physical attributes, and the complete nature of the care and services they are organized to provide.  Comparing facility performance solely by patient diagnoses or cost provides an extremely limited picture of the patients treated in these settings, the nature and value of the care they receive.  One must look at additional patient information to truly appreciate the patients and their characteristics.  

IRFs provide programs of care that utilize skilled rehabilitation services to Medicare patients at a pace, intensity, and sophistication that cannot be obtained in other health care settings.  IRFs provide intense rehabilitation medicine and therapy to patients with 24-hour nursing and physician services.   Patients receive a high-quality, coordinated program of care with the goal of achieving the maximum level of function possible and a rapid return to the community.

ARMPA shares the Committee’s interest in examining the complicated issues surrounding assessment tools and looking at other ways to address payment across post acute providers, and we appreciate the opportunity to present our recommendations to the Committee.

75% Rule

One overarching concern facing all post-acute care rehabilitation providers is the dramatic impact implementation of the 75 Percent Rule on patient access to rehabilitative care.  The 75% Rule is unquestionably having a more severe impact on patients and providers than CMS or OMB originally estimated.  The Medicare program originally estimated that implementation of the 75% Rule would reduce payments to IRFs by $10 million in FY 2005 and $30 million in FY 2006.  However, the President’s FY 2006 Budget revised these estimates to show a savings of $50 million in FY 2005 and $70 million in FY 2006.   AMRPA’s own data suggest that Medicare is likely to save $165 million dollars in the first year alone.   Clearly, CMS did not anticipate such a dramatic decline in patient services as a result of implementing this regulation. 

Most alarming is the impact the rule is having on patients’ access to treatment.  Clear evidence now exists that IRF discharges have started to decline, and this change is orders of magnitude greater than CMS estimated. ERehabData®, AMRPA’s data service, estimates that in the first year alone, over 39,000 patients will be refused admission to inpatient rehabilitation facilities in order for hospitals to maintain compliance with the new 75% Rule.  For the first three quarters under the new 75% Rule, volume is down 5.8% from the comparable three quarters in 2003 and 2004, meaning that approximately 20,000 Medicare patients have been denied admission since July 1, 2004.  By the fourth year of the 75% Rule, IRFs will be forced to turn away one out of every three patients in order to remain compliant.  As noted in the GAO Report entitled “More Specific Criteria Needed to Classify Inpatient Rehabilitation Facilities,” only 6 percent of IRFs will be able to meet the 75 percent threshold required at full implementation ofthe rule at the end of the transition period.  Without any direction from Congress, the 75% Rule is eliminating intensive inpatient rehabilitation as a treatment option for a significant number of Medicare beneficiaries. 

At the core of the 75% Rule seems to be a mistaken reliance on the assumption that one site of care can be substituted for another with no impact on quality or outcomes.  In particular, CMS, in promulgating changes to classification criteria for IRFs, assumed that SNF and other post-acute care settings can be substituted for IRFs if patients are denied care due to the exclusion criterion in the 75% Rule, and that this is clinically acceptable and economically desirable.  AMRPA strongly disagrees with this premise. IRFs provide a very unique, specialized, intensive form of rehabilitative care that cannot be duplicated in other Medicare settings.  Given the enormous impact the 75 Percent Rule has had on inpatient rehabilitative care, AMRPA urges the Ways and Means Committee to consider legislation that would hold the 50% threshold for compliance for two additional years.  Moreover, to facilitate collaborative relationships with federal policymakers, AMRPA urges consideration of a federal advisory council on medical rehabilitation that would work with CMS to properly characterize IRFs and separately establish workable guidelines to distinguish appropriate patient selection criteria.

