The Subcommittee on Population-Specific Issues and the Subcommittee on Data Needs, Standards and Security of the National Committee on Vital and Health Statistics (NCVHS) was convened on Monday, March 2 at the Health Care Financing Administration in Baltimore, Maryland. The meeting was open to the public. Present:
Lisa Iezzoni, M.D., Chair
M. Elizabeth Ward
Barbara Starfield, M.D., Chair
Jeffrey Blair, M.B.A.
Simon
Cohn, M.D., M.P.H., FACP
Kathleen Fyffe, M.F.A.
Vincent Mor, Ph.D.
George Van Amburg
Carolyn M. Rimes
William Braithwaite, M.D., Ph.D.
Robert Moore
Judy Ball
Marjorie Greenberg
Gail Janes, Ph.D., M.S.
Paul Newacheck, Dr.P.H., NCVHS
Robert Gellman, J.D., NCVHS
James
Scanlon, Office of the Assistant Secretary for Planning and Evaluation
(ASPE)
Jackie Adler, National Center for Health Statistics (NCHS)
Lynnette Araki, NCHS
Marilyn Field, Institute of Medicine
Ye-Fan
Glavin, New Health Management
Glenn Levy, New Health Management, Inc.
William Buczko, Ph.D., Health Care Financing Administration (HCFA)
Steven Clauser, Ph.D., HCFA
Helene Fredeking, HCFA
Marc
Freiman, Agency for Health Care Policy and Research (AHCPR)
Nell
Wood, Maryland Hospital Association
Catherine Hawes, Ph.D., Meyers
Research Institute
Tom Hoyer, HCFA
Korbin Liu, Sc.D., The Urban
Institute
Nancy Miller, Ph.D., HCFA
John Morris, Ph.D., Hebrew
Rehabilitation Center for the Aged
Harvey Schwartz, Ph.D., AHCPR
William Spector, AHCPR
Joan Buchanan, Harvard Medical School
Kathryn Crisler, Center for Health Services and Policy Research
(CHSPR)
James Phillips, State University of New York at Buffalo
Armen
Thoumaian, Ph.D.
Robert Burke, Muse & Associates
Bargara
Gage, Urban Institute
Brenda Spillman, AHCPR
Liliana Pezzin,
AHCPR
Karen Seitz, Veterans Affairs
Suzanne Niemeyer, Aspen
Publishers Inc.
Mary Waid, HCFA
Susan Levy, HCFA
Jennifer
Carter, HCFA
Linda Lebovic, HCFA
Karen Reilly, HCFA
Tim
O-Neill, Thompson's Subacute Guide
Fred Decker, American Health Care
Association
David Rabin, M.D., M.P.H., Georgetown University Medical
Center
Judy Sangl, AHCPR
Donna Pickett, NCHS
Krista
Robinson, EDS
Mike Grambo, St. Anthony Publishing
Elizabeth
Cornelius, OTR
Bill Sobaski, HCFA
Terry Moore, R.N., M.P.H., Abt
Associates
Helene Guilfoy
Chris Murtaugh, Ph.D., Center for Home
Care Policy and Research
Raymond Frederick, HCFA
Sue Nonemaker,
HCFA
Mary Wheller, HCFA
Bob Connely, HCFA
Peter
Shaughnessy, CHSPR
Janet Spinks, AQUAS, Inc.
Dorothy Henemann,
HCFA
Gerry Adler, HCFA
Peggy Parks, HCFA
On March 2, 1998, the Subcommittee on Data Needs, Standards and Security and the Subcommittee on Population Specific Issues held a special joint meeting at HCFA headquarters on the topic of data systems for long term care, home health care, rehabilitation care, and skilled nursing care.
Mr. Hoyer explained that HCFA is trying to think conceptually about long-range issues in post-acute care while simultaneously moving forward with practical activities in the near term. In the short range, HCFA must put in place several prospective payment systems (PPS) for home health, skilled nursing facility care, rehabilitation hospitals, and long-term care hospitals. The goal is to have a PPS for each setting by the year 2000. They will use different assessment instruments while HCFA looks toward building one that is more universally applicable, with the goal of having tools that serve multiple purposes and evolve along with the health care system.
Dr. Clauser reported that the Balanced Budget Act has increased the pace and focus of HCFA's work in this area, and he described some of specific projects.
Ms. Fredeking described HCFA's National Information System of Clinical Data Sets and noted these timetables:
The data sets described by Dr. Clauser and Ms. Fredeking will be made available to the Committee.
Committee members expressed interest in supporting HCFA's effort to build systems and instruments that permit the use of data across settings.
Dr. Field described a new IOM study project on long-term care that will look at the instruments for measuring quality and outcomes, the characteristics of persons requiring long- term care, and the role of various settings of care and their relation to other components of health care systems. Dr. Mor is on the IOM Committee.
