The Subcommittee on Population-Specific Issues of the National Committee on Vital and Health Statistics convened at 10:40 a.m., Monday, February 9, at the Wyndham Metrocenter Hotel, Phoenix, Arizona. They reconvened at the Arizona Health Care Cost Containment System on February 10. The meeting was open to the public. Present:
Lisa I. Iezzoni, M.D., M.S., Chair
Hortensia Amaro, Ph.D.
George
H. Van Amburg
M. Elizabeth Ward
Dale C. Hitchcock
Lynnette Araki, National Center for Health Statistics (NCHS)
Jason
Goldwater, NCHS
Kathryn Coltin, NCVHS Member
Don Detmer, M.D.,
NCVHS Member
Sundin Applegate, M.D., Arizona Department of Health
Services
Mary Moien, Consultant, NCHS
Diane Greeneich, D.N.Sc.,
Fox Systems, Inc.
Carol Lockhart, Ph.D., Lockhart Associates
Mary
Temm, Temm and Associates, Inc.
Betsy Trombino, Arizona Family Voices
Amy Burns, NCHS
Donnie Hall, Health Systems, Inc.
Brian
Lensch, Arizona Department of Economic Security
Cathy Valle
Alan
Schafer, Arizona Health Care Cost Containment System (AHCCCS)
Brent
Ratterree, AHCCCS
Joe Gaudio, Arizona Physicians IPA
Nelda
McCall, Laguna Research Associates
Gayle Averitte, The Medstat Group
John Murphy, Flinn Foundation
Representative Susan Gerard,
Arizona State Legislature
Jan Hart, AHCCCS
Kari Price, AHCCCS
Terri Keagle, AHCCCS
Susan Cypert, AHCCCS
In visiting Arizona, the Subcommittee sought to learn how Medicaid managed care has worked there, particularly in respect to information systems and data. They heard from two panels on the first day. On the second day, they heard from another panel and then visited the Arizona Health Care Cost Containment System (AHCCCS) where they met with three panels of AHCCCS staff members.
Mary Temm (Temm and Associates) described problems with Medicaid managed care data in both the acute and long-term care arenas, focusing on quality and timeliness of encounter data. She emphasized that Arizona's information is superior to that of many states. She emphasized that it is more useful to compare utilization patterns between the Medicaid programs of different health plans than to compare a plan's Medicaid enrollees with its own commercial plan populations. She stressed the differences between the capacities of large and small plans in Arizona and noted data collection and analysis difficulties caused by "an ever- increasing contractual hierarchy." She recommended direct submission by the provider to the Medicaid oversight agency. She noted that sanctions have drastically improved compliance in encounter reporting, as well as the reliability of the data. Finally, she recommended the standardization of data collection and reporting systems for Medicaid managed care.
In discussion, the Subcommittee touched on race/ethnicity data collection, language issues, validation studies, the use of satisfaction surveys, and the gaps caused by behavioral health carve-outs and "churning."
The second panelist was Betsy Trombino. Her son (now an adult) has special health care needs, and she works with Family Voices, a family support and advocacy organization. She affirmed the need for data to improve services, but raised two overarching questions: Why is the information being collected? and, is it really going to make a difference for the people served?
She stressed the need to improve coordination in data collection: for example, the same information for the same individual is gathered many times, by the same agency. This also can be threatening and confusing for families, as they fear benefits loss for their children. She reported that Family Voices, in collaboration with Brandeis University, is currently conducting a nationwide survey of families of children with special health care needs and their experiences in managed care. In regard to data dissemination, she recommended that data should be made available to families, and to organizations like Family Voices so they can assist in interpreting the data and provide a family perspective. She also recommended that agencies and providers establish family advisory groups.
In the discussion period, Subcommittee members noted the challenge of reconciling families' confidentiality interests with their desire not to be asked repeatedly for the same information. Ms. Trombino urged that families be asked to advise on the way questions are phrased and who asks them.
Dr. Carol Lockhart, a health systems consultant with long experience with Medicaid managed care, focused on policy and design issues. She said that committed providers have minimized the number of people who have "fallen through the cracks" in the transition to managed care. Legislatures' unwillingness to adequately fund information systems is a major source of data problems. She stressed that Medicaid agencies should use their purchasing power to influence outcomes, and regulatory controls are also needed.
Like Dr. Lockhart, Dr. Diane Greeneich (Fox Systems) formerly worked for AHCCCS. She called particular attention to the need to associate Medicaid managed care data with other data systems, e.g., a pilot program combining immunizations and asthma information. She commented on a range of problems with encounter and enrollment data and on the importance of standardization. Her recommendations included a minimum data set for encounter data and a national Medicaid managed care database maintained by the Health Care Financing Administration (HCFA).