Current Financing for Post-Acute Care Services

Current Medicare program post-acute care policy is focused on providing care based on types of providers, with the key post-acute care institutional providers being LTCHs, IRFs and SNFs.  While all of these sites provide post-acute care to Medicare beneficiaries, each site of care currently utilizes its own prospective payment system.  The SNF PPS began in 1998 and is based on a per diem payment unit.  SNFs use a patient classification system called resource utilization groups (RUGs), of which there are 44 groups.  On May 19, CMS issued a proposed rule to change the RUGs and increase the number to 53.  In contrast, the LTCH PPS is based on a per discharge payment unit and uses LTCH DRGs, of which there are currently 550.  The LTCH PPS is being phased in over 5 years.  Finally, the IRF PPS was initiated in January 2002 and is also based on a per discharge payment unit.  There are 21 Rehabilitation Impairment Categories (RICs) and 95 case mix groups (CMGs) with four payment tiers, for a total of 380 possible CMGs and separate HIPPS codes.  Each system is based on research reflective of the costs of care in a base year used to calculate the payment rates. 

CMS, MedPAC and others have expressed concern that the post-acute care payment systems provide incentives for engaging in behavior solely to enhance reimbursement, without regard to quality or appropriateness of care, patient outcomes or cost.  Policymakers must realize that looking at payments in the context of diagnoses only, without looking at other factors, can be quite startling but does not reveal much about patient differences and reasons why a particular setting (1) best suits the need of that patient and/or (2) contains the resources necessary to obtain the optimum patient outcome.  For example, payment for a stroke case may vary from $31,496.00 in an LTCH to $8,905 in a SNF according to a MedPAC report in June 2004 examining the most severe stroke cases (Chapter 5, June 2004 report on LTCHs).  However, since those figures are for the most severely ill types of patients in that diagnosis, the numbers cited do not reflect the average payment, which is considerably lower.  For example, the average Medicare payment for a stroke in an IRF in 2003 was $16,769.00 according to AMRPA’s eRehabData®.  

While federal policymakers understandably look closely at payment differentials, these payments encompass costs that are larger than the individual patient being treated.  All of the payment systems discussed are based on historical costs that reflect not only patient care but also the setting-specific requirements and different Medicare Conditions of Participation each type of entity must meet. These requirements vary considerably by setting in the length, depth, scope and cost of compliance.  Each system also relies on some patient’s diagnosis information and varying amounts of functional information. 

AMRPA has closely analyzed cost reports for SNFs and IRFs, examining both routine costs and ancillary costs in order to determine any differences between the two settings and whether such differences are representative of varying levels of services delivered.  When the SNF PPS and IRF PPS were under development in 1998, AMRPA analyzed the available costs reports for 1996 to see what the impact of a prospective payment system would be on SNFs.  AMRPA found that there were higher costs in hospital-based SNFs than freestanding SNFs, a finding later reaffirmed by MedPAC reports.  These findings suggested that a different type of patient was being treated with more complex needs in the hospital-based SNF setting.  At the time of the analysis, the average length-of-stay (ALOS) for the hospital-based SNFs was 16.56 days, in contrast to 45.03 days in the freestanding SNFs. 

AMRPA also examined routine and ancillary cost differences between IRFs and SNFs.  It was clear that both the routine costs and ancillary costs were higher in the IRF setting, reflecting the greater intensity of care.  IRFs had higher ancillary costs per day ($274 per day for rehab units; $134.74 for SNF hospital based units; $268 for rehab hospitals; and $118.96 for freestanding SNFs), as were specific therapy charges.  However, we believe that ancillary costs have decreased in response to the SNF cuts and therapy cuts in the Balanced Budget Act of 1997 and the implementation of the SNF PPS.  Such a decrease would reflect a reduction in the amount of therapy delivered and the intensity of care.  AMRPA is currently working on updating this information using 2002 costs reports. 

The cost differential between SNFs and IRFs is significant, but the cost variation represents differences in prospective payment systems and the greater intensity of care provided in the inpatient rehabilitation setting.    Thus, the faulty belief that care is equivalent among post-acute care settings is also leading CMS to argue that Medicare is paying too much for some patient care provided in IRFs.  In its September 9, 2003 proposed IRF rule, CMS assumed that the average payment for an IRF was $12,525 and that by substituting care at a payment of $7,000 per case it would “save” approximately $5,525 per case.  It is clear now that the cases being denied access to IRF care due to the 75% Rule are primarily lower extremity joint replacement cases whose payments on average in 2004, based on eRehabData®, were approximately $9,151.  Hence the actual difference in payments is only $2,151 per case.  Additionally, these numbers may also be misleading because of differences in lengths of stay.  If the average Medicare SNF stay for similar cases is 31 in 2001 and 33 days in 2003 according to MedPAC, at an average daily rate of approximately $400, then the payment is closer to $12,000 thereby further reducing Medicare’s alleged savings.  We would be pleased to provide the Committee with the AMRPA analysis.