Ms. Seitz reported that long-term care is the fastest growing segment within the VA system, and they are working to improve information in this area. They expect to have the RAI-MDS up and running by the end of 1998.
Mr. Freiman discussed the nursing home component of the 1996 Medical Expenditure Panel Survey.
Committee members and panelists discussed the fragmentation exhibited by the first set of presentations, noting that Congressional mandates have exacerbated the compartmentaliza- tion. They agreed on the need to think about care in the context of the person over time, a conceptualization for which the chronic disease model makes sense. The difficulty of applying the prospective payment approach to long-term care was also noted, as well as the infeasibility of Congressional time frames. Mr. Scanlon invited the Committee's advice as the Department tries to rethink its approach to provider-based surveys, and Dr. Iezzoni suggested that the Data Subcommittee working group take a look at that question.
Dr. Liu described two Urban Institute studies on post-hospital skilled nursing facilities (SNFs) and home health care. He said that very few variables were significant in predicting episode costs for post-acute care, and he predicted that research will not offer much guidance for policy formation in the area of integrating various forms of post-acute care. The study using the MCBS provides a baseline for looking at the impact of the BBA in the future. A focal group to watch is those using post-acute care who are especially frail and poor.
The Visiting Nurse Service of New York is the parent agency for the Center for Home Care Policy and research is an OASIS demonstration site, and Dr. Murtaugh noted some concerns about the reliability and validity of the data set.
Dr. Rabin described a study by a Georgetown group of whether the use of post-hospital home health care affects Medicare spending in the post-hospital period, or if it improves health [sic]. The study found that use of home health care improves patients' functional status and their general sense of well-being, especially if there is high use of home health care. It also increases program costs, but "less than would be seen just looking at the observed data." He stressed the importance of continuing the MCBS and other ways of assessing personal characteristics at the beginning of episodes of care as well as at discharge and some time thereafter.
Dr. Miller described a HCFA design project aimed at long-term data strategies for post-acute care. The goal is a system focused on the patient's health care needs and service preferences. The design phase will take two years, and then HCFA will decide whether to move on to a demonstration phase.
Asked how the Committee could best be involved and supportive, Dr. Miller said she would share design reports as they are released and would consider having a liaison on the technical advisory group.
In the discussion period, Committee members were concerned about moving forward as quickly as possible with the project described by Dr. Miller. Dr. Mor urged that HCFA commit to an overall strategy for developing a pan-instrument for post-acute care. Dr. Iezzoni expressed concerns about cream-skimming and the development of a situation analogous to DRG-creep.
Dr. Buczko discussed a project exploring the use of the Functional Independence Measure (FIM) and FIM functional related groups (FRGs) as a case mix measure. Overall, the study showed that the system was workable as a PPS for patient rehabilitation. He later explained that for now, at least, HCFA has decided not to use this system.
Dr. Hawes described 1) a comparison of Medicare utilization in a 10-state study of board and care homes and a 10-state sample of nursing home residents, and 2) a national study of assisted living for the frail elderly.
The group discussed the relative data gathering burden of the FIM and a RUGS-based system for rehabilitation and their relative merits. HCFA representatives explained the effort to minimize the number of systems being implemented at present. It was agreed that the Committee should stay focused on the long-term approach described by Dr. Miller.
Dr. Morris reported on the minimum data set for post-acute care (MDS-PAC) and its development. Version 7 will be released in April, and there will be a big data gathering for that iteration.
Dr. Glavin spoke about state policy, providers, operations, and data gathering for post- acute care. In Ohio, implementation of MDS has resulted in better provider care planning and quality care, increasing the facility's ability to do budget planning and projection. MDS offers a common language for a multidisciplinary staff and standardizes resource allocation based on patient acuity. The major challenges are system abuse and the consistency of Medicare- Medicaid MDS assessment and classification. Dr. Glavin recommended that the many demonstration projects be linked under a single conceptual framework.
Ms. Wood described her association's Quality Indicator Project, in which 1100 acute care provider sites now participate. It was developed in response to members' needs and has been adapted for various sites and types of treatment. A pilot test is being conducted in 67 sites in 18 states.
Discussion focused on the fact that HCFA does not currently envision a single data set for home health, and the inevitability of its having to decide on one eventually.
Dr. Shaughnessy discussed the Outcome-Based Quality Improvement (OBQI) for home health care, which will be implemented in late 1999, and OASIS, which he described as a small part of the system for assessing home health care outcomes. Ms. Crisler discussed the internal and external uses of this system and a 54-agency demonstration funded by HCFA. Committee members noted that the instrument does not facilitate understanding of a complete individual, but rather chops people up into body parts. Dr. Shaughnessy said the OBQI is having a positive impact on outcomes. He noted the commonality of applications frameworks among the different systems and suggested that the guiding principles and specific applications be made more explicit as data sets are developed. The Committee could facilitate the development of such a cross-cutting data set. He noted that OASIS was a spin-off of the process of specifying outcomes in OBQI.