Dr. Sundin Applegate (Arizona Department of Health Services, Community and Family Health Services) discussed issues related to the interface between public health and Medicaid, especially as they concern children with special needs. He recommended a move to "specialized managed care" for such children; “carving out” both primary care and special services. Pilot programs in California, Washington, D.C., and elsewhere are experimenting with this approach. He noted the need for a clear definition of a "child with special health care needs" and for better functional assessment tools. Finally, he noted the potential benefits of managed care, in terms of both providing points of entry for public health interventions and improving continuity of care.
Much of the discussion period was spent exploring ways to leverage managed care resources for public health improvements. The group considered possible incentives as well as the state's ability to build requirements into contracts. It was noted that the public health sector can teach plans how to do prevention, and this is beginning to happen. Roz Lasker's work in this area was cited, and the Subcommittee agreed to ask her for a briefing.
Before recessing for the evening, members discussed the themes emerging from this and the previous hearing and their implications for the Subcommittee's report. It was noted that population-based data are not receiving attention in Medicaid managed care, as they should be; nor is anyone looking at the Medicaid population as a whole. Gaps in services for specific populations were raised as another issue, along with limitations in race/ethnicity data.
There was general agreement that the purpose of information should be clearly identified before plans are asked to provide it. Moreover, the Subcommittee's report should focus constructively on the general principles underlying desired changes, especially in terms of what the proposed information-gathering will enable plans and providers to do. They reviewed the idea of a minimum data set with flexibility for state adaptations.
The group outlined a list of questions to be addressed in the report, following a pattern like that in the Committee's metaphorical camera. They agreed that HCFA's help should be enlisted in addressing these questions. They also agreed to ask Dr. Rosenbaum et al. to look at enrollee satisfaction, service denials and complaints.
The first panelist to speak was Nelda McCall of Laguna Research Associates. She gave a brief overview of AHCCCS and its management information system. She indicated that this provides considerable accessible on-line information.
John Murphy of the Flinn Foundation then described a series of surveys conducted by the foundation, with the support of the Robert Wood Johnson Foundation to look at the Medicaid managed care (AHCCCS) population over time. Generally, the findings show satisfaction and adequate coverage. Representative Susan Gerard expressed concern about the data collected on managed care, lack of comparison with Healthy People 2000 goals and enforcing contact compliance to collect outcome information.
Joe Gaudio, chief financial officer of Arizona Physicians IPA, described the information requirements for managing his organization, to assure quality of care, manage liability and prepare bids.
Panelist Brian Lensch works with Medicaid and non-Medicaid clients in the developmental disability and long-term care area. He noted that data requirements can sometimes undermine program efforts and then described efforts to reduce the number of times families are asked for information. He supported earlier comments about linking public health and encounter data.
The discussion session focused on the data and processes used to monitor quality of care. The group also talked briefly about the public health implications of the large number of illegal immigrants who are unwilling to be registered in the state's immunization registry.
In the afternoon, the Subcommittee met with three panels for an overview of AHCCCS program and data requirements and presentations on encounter data, validation studies, and quality indicators. All presenters were AHCCCS staff members.
Office of Managed Care managers, Jan Hart and Kari Price, use plan data primarily for financial viability and capitation rate development. They described the information systems as excellent, acknowledging that it would be better if it were possible to download data and address special queries. AHCCCS is developing an RFP to create a data warehouse. This will provide query capability. The absence of pharmacy data was raised as an issue, one that was revisited with the other panels. The methodology for setting capitation rates was discussed. In the acute care portion of the program, AHCCCS risk adjusts for: age, gender, and service area. AHCCCS also has a stop-loss reinsurance mechanism. One of the gaps mentioned was information on services and enrollment data on dual-eligibles. Ms. Hart called attention to the significant amount of oversight in Arizona and the strong familiarity of AHCCCS staff members with the plans.
The group looked at several aspects of the disparities between large and small plans and the tension to preserve choice for members, and the tendency of larger plans to crowd out the smaller ones. Ms. Hart described the Readiness Assessment Tool with which the state evaluates plans before contracting with them, to determine their viability and ability to comply with requirements. Asked for recommendations, Ms. Hart stressed the importance of being clear at the outset about what data are to be reported, and how these are reported.