Services Provided in IRFs Compared to Other Post-Acute Care Settings

One frequent discussion in comparing settings is whether a nursing home or skilled nursing facility can substitute for IRF care and provide equivalent services and outcomes.  Practitioners find that in general, nursing homes and skilled facilities do not have all the characteristics of an IRF.  Facilities may share some characteristics with IRFs, but this varies widely geographically.  IRFs are subject to a number of standards that no other post-acute care setting must meet, including: (1) close medical supervision by a physician with specialized training in rehabilitation; (2) patients must undergo at least 3 hours a day of physical and/or occupational therapy; and (3) a multidisciplinary approach to delivery of the rehabilitation program.  (Please find attached a chart delineating a comparative analysis of SNF and IRF coverage criteria).  There are no comparable specific standards for other facilities relating to rehabilitation services (such as the ‘three hour rule” for IRFs), and, therefore, each nursing home or SNF must be evaluated individually.

A good illustration of the difference in services provided in these rehabilitation settings can be seen in the Spring 2005 MedPAC analysis examining single hip and knee joint replacements in IRFs and SNFs.  MedPAC commissioned the RAND Corporation to study outcomes across settings for hip and knee replacement cases in response to changes to the 75% Rule that would force fewer hip or knee replacement patients to be treated in IRFs each year.  MedPAC staff conducted two studies and presented the results at the April 2005 meeting.  The first study involved a physician panel of six (6) orthopedic surgeons and five (5) specialists in physical medicine and rehabilitation.  The physician panel noted that close to 50-80% of such patients go home with home health care or outpatient services, and therefore not to institutional settings. The panel said that patients who could not go home should have the following characteristics for referral to a SNF or IRF:

·        Be limited in weight bearing or unable to walk 100 feet;

·        Be obese or have comorbidities;

·        Have an impairment of one or more joints (not replaced);

·        Have diminished pre-surgery functioning; or

·        Have architectural barriers or no informal caregiver at home.

Panelists also said that patients who need extra medical attention should go to IRFs, while patients who need convalescent care or cannot tolerate 3 hours per day of therapy should go to SNFs.  In some communities, surgeons refer based on the qualifications of specific facilities that are available, such as how the facilities are staffed, whether they follow rehabilitation protocols or are convenient for the surgeon to follow-up.     

Another point MedPAC has clearly established is that the types of patients treated in each setting are considerably different.  MedPAC recently examined the types of patients in SNFs, IRFs and home health agencies (HHAs) receiving care for single joint replacements.  Specifically, it found that:

*  MedPAC Staff Handout, April 2005 Meeting

RAND presented a number of conclusions about the differences in cost and care among settings.  Generally, RAND found that the functional level of patients in IRFs was lower at admission than in SNFs, but patients ultimately had greater functional gains, suggesting that the greater intensity of therapy in IRFs improves functional status.  In addition, after controlling for a number of variables, RAND found that SNF and IRF patients were more likely to be institutionalized compared to patients sent home.  However, 2.5 times more patients in SNFs were institutionalized or died (0.46%) than those in IRFs (0.18%).  Further, as expected, SNFs and IRFs were paid more than patients discharged home.  RAND found that SNFs cost $3578 and IRFs cost $8,023 for total post-acute payments as opposed to home care.  Note, however, that these figures are misleading and understated for home health costs and SNF costs because they do not include any Part B outpatient services provided.