Abt Associates has three major case mix-related projects funded by HCFA, two on nursing homes and one on home health care. Ms. Moore described the use of OASIS in the latter project, which uses OASIS-plus, an adaptation developed for greater utility as a case mix model. The research is testing whether OASIS can be used both for outcome-based quality improvement and to predict resource use. The first report is due in June 1998. Ms. Moore took issue with Ms. Crisler's characterization of the data collection burden as minimal, as the agencies in Abt's study have "struggled" with it. She noted that the OASIS is an evolving tool.
In the discussion, Committee members focused on concerns that OASIS does not take into account a population focus and the reality that people move in and out of home care. Other concerns about OASIS are the potential for gaming, the likelihood of a greater burden on patients in conjunction with audits, and possible problems akin to DRG creep.
Dr. Iezzoni chaired this special joint meeting, held at HCFA headquarters and focusing on data systems for long term care, home health care, rehabilitation care, and skilled nursing care. She noted some of the problems with the term "post acute care" and noted that this rapidly growing area of health care gets relatively little attention in many policy circles. The Committee's purpose is to get a systematic overview of what is going on in this area. She then introduced the first of six panels on the agenda.
HCFA's post-acute care project is an attempt to think conceptually about long-range issues in post-acute care while simultaneously moving forward with practical activities in the near term. In the short range, HCFA must put in place several prospective payment systems (PPS) for home health, skilled nursing facility care, rehabilitation hospitals, and long-term care hospitals. The goal is to have a PPS for each setting by the year 2000. They will use different assessment instruments while HCFA looks toward building one that is more universally applicable.
He reviewed the instruments used now or slated for use in each setting: The Minimum Data Set (MDS) with nursing homes, OASIS in home health (as of 10/99), RUGS-3 for skilled nursing facilities. For rehabilitation hospitals, the preliminary decision is to expand the resource utilization groups and the MDS. The vision of the future involves a system that is "more widely bundled" and that may involve both Medicare and Medicaid.
This office is responsible for integrating all the quality care and quality improvement strategies. The idea is to use clinical data sets for developing quality indicators, performance measures, surveillance and evaluation activities, and program accountability tools. There is growing appreciation for the use of clinical data for core business functions.
The Balanced Budget Act has increased the pace and focus of HCFA's work in this area. It imposes short deadlines and high expectations for program improvements. In response, the Office of Strategic Planning is studying the possibility of building payment system initiatives out of the OASIS information set for home health. It and the MDS (which is being extended to make it more applicable) are the cornerstones of the new initiatives in post- acute care. The idea is to have tools that serve multiple purposes and that evolve along with the health care system.
In addition to adapting and extending the MDS and OASIS, HCFA is working on the Uniform Needs Assessment Instrument (UNAI), which was developed to improve discharge planning. It will be refined to be more compatible with other HCFA assessment instruments. HCFA is also developing a clinical data architecture to assist the effort to make data more flexible. This is being done with HIPAA requirements for content and transmission standardization in mind.
Ms. Fredeking focused on the "nuts and bolts" of implementation, particularly of HCFA's National Information System of Clinical Data Sets. She noted these timetables:
The aforementioned data system involves a feedback loop among providers, their states, and HCFA. A national repository is being built that will reside at HCFA starting in July, for use in monitoring PPS. A free software for the MDS system, Raven, will help providers with implementation. Eventually there will be similar software for OASIS as well. There is also an 800 number for states to use. The Web site provides a major link for communicating with providers. Providers will be able to use the information for their own improvement activities, and it will also be used for surveys and certification. Eventually information will be made available to consumers.
Key issues in implementing this system include the scale of implementation, which Ms. Fredeking outlined; data confidentiality and privacy issues; and the accuracy of the input data. A plausibility index to look for data trends and analytical approaches for auditing the information are being developed for HCFA.
On request, Ms. Fredeking and Dr. Clauser said they would supply the Committee with copies of the data sets, which are still in development.
Dr. Cohn noted that in the existing HCFA data systems, patients seem to be in different boxes. Dr. Clauser agreed, but said the work toward a more comprehensive assessment instrument is moving along well in institutional settings; it is more problematic in ambulatory settings.
Ms. Fyffe asked for a written plan, but Mr. Hoyer said a plan at this stage of conceptualization would be artificial and not useful. He stressed the goal of using data across settings as the principle of the long-range effort.
Ms. Ward stated that the Committee shares this long-term goal, and she suggested that they work together.
