Lisa I. Iezzoni, M.D., M.Sc., Chair
Hortensia Amaro, Ph.D.
Vincent
Mor, Ph.D.
George H. Van Amburg, M.P.H., M.S.
M. Elizabeth Ward,
M.N.
Carolyn M. Rimes, Lead Staff
Dale C. Hitchcock
Sam Shekar, U.S. Public Health Service, Region I
Tony Robbins,
M.D., U.S. Public Health Service, Region I
Jack Fowler, Jr., Ph.D.,
University of Massachusetts (UM)
Philip Caper, M.D., Codman Research
Group, Inc. (CRG)
Jo Anne Lutz, CRG
Manon Spitzer Ruben, CRG
Brian Burwell, MEDSTAT Group
Barbara Fain, Office of the
Attorney General
Wendy Mariner, J.D., M.P.H., Boston University
School of Public Health
Nora Wells, M.S.Ed., Family Voices
Elliot
Stone, Massachusetts Health Data Consortium, Inc.
Robert Restuccia,
Health Care for All
Phyllis Freeman, J.D., UM Law Center
Kate
Ackerman, M.D., Boston Medical Center Health Net Plan
Castellano
Turner, Ph.D., UM
Sally Bachman, Ph.D., Boston University School of
Social Work
Debbie Thompson, J.D., Alzheimer’s Association of
Eastern Massachusetts
Zolia Torres Feldman, M.S., R.N., Great Brook
Valley Health Center
Lisa Federman, Neighborhood Health Plan
Carol
Tobias, Medicaid Working Group
Diane Flanders, Division of Medical
Assistance (DMA)
Bruce Landon, Ph.D., Harvard Medical School (HMS)
Representative Harriette Chandler, Joint Committee on Health Care of
the MA Legislature
Richard Frank, Ph.D., HMS
James Michel, M.A.,
M.B.A., Massachusetts Mental Health Substance Abuse Carve Out
Barbara
Dickey, Ph.D., HMS
Mary Beth Fiske, M.P.P., DMA
Anthony
Asciutto, M.A., M.P.H., DMA
Marjorie Porell, M.S., DMA
Andrew
Bergman, Tufts Associated Health Plan
Carol Upshur, Ed.D., UM
Frank
Wong, Institute of Asian American Studies
Paul Newacheck, Dr.P.H., NCVHS Member
Kathryn Coltin, M.P.H., NCVHS
Member
Daniel Friedman, Ph.D., NCVHS Member
Marjorie Greenberg,
Executive Secretary, NCVHS
Lynnette Araki, M.P.H., National Center
for Health Statistics (NCHS)
Mary Moien, Health Survey Consultant
Sharon Sokoloff, Medical Outcomes Trust
Kevin O'Toole,
Massachusetts League of Community Health Centers
Jason Goldwater,
NCHS
Alba Cruz, HCFA
Lein Han, HCFA
Melissa Sinrich, Boston
University
Elinor Sacholitzky, Tufts Health Plan
Rachel Block,
HCFA
The Subcommittee on Population-Specific Issues met in Boston on April 14-16 for the second of its site visits involved in its study of data issues related to Medicaid managed care. They held discussions with nine panels over two days, and then met in executive session for a day. The need for better data on race, ethnicity and language was a major and recurring theme in the meeting. Other themes included a variety of quality issues, the paucity of data on people with disabilities, and the need for more information on substance abuse and mental health patients and services. The confidentiality dimension of these and other information issues was another prominent theme.
The organizations represented on this panel work with various states on data and information. The panelists reviewed the data problems they encounter with Medicaid populations and described their respective efforts to get better data --e.g., by participating in the CAHPS project. Arguments were made both for encounter data and for supplementing them with population-based data. The underlying issue in all cases was data quality.
The discussion period focused on the difficulty of comparing states, concerns about response rates, and race and ethnicity data collection and analysis issues. There was considerable discussion of the latter topic by this and other panels. One point that emerged was the difference among states in the amount of information they have on this variable, partly because of differing policies toward collecting race/ethnicity data. On another subject, the group agreed on the unsuitability of the CAHPS instrument for people with cognitive limitations, and on the need to find better approaches. The difficulty of getting information on people with disabilities also was discussed at some length, including the confidentiality issues related to gathering data on them.
Ms. Fain described the active efforts of her office to protect consumers' interests in managed care. They receive and follow up on complaints, recommend particular types of data collection, and initiate policy. The Attorney General's office uses regulatory oversight to make sure that standards are followed, with the goal of creating a level playing field and preserving the positive qualities of managed care. Dr. Mariner pointed out that data must be relevant to both "consumers" and "patients." She also encouraged efforts to gather data on what did not happen in addition to what did happen. Ms. Wells talked about the problems families with children with special needs are having with managed care, particularly when mental health services are needed. She stressed the importance of involving consumers in planning and carrying out data collection and analysis. Family Voices is launching a national survey of families of children with special needs.
In the discussion period, the Subcommittee members were interested in the Attorney General's input into data specifications for quality control, and in Family Voices' solutions to getting good survey response rates from families.
Mr. Restuccia discussed the use of data by and for consumers, especially in a Medicaid managed care context. He stressed the need to supplement data with other factors for quality control. He also called for interpretive measures to help consumers use data. Mr. Stone stressed the goal of standardizing data repositories to assist research and analysis, encompassing insurers, purchasers, and health plans and providers and taking care not to isolate Medicaid managed care from other sectors. He urged the development by NCVHS and the National Library of Medicine of a "meta-dictionary" to be used across the data dictionaries held in public and private agencies. He also had some caveats about the need for flexibility and the reality of change in core data elements. He used a network image to describe the kind of data repository needed for medical records, given that patients change providers several times in a lifetime. Finally, he urged the Committee to "let the market prevail" rather than trying to select single definitions.
The discussion period began with a recounting of a conversation between Dr. Iezzoni and Mr. Burwell which revealed that MEDSTAT responsibility to ensure compliance with HIPAA for its State Medicaid client, was not included as a part of their contract with Massachusetts Division of Medical Assistance. The group then discussed the lagging awareness (or acceptance) of HIPAA in the country, notably by Medicaid programs. It was noted that every RFP should require compliance with HIPAA. Ms. Greenberg reported on the effort to bring more public health participation into standards activities, with a workshop planned for November.
Much of the discussion then focused on the need for more interpreter services and more information about the languages of patients and providers in Medicaid managed care. The group also explored Mr. Stone's ideas about a compendium and discussed the need for non-traditional types of data on Medicaid managed care populations.
Ms. Freeman discussed Medicaid in terms of its public health function, noting that society expects much more from medical care than it can deliver and much less from public health than it can accomplish. She echoed a recurring theme--the special opportunity Medicaid managed care offers as a set of public programs without proprietary restrictions and with the ability to work with private providers in new ways. She cited four examples from different states of the use of reporting requirements to yield useful information to improve the health of the Medicaid population. She then discussed her concern that fears about abuses of health records will overshadow their good uses for improving health, referring to a paper she co-authored with Dr. Robbins that outlines the issues and ways to address them constructively.
Dr. Ackerman described her experience as medical director of a Medicaid managed care plan, and in particular the difficulty of getting and using data. She also talked about what she looks for in data to support clinical processes. Among other things, she noted the preponderance of substance abuse as a medical issue, and the need for better data about it.
Dr. Turner discussed the McCormick Institute's contract with the state Medicaid program, looking at the impact of race, ethnicity, and language on access and providing some education for DMA staff. He described his own learning process about the diversity among blacks, 35 percent of whom in Massachusetts are foreign born and many of whom speak Creole. Among other factors, the investigators studied the disparities in life expectancy and whether these differences can be correlated to lack of access to medical services. They found that this appears to one of many causes. He noted the difficulty of reaching conclusions because of the limited availability and quality of race and ethnicity data.
In the discussion period, the group focused primarily on issues surrounding race and ethnicity data and the shortage thereof for HMO use. Dr. Robbins' assertion that race is often mistakenly used as a proxy for social class generated some lively discussion, with Dr. Turner observing that some differences in health status seem to be explainable only in reference to race or ethnicity, independent of class. The Subcommittee's relevant comments on the OMB Directive 15 revisions, authored by Dr. Amaro, were referenced. On another topic, Dr. Ackerman was questioned about what information she would like to have on enrollees' substance abuse problems.
Dr. Bachman proposed that Medicaid managed care be viewed as a form of privatization, with the same incentives for both the public and private sectors. This raises the question of what kind of information, and how much, the government needs to fulfill its responsibilities. She noted that the Medicaid program now will be working in alliances with other funders and providers.
Ms. Thompson discussed the dearth of data on long-term care and recommended that the National Committee encourage the government to look for effective ways to measure long-term care, including quality measures for disabled populations. Dually eligible populations present a special challenge for data collection and program management, and she also recommended attention to models for integrating managed care with Medicaid for dually eligible people. She called particular attention to the need for better mental health measures for disabled people and seniors.
Ms. Federman stressed the need for standardized indicators across states, for service to vulnerable populations. She also talked about the trouble safety net HMOs and providers are having in assessing the uninsured population when Medicaid people are moved into managed care. Also, plans need more data in order to protect the interests of people with disabilities.
During the discussion period, the group talked about the data collection burden on local providers, the issue of integration versus carve-outs for behavioral health, and confidentiality issues.
Ms. Tobias spoke about the kinds of data they are looking at and using, focusing on three issues: benefits and service utilization, cost of health plans, and people with disabilities. They have worked with Medicaid claims data to develop risk adjusted payment systems for people 65 or older with disabilities. She emphasized the importance of paying adequate rates, especially for people with disabilities, to enable plans to provide care and develop operate their information systems.
Dr. Landon noted the variability across states in the implementation of managed care programs. He emphasized the need for technical assistance in standardizing data collection and dissemination.
Rep. Chandler said the legislature often is forced to make policy based on anecdotal information because of the insufficiency of good data, and she talked about what kind of data are desirable to inform policy. She discussed the likely consequences of shrinking Medicare resources for seniors and others, and predicted unsustainable pressures on the Medicaid program as a result of the Balanced Budget Amendment and other stresses. She described the legislature's work on reforming Medicaid managed care was one of several panelists to call for national standards that would make inter-state comparisons possible. She also discussed the work on State confidentiality legislation, and was urged to follow a balanced approach that recognizes the positive uses of information. In answer to a question, she speculated that the State's leaders had the resolve to pay for the kind of information system the state needs to deliver quality care. She stressed that the federal government must help pay for the information requirements it is heaping onto the States.
Diane Flanders from the Division of Medical Assistance at the State Medicaid Office described their involvement with the Program for All Inclusive Care for the Elderly (PACE).