Terri Keagle and Brent Ratterree then reported on encounter data and validation studies. The state has collected encounter data for many years, and reports these data to plans and HCFA. Ms. Keagle and Dr. Iezzoni discussed issues surrounding pharmacy data collection. Ms. Keagle then described the state's staffing for providing technical assistance and processing encounter data. The timeframe for reporting encountered data (i.e., up to 240 days) was discussed in terms of how this lag undermines quality monitoring. Mr. Ratterree described the validation methodologies used by his office. As recommendations, the panelists said states should know what “correct” information looks like before they begin to collect it.
Susan Cypert and Alan Schafer then discussed the quality indicators used for performance measurement, providing the Subcommittee with extensive documentation. They returned to the issue of the 240 days allowed for encounter data submission, which means they sometimes have to resort to chart review for more timely quality assessment. Another barrier is the unwillingness of sister agencies to share data, largely because of confidentiality constraints. They hope to get the legislature to change the state statute in this regard. Other gaps are: absence of the pharmacy data, and benchmark data from other states to which they can compare themselves. (It was explained that the state's data are not comparable with those of others in a number of ways, partly because Arizona has used encounter data for a longer period of time.
The group then discussed immunization data and the potential uses and limitations of the state's immunization registry. Ms. Cypert indicated that they are developing prenatal data within the encounter file format. Asked about risk-adjustment, she said they are starting to think it will be necessary to risk adjust; they have not figured out how.
Finally, after describing several member satisfaction surveys conducted by AHCCCS, the panelists were questioned about staffing. They described it as extensive and provided details. Mr. Van Amburg concluded from the descriptions that if states moving from fee for service to managed care follow Arizona's example, reductions in staffing will not necessarily occur if it’s done properly.
Dr. Iezzoni remarked that the Subcommittee's aim in visiting Arizona was to learn how Medicaid managed care has worked there, particularly in regard to data and information. After the participants identified themselves, she thanked Jason Goldwater and Carolyn Rimes for preparing the meeting and background materials.
Mary Temm
The first speaker, Mary Temm, has been in the managed care industry for 13 years. She has been involved in two components of the Arizona Medicaid agency: the acute care Arizona Health Care Cost Containment (AHCCCS), and the long-term care Medicaid managed care program.
She noted that data issues vary among the health plans, depending on the plan size, ranging from about 500 to more than 400,000 covered lives. She began discussing the type of data available to help health plan program developers to determine the feasibility of entering into the Medicaid arena. Specifically, she cited the flaws in encounter data, i.e. these data do not accurately portray current program activities. The data must be adjusted because they are relatively old when submitted to AHCCCS. Their accuracy and completeness are also questionable. Still, in comparison to data she has seen for other states, Arizona’s Medicaid managed care program is "stellar."
In Arizona, all the health plans and both the acute and long-long-term Medicaid care programs are required to submit encounter data using HCFA 1500 data and UB-92 data. The health plans analyze the data to evaluate the quality of care and cost containment performance of their subcontracted providers.
The long-term care programs use encounter data to place recipients. All Arizona Medicaid programs require that data collected through member satisfaction surveys be used in conjunction with encounter data. She noted that because of the large differences between their enrolled populations, it is more useful to compare utilization patterns between the Medicaid programs of different health plans than to compare a plan's Medicaid enrollees with its own commercial plan populations.
While larger health plans are exploring outcomes for Medicaid managed care beneficiaries, smaller plans lack data management capabilities for these analyses and they focus on: primary care data and utilization trends.
Ms. Temm cited as the most difficult part of data collection and analysis is created by the increasing contractual hierarchy. For example, direct submission by the provider to the Medicaid oversight agency would drastically reduce processing expenses. Her personal preference is a system to allow a direct flow of data from providers into the oversight agency, with provisions for other entities to access the data. Of course, confidentiality safeguards would be necessary.
The Arizona Medicaid system imposed sanctions upon health plans for timeliness, correctness, and submission of data based upon encounter AHCCCS validation studies. Most of the health plans pass along the sanctions to their contracted providers. The assessment of sanctions has improved compliance related to encounters, and increased the reliability of encounter data. However, these sanctions are not monitored in a timely way.
She recommended that data collection and reporting system on Medicaid managed care be standardized, and address Medicaid needs. Few systems are able to accommodate the nuances of Medicaid in collecting, processing, and submitting encounter data. If Medicaid managed care data collection and reporting were standardized, information system vendors may be able to accommodate the requirements, which would be a "real win" for smaller health plans, especially for rural health plans. The standardization also should apply to procedure codes. These vary across states.
She also suggested that the Subcommittee consider differences between commercially licensed HMOs entering the Medicaid program and HMOs developed specifically to serve Medicaid populations. In Arizona small plans lack necessary resources to accommodate major changes in their information systems. In conclusion, she said it will be imperative to put in place data systems that allow for analysis of data for Medicaid managed care.
Discussion
Ms. Temm said that race and ethnicity data are not easily accommodated. The more sophisticated health plans collect those data at enrollment and try to merge them into the encounter information. This is difficult in the absence of information systems that can accommodate demographic information.
She was then asked how they track cultural/linguistic performance as part of their oversight of quality of health care in Arizona, given the significant size of the Spanish speaking population. She said that most plans survey their providers at the time of contract initiation and then annually to ensure they have updated information about linguistic capabilities in their offices. They are aiming at a dual profile of both providers and members. Most of the information systems in use allow for tracking languages related to specific providers. The difficult part is collecting plan membership information. In some of the smaller plans, case management has been helpful in eliminating problems of this kind. She reiterated that demographic information collected at each encounter does not include race and ethnicity, and that most of their cultural/linguistic information comes from surveys.
At Mr. Van Amburg's request, she described the process for validation studies as a collaborative effort between AHCCCS and the health plans. AHCCCS staff audit the records and present the results to the plans, which are given an opportunity for rebuttal.
Asked if there is a similar data collection process in long-term care, she affirmed that the process is the same in terms of cycles and audits. However, sanctions have been minimal for the long-term care programs, because they are small and the state is fully aware of the account information.
As to whether they need an encounter data system, Ms. Temm said, "definitely ... for rate setting and for quality and performance measures for individual providers." In response to the question of whether they could extrapolate data on a sampling basis, Ms. Temm said it is very difficult because current systems do not allow set matching, e.g., to look at age- specific mortality versus diagnoses.
Turning to the problem of coding, Dr. Iezzoni noted the consensus on the need to systematize and make all the codings equivalent nationally, and asked Ms. Temm's opinion on impediments to this. Ms. Temm identified the major impediment the need to change state contracts and reimbursement schemes. As for what "carrot" could be held out to lead to a more systematic way of coding, Ms. Temm replied that the incentive to change would depend upon the entities involved in multiple states, and that it would be hard to find a carrot for smaller plans.
Asked to amplify the issue of information system duplication for very small companies, and how they set up systems to report to the state, Ms. Temm mentioned that in Arizona three to four health plans are utilizing the same "home-grown" vendor who developed an information system to accommodate AHCCCS. The development of a national coding system could affect this single vendor and the health plans because of their size and rural nature. Any major changes should consider protection of these small entities.
Asked to comment on use and cost of statisticians in processing data, Ms. Temm noted that the larger health plans have their own statisticians, but the smaller plans must contract for those services.
Returning to the issue of quality, Ms. Temm cited the type of analysis used in an examination of over and under-utilization. They try to work with plans to develop their own benchmarks, and then analyze the data specific to the provider level.
In response to another question, Ms. Temm said that satisfaction survey data are rolled into their quality assessments. Each plan does things a bit differently. Other satisfaction surveys look at providers in relation to access and availability of services. The health plans may or may not collect information on gender, race, ethnicity, age, and disability status, and not all plans have the capacity to do analyses across population sectors.
Asked why she considers the Arizona data systems for Medicaid managed care "light years" ahead of other states, Ms. Temm said it had taken a long time to get where they are today. The imposition of sanctions is a key factor, plus the fact that the data are edited and corrected before they are accepted into the ultimate repository. The edit process has also helped greatly with data integrity.
Dr. Iezzoni asked how the people in Arizona anticipate the impact of the Balanced Budget Act on Medicaid managed care and data reporting. So far, Ms. Temm said, it has been quiet in Arizona, unlike California where "it's all a-buzz."
Dr. Amaro, asked about coverage gaps in behavioral health, where people are going for those services, and who funds them. Ms. Temms acknowledged that those people are falling outside the Medicaid managed care arena, and consequently those data are not being analyzed. There is not a mechanism to pull together all public health services being provided for these persons. Therefore, there is no way to really know what public health is covering and how that contributes to outcomes.
Asked about the "churning" issue and whether there is a system for tracking people who are continually going in and out of the Medicaid system, Ms. Temm said that, at the state level, AHCCCS monitors those data more than the health plans do. However, some of the plans do monitor the average length of time for their Medicaid recipients. As for a way to study outcomes for those people who go in and out of Medicaid, there may be information at the individual health plan level, but Ms. Temm is not aware of any system for tracking those people at the state level.
Betsy Trombino
Ms. Trombino is state coordinator for Family Voices, a national grassroots network of family and friends who advocate for children with special health care needs. Her son grew up in the managed care system as a child with special health care needs, so her experience with commercial managed care and with Medicaid managed care is first hand. She said that