AMRPA is particularly concerned that patients referred to LTCHs and IRFs are being pressured by Medicare into staying in acute care longer or treated in SNFs.  This view has become much more prevalent as CMS issues regulations that are detrimental to certain sites of care, such as CMS’s FY 2005 LTCH rate year update, the IPPS FY 2005 proposed rule proposal pertaining to hospitals within hospitals, and the various proposed and final rules pertaining to the 75% Rule for IRFs.  Many post-acute care LTCH and IRF providers are left with the impression that a federal bias in defining LTCHs and IRFs more narrowly is designed to: (1) close many of these facilities; and (2) force patients to be treated in skilled nursing facilities (SNFs).  Many post-acute care providers and physicians believe that while SNFs may be able to treat a percentage of such patients successfully with respect to outcomes, many are not able to successfully treat these patients because of serious differences in a patient’s medical and functional abilities and the significantly more limited resources provided in SNFs. 

CMS and Congress should actively initiate research on how these sites of care provide treatment to Medicare beneficiaries and how each site’s functional outcomes vary by patient diagnosis.  As noted by the National Institutes of Health’s February 2005 panel on medical rehab and by MedPAC, there is little evidence on the different care provided by these entities and how outcomes differ by site of care.  The Agency for Healthcare Research and Quality (ARHQ) conducted a literature review and found after reviewing 4600 studies, few studies are available on this topic.[1]  We call the Committee’s attention to one timely published study that compared the outcomes of hip fracture patients treated in SNFs and IRFs.  The study, “Effect of Rehabilitation Site on Functional Recover After Hip Fracture,” by Munin et. al[2] found that IRF patients had superior functional outcomes compared to those treated in SNFs when the same measurement tool was used.  The improved outcomes occurred during a significantly shorter rehabilitation length of stay and remained even when statistically controlling for baseline differences between groups.  The study called for further research to more fully understand the differences between rehabilitation treatment settings. Notwithstanding current available research, there is a significant need for prospective studies examining the provision of care among various settings providing medical rehabilitation services, SNFs, IRFs and LTCHs, to better determine how outcomes and treatment differ among these settings.  We would be pleased to work with the Committee in developing these studies as well as working with our colleagues in the medical rehabilitation field to engage in research efforts.

Patient Assessment Instruments

While post-acute care payment systems generate considerable data about each setting of care, the data is difficult to compare because each payment system uses a different data collection tool.  At its March 2005 meeting, MedPAC examined the various data sets and realized that they cannot be easily cross-walked with each other in order to compare the patients, outcomes, and costs, other than to observe broad outcomes such as mortality and readmission to acute care.   The LTCH PPS uses the standard UB 92 claim form.  The IRF PPS requires each facility to complete the inpatient rehabilitation facility patient assessment instrument (IRF PAI) as well as the UB 92 for each case.  The SNF PPS requires each facility to complete the Minimum Data Set (MDS) form for each patient and the UB 92.  The UB 92 form, while common to all settings, collects information solely on diagnosis codes and does not include any functional information.

Because these settings serve different populations and do not have any common functional assessment tools, outcomes at this point can only be measured at a broad level that is not truly representative and fails to measure the full impact of a rehabilitation program. As noted above, certain observations can be made about mortality, readmission to acute care and institutionalization of patients for the long term when referred to certain settings, such as SNFs.  However, in comparing these settings, there are significant limitations that were studied and acknowledged by MedPAC in its March 2005 discussion of post-acute care and patient assessment tools.  RAND repeatedly cautioned about some significant deficiencies in the obtainable data that limited the findings of the study.  First, controlling fully for selection is extremely difficult, and it is unclear whether the models capture this data in an accurate manner.  Second, RAND was unable to conduct a substantive analysis of patient function; thus, the outcomes analyzed are not the ideal outcomes measures for joint replacement patients. 

Similar to variances discussed in conjunction with the different payment systems, each tool used to assess diagnoses, comorbidities and medical functional status and cognitive status uses significantly different measurement items.  As a result, today it is simply impossible to assess outcomes and quality of care at the level necessary to accurately and fairly compare the various sites of care.

Recommendations

We think the issues facing policy makers, providers and patients relating to post-acute care payment and services would best be addressed through a broad, cross-site prospective study of these sites of care and the outcomes provided by their distinct treatment resources.  Not only do Congress and CMS need to have comprehensive and accurate data before engaging in any sweeping payment structure changes, such data will be crucial if the federal government intends to take any substantive, meaningful action that will save the Medicare system money while still protecting the quality of care given to beneficiaries nationwide. 