Dr. Frank advises states on procurement and the bidding process. In this role, he has experienced the ways that data gaps undermine rational bidding and the ability to compare bidders objectively. The same data gaps can ruin businesses, because the assumptions they have to make in their bids may prove overly optimistic. He discussed the impact of selection and the need to analyze the factors that drive it -- something that can be watched in an experiment underway in the Maryland Medicaid program.
Mr. Michel spoke from the perspective of a public purchaser of mental health and substance abuse treatment services. He recommends the use of simple report card formats for performance monitoring, and he offered some of the questions they should answer. Noting that Medicaid is a minority funder in this arena in most states, he stressed the importance of knowing the status of interagency relationships between Medicaid and other agencies that are co-funding services for vulnerable populations.
Dr. Dickey spoke of the data aspects of her research on Medicaid managed care in Massachusetts and also from the provider's view about how McLean Hospital responds to requests for data. She emphasized quality of care issues for behavioral health clients and the difficulty of getting adequate data to measure and monitor this. She noted the need for clinical as well as administrative data to get a quality of care issues. She concluded with several recommendations to the Subcommittee.
The discussion period pursued several themes relating to data issues involved with substance abuse and mental health services, particularly in regard to measuring quality of care for those inside and outside the treatment system.
Ms. Porell cited six categories of data issues: sources, collection approaches, data elements and assessment methods, data quality, resource constraints, and intra- and inter-agency cooperation.
Mr. Asciutto gave an overview of the three Medicaid managed care programs in Massachusetts and some of the data collection practices related to each for the purposes of quality management. They use both HEDIS instrument, which they occasionally adapt, but have found it difficult to find benchmarks in HEDIS appropriate for the population. They are developing uses of the CAHPS.
Ms. Fiske commented further on DMA's data collection for quality management in DMA, focusing on activities specific to the HMO program. Their quality improvement initiatives are linked to contract requirements and designed to determine whether the plans are meeting those requirements. She described the DMA's development of quality improvement initiatives with HMOs to move the plans toward desired management conditions, as well as the consequences if they do not improve. She stressed the need to negotiate goals and performance measures with plans at the outset.
In the discussion period, Subcommittee members praised the impressive information activities underway in Massachusetts. The panelists from DMA noted that the state has had a Medicaid managed care program for longer than most states, and it has good information resources to draw upon. The group returned to the issue of becoming HIPAA compliant. There was further discussion of the problems with race/ethnicity data, which partly stem from the fact that the question is optional on the enrollment form. This information is reportedly most reliable when self-reported on member satisfaction surveys. The challenges of enrollees' multiple languages also were noted.
See detailed meeting summary for day two.
Dr. Iezzoni thanked the staff of the National Committee for their work preparing the meeting, and members of the Subcommittee introduced themselves. Sam Shekar and Dr. Tony Robbins, representing the New England regional office, welcomed the Subcommittee and spoke briefly. Mr. Shekar noted the importance of the discussion in New England, where over 55 percent of the Medicaid population is under some kind of managed care. The issue of public accountability and the intersection of managed care and traditional health care delivery is of particular interest to the region.
Dr. Robbins, editor of the Regional Public Health Reports, said that the region has strong state health departments that take seriously the question of how managed care could contribute to public health. Earlier this year, regional health officers met with major managed care plans to work toward collaboration. He called attention to a recent article in his journal that makes a strong argument for protecting the public uses of medical care data.
The Center has been working with the Department of Medical Assistance (DMA), the state agency in charge of the Medicaid program, to get feedback from enrollees about their experience with managed and "not so managed" care. Part of the national Consumer Assessment of Health Plans (CAHPS) project, DMA and CSR are looking for the best ways to access information and test the CAHPS instrument under development. They are attempting to develop a set of instruments for different populations that will get comparable data from recipients of Medicaid and privately insured people. Areas in which they want to monitor include access, interactions with providers, and interactions with plans.
One of the challenges they have found is in obtaining information about costs from those who are served. Often patients and their families on Medicaid, like others are unaware of what is covered by plans. A goal of the CAPHS project is to identify areas in which the respondents can provide information. Collecting data from the Medicaid populations is also a challenge. They are testing different modes of data collection to determine which are most effective, as well as which questions and measures behave consistently in different modes.
Another problem they encounter is the many languages spoken by Medicaid populations. Questionnaires are written both in English and Spanish, but it is not possible to know in advance which instrument is appropriate. There is the potential that questionnaires presented in two languages could have a suppressive effect on English speaking people. However, they are testing the Canadian approach of sending a questionnaire hat is in Spanish and English. When asked whether CAHPS is presently translating its instruments into languages other than English and Spanish, Dr. Fowler replied that eventually there may be other languages, but not at present.
In addition, CAPHS is trying to collect essential data from adolescents. Dr. Fowler explained that the CAHPS vision is to have a core of questions applicable to all patients, with supplemental questions for adolescents and persons with special needs or in special circumstances.
Dr. Caper characterized CRG as a vehicle for developing scientifically rigorous population-based tools for medical management. The company assists clients in converting raw data into statistical information for monitoring, surveillance and management. Their Medicaid clients include several states for whom they convert data from client claims and enrollment information.
Dr. Caper observed that unlike patient-based clinical studies, population-based statistical analysis is a relatively "forgiving" methodology. Their analytic efforts provide data on status, access to care, resource use and costs of defined subsets of the Medicaid population. CRG assumes that appropriate safeguards will be used in collecting and reporting detailed medical data. Dr. Caper listed three reasons a comprehensive view of medical care is vital to Medicaid programs: (1) only a population-focused model can detect under-utilization by monitoring rates of events; (2) the population-oriented view is more representative of care as it is experienced by program recipients than conventional provider-center views; and (3) necessary health risk assessments can only be applied through comprehensive population-based measures.
Since methods for grouping and assessing morbidity are evolving rapidly and will continue to improve, they must continually incorporate improvements into their tool kit.
Ms. Lutz presented examples of transitions from fee-for- service to managed care organizations now serving state Medicaid populations. Knowledge about the composition of the eligible group for managed care plans can now be used in a systematic approach to managing clients' health care -- for example, by identifying the percentage of services supplied by managed care organizations. Currently, MCOs are spending $450 less per year than the statewide average, yet their health status indicates that they are expected to spend seven percent more than the statewide average. Data from claims and encounters, coupled with health survey instruments, are increasingly useful to assess the point at which levels of care meet state guidelines.
Admission rates continue to be a good way of measuring the continuity and availability of primary ambulatory care. There may be a correlation between improved continuity of ambulatory care through managed care and reduction of hospital admissions. She concluded with a remark about the need for states to bring their encounter data up to the level of traditional fee-for-service claims data. She stressed the importance of ongoing dialogues between state Medicaid organizations and managed health plans to monitor the quality of care delivery, contractual requirements and sanctions.
MEDSTAT is a health care information technology company that develops and designs proprietary analytical tools for looking at data and decision support products. They also provide encounter data technical assistance. Their current client list includes several state Medicaid programs. They are now building an encounter data base for the Massachusetts state health program, which includes the state HMO program and primary clinician plan as well as the state's behavioral health program, a carve out. Also, MEDSTAT will access and integrate fee-for-service pharmacy claims and encounter records, which will be linked with the Medicaid eligibility file. Their contract also involves transmitting a minimum data set to HCFA, using this administrative data set to generate quality indicators for Massachusetts health program enrollees. The state is interested in the possibility of risk adjusting capitation payments to their HMO plans, for which this data base will be an information source for adjusting rates.
Mr. Burwell discussed their approach to building data bases from encounter records, in which they use a proactive approach to ensure quality of information, providing detailed data specifications to each plan. They are currently receiving test data. He identified as a problem system deficiencies in managed care plans that prevent the capture of encounter data within plans. Another area that needs work is standard processes for validating encounter data, in the absence of standard procedures for identifying points along the data generation and transmittal path where quality problems are more likely to occur. Interventions that could be targeted toward those problems are also needed.
Their group feels strongly that person-level encounter data is essential for effective management of state Medicaid programs and that an important part of ensuring compliance with data requests is talking directly to plan representatives about how the data will be used. It is also important to provide valuable feedback to the plans from the data collection process and to include them in the process. Though not easy, it is possible to develop Medicaid encounter data systems that contain complete and reliable data. For this, a continuous process for data quality improvement is necessary. He described how they use the same programs to generate similar data sets for Medicaid managed care plans they send to the Group Insurance Commission.
In response to a question about variations among states, Ms. Lutz said that one difficulty in cross-state comparisons is that socioeconomic differences must be taken into consideration. CRG has found that morbidity and health status are directly correlated with levels of poverty. Her colleague Ms. Ruben added that less variation occurs in referrals for sensitive surgeries, and the greatest variation is in ambulatory care.
Responding to a question about whether race/ethnicity data are collected as part of MEDSTAT's analysis and whether they look at issues of access and other measures of quality of care, Mr. Burwell answered yes, in some states, but states have different policies regarding collection of race and ethnicity data during the eligibility process.
Dr. Fowler, asked about response rates, estimated that Massachusetts would be in the mid-range, with about 40 percent response rate to mail alone. That ratio would increase by 15 or 20 percentage points for telephone surveys. They hope to get the response rate up to the 70 percent level, above the minimum standard for response rates of 60 percent. Regarding the question of relative costs of different modes, he estimated that telephone interviews cost twice as much as mail responses. Answering the question as to what can be done to reverse the trend of decreasing response rates, Dr. Fowler said their experience with the CAHPS surveys was that telephone refusals are rare, while mail is only a fair way to enlist cooperation. Medicaid populations have been more resistant than privately insured respondents. He thought that surveys of plan members were feasible, particularly for sampling people with specific conditions as they are likely to cooperate if something relevant to their well-being is being done.
To a follow-up question on whether their analyses of race and ethnicity factors had looked at issues of equity and access across populations, Mr. Burwell could not provide precise information but offered to follow up with examples after the meeting.
Ms. Ruben added that all five states in which her company works have done analyses based on race, ethnicity, and income qualifications. Ms. Lutz pointed out that in a fully capitated managed care situation like Tennessee, each of the 12 MCOs also are segmented by the dominant ethnic category as well as age category, which has been helpful in providing answers to advocacy groups. Managing data with these population-based methods today offers a significant public health opportunity to Medicaid programs that have moved away from being simply fiscal programs to become health care delivery programs.
A question from George Van Amburg regarding the use of ambulatory care sensitive conditions to analyze the performance of managed care programs. Ms. Lutz indicated that they are incorporating this into their analysis for a number of states.
A question from Dr. Friedman generated discussion of possible linkages of data sets.
Dr. Iezzoni asked the panelists to comment on how the CAHPS instruments can get information from people with cognitive and other impairments that make it difficult for them to respond effectively to mail and phone surveys. Dr. Fowler explained that when people can't speak for themselves or write, there is no special strategy beyond what is used for elderly people, the proxy approach. He agreed that the problem should be addressed more systematically.
Dr. Iezzoni asked what they do to supplement the information they get from ICD-9-CM codes, which are notoriously bad at identifying disabilities. Dr. Fowler confirmed that some people who may have disabilities are not so labeled in administrative files because their eligibility depends on some other factor. His organization is going to estimate the rate at which a cross- section of AFDC-eligible people have disabilities. Ms. Ruben added that the Social Security file contains information on what qualified a person for disability, and at least two states have facilitated access to that file. They are making some progress in identifying disabilities.
Mr. Burwell mentioned their recently completed study of children with severe disabilities in which they linked SSI disability data with Medicaid claims data. He expressed skepticism that linkages can be made on an ongoing basis because of confidentiality concerns at the Social Security Administration. He noted that we are entering a phase in which state Medicaid programs would very much like to have that information.
Dr. Caper noted that data confidentiality issues are a debate that is "going to go on forever." He asserted that isolating these issues in civil rights terms is a mistake in view of other considerations.
Ms. Fain's particular interest is health insurance specific to managed care and long-term care. She provided a brief overview of the role of the Attorney General's office in Massachusetts, and its limited role in Medicaid Managed Care in terms of data collection. One facet of the Attorney General's broad responsibility is to protect the interests of consumers. The Regulated Industries Division represents rate payers, and this role is somewhat unique. For the past seven years health insurance coverage has been a very high priority for the Attorney General, a key player in securing passage of several relevant pieces of state legislation. Her division's role is to scrutinize not-for-profit institutions and to establish community benefit guidelines for both hospitals and HMOs.
She enumerated some of the Division's activities, including the launching of a bipartisan task force for long-term care financing to help elders avoid impoverishment if they need long-term care, and to relieve the strain on the state Medicaid budget. They also perform ongoing ombudsman functions to mediate formal complaints from consumers about health insurance. Although they have only received about ten complaints from Medicaid beneficiaries over the past two years (indicating that many may be going elsewhere), they have provided perspectives on the issues consumers face. One of the categories of complaints is called "rationing." Others are discontinuity of care and lack of choice or control over who provides medical services. One major concern stems from the changes, perceived or real, in health care. Many consumers are confused and don't know how to navigate the system. Ms. Fain noted that their complaints are probably similar to those of others who have moved from fee-for-service to managed care plans.
In response to complaints, a study of these issues was launched two years ago, resulting last year in the report "Attorney General's Report: Managed Care in Massachusetts." It made recommendations for state regulation of managed care organizations. The Attorney General's office favors regulatory oversight to make sure that standards are followed. The idea is to create a level playing field that will preserve what's good about managed care. Ms. Fain explained that any state regulation would apply only to people not in ERISA plans.
Ms. Mariner noted that considerable research on data collection methodology has been done in Massachusetts. She observed that methodology problems can sometimes overshadow the purpose of the collected data. She reviewed several uses of the data, noting the differences between the information needs of consumers (people in the capacity of making purchasing decisions) and of patients. Noting the increasing emphasis on consumer information, she hoped that patients would not be overlooked, since everyone is a patient at one time or another. The relationship between patient and provider and patient and health plan is fundamentally different from the relationship of a consumer choosing a health plan. For that reason, patient rights are a major concern.
In data collection, it is important to look also at what did not happen, a type of information that is not easily collected. Acknowledging problems with "standardized exams," she encouraged data collectors to "act on what you already know and get the information you need." She also encouraged data collectors to look at such sources as records of grievance mechanisms and patient surveys conducted in connection with physician incentive plans.
Ms. Wells is national coordinator of Family Voices, founded five years ago to advocate for adequate health care for all children in the country. The organization realized that the great changes in health care make it vital for consumers of health care services to know how that care is affecting people, and especially children with special health care needs. Mental health issues are among the family problems of major concern. Their organization, with offices in four states and volunteer coordinators in every state, is active in state policy. Nationally, they have commented frequently on HEDIS and have worked closely with the CAHPS project.
She asserted that consumers need a larger role in data collection and in determining how to report information on managed care for both Medicaid and privately insured children. Family Voices conducted a small consumer survey of families in four locations, which found that families are generally satisfied with routine primary care. However, they registered concerns about home nursing care, hospital services, and mental health services. Almost all the families who reported mental health care needs felt they were not adequately provided for, although many of those needs would not be difficult to meet.
Families not in managed care were almost twice as likely to be highly satisfied with services for children with special needs. A big issue for families with children with special needs is the kind of restrictions placed on providers by certain managed care plans. Another issue was the need for information. The survey indicated that satisfaction with the plan was a function of choice: where choice of provider was limited, satisfaction decreased. Surveys of families can be particularly helpful to the health care system because children with special needs are markers for all children in the system.
Family Voices has been funded to do a broader national study of families with children with special needs, in conjunction with Brandeis University, using a survey tool based on the original one. Some important information comes from outside any system, because these children are served by many systems. Many of the tools now collecting data are not good indicators of quality health care for children. Access to first line medications for those with special needs is very important, as is home care. She also noted the lack of a national standard of medical necessity. Another issue is the role of families in decision making, and the need for families to understand the mechanisms within managed care for making medical decisions.
Ms. Wells recommended improvements in data analysis, particularly regarding children on SSI and fee-for-service plans; monitoring services over time (e.g., physical therapy in Medicaid programs); and expanding the role of families in data collection policy and reporting data that families can use. She urged those in federal health agencies to think about including consumers in designing tools and programs, as they can be sophisticated data users.
Ms. Ward noted the importance of ensuring that data reported on surveys is not identifiable so that it cannot be used to the disadvantage of families. She suggested that the reluctance of parents to report the health status of children is a specific research issue for health plans in providing services to their enrollees.
Ms. Wells was asked whether Family Voices considers the CAHPS instrument appropriate and adequate. Her answer was "no, not quite": More questions about children should be included, but it is going in the right direction.
In response to a question, Ms. Fain said that the Attorney General's office has not tried to describe the kind of data needed to evaluate performance with respect to quality and access. However, they have made recommendations regarding what needs to be collected before informed policy decisions can be made, e.g., about the effects of incentives. They have also made recommendations about quality data of the HEDIS kind, as well as the need to focus NCQA and other organizations on collecting data to better serve children and other underserved and at-risk populations. The Attorney General's office does not get into specific data elements. However, they do articulate the kind of data they need to carry out the mission of guarding quality and equity across populations.
Asked if their survey found any differences in satisfaction with services across different racial or ethnic backgrounds or other economic indicators, Ms. Wells said that they didn't see any differences, but that the small number of interviews conducted may have been a factor. They indicated that there may be differences in a larger national sample.
Regarding the risk of over-surveying populations with children with special health care needs and its impact on response rates, Ms. Wells explained how they work collaboratively to avoid overlaps with other surveys, such as New England Search. She emphasized their role as guides in getting people to coordinate with each other. She also stressed the importance of refining questions, asking only those that will elicit the most useful data, and informing families about how the data will be used. For their survey they identify families with children with special needs in two ways: through the membership list of Family Voices, from which they randomly selected a sample; and through the special needs health care department, which sends out the survey and maintains confidentiality.
Health Care for All, a consumer advocacy agency, provides information about access to health care and helps people get the care they need. Mr. Restuccia expressed his belief that managed care can result in higher quality care and improved access for Medicaid beneficiaries, given careful planning and strong state oversight. He also believes that quality assurance efforts should focus on high-risk populations. He noted that the Medicaid program in Massachusetts has expanded dramatically and now includes the homeless population and the dual eligible population.
His organization's advocacy efforts extend to legislative and policy issues. Although data are necessary to protect consumers and to assist them in choosing the best plan, their view is that data alone to measure quality of care will not make the market work more effectively. Comparative indicators that come out of HEDIS, for example, are not geared to the individual, but to the benefit manager. So, data must be viewed as a component in an overall quality control system.
The speaker said Massachusetts needs to develop a stronger program to help explain patterns suggested by the data. Readily available data tend to lump together seriously and chronically ill people with those with minor illnesses, making it hard to tease out important data on those who are the sickest. In certain key areas (home care, dental care, mental health) data are scarce. The homeless population presents special challenges, but even within that population it may be possible to look at preventible hospitalizations and evaluate whether people have appropriate access to primary care.
The Consortium is a private sector not-for-profit regional data repository. Mr. Stone's remarks focused on the process of data collection and the goal of creating standardized data repositories to assist research and analysis, with data standards for insurers, purchasers, and health plans and providers. Multiple activities are going on in all the states related to Medicaid research, and as Medicaid recipients are increasingly enrolled in managed care plans, data collection efforts will be eased only if collected and reported in a standardized format. He noted the relevance of HIPAA to Medicaid managed care data sets. These requirements must be adopted across the sectors, taking care not to isolate Medicaid managed care from other sectors.
Noting that the National Committee has been concerned with standardization for a long while, he urged prompt action by the Committee and the National Library of Medicine to develop categories of data elements, a "meta-dictionary" to be used across the data dictionaries held in public and private agencies. A set of core health data elements on persons and encounters or events that can serve multiple purposes and would benefit from standardization is possible, provided the approach acknowledges three concepts: (1) core data elements will have many and unanticipated uses; (2) they will have definitions that will be changed continuously; and (3) it will take a long time for them to be used in general practice.
He mentioned the local planning around seven categories identified by IOM and essential to DHHS's responsibilities: (1) demographic data (2) administrative data (3) health risks (4) health status (5) patient medical history (6) current management of health conditions and (7) outcomes data. He likened the standardized categories to a chart of accounts for data. There are gaps in many areas they would like to look at, not just in centralized data warehouses but in all of the holders, including hospitals and health plans.
The approach they've been using since 1994, the Affiliated Health Networks Projects, focuses on commonality of the elements, terms, and definition. They're trying to facilitate connectivity between and among competing organizations. Since most people will have on average seven health plans over a lifetime, they think the future is going to be more about networks of networks than centralized data repositories, with data aggregated on a need-to-know basis.
They have active work groups trying to implement locally what the National Committee recommends on standardization, and every year they have a public meeting to discuss how they are implementing locally the Committee's standards. At that meeting they continually raise awareness of what the National Committee and HIPAA are about.
Mr. Stone made the following recommendations: (1) Holders of national databases should take a leadership role in data set standardization, e.g., list themselves in a national compendium (meta-dictionary) on the basis of elements needed (acute care, chronic care, outcomes research, etc.). He cited the CDC's Data Elements for Emergency Room Department Systems (DEEDS) as an example. With this kind of compendium, each state would have accessible, on-line data elements on emergency room data, for example. (2) As the compendium evolves, the National Committee could reach consensus on minimum data sets, definitions, and guidelines. (3) The Committee should recommend to ANSI and CDC that a public health work be established to look at data set standardization. (4) The National Committee should avoid the chaos of trying to choose one winning definition, but instead let the market prevail to some extent and let everyone see what's being done. In summary, he acknowledged that Medicaid managed care data sets "are going to be messy for a long, long time," but small incremental steps forward toward an on-line meta-dictionary would help the health care industry evolve toward consensus on core data sets.
Before engaging this panel, Dr. Iezzoni related a conversation she had with Mr. Burwell about HIPAA after the earlier session. She expressed surprise that when she asked him whether MEDSTAT's reporting requirements were consistent with HIPAA, his answer was "no." She said she reminded him that the HIPAA standards will apply to every health care transaction of the specified types regardless of the payer, and it should be under consideration. She observed to the group that this dialogue is instructive to the Committee about how the HIPAA issues still have not reached the awareness of all those who need to know about it.
Mr. Stone observed that Medicaid, in Massachusetts and elsewhere, "has not totally embraced HIPAA" yet. He recommended that RFPs include a sentence from Medicaid requiring HIPAA compliance in proposals, and that every health plan and provider should be encouraged to expect compliance with HIPAA. He speculated that there are concerns that it could be perceived as an unfunded mandate and that if Medicaid asks for compliance, there may be additional costs to Medicaid.
Ms. Greenberg mentioned that the HIPAA standards have had little input from public health constituencies, and they should be talking to one another about how they can contribute to standards development in the context of HIPAA. She noted plans for a workshop in November on this topic. She agreed with Mr. Stone's concept of a meta-dictionary and mentioned that Bob Mays at HCFA is very much involved in trying to sell the concept to the Department. She asked how the meta-dictionary would help to get people to implement the HIPAA standards. Mr. Stone noted that it might not be as "scary" as the manuals for HIPAA implementation.
Dr. Amaro asked for further discussion concerning access and quality of care and how it is affecting those in managed care. She described the frustration of one woman who reported that after waiting three months to get an appointment at her health plan, she was kept several more hours past her appointment time and finally told to take a Tylenol.
Mr. Restuccia said that although this is not the focus of his organization, the lack of interpreter services continually arises for the Medicaid population. Dental care problems are also enormous, as are issues related to home care. He noted that data are probably more useful to the agencies enforcing the contracts than to the consumer making the choice. He speculated that it might be worth giving people less choice to raise the bar and eliminate HMOs that aren't providing needed services.
On the language issue, Mr. Stone said his Consortium has asked all the health plans they work with what data elements they have about physicians, and only two list the language(s) spoken by the physician. So, they intend to do a study this summer in which they ask patients to look at a long list of elements and choose the ones they consider most important. Their Consortium wants to give prominence to what questions Medicaid beneficiaries ask when they choose a doctor from a list of data elements.
Dr. Iezzoni observed that some hospitals are creating websites where much of this kind of information is available, but she questioned how many pregnant women are logging on to the Internet to choose their doctors. Mr. Restuccia replied that people choose their physician or their health plan mainly through networks of friends. It is important to provide information to people about providers' cultural competence, an attribute for which standards are needed. He proposed the idea of new organizational forms that relate to different populations. Investing in that kind of network could make a difference in how people choose their Medicaid provider.
On the idea of looking beyond data systems for data on sub-populations, Mr. Restuccia observed that the diversity occurring in the expanding Medicaid population requires new ways of characterizing populations. He noted that the dually eligible population is going to present tremendous issues, since they are not well served by either Medicaid or Medicare now.
Returning to his vision of the compendium of data sets, Mr. Stone was asked whether registering a data set includes definitions of the collective microdata elements or just the distributed and aggregate data elements. He replied that they need definitions of every little component. Because this causes confusion in respect to standards, his organization does not recommend one. Rather, they tell people what they can collect, what most other people are doing, and what his organization wants to know. He noted that broad categories would allow people to share ways they find effective to get needed data; then "we could slowly inform the whole marketplace how some of the best and brightest are collecting data" -- on ethnicity or wait times, for example.
Asked how they decide the relative utility of various data elements, Mr. Stone said they consult with people who do the end-use analysis, such as the outcome researchers and resource planners.
Mr. Restuccia observed that Medicaid managed care has provided the user an opportunity to change the system for some of the underserved subgroups. The importance of looking at populations that have historically been underserved is becoming apparent, in order to figure out how to collect and use data in new ways. There are ways to be creative and to see things from the consumer standpoint to make the health care system more responsive.
Dr. Iezzoni explained to Mr. Stone that when NCVHS developed its core data elements recommendations, the first thing they did was a market survey regarding the ways different organizations were collecting data. However, many proprietary organizations refused to share this information.
Ms. Freeman is Professor and Chair at the University of Massachusetts and a senior fellow at the McCormick Institute of Public Affairs. She praised the Committee's website and said she finds the frequent postings very useful in her work.
She observed that there are many reasons to collect Medicaid data, not only for directing medical services but also for directing community, school, and work-site interventions -- reasons not often mentioned in conversations about health data. She commented that society expects much more from medical care than it can deliver, and much less from public health than it can accomplish. On this basis, she expressed concern that the Massachusetts Medicaid program is not operating in its full capacity as a public health organization. Some states have done better at combining the capacities of its public health organizations to improve the health of the whole population.
She echoed a recurring theme--the special opportunity Medicaid managed care offers as a set of public programs without proprietary restrictions and with the ability to work with private providers in new ways. She stressed the importance of Medicaid's using public leverage to elicit the data that help people understand the health status of the population, and to improve it. Although there is a lot of fear about the impact of the shift to managed care, the important thing is what services are provided, not who provides them. In this regard she noted that it is not helpful just to document unfavorable outcomes; the reasons for them must be documented.
She cited a report done by the George Washington Center for Health Policy, "Negotiating the New Health System," from which she selected four examples of how collecting and reporting requirements can yield important and useful information that could improve the health of the Medicaid population. In Iowa, a collaboration between the managed care organizations under contract with the Medicaid agency and the WIC program evolved into an accountability system. The examples included an immunization registry in Oklahoma, a pregnancy risk assessment program in South Carolina, and quality indicators on four conditions in Missouri. All are described in the written testimony submitted by the speaker.
Ms. Freeman applauded the Committee for its role in the privacy debate, and expressed fears that "we're going to have a terrible privacy law." She cited a paper she co-authored, "The Health Privacy Debate: Can We Achieve Comprehension Before Closure?" In her view, the public discussions of the privacy of personal health data have shown a lack of clarity about the purposes for which data will be used. She expressed concern about undue restrictions on the use of data that could improve the health of Medicaid and other populations. The justifiable fears about the abuse of information for commercial gain tend to eclipse any awareness of the socially important uses of data for public health, health research and licensure.
The aforementioned paper, which she wrote with Tony Robbins, suggests ways to carry on the debate so that more people can be engaged in the understanding the issues and giving voice to their views. She acknowledged that it is rare for any enterprise to describe its use of data without claiming the most altruistic motives. However, the arguments can be subjected to a simple analysis that makes it fairly easy to sort out the primary beneficiaries of any data transaction. She lamented that the recent study from the National Research Council "carefully skirted the issue of economic advantage of the use of data," adding that she did not "think that's a good way to run a democracy." Her paper suggests how people who may not be familiar with or comfortable in the world of policy formulation can understand the trade-offs. It also outlines a way in which to talk more directly, "but still civilly." She noted that people react very differently to the prospect of data use depending on what is being traded off.
Ms. Freeman asserted that this democracy should not treat the privacy of Medicaid patients differently than that of others, which means that society must take on this debate across all sectors of society and work it out in the broadest possible framework. This, in turn, means "using the public leverage with managed care organizations to get the data that allow us to improve everybody's health."
Dr. Ackerman presented personal background to document her "bias" in favor of managed care. She described the challenges of discovering, after joining the managed Medicaid program at Boston Medical Center in January of this year, that much of the data they are asked to collect and report do not easily or usefully come together. As a medical director, she needs data that will allow her to manage resources and plan for the longer term, and that can teach and motivate. The data also should be clear, usable, and economical to produce, as well as "real-time" and related to goals.
She observed that in her present position, she has less information available about Medicaid than when she was dealing mainly with Medicare and commercial populations. There also are fewer official reports, trade group newsletters, brochures, advocacy groups, and thematic journals. The information may be available, but it is not coming across her desk. As a result, the data acquisition process is more difficult. She noted also the excess of process measures and "proxies for outcomes," including HEDIS measures.
Another problem is the lack of stability of denominators and numerators for the Medicaid population, with patients moving in and out of the plan who are difficult to track. Dr. Ackerman said she wished there were a way to track data without intruding on confidentiality and privacy. Lack of standard formats is another problem.
Her recommendations were for a clear articulation of goals for data collection and measurement that are categorized and prioritized, since not all data are equal in importance or quality. Guidance is needed in order to know the validity, reliability, and usefulness of the data. There is need for user-friendly formats, identifications of best practices, and flexibility. Multi- disciplinary teams are extremely important in terms of intensity of resource use. Dr. Ackerman noted the repeated occurrence of substance abuse problems and the need for better information in order to direct patients to the optimal providers.
Dr. Turner's interests are largely in the mental health area, especially in regard to access issues for racial and ethnic minority populations. He discussed his role in the McCormick Institute's contract with the state Medicaid program in the Department of Medical Assistance. They have been looking at the impact of race, ethnicity, and language on access. Their report to DMA summarizes recommendations concerning these issues as they effect minority populations in the state. They have also arranged a series of colloquia for the DMA.
A major issue addressed is the barriers that prevent minority populations from getting health care, many related to language. Dr. Turner pointed out that black Americans are not a homogenous population; including French-speaking Haitians, about 35 percent of the black population is foreign-born. Understanding this fact has enabled them to better address the issue of health in major ethnic minority populations.
One of the issues they studied was the discrepancy in life expectancy and whether these differences can be correlated to lack of access to medical services. The tentative conclusion was that shorter life expectancy among blacks in Massachusetts has multiple causes, a finding that mirrors the pattern in the nation. The infant mortality rate among blacks is much higher than for white populations, and may be attributed to such factors as younger mothers, less prenatal care, and less medical follow-up, as examples. But Dr. Turner said the statistics don't completely describe what is taking place.
In terms of gathering additional information, they found the data quite poor. There is need for sound data to reflect the heterogeneity of the black population, for instance.
In answer to a question, Dr. Ackerman said they are not currently using race/ethnicity data. In general, they are very limited in the descriptors of the population because there is no way to incorporate such information in claims data. High on her wish list is more descriptive data about the populations they serve. As a new health plan, they are having to build their own data bases with only limited information coming from the state about the enrollees. She said they cannot afford to get this information themselves, but they might invest in it collaboratively with others. They also need more descriptive data on dual diagnoses and stages of diseases, as well as more about individual patients and what they want or don't want.
Regarding her earlier comment on the prominence of substance abuse as a clinical issue, Dr. Ackerman was asked what kind of information would be most useful related to addictions and substance abuse. She said they would like to know the basics: whether the patients use drugs, and their use patterns. However, she doubts that they will be able to get that kind of information. She noted that behavioral health is general requires that confidentiality be respected and preserved.
Asked to comment on whether there should be a different standard for protection of privacy or confidentiality relating to services for mental health and substance abuse, both Ms. Freeman and Dr. Turner recommended a single standard for all medical conditions.
Dr. Ackerman was asked what cautionary details from her own experience might be pertinent to evaluating Medicaid surveys, e.g., in regard to language. She replied that she would be very concerned about a survey conducted in English only, yet they have 140 languages listed in their interpretative services directory. Concerned about low rates of response to surveys, Dr. Turner reasoned that the more responses the better, in order to get as close to the population universe as possible, but mail surveys are not going to get adequate information. As a strategy for improving information from Medicaid populations, Dr. Turner replied that money would help; for example, adequate resources would make it possible to go house-to-house to survey people, but that is not plausible with limited resources.
Ms. Coltin noted that the CAHP member-based survey contains very few questions pertinent to people who have not used services, so there is a great potential for non-response bias relative to people who have not used services. She mentioned there are other ways one can use administrative data sets and enrollment files to identify the population that hasn't used services and, perhaps, to look at other types of issues in that population. She noted that there will be issues with any alternative.
Dr. Robbins cited the latest issue of Public Health Reports which included his comments on the new report on health inequalities. Speaking as a citizen (not as representative of the Regional Office), he said, "We are participating in a very racist exercise if we use race without understanding why we are doing it and asking reasonable questions." He continued, "Maybe if the U.S. had been more social class conscious and a little less race conscious, we might have described very different phenomena in this country than what it looks like when we read most epidemiological studies. . . . The best single proxy in the U.S. for social class is the value of the real estate in which people live." He noted that this proxy picks up both education and income, and cultural aspects as well.
Dr. Newacheck asked the panel for particular thoughts about specific sub-populations that should be included in surveys of the Medicaid population that may not have been adequately sampled. Ms. Freeman mentioned women who have been subjected to abuse.
Dr. Iezzoni referred to the favorable response to Dr. Amaro's report to the OMB directive, #15: “Changes in the Race and Ethnic Classification." In response to a question, Dr. Amaro said that a limited number of people who write in public health and medical research continue to use race and ethnicity without providing a framework or explaining what the terms serve as a proxy for; however, the field is making some progress in articulating the many things in a more explicit way that race and ethnicity have stood in for. So "we're beginning to push the traditional edges, but as a whole, the field has a long way to go."
Dr. Turner pointed to the mistaken impression that there is a one-to-one surrogate relationship between social class and race, and that if we understand the former, the latter will disappear as a factor. In fact, we need to know what part of the variance is being accounted for by something that has nothing to do with how much money one has.
After thanking the panelists for the thought-provoking discussion, Dr. Iezzoni adjourned the hearing for the day.
Dr. Iezzoni opening the meeting with brief self-introductions.
Dr. Bachman spoke from her experience in Medicaid managed care. Points of emphasis included data requirements, issues in collecting encounter data, data analysis, and alliance. Viewing Medicaid managed care as a privatization issue, she said that state policy makers want to use Medicaid because this type of privatization off-loads both financial risk and responsibility to a private organization, so not only are they able to mainstream Medicaid recipients but they also are able to off-load some of their "thorny policy-making." Viewing data collection within this context, the basic questions are: how much do Medicaid policy makers really want to know, and how responsible should they be in staying on top of what is happening in health care transactions?
Traditionally, Medicaid has kept information about health care transactions only in the record of monthly payments, Dr. Bachman said. She then touched upon some of the types of data now available to Medicaid policy makers:
§ Traditional claims data linked to eligibility systems. The MMIS hasn't always worked flawlessly and she acknowledged that animosity around this weakness has built up over time.
§ Encounter data. Although the data are probably out there, getting them in a form that can be used creates problems.
§ HEDIS data. There are also problems in using those data, including issues of quality and utility.
§ Medical records. Medicaid programs conduct medical record audits that provide valuable information, but in addition to the expense to collect these data, there is the question of what to look for in the records.
§ Consumer surveys. This is a promising source for useful information about Medicaid managed care programs.
§ Miscellaneous sources. A general catch-all for information about transportation, health, services provided by other state agencies, etc., all of which could add to the understanding of what is happening in Medicaid managed care plans.
Once all these types of data are gathered, the hope is that the state agency would be able to use them--which is not as easy as one might think. Medicaid agencies need staff sophisticated enough to analyze the data, and they also need a clear understanding about the purpose of data analysis.
On the topic of alliances, Dr. Bachman observed that Medicaid financing has become a cornerstone of state health reform efforts, so now the linkages between Medicaid agencies and other state agencies have become more powerful than they have ever been. This raises the question of who has access to these data and how Medicaid data are used by other state agencies to manage their programs. Data needs are not determined only by what Medicaid has to know but by the information needs of other state agencies and legislators.
Questions posed by Dr. Bachman: What are the limits of privatization? What is the government's responsibility to safeguard information even as it is off-loading risk and responsibility for service delivery? Where is the state going in its overall health reform efforts? How will Medicaid data be viewed within that larger context? A key question is where safety net providers are to get the resources and support they need to provide data that are going to be mandated or requested by state agencies. In order to gather appropriate data, support is needed to create, sometimes for the first time, automated data systems.
Dr. Bachman urged the Subcommittee to give more thought to standardizing data requirements, and to consider what the mandates are going to be.
Ms. Thompson said she is trained as a lawyer and by predilection an advocate, describing her approach as "consumer-driven." She cited an experiment now underway, an 1115 waiver proposal that would enroll dually eligible people in managed care plans funded by both Medicare and Medicaid. While such arrangements are being put in place, an appropriate approach to the dually eligible population is to focus on managed care quality and data collection issues.
Ms. Thompson identified long-term care as an area that needs work, using the term to include care in homes and institutional settings. Little exists in the way of quality measures for such care, since even HEDIS measures for Medicaid managed care neglect chronic illness and long-term care in both home and institutional settings. Therefore, she recommended that the National Committee encourage the government to look for effective ways to measure long-term care, including quality measures for disabled populations.
She also discussed the difficulty of measuring accurately the quality of care for dually eligible populations across the spectrum of services, primarily because Medicare and Medicaid are separate systems. Medicaid is combining data from HCFA to try to project accurate patterns of usage for this population; and they are using these data in determining capitation rates for their waiver proposal. This may be the first time this has been done on a large scale, and she hopes it may be a model. She further recommended that thought be given to models for integrating managed care with Medicaid for dually eligible people, to allow active assessment of data by both the state and federal governments. She also recommended that the Subcommittee encourage the involvement of external review entities in assessing quality of care for the dually eligible population.
Another aspect of care for the dually eligible population that does not get adequate allocation is mental health. She recommended development of measures to ensure that mental health needs are properly screened for and treated when identified. It is especially important to integrate mental health services for seniors and disabled people, she emphasized.
Ms. Thompson mentioned several quality measures and sources helpful in gathering data for assessing quality:
§ Consumers, family members, and caregivers. In addition to involving them in assessing the quality of care, they can provide information about barriers to accessing care.
§ The disenrolled. Objective measures of consumer satisfaction can be gained from disenrolled people, as well as from the records of grievances and appeals. In regard to grievances, Dr. Bachman recommended that plans include an ombudsman, perhaps one for internal and one for external complaints.
§ Staffing. This is another quality measure for institutional settings, and though salary is not the decisive factor, the connection between what providers are paid and the quality of care they provide is a fairly reliable indicator.
§ On-going care plans. Choice of provider is an important quality measure. She suggested looking at mechanisms in place that screen for and allow planning for on-going care.
§ Facility report cards. She said that mechanisms are needed for reporting quality assessment and data collection measures to the public, e.g., nursing home report cards.
Ms. Thompson saw a major challenge for the future in the integration of Medicare-plus- choice plans with Medicaid managed care.
In her role as director of the center, she provided examples specific to their health center, but said they are representative of health centers in general. She described Medicaid managed care as an opportunity to improve the health status of vunerable populations, if implemented properly. She described the performance improvements at her health center in the areas of hospital admissions and emergency room visits. She stressed the importance of obtaining data in a timely manner, described the level of access required for this information, discussed possible redistribution of resources (to community health centers), and discussed the importance of identifying special populations (e.g., racial, ethnic, and language). In addition, she stressed the need to include access to the model, rather than continuing to rely on the insurance or payment component.
The Neighborhood Health Plan is an HMO that serves a membership of 50,000 predominantly Medicaid enrollees. Ms. Federman's comments reflected views of both their plan and her experience in public policy making. She observed that for HMOs to perform well, Medicaid managed care needs adequate funding. Data collection is important in determining whether the payment methodology actually works. Since different methods are used in different states and many more people are moving into managed care, data collection and analysis methodology becomes less clear. Both the state and the HMOs need help in sorting out the impact of managed care on new populations coming into Medicaid.
Ms. Federman asserted that a lot of work needs to be done on the federal level to standardize the indicators that would make it possible for state agencies and the HMOs to work out these problems. For example, their plan has difficulty in learning from the experience of other states, and all of the safety net HMOs and providers are having trouble assessing the uninsured population when Medicaid people are moved into managed care.
A major issue is that health plans, although serving large urban populations, get almost no information from Medicaid to identify the disabled, and Medicaid also has a difficult time getting that information. With more and better data, the plans could be much more protective of those with disabilities.
Sanctions is another issue that deserves careful thought. Many providers lack resources and access to additional funding, which makes it difficult for plans to impose sanctions on them. Finally, she noted the inability of HMOs serving low-income populations to generate resources and she reiterated the importance of finding ways to standardize indicators in order to make reasonable comparisons.
In response to a question from Dr. Newacheck, Ms. Feldman said that about 32 percent of their enrollment is covered by Medicaid and 44 percent of their clients are uninsured. Their data collection and analysis are a financial burden that she thinks should be carried by state and federal governments.
Asked if other health centers in the area are able to collect and use data adequately, she observed that the physicians and other clinicians who choose to work at HMOs serving low- income populations are committed to providing medical services to the patients. For this they need not only individual but aggregate data. Asked specifically whether their plan is putting mechanisms into place for using HIPAA standards for data transmission, she replied that they are just starting to do so, and that their local community hospital has assisted them in making sure that their vendors have an understanding of the requirements.
Asked to comment on issues of behavioral health care and the need for appropriate rates of reimbursement and cost calculations based on the population they're serving, she briefly described the benefit package. Ms. Feldman said that because they use an integrated model they have not had a problem with the methodology Medicaid uses. Information has been gathered on Medicaid and analyzed in terms of cost effectiveness of medications and case management for the chronically mentally ill. One of the benefits of Medicaid managed care is that they are able to track chronically mentally ill as a group as well as individually. However, in the area of substance abuse, much more work needs to be done. They do not have adequate access to services, which has a lot to do with the level of financing but also with contracting with competent providers, because there are simply not enough of them.
Ms. Bachman has done some work on carve-out arrangements and said that in general, state policy makers think that eventually all services will be integrated under one capitated arrangement for most populations. She described a variety of arrangements in Massachusetts, which has a statewide carve-out for behavioral health services. There could be basic outcome measures that would give a sense of what is happening to individuals receiving behavioral health services, and some work has been done toward identifying those measures.
Ms. Feldman said their integrated system works well because it is possible for clinicians to easily and quickly talk to each other. They also do joint case management. Data collection is relatively simple, although it would be helpful to have standard measures. One of her concerns is the large population that needs a lot of support--those with AIDS and substance abuse problems, for example. It would be helpful to have some way of monitoring utilization of services by that population group. She also pointed to the need for better measures for persons who are dually diagnosed.
Ms. Coltin commented that Harvard/Pilgrim's integrated system affords the advantage of having complete data across populations for both mental health and non-mental health services, but many plans do not have that linkage capability. Having heard from others who have also testified about privacy and confidentiality related to mental health and substance abuse data, she noted that when the health plan does not have a relationship with the carve-out vendor, there is no obligation for that vendor to share information, particularly of a sensitive and confidential nature, and then the vital but informal communication among the caregivers is also compromised.
Pursuing further the privacy and confidentiality issue for people with mental illness or substance abuse, Dr. Iezzoni asked the panel if this issue has become more acute or whether people are now talking more about it. Recalling yesterday's testimony that family members are increasingly concerned about responding honestly to requests for data of this kind out of fear that they will be dropped from the plan, she asked if this is something the panel is hearing more about.
Ms. Thompson said that in their experience most of their people are concerned with cutbacks in Medicare home health services, which is now a real problem. Also, caregivers who are not sophisticated about eligibility criteria may wonder what they should be saying to surveys. Their answer may in fact result in a denial of services, when that is not what they intended. She reported that they have heard that family members are being asked questions which they do not know how to answer, which is more of a concern than being afraid to answer.
Asked for suggestions, Ms. Thompson recommended assistance in finding and using screening tools that would provide a standardized process for collecting accurate data about the ability of persons to function, noting the element of pride that tends to interfere with accurate reporting on the part of patients. In addition to having an accurate assessment tool, it is important to have an assessor trained to work with these diverse populations who understands where they are coming from, including the understanding of cultural issues that affect how particular disabilities are viewed.
Ms. Bachman mentioned examples of managed care plans that have instituted assessment systems: the Intermountain Health Plan in Utah, for one, has an enrollment and assessment system linked to functional status. These plans are able to use the information to develop care plans that reflect the individual's need for services.
Asked to talk about the locus of responsibility for data collection and analysis that depend on decisions made on benefit packages, Ms. Bachman said that some of the responsibility must reside at the state level, especially for making it clear what data they need and how the data will be used for monitoring health plans. In addition, more dialogue is needed between the plans and the state about benefit exceptions so that creative and flexible care plans could be developed.
Ms. Thompson mentioned that the dually eligible waiver that Massachusetts is now developing provides discretion for plans to provide additional services.
Ms. Tobias spoke briefly about the kinds of data they are looking at and using, focusing on three issues: benefits and service utilization, cost of health plans, and people with disabilities. They have worked with Medicaid claims data to develop risk adjusted payment systems for people 65 or older with disabilities. A major shortcoming in these data is lack of information on functional status, to know if people with life-long chronic needs are getting access to the right amount and type of benefits. When states begin contracting primarily with health plans to serve people with disabilities, the scope of benefits and utilization will be very different from anything the health plans have ever experienced. Monitoring this process will be essential, along with communicating information to consumers.
An example of good oversight is the state of Oregon's managed care for the elderly. Another is the experience of Massachusetts regarding mental health and substance abuse treatment utilization. Most states probably do not have the monitoring system and data collection tools necessary to identify utilization rates as quickly as Oregon and Massachusetts.
Regarding the issue of capitation rates and financing of care for populations with disabilities, Ms. Tobias described a pattern observed around the country in which behavioral health plans are pulling out or going under, claiming they are underfunded. She questioned whether the necessary data are being collected to understand why this is happening. There are probably many reasons, but the plan failures are early warning signals that should be heeded.
Another point about data collection for Medicaid managed care is whether there is any way to piggyback on what is being done currently for other Medicaid populations, to address people with disabilities. Ms. Tobias said when they asked plans what they are doing to monitor the quality of care for people with disabilities, they were generally told, "Well, we haven't got to that yet." One early step that could be taken: states could collect data about how many people are getting mammograms and immunizations.
Diane Flander’s office is involved with the PACE program. She summarized the types of studies that they have done regarding the effectiveness of PACE and its integrated approach to service delivery. She indicated that in the future they expect to use HEDIS, MDS, enrollment, and family surveys. In addition, they are planning to set up a complaint and appeals process focusing on selected measures of encounter and utilization data, e.g., instituional stays and pharmacy encounters.
In addition to his work at Harvard, Dr. Landon is associated with Beth Israel Deaconess Medical Center. In sharing some of the data they have collected, he noted the variability across states in the implementation of managed care programs, the ways states are implementing the programs, and what some states are planning. He emphasized the need for technical assistance in standardizing data collection and dissemination.
As chair of the State legislative Joint Committee on Health Care, Rep. Chandler described herself as a case study of "where the rubber meets the road." She explained that they do not collect data themselves, but depend on other sources. Unfortunately, they find that so little data is being collected that they often are forced to make policy based on anecdotal information. As an example of their "wandering in the dark," two years ago they developed a program to provide prescription assistance to seniors based on a faulty estimate of the target group, and the program consequently has been under-subscribed. So they have $30 million sitting, waiting for people they hope they can identify--clearly not the way to make policy, she said.
Another example she cited: they expanded Medicaid this year for the child care program based on guesswork that they hope is correct; but "guessing at numbers is not good planning." She stressed that as the federal government puts more demands and more responsibilities on the states, the need for data becomes greater.
For the first time in 22 years, Massachusetts is looking at managed care companies and trying to provide reforms that reflect consumer input in the process of managed care. If substantial numbers of the population are going to be put into managed care through the Medicaid program, they need a solid basis.
Although Ms. Chandler considers that the Medicaid program is administered with great skill in Massachusetts, she foresees a real crisis for that program. State legislators are very concerned that cuts in Medicare will have serious effects in Massachusetts and every state. She predicted that 1998 will be the year of crisis in Medicare for Massachusetts, which will in turn become a crisis in Medicaid.
Massachusetts concerns about home health services are now reaching monumental proportions, she warned. There are also major concerns about what this will mean in terms of residency programs. Because a great deal of funding is through the Medicaid program, they expect a trickle-down effect on the state "that will take one of the shining lights and perhaps diminish it." A special concern is that many of the population are seniors. She charged that "the Balanced Budget Act has been transacted on the backs of some of the most vulnerable members of our population." A survey of 40 home health agencies found that 60 percent expect to lay off staff within the year, and about 10 percent said they already have initiated layoffs. People who normally would have been able to stay in their homes are going to start looking for other sources, ultimately putting pressure on Medicaid funds. Although they think the state is doing a good job as it moves into Medicaid managed care, there is considerable uncertainty.
The legislative committee is taking steps to ensure that medical records have safeguards to preserve the confidentiality of patients. They are working to provide a bill to protect patients' privacy in Massachusetts, but they are especially worried about how it will play out in the mental health and substance abuse areas.
Ms. Chandler said they need to do a great deal more with outcomes because most of the current information is slanted toward patient satisfaction. They need to educate the patient population as to what they can and should expect in their medical care. Other serious issues: the impact of capitation, and whether the rates the state is paying are high enough. They want to keep seniors independent and at home as long as possible, but money is needed to make that happen.
In response to requests for additional comments related to specific mental health and substance abuse treatment, Ms. Chandler said they don't have enough State-based data or comparative data on other states, and a national standard would be helpful. She said she believes Medicaid managed care will have high standards and provide a critical mass that will draw attention to Medicaid patients.
Ms. Coltin asked whether the state has given thought to using data tapes with identifiers that could be used by researchers to help the state Medicaid agency expand their resources for analyzing data and informing policy. Noting the magnitude of the confidentiality issues involved, Ms. Chandler cited the report card being developed by large purchasers of health insurance and managed care using HEDIS data. She believes report cards for managed care should be available to everyone in the state, and she expects to see improvements in Medicaid across the board.
Ms. Ward observed that people are concerned the SSI enrollees and others are being moved into managed care before any quality indicators are in place. Encounter data may need to serve as a warning system because it is possible to look across plans and develop a critical mass. Ms. Chandler said before she could comment, she would need to know whether confidentiality issues would be a problem.
Mr. Van Amburg noted that states have not invested in health information systems to the same extent as private industry. He asked whether Massachusetts is willing to invest enough money in their information systems and data analysis to answer the questions that need to be answered on Medicaid managed care. Ms. Chandler said she thinks the State has the resolve to do so.
Dr. Iezzoni recalled testimony from the Neighborhood Health Plan and contrasted their information systems, which are "not up to speed," with those of Mr. Stone's Massachusetts Health Data Consortium. She noted that those "high on the food chain" are doing fine while those on the bottom "are being eaten" because they lack the needed capacity to remain in the market. Ms. Chandler agreed that this raises issues about the adequacy of rate payments for Medicaid to support not only the care given to patients but also documentation of what happens to them. She pointed out that Massachusetts is the birthplace of community health centers and it is critical to provide the same data in that arena as elsewhere. She added that the issue will continue to be raised since rates are becoming increasingly important.
Asked for further comment about the relative responsibilities of the state and federal governments in respect to data systems, Ms. Chandler reiterated that the states cannot meet the responsibilities being imposed on them within their current budgets; they will need federal help.
Returning to the confidentiality issue, Dr. Iezzoni asked how the legislature is going to balance this value with the equally important need for health status information to plan effective service delivery. She urged that the State work with knowledgeable people to find alternatives to the "knee-jerk response" that uses the privacy issue to avoid releasing any kind of information.
Ms. Chandler said that approach makes good sense. She described the work on a confidentiality bill, for which her committee brought together a working committee with broad representation that has been meeting every two weeks for ten months. The bill will enumerate conditions for which informed consent would be required. The legislators also are trying to balance the issues of confidentiality with the need for information, e.g., to protect Medicaid against fraudulent claims. She thinks that they can have "at least a modicum of confidence that we are just not giving it all away."
As a professor of health economics, Dr. Frank researches management health issues, including substance abuse care, as they relate to the managed care industry. His research has involved him in working with states to design payment systems for their Medicaid plans, as well as on procurement. He explained that three main goals guide the procurement process for states: (1) getting a good price; (2) paying fairly; and (3) being able to monitor and try to assure quality. These apply in general, but particularly in the area of mental health and substance abuse services.
He briefly reviewed steps in the procurement process: a data book goes out to prospective bidders, and bidders submit bids to provide services. The process involves information of various sorts, such as major classes of services and prices paid--what he called "the basic building blocks." However, this key information is not readily available.
As an example, he said he was asked by the Auditor General of Arizona to review the bidding and procurement processes for the Maricopa County Behavioral Health and Human Services. In drafting a model bid, Dr. Frank found that he had to make certain assumptions, not trivial in nature, that could "swing things around 10 or 20 percent." Such assumptions can mean the difference between winning and losing the bid, but more importantly, also between "winning or losing your shirt." Having mismatches in the data creates uncertainty for the bidder, which results in higher bids and causes bidders to withdraw. In Arizona, bidders frequently explained their withdrawal of bids in terms of their concerns about the uncertainty of the data. He experienced this first hand when helping the state staff in Massachusetts to devise a model bid based on the data book, in order to better judge incoming bids. They had to make many guesses and assumptions for this "mock" capitation bid because of all the holes in the data. Problems such as these make it difficult to compare bids, because both the bidding process and the awards process are subjective.
Another problematic area is monitoring. One of the fundamental policy issues in mental health and substance abuse is whether to carve in or carve out. This "raging debate" turns on whether people are more afraid of adverse selection, coordination of care, or potential opportunistic behavior when new boundaries are set up. This is a balancing act. One has to study carefully who enrolls, why, and in what plans; and this obviously requires good data.
Dr. Frank cited the case of the medical director of a large not-for-profit HMO in New England who went into the business to change the world but learned over time that "the market will not reward you for being the best mental health plan in town." Finding out the factors that drive selection is important, as is finding out the consequences of selection and whether integrated plans coordinate care better. With adequate data, it is possible to then lay out the options for states to consider.
He cited a national Medicaid experiment worth watching in Maryland, where Medicaid assigns a time period for people to enroll to make a choice and if they wait beyond that period, Medicaid will assign them. These conditions provide an opportunity to look at the results of selection.
Asked if Massachusetts has enough information to take advantage of this national experiment, Dr. Frank said he thinks they can make progress with a modest investment on the part of the state that would go a long way toward answering that question. He mentioned that people are analyzing the data, e.g., SAMHSA's evaluation of substance abuse managed care reform in several states. The Subcommittee asked staff to check the status of this CSAP project with the project officer, Joan DeLeonardo.
Mr. Michel described himself as a public policy maker interested in working with public purchasers in the mental health and substance abuse fields. He spoke from the perspective of the public purchaser.
He asserted that managed care, when done well, brings to bear a prevention and quality management orientation as well as realism about what society can afford. The task for public purchasers is to be sophisticated, knowledgeable, and focused about the top priorities in performance management. He favors organizing information into meaningful and manageable report card formats, rather than a single report card. A few models are emerging.
Mr. Michel recommended keeping the report cards simple at the outset, with such key questions as:
§ Who is being served out of the potentially eligible population?
§ What do they get?
§ Are there any meaningful ways in which we can know the correlation between what services they get and how they fare?
In terms of outcome, access is the issue, and penetration is the best measure to stratify meaningful population characteristics, type of services, and so on. The field is far from having consensus on what constitutes quality of care. Readmission rates are meaningful process measures, as is the lag time between hospital discharge and first outpatient session.
A reliable source of data is claims, because people want to get paid for their services. His advice is: if you are required or try to collect non-claims based information, pick what you know you can do well, so there is consensus that the information is worth collecting on the part of all the moving parts of a large system. It is important to try and document barriers and impediments.
Mr. Michel pointed out that Medicaid is only a minority funder of the mental health, substance abuse and other systems in most states. Thus, it is important to know the status of interagency relationships between Medicaid and the sister agencies that are co-funding services for vulnerable populations.
Dr. Dickey, professor of psychology at Harvard, also works at McLean Hospital. She spoke of the data aspects of her research on Medicaid managed care in Massachusetts and also from the provider's view about how McLean Hospital responds to requests for data.
For her, the big issue is quality of care, particularly for the seriously mentally ill under Medicaid managed care. The total number of admissions and the length of stay of this population have dropped. The shift from inpatient care to community-based care is consistent with the federal and state policy, but many advocates have expressed concern about this decline. However, her data do not show whether or not somebody was denied admission. She stressed the difficulty of answering questions about quality based only on paid claims data; access to clinically-driven data is also needed.
Another issue looms around the provider community and labor market. Here, too, almost nothing is known about what happens when performance indicators are used and whether they improve performance. Again, clinical data are needed.
Dr. Dickey noted a shift in the last year or so in the climate of access to data because of the confidentiality issue. Computerizing clinical data and providing confidentiality protection, though expensive, would enable researchers to look at a medical record without knowing whose it was; now that is almost impossible. For the record, she said that all her work is done only after it has been cleared by the DMH/IRB, by counsel at Medicaid, by Harvard Medical School, and, in her current study, by quite a few other people as well. In addition, they take many measures to ensure confidentiality of the data they work with.
Dr. Dickey and colleagues are now funded by NIMH to study quality of care for seriously mentally ill patients in the Massachusetts managed care plan, for which they are collecting primary data--again, with the appropriate safeguards for maintaining confidentiality, which has made their job difficult. Since they have to go through providers, they will not have as many enrolled in their study as they would like, which she said would probably result in a biased sample.
About performance standards and indicators, and speaking as a provider, she said the cost to collect and process outcome data in staff time is about $90,000 a year plus the cost of computer time for data analysis and reporting.
Dr. Dickey cited risk adjustment as an area that has been woefully underdeveloped in mental health. Sophisticated statistical techniques are needed to carry out this work since there has clearly been selection bias. Another issue is collecting primary data from the seriously mentally ill. Although there is concern about the reliability and validity of the information, they studied this problem and found that the information worked well.
As for what is needed, her recommendations were: (1) find a way to encrypt ID numbers in order to be able to share data across agencies within the state without concerns of confidentiality; (2) provide better computerized medical data; (3) think carefully about carving out the mental health care of seriously mentally ill people, whose medical problems are far in excess of medical problems of other Medicaid beneficiaries.
Dr. Amaro raised the question of how to integrate behavioral health care into the Subcommittee's recommendations, given that most mental health and substance abuse treatment is provided outside the Medicaid managed care system. She asked about the implications of this situation for assessing substance abuse within managed care. One challenge is how to determine the outcomes attributable to services not funded under managed care.
Mr. Michel noted the "big national argument" in the behavioral health field about the extent that mental health and substance abuse should be integrated, with about 60 percent of the people diagnosed with serious, persistent mental illness also having substance abuse culpability. Medicaid is a co-funder of these two separate systems. It is important to know about outcomes, and a better way must be found to do so. Dr. Iezzoni commented that some of the outcomes tracked may be mis-attributed to Medicaid managed care. Mr. Michel added that very few people in this population are pure Medicaid clients.
Dr. Newacheck asked if the panel was aware of any states that have made progress in linking various data on mental health services. Dr. Frank mentioned Arizona and Massachusetts as two that others can learn from.
Dr. Amaro asked if there were interest in looking at cost offsets from treating substance abuse and mental health in terms of reduced health care expenses. Mr. Michel said that although this makes intuitive sense, medical offsets have been very difficult to show. Dr. Frank added that a lot of people are trying to study this, but the evidence so far is not persuasive. Dr. Dickey is persuaded from her data that people with substance abuse and mental illness have far higher psychiatric costs and much higher medical costs. She agreed with Dr. Frank that this is different from saying there is an offset.
Readmission rates for both mental health and substance abuse, were discussed by Ms. Coltin. She noted a proposal released for public comment that would delete both of these from quality measures because their usefulness for comparative purposes is low. Dr. Dickey agreed that the measure is sometimes not meaningful.
Dr. Frank raised the point that for the time being, states are in a strong bargaining position with managed care companies. He noted the possibility of using utilization review data and screens to get clinical data, a process that is less intrusive than going into charts or talking to patients.
Dr. Iezzoni asked how to think about quality of care for those who may not be in the mental health system but could potentially benefit from it. Dr. Frank suggested using markers of care in terms of people who are not treated. Dr. Iezzoni commented that the disability model provides tools for assessing mental health services. Dr. Frank noted that screening in mental health, except for depression, is limited. Mr. Michel mentioned that everybody uses the SF36 upon enrollment. Dr. Frank said the Maryland Health Choices program uses it for substance abuse. Mr. Michel noted that in Massachusetts, only about 5 percent of those receiving services are classified as seriously mentally ill.
Ms. Porell works in the information analysis unit of the DMA's systems department, and she spoke from the viewpoint of the data facilitator. She cited the following problems:
§ Data sources. A major problem is the tendency to rely on too few data sources.
§ Collection approaches. She would like to promote uniformity, not standardization; data collection needs a uniform process.
§ Data elements and assessment methods. A particular issue is that instruments should use standardized fields for data gathering. Indicators are needed to study the effects of managed care, and should be the driving force for developing results.
§ Data quality (protocols and processes to ensure quality).
Simple but effective protocols are needed for calculation of standard utilization indicators for very low or very high levels of quality.
§ Resource constraints. The issue is that resources are very limited for adequately monitoring managed care and ensuring data integrity and efficiency.
§ Intra- and inter-agency cooperation. (This is one of the most important but least talked about issues.) Cooperative efforts in data collection are often ignored or hampered by obstacles within agencies and legal obstacles. Although cooperation between the DMA and MCOs has improved, more assistance would be highly appreciated.
Mr. Asciutto, Director of Quality Management for the PPC plan at the DMA, provided a brief overview of ways in which they rely on data to help them manage their program. In Massachusetts Medicaid, there are three different managed care programs. The primary care case management program covers about 75 percent of their population. Their MCO or HMO program has about 160,000 enrollees or between 8 and 11 percent of the population. Third is the behavioral health program (a capitated program), most of whose members are PCC plan members. Some data activities are common to members across those three programs and some help only one. One data piece common to all programs is a benefit plan report that provides information on a quarterly basis.
The DMA has an annual member satisfaction survey that has one instrument and one data collection methodology. Another activity they are now highlighting is the CAHPS survey on the PCC plan population. They tested the CAHPS instrument on their population, surveying both adult and pediatric samples. They will be able to delve into more detail about people within the plan who have disabilities. They are doing some testing of the survey methodology as well, trying to increase response rates. They also can compare results within the CAHPS survey to the general member satisfaction survey. The Division is in its fourth year of collecting HEDIS information, focusing largely on DMA's agency goals and data. A rotation of measures allows them to focus on a small set of measures in a given year. Sometimes it makes sense for the agency to deviate from HEDIS specifications, e.g. the child immunization program. They here found it difficult to find benchmarks in HEDIS that are appropriate for the population.
The focus of their first round of data collection and database development has been on the reliability and validity of the data. They expect to complete their first round by June, after which their focus will be on the production of a minimum data set and a report on quality indicators.
Mr. Asciutto mentioned that to the extent possible, they collect data from the administrative data in the encounter system to ease the burden on the plans. In regard to specific PCC plan activities, they have developed a series of reports to support quality improvement activities, including rates and member-specific detail. At the plan level, they have an asthma quarterly report and also have developed a similar report for utilization of emergency services.
Amplifying on Mr. Asciutto's remarks, Ms. Fiske explained that use of HEDIS and the member satisfaction collection cuts across all the HMOs and the PCC plan. Her remarks focused on activities specific to the HMO program and the way DMA manages contracts with vendors. Their quality improvement initiatives are linked to their contract requirements and designed to determine whether the plans are meeting those requirements. Over time, they will try to develop quality improvement initiatives with HMOs to move toward desired management conditions. All the HMOs must perform annual quality improvement activities. Recently, the DMA set about eight quality improvement goals, four that are standard for all the HMOs.
Ms. Fiske described how they come up with these quality improvement goals. They examined the previous year’s goals and looked at contract management challenges in areas where the HMOs are having difficulties, e.g., to make sure that people are getting consistent access to benefits. She added that they have limited information about the health status and needs of enrollees when they enter an HMO. Getting better information sooner is also a challenge.
They compare plans using HEDIS data, and also look at disenrollment information (both the rates and reasons). Once goals are established for the plans, they develop performance measures, which may vary from one HMO to another. HMOs are measured every six months. It is a labor-intensive process to negotiate these goals. Unless ample time for this process is allowed up-front, the result is confusion for everybody concerned. With input from clinicians, they hold regular meetings to discuss and clarify the process. It is difficult to conduct this review on a six- month basis since a year is needed to get a look at activities across the board. HMOs are given an overall ranking compared to one another. If an HMO fails to meet a number of the goals across the board, enrollment is frozen until the HMO is able to provide additional information. Incentives are provided for plans that score well. She added that not all of the information is objective and easy to score, nor is it easy to determine whether one plan has done significantly better than another plan on some of the goals.
Ms. Fiske then reviewed some lessons learned through this approach. They would like to have more frequent meetings with the HMOs, perhaps quarterly, to deal with specific problems and avoid surprises when goals are not met. She stressed that close collaboration with the HMO as essential.
The Division of Medical Assistance has difficulty collecting reliable data on race and ethnicity. The question is included on applications, but responses are optional. It is a challenge to come up with an appropriate way to capture race/ethnicity information that allows people to identify with the group they feel comfortable with while at the same time providing usable data. They now seem to be getting more detailed information on language through the member satisfaction surveys.
She cited some of the ways they try to identify persons with special needs. They do not get much information from SSI on disabling conditions, but have been working with SSI to try to get more on conditions that make people eligible for SSI.
Dr. Iezzoni said the Subcommittee now understands why they were told that Massachusetts was a national leader in approaches to data for Medicaid managed care. Dr. Newacheck, too, expressed appreciation for their impressive efforts. He asked how they thought they compared to other states in terms of the sophistication of their approaches. Ms. Fiske speculated that Massachusetts is more advanced than a number of states in the ability to collect and analyze data for managed care, probably because they have been doing it longer and have a number of active collaborative groups in New England. The HMOs they have contracts with have been working with the Massachusetts Medicaid program since the early 1980s. They also may have a larger staff than other state Medicaid agencies.
Another question from Dr. Newacheck concerned the goals and ranking system they have developed and how receptive the HMOs have been to that approach. Ms. Fiske said it depends on the HMO, but that for the most part, HMOs have found the process useful -- e.g., they have been able to extrapolate and apply some things for their commercial membership. However, there are debates about scoring, and suggestions for improvements. Those who were not responsive have left the program.
Mr. Van Amburg asked Ms. Porell to elaborate on her emphasis upon uniformity in data collection rather than standardization. Ms. Porell explained the need for flexibility, e.g. DMA uses their own provider ID. His follow-up question concerned their progress in responding to the forthcoming HIPAA standards for transactions. Dr. Iezzoni emphasized the importance of the issue, explaining that the standards are to be implemented throughout every health care transaction. Ms. Porell admitted they have work to do internally to set up protocols. Dr. Iezzoni underscored the point that it is probably going to be hard for some of their plans if the rest of the system asks them to use the HIPAA standards in reporting and Medicaid does not.
Mr. Van Amburg asked Mr. Asciutto what quality indicators they use, and Dr. Iezzoni asked if it were possible to get a copy. Mr. Asciutto said they could get a draft of the indicators.
Mr. Van Amburg asked Ms. Porell about their staff composition. The staff has one statistician and the other members are MS-level senior analysts and Natural programmers (claims data are stored in Natural). Additionally, DMA has contractual relationships with different groups for special projects.
Dr. Amaro asked Ms. Fiske to elaborate further on problems in collecting race/ethnicity data, and she replied that they are still trying to learn the best way to ask the questions. People are much less willing to fill out that type of information on applications, perhaps fearing it will either be linked to their immigration status or to their benefits determination.
Ms. Porell said that the optionality of the race/ethnicity question on enrollments brings the reliability of those data into question. Ms. Fiske added that they cannot deny benefits based on omission of information about race/ethnicity. If a person fails to answer, staff will try to fill in the answer on the enrollment record. Ms. Fiske said the questions have been improved. Mr. Asciutto said that if they have to rely on information found in eligibility files, they feel less comfortable with the analysis. They are more comfortable looking at results across different groups when information is self-reported, such as on the member satisfaction survey.
Dr. Iezzoni, referring to the many people in Boston who speak Creole (Haitians) or Portuguese, asked if the satisfaction survey is only in English and Spanish and whether they have any initiatives to try to address access problems for those populations that do not speak Spanish or English. Mr. Asciutto said they are trying to get information on everyone's language using "Babel" cards, which lists 13 languages. Ms. Fiske added that the HMOs also require multilingual cards that direct them to get assistance in translating a particular language. All of their HMOs also have access to the AT&T language line for people calling on the phone. One of the health plans has a large Khmer population, so their providers have to perform translation services for that group.
Dr. Iezzoni thanked the panel for their extremely helpful remarks. She reminded them to supply the Subcommittee with the MEDSTAT specs, as requested, and said that more information about HIPAA is available on the Committee's website.
Dr. Bergman described the Tufts Health Plan. Tufts is a network model with 925,000 members, 16,000 of the members are Medicaid recipients. He indicated that they currently collect encounter data, selected HEDIS measures, satisfaction surveys, selected outcomes measures, health history and risk assessments. Tufts compares the Medicaid population to other population groups using: utilization and quality trends, as well as HEDIS and satisfaction surveys.
Dr. Bergman described gaps that exist in data that is currently available.. The gaps in existing data include: clinical practice guideline compliance, health and educational status, primary language and disease incidence in the Medicaid population currently not in enrolled in managed care. These gaps restrict restrict comparisons across the population groups. In addition to gaps he indicated that there are impediments to collecting data from the Medicaid population, including: difficulty in location (i.e., lack of addresses and telephone numbers), state specific modification to HEDIS, lack of access and costliness of reviewing medical records and specific restrictions related to privacy of the medical record and requirements regarding patient consent to release information. Dr. Berman indicated that their future plans for data collection include increasing the focus on risk assessments.
Dr. Upshur described her experience as a health care researcher including her membership on the Division of Medical Assistance Medical Advisory Board. She discussed the difficulties in obtaining accurate information on racial and ethnic populations and reminded the audience that race and ethnicity are cultural, not biological or genetic constructs. She cited an example of outcomes from a number of focus groups that were designed to assess community health issues for the Latino and Black population in MA. She indicated that a recurring issue in the focus group was the difficulty they were experiencing in understanding and using managed care. These included: complicated enrollment and managed care provider and physician selection. In addition, Dr. Upshur described the difficulties resulting from inadequate coding of racial and ethnic minority status.
Dr. Upshur recommended that national policy be established requiring state Medicaid programs to collect basic information that can be used to monitor health care access, utilization and outcomes for racial and ethnic groups including English language proficiency and language barriers in health care.
Mr. Wong indicated his support for Dr. Upshur’s comments on the importance of collecting accurate racial and ethnic identifies as a means to reduce disparities in health. Mr. Wong spoke about the Asian and Pacific Islander population, their growth rate, and barriers to health care that newly arriving Asians experience. Mr. Wong stressed that implementation of the OMB Directive #15, identifying race and ethnicity is critical. In addition he indicated that more must be done to collect additional sociodemographic data, including employment, health insurance and type of payments for health care (including variables tha measure public assistance). Mr. Wong stressed that sociodemographic data, coupled with racial and ethnic indicators, can provide a more systematic approach to monitoring the health of the United States population.
Dr. Amaro asked the panel about the difficulties in reporting race and ethnicity data and how this related to the need for information that insures that public dollars are purchasing access to health care. Mr. Wong recommended the utilization of th OMB Directive #15 combined with specific training for interviewers. Dr. Bergman recommended face to face interviews to collect this information for Medicaid enrollees. Ms. Greenberg probed the panel regarding the timing and placement of questions on race and ethnicity. Mr. Wong indicated that he had not experienced a problem with this, and Dr. Upshur stated that the best time to ask these questions is at the time of enrollment.
Dr. Iezzoni then adjourned the meeting.
I hereby certify that, to the best of my knowledge, the foregoing summary of minutes is accurate and complete.
/s/ Lisa I. Iezzoni August 19, 1998
Lisa I. Iezzoni, Chair Date