Dr. Field described a project begun at the request of the Robert Wood Johnson (RWJ) Foundation "to look at the world 10 years after the Institute’s 1986 report on improving quality in long-term care." The current focus is on the shift of patients and care from nursing homes to a variety of settings. Information about the work of the IOM committee, is available on the National Academy of Sciences Web site (www.hap.edu -- click on current projects). The committee uses the term "long-term care" and will focus on conceptual issues, with a patient focus. The three aims are to look at strategies for measuring and improving quality and outcomes, the characteristics of persons requiring long-term care, and the role of various settings of care and how they relate to other components of health care systems.
Ms. Seitz works on geriatrics and extended care, the fastest-growing segment of the VA system. A significant portion of this care is provided on a contract and fee basis. Information systems in this area are not strong, and the MDS has not been implemented. The incentive to have comparable data as veterans move from setting to setting has led to the goal of implementing the RAI in VA nursing home care programs, and in the future obtaining MDS data on veterans in post-acute care settings. They expect to have the RAI-MDS up and running by the end of 1998, interfacing with the VA's Vista programs.
Mr. Freiman discussed the nursing home component of the 1996 Medical Expenditure Panel Survey. This component gathered data on nursing home use and expenditures, other health care use, the health and behavior of the residents, and characteristics of the facilities. The major focus is the use of nursing facilities and the costs of that care. There was a two- stage nationally representative sample of 800 nursing facilities, after which a sample of people using the facilities was followed. The MDS, when it was in place, was very helpful. Prescription drug data were collected every month through 1996, which adds richness to the data.
The first round of data were released in March 1997, applied to the first three rounds of the survey. The initial data sets will be out in August 1998, and a richer data set will be put out for public use in November 1998.
Finally, Mr. Freiman noted that the household survey also contains data that are relevant to the Committee's interests.
In response to a query about a 5-year plan for the VA, Dr. Mor cited the report of the Committee on the Future of Long-Term Care in the VA.
In response to another question, Mr. Freiman explained that the MEPS does not include patient satisfaction information because of the resistance of nursing home staff to having AHCPR deal with residents. Nevertheless, he expressed confidence in the data, which do not address patient satisfaction.
Dr. Mor noted the interest of many agencies and the National Committee in understanding people's movements from one setting to another. Mr. Freiman gave him some details about when certain MEPS data will be available for public use. Data from the longitudinal file will appear this fall, but the entire file will appear later. There are no plans to construct a timeline from the data in the larger community sample.
Dr. Iezzoni asked how the information from MEPS will inform future HHS policy decisions. Mr. Freiman said that while there is a lot of feedback among the agencies, there is no structured process in this regard. He added that there probably should be more involvement of government agencies.
Dr. Starfield asked how, or whether, the IOM committee would address HIPAA requirements and the core data elements. Dr. Mor said the committee is not likely to get into detail about specific data elements. As to what percentage of long-term care is being paid for by HCFA versus private payers and the states, Dr. Field said the IOM has commissioned a paper to sort out this question.
Dr. Iezzoni reiterated Ms. Ward's earlier point about the Committee's interest in helping people think through the issues addressed in this meeting. She noted that private payers and others need to be at the table for this process.
Dr. Cohn expressed his confusion at the fragmentation exhibited by the first set of presentations, which described data processes focused not on patients but on pieces of them or of the care process. Dr. Iezzoni agreed, noting that Congressional mandates have exacerbated the compartmentalization. Dr. Starfield observed that people need to think about care in the context of the person over time. Dr. Mor commented that the chronic disease model makes sense for thinking about the whole person over time. A variety of settings often contribute something to outcomes, so a provider-specific vision will not work. Dr. Starfield noted that HCFA's need for a prospective payment mechanism is a strong motivation, but one that is difficult to apply to long-term care. Dr. Iezzoni added that the time frames for producing the prospective payment systems are unrealistic.
Mr. Blair raised the question of whether the right data are being captured to identify failures that cause patients to cycle back to more expensive health care settings. The group agreed to keep this question in mind, as well.
Mr. Scanlon invited the Committee's advice as the Department tries to rethink its approach to provider-based surveys. These have traditionally been confined to single settings. Dr. Iezzoni suggested that the Data Subcommittee working group take a look at that question.
Dr. Liu described two Urban Institute studies on post-hospital skilled nursing facilities (SNFs) and home health care. In the first, they are using the Medicare Current Beneficiary Survey to look at who uses post-acute care, what distinguishes people who use SNFs from those using home health, and what are the episode costs for this care after a hospital stay. A related question was whether supply is driving post-acute care use. In addition to Medicare data, they used Medicaid related policy variables.
They found that people using post-acute care were predominantly older, female, on Medicaid, and disabled. The people for whom these characteristics (not including gender) were the greatest were more likely to use SNFs than home health. Need, not supply, was driving post-acute care. Medicaid variables were found to be significant, with some substitution between Medicare and Medicaid. The characteristics that predict SNF use also predict nursing home use.
Very few variables were significant in predicting episode costs for post-acute care, and the models explained little of the variance or heterogeneity. Dr. Liu concluded from this that research will not offer much guidance for policy formation in the area of integrating various forms of post-acute care. The study using the MCBS provides a baseline for looking at the impact of the BBA in the future. A focal group to watch is those using post-acute care who are especially frail and poor.
This study looked only at Medicare Part A services. Barbara Gauge of the Urban Institute will be analyzing A and B use, and there are plans for parallel analyses on episodes. The researchers were unable to answer questions of hospital ownership of post-acute care providers, which will be important to figure out. They did not look at the interface between post-acute care and long-term care.
The second study is a synthesis and analysis of information on costs, use and outcomes in post-acute care, based on a literature review. It focuses on the effects of Medicare coverage and payment policies on post-acute care use, and it will include a look at managed care organizations.
The Center's parent agency is the Visiting Nurse Service of New York (VNSNY) which is an OASIS demonstration site. The Center collects additional data on VNSNY home care clients in studies that it conducts and has found OASIS data limited in terms of services provided to home care clients. They have concerns about the reliability of the data from the OASIS instrument and wonder whether staff with different professional training, i.e., RNs PTs will differ in the way they record information. They also are concerned about the validity of the outcome data if this instrument is used alone. For example, persons in managed care tend to be discharged from home care sooner than others because fewer visits are authorized. Differences between managed care enrollees and others at discharge may reflect differences in where persons are in the course of their disease and not true differences in outcomes attributable to home health care. Information on the home environment and informal care is critical for assessing outcomes.
The Center is planning a study of managed care and the use of post-acute care. Access to data is an issue, along with data quality. A study of payment system changes is being considered, for example to look at whether the interim payment system in the BBA financially favors hospital-based home health agencies and their patients.
A Georgetown group is studying whether the use of post-hospital home health care affects Medicare spending in the post-hospital period, or if it improves health [sic]. The MCBS is the information source, with data from 3 years. They looked at Medicare spending for all services for the 6 months following discharge and at changes in health status between the baseline pre-hospitalization and 1 year later, and asked about the self-assessed general health status and the number of ADLs and IADLs. A problem in using the data on users is the inability to precisely measure the person's health status when the episode begins. In lieu of this, they used an instrumental variable method to identify factors that identify factors that influence home health care use but are not related to a person's health status.
The study concluded that the use of Medicare post-hospital home health care clearly increases program costs, but "less than would be seen just looking at the observed data." Use of home health care improves patients' functional status and their general sense of well-being, especially if there is high use of home health care.
Dr. Rabin noted that reducing access to and intensity of home health care to try and save money may be associated with a decrease in the patient's functional status. He stressed the importance of continuing the MCBS and other ways of assessing personal characteristics at the beginning of episodes of care as well as at discharge and some time thereafter. He also recommended that the pattern of service use and cost factors should be looked at in managed care. Finally, he observed that home health care is clearly being used by people with chronic illness who have an acute exacerbation, implying that home health care is to a considerable and increasing degree a long-term care benefit.
Dr. Miller described a design project being conducted by Bob Kane at the University of Minnesota. She reminded the group that HCFA is simultaneously pursuing short-term and long-term strategies for post-acute care; the first panel talked about the former, and her design project deals with the latter. Currently, the post-acute care system is benefit- and provider- driven. The aim is to design a system that is more focused on the patient's health care needs and service preferences.
She noted the relationship between Medicare post-acute care services and Medicaid long-term care: 35 percent of post-acute care expenditures are attributable to people who are also eligible for Medicaid. So researchers are looking at this relationship. The initial focus is on post-hospital post-acute care in SNFs, home health agencies, inpatient rehabilitation facilities and long-term care hospitals. Thereafter, they hope to address post-acute care that starts in the community.
This is design work for a demonstration, a phase that is expected to take two years. The contractors are developing an assessment tool to determine needed services and to engage patients and families in thinking about where they want to receive them. It builds from the UNAI and is coordinated with the work on the MDS. The other pieces are a beneficiary needs- based payment system that would be given to the patient; a case management component for helping the patient and family make decisions; and outcome measures with "some fairly generic constructs."
After the design phase is completed in two years, a demonstration will be designed and mounted. The first technical advisory panel meeting will take place in April.
Asked how the Committee could best be involved and supportive, Dr. Miller said she would share design reports as they are released, and would consider having a liaison on the technical advisory group.
Dr. Cohn praised the thinking behind the project, and noted the "disconnect" with current HCFA policy. He asked whether this project could be expedited in order to inform HCFA policy as soon as possible. Dr. Miller said the BBA directs HCFA to move from benefit-specific PPS systems to a more integrated system, but that process will take years. Meanwhile, there is an effort to coordinate instrumentation. Asked if this is "happening with all due speed," she said the design work will take 24 months; after that, HCFA will decide whether to move forward with a demonstration. Conceivably, it could skip over that phase, but that would leave many unknowns.
Noting the need for linkage as people move, Dr. Starfield asked the panelists to comment on what the various settings and data systems have in common. Dr. Rabin mentioned function and functional measurements, and the outcome measures based on them and on social functioning. Dr. Miller noted the need to reconcile the basic measures in OASIS and MDS, which need a consistent core and process of collecting the data. Dr. Liu observed that the problem with other secondary data sources like the MCBS is the limited number of variables. He stressed the need for coordinated multiple efforts.
Dr. Mor asked how the Committee could be useful in achieving HCFA's goal of coming up with a pan-instrument, and what its overall organizational strategy was. Dr. Miller described an incremental process through an internal HCFA work group, and Dr. Mor stressed the importance of coming up with a strategy.
Dr. Iezzoni expressed concern about the potential for cream-skimming and gaming, to the detriment of vulnerable populations, and she asked what policy guidance the Committee might give HCFA in this regard. She recalled the problem of DRG creep and predicted an analogous situation. Dr. Rabin said particular attention should be paid to what happens to people in home health who have substantial disability. Dr. Liu noted that survey data will give insights into who has access to post-acute care services, but he acknowledged that these data take a long time to gestate. Dr. Murtaugh noted that dual eligibility is a marker for frailty, poverty, and an inadequate supply of services. Dr. Miller said HCFA is concerned about these problems. The nursing home demonstration is monitoring how people do. Dr. Iezzoni noted the problems with having outcomes monitored by people with an interest in showing positive ones. Dr. Liu agreed that a problem like DRG-creep is likely.
Dr. Mor noted the potential benefits of linking HCFA's work with that of the New England States looking at dual-eligible packaging.
Dr. Buczko discussed a project on functional assessment measurement using the Functional Independence Measure (FIM). The growing demand for PPS for inpatient rehabilitation has led to exploration of using a functional status measure as a case mix measure for that purpose. HCFA let a contract to examine the feasibility of using the FIM and FIM FRGs (functional related groups) for a PPS, using data obtained from UDSMR because HCFA has not mandated functional status data reporting from rehabilitation facilities.
In the evaluation, the panel supported using the FIM as the instrument and suggested possible improvements for the cognitive scale and study of the reliability of FIM and the effect of comorbidities on classification accuracy and variance. Overall, the study showed that the system was workable as a PPS for in-patient rehabilitation.
Dr. Iezzoni expressed confusion about the fact that the first panel said HCFA is planning to use RUGS or a RUGS analogue rather than this one. Dr. Buczko explained that indeed, the system described above is "an alternative that has been relegated to secondary status in the current debate."
Dr. Hawes focused on assisted living and residential care, which she noted varies greatly among facilities and states. Residential care is called by more than 30 names in the U.S. She described two studies: 1) a comparison of Medicare utilization in a 10-state study of board and care homes and a 10-state sample of nursing home residents, and 2) a national study of assisted living for the frail elderly.
The studies found that nursing home residents are much older and more than two-thirds have moderate to severe cognitive impairment. 60 percent are receiving assistance with six to seven ADLs, compared to 3 to 5 percent of residential care residents. Residential care has much higher use of inpatient care and hospital ER; it appears that the hospital is used to provide low-level nursing care, and these residents are much heavier and more expensive users of acute and post-acute care services.
Dr. Hawes stressed that it is wrong for the National Health Provider Inventory to exclude residential care from its listing of facilities. Also, standardized data on resident status are needed, as well as more information on cost and quality of care. It appears that the acuity of residents' conditions is growing in residential care. Finally, she noted that states have been asking for a uniform resident assessment system.
Asked how states are choosing to use the RAI for assisted living in adult care, Dr. Hawes said the process is haphazard. Many states are moving toward requirements for assessment and care planning because of the increased acuity of residents' conditions.
The group discussed the relative data gathering burden of the FIM and a RUGS-based system for rehabilitation. Dr. Buczko said the burden of a RUGS system would probably be greater because it uses a larger instrument. Dr. Hawes noted that it explains more variance in resource use because it catches comorbidities. Dr. Mor said the R-squared of the RUGS is higher for all populations. Ms. Rimes said the FIM FRG has not been eliminated as a possibility for the future. Dr. Miller explained that the reason for using the RUGs in the short run was to minimize the number of data systems. Ultimately, it was agreed that the Committee should stay focused on the long-term approach described by Dr. Miller.
Dr. Phillips (Uniform Data System for Medical Rehabilitation) described the architecture already in place in the VA system, which uses Web technology and collects FIMs across the continuum of care and has been in operation for 18 months.
Dr. Morris reported on the minimum data set for post-acute care (MDS-PAC) and its development. He began by explaining that the resident assessment instrument (RAI) is an encompassing system that includes the MDS, which is itself a series of standardized assessment items. Several tools have been built off the RAI, including the RUGS-III system, the MDS- PAC and the MDS 2.0. Other components include resident assessment protocols (RAPs) and a series of quality indicators. The system was created from the point of view of the resident. There is an effort to capture clinically relevant information.
The developmental process has included several focus group meetings with industry segments. The current draft of the MDS-PAC is #6, and it has been sent to some 700 people for comments. The expert panel is about to have its second meeting. Version 7 will be released in April, and there will be a big data gathering for that iteration.
Dr. Morris explained that MDS 2.0 provides a lot of computerized data for understanding the patient population, but there is no comparable data from rehab facilities and specialty care hospitals. The Hebrew Rehabilitation Center has a contract with HCFA to gather longitudinal data on a sample of that population. Four data-gathering efforts will be concluded by the end of 1998.
Dr. Glavin spoke about state policy, providers, operations, and data gathering for post- acute care. Ohio initiated the use of MDS in 1990 and implemented payment systems in 1992. It now uses MDS-plus as well as other programs. They have found that occupancy in the long- term care skilled nursing facilities has declined. Implementation of MDS has resulted in better provider care planning and quality care, increasing the facility's ability to do budget planning and projection. The major challenges are system abuse and the consistency of Medicare- Medicaid MDS assessment and classification.
Dr. Glavin showed statistics from the Benjamin Rose Institute, which has a variety of post-acute and long-term care services. It is the Northeast Ohio PACE site. It downloads MDS every week for quality review and resource management. The key conclusions are that MDS offers a common language for a multidisciplinary staff, and it standardizes resource allocation based on patient acuity. There is a 90 to 95 percent accuracy in predicting resource allocation. The accuracy and projection power could use more work, however.
The Committee received a handout on a future post-acute care and integrated care system. Dr. Glavin recommended that data collection be driven by good definitions of post- acute care and a framework of post-acute or integrated care, and that the many demonstration projects be linked together in a conceptual framework.
Ms. Wood reported on the Quality Indicator Project of the Maryland Hospital Association (MHA). It was begun 12 years ago for acute care facilities and expanded at the request of members who needed to compare clinical performance across organizations. A three-year research and evaluation project was funded by the Robert Wood Johnson Foundation. Today, about 1100 post-acute care facilities participate in the QI project. Ms. Wood noted that the Maryland Hospital Association is governed by users, and they are "right on the mark" in knowing what they need. The QI project has expanded into psychiatric care and long-term care using the same philosophy, which she characterized as "essentially, parsimony."
In consultation with members, they developed a list of 100 measures which they grouped into seven indicator domains. She listed the criteria for selecting the indicators and said they tried to stay close to the MDS. In regard to risk adjustment, she noted that while many organizations insist on risk-adjusted indicators, there is "loud debate" over the most appropriate methodology for this purpose. The technological sophistication of the facilities is another issue in this regard.
The MHA stresses the need for commitment in implementing the QI project, and requires a letter from the CEO of institutions committing to the pilot test process. 150 organizations applied to be in the pilot test, and 67 were selected, representing 18 states. The program has also stressed data reliability and a commitment to participate in reliability surveys and conformance assessment surveys.
Finally, Ms. Wood reviewed the long-term care indicators, some but not all of which are obtainable from the MDS.
In response to a question, Dr. Glavin said the MDS is the most common information set among the post-acute care facilities in Cleveland. Some doubts were raised by Committee members about the reported 90-95 percent accuracy in predicting resource use.
In response to another question, Dr. Morris said HCFA will decide whether to use the MDS-PAC for home health care. It has been shown that MDS elements have the same reliability in home care as in nursing homes. Dr. Iezzoni observed that HCFA's arguments for using a single approach with rehab hospitals does not seem to apply to home health. Mr. Moore responded that eventually HCFA would have to think about the minimally necessary information that cuts across the continuum of care, and to compare the existing vehicles and build the necessary crosswalks. He noted that the current ones have different measurement philosophies and serve different purposes, and each has strengths and weaknesses.
Dr. Shaughnessy framed his talk in terms of the Outcome-Based Quality Improvement (OBQI) for home health care, which meets HCFA's need for developing a prospective payment system and will be implemented in late 1999. OASIS is a small part of the system for assessing home health care outcomes. He agreed on the need for more convergence in the long run to generate common data items among the various data sets.
OBQI has been under development for 15 years. It can be used by home health agencies and by government and regulators. Ms. Crisler explained that agencies develop outcome enhancement activities in response to outcome reports. This is happening in a 54- agency demonstration funded by HCFA and in OBQI projects in the State of New York and elsewhere. The 54 demonstration agencies are about to receive their outcome report for the second year.
Dr. Shaughnessy described the OBQI's 15-16 quality indicator groups, for each of which outcomes are reported. The goal of the system is a growing body of knowledge with which care providers can enhance outcomes. He showed a hypothetical report as an illustration, noting the absence of cross-cutting aggregate outcomes because care providers need more specific outcome indicators. Dr. Starfield observed that the instrument does not facilitate understanding of a complete individual, but rather "chops people up into parts of bodies."
The OBQI is having a significant impact in terms of improved outcomes. Dr. Shaughnessy noted the commonality of applications frameworks among the different systems, and he suggested that the guiding principles and specific applications be made more explicit as data sets are developed. The Committee could facilitate the development of such a cross-cutting data set.
Underpinning the OBQI are some 10 studies, funded largely by HCFA and the Robert Wood Johnson Foundation, looking at questions of feasibility and of what the outcomes should look like. The process began by specifying outcomes, not data items. Considerable thought has been given to the risk factors that are important to each outcome. OASIS was a spin-off of this process of specifying outcomes in OBQI, and there was sensitivity to the needs of home health care providers. The developers consider OASIS a data set, not an instrument. A decision has been made not to mandate a comprehensive assessment instrument for home health care.
Ms. Crisler said that the OASIS has been well received by the industry. A data collection burden "does not exist," because the items are integrated into agency records and need only be cut and pasted. The data are then transmitted electronically.
Dr. Shaughnessy said that in the future, Medicare will move toward using a system such as this. OASIS and OBQI can adapt core data items that will cut across the post-acute care setting.
Abt Associates has three major case mix-related projects funded by HCFA, two on nursing homes and one on home health care. Ms. Moore described the use of OASIS in the latter project. They recruited and trained 90 agencies in eight states to collect a version of OASIS called OASIS-plus. They adapted OASIS to make it more useful for developing a case mix model. The research is testing whether OASIS can be used both for outcome-based quality improvement and to predict resource use.
Data collection began in October 1997 and will be collected for 18 months. The interim report is due in June, but Abt does not expect to know by then which items are predictive of resource use. Ms. Moore described the process of developing the tool, for which they convened a clinical panel to talk about what predicts resource use in home care. They were seeking clinically meaningful data. There was a decision to add 41 items on cognition and in other domains. The OASIS-plus is a combined tool with measures from both OASIS and the MDS for home care.
Ms. Moore took issue with Ms. Crisler's characterization of the data collection burden as minimal, as the agencies in Abt's study have "struggled" with it. She noted that the OASIS is an evolving tool and agreed that it can be made consistent with other measures used with the post-acute care population. The major problem is not knowing how frequently characteristic information on home care needs to be collected to measure resource use. In addition, some of the administrative requirements for Medicare-certified agencies complicate things. The reliability of the data the agencies will submit is not yet known. There will be more answers in a year.
Dr. Starfield observed that no other country requires payers to look at quality in the kind of detail described by the panelists. She characterized OASIS as an academic exercise the relevance of which she questioned. It does not take into account a population focus and the reality that people move in and out of home care. Dr. Shaughnessy agreed that neither the OASIS nor any other existing approach to outcomes is population based. He agreed that ideally, information on patients should be longitudinal, but he asserted that this will not happen for at least a decade. In the meantime, it is necessary to make sure that the effectiveness of care providers needs to be monitored.
Ms. Crisler commented that the OBQI works well with agencies' own internal performance improvement activities, and the quality indicators are more scientifically sound than anything they have had in the past. She said the home health agencies in the study are "a leadership sample" and were not randomly selected. They represent a range of sizes. Dr. Iezzoni commented that it is natural for home health agencies to like OASIS, because they and their staffs are reporting on their own performance. Ms. Crisler agreed that agency employees have a motivation to make themselves and their agencies look good, and audits are critical. Dr. Iezzoni expressed concern about the added burden on patients necessitated by audits.
She also noted the potential for gaming when this is used for reimbursement, but agreed with Dr. Shaughnessy that this is a generic issue that is not particular to OASIS or home care. He explained that they went with capacity-based rather than performance-based measures of functional health status because capacity seems to have greater validity. However, there is a tradeoff with reliability. Dr. Mor said it would be desirable to look at the conceptual split in the interpretation of the relative valence given to capacity versus performance in a variety of ADLs and IADLs, and the implications for building crosswalks.
Finally, Dr. Iezzoni urged that priority be given to patients' perceptions of their functional status, rather than having the care provider make the determination. She also noted patients' and families' concerns about telling the truth about their status because of the possible ramifications for insurance coverage. On behalf of both Subcommittees, she thanked HCFA for hosting this meeting. After also thanking all the panelists and staff for their contributions, she adjourned the meeting.
We hereby certify that, to the best of our knowledge, the foregoing summary of minutes is accurate and complete.
/s/ Lisa I. Iezzoni July 5, 1998
Lisa I. Iezzoni, M.D., M.S., Chair Date
/s/ Barbara Starfield July 15, 1998
Barbara Starfield, M.D., Chair Date