We recommend a multi-step approach to evaluating the state of post-acute care across settings for rehabilitation patients and implementing a new payment structure to capture the true costs of patient care.  As mentioned above, measuring function is the critical aspect of understanding a patient’s rehabilitation needs.  The approach outlined below should be viewed as a framework and could be amended or added to other studies designed to lead to creation of a new payment structure:

1.     Data Collection: CMS should use the IRF-PAI for data collection throughout the treatment sites in order to collect data and compare costs, patient characteristics, and medical and functional outcomes across sites.  Such a uniform data collection tool is necessary to eliminate the problems with the various existing tools and create one assessment instrument to cross walk to the three different tools currently used in post-acute settings.  Data should be collected at admission, discharge, and for a follow-up period.

2.     Creation of new Rehab Post-acute Care Groups (RPACGs):  New patient groups would be created using an expanded version of CMGs that would reflect function, age, diagnosis, LOS, and comorbidities for medical status, and the ICF conceptual approach.   Expanded CMGs would then be matched with costs to create new Rehab Post-acute Care Groups (RPACGs) and to develop appropriate weights.  The RPACGs would use a per discharge model using a discharge as the payment unit and episode of care.  SNF and LTCH patients who are not discharged and who exhaust their Medicare days should be tracked separately even after they exhaust their care and go on private pay or Medicaid for one year in order to establish total costs for that period.  Facility adjusters would be provided (wages, low income, rural, others), as well as special payment rules such as transfers, short stay, interrupted stay and outliers.  The groups would be matched with cost to develop the complete set of new payment groups reflecting payment rates for various types of patients receiving medical rehabilitation.  Payments would reflect patient characteristics (such as age, diagnosis, function, comorbidities, complications, length of stay, etc.) and resource use in whatever setting, eliminating the need to distinguish patients by current institutional sites or “silos” of treatment. 

3.     Adjustments: Adjustments would be made for facility specific costs as are currently recognized in all prospective payment systems (e.g. wages). 

4.     Revision of Payment System: Finally, after initial implementation, revision of the payment system would take place in order to provide bonuses for better functional outcomes.

As we know, therapy services, physician services and nursing services of varying intensity, length and costs are provided currently in these three inpatient settings.  These three sets of services, especially the intensity of therapy services, are key to the success of a rehabilitation program.  From a policy perspective it makes the most sense to reexamine this situation and realign the policies with the providers, payers and, most importantly, patients in mind.

AMRPA acknowledges that these ideas may appear quite radical, but we firmly believe that the study recommended here would help settle the current debates and assumptions and remedy recent action by CMS that is jeopardizing patient care.  CMS and Congress should continue its efforts to engage all the stakeholders, public and private, state and national, involved in this issue.  Each such entity has its own priorities and perceptions that will need to be addressed for any proposal to be effective and successful.

Conclusion

AMRPA cautions against adopting a simplistic viewpoint that growth in post-acute care is simply a function of substitution of care,  or adopting the attitude that “if you build it they will come.”  CMS’s rationale to date in making these assertions about substitutability has been that since there are few studies on point, the assumption must be correct (e.g., the absence of proof is the proof of absence).  Most post-acute care providers vehemently disagree.  We urge Congress to recognize that the federal government cannot adopt the improper assumption that these settings can provide similar outcomes at similar costs.  One only needs to look at the enormously detrimental effects of the 75 Percent Rule to see that such a policy will ultimately be grossly adverse to patient outcomes. 

We again commend the Committee for its interest in rehabilitation and post-acute care, and we look forward to working with you and your staff on these issues.

Inpatient Rehabilitation Facilities Provide a Rehabilitation Setting  Distinguishable from Skilled Nursing Facilities

COVERAGE CRITERIA

CMS assumes that post-acute rehabilitation care settings are readily interchangeable. In doing so, CMS ignores the enormous difference between the two care settings and the improved outcomes that occur at IRFs.

Sources:  IRF – Medicare Benefit Policy Manual §110.4 (Rehabilitation Hospital Screen Criteria)

   SNF – Skilled Nursing Facility Manual, Pub. 12, §214 (Covered Level of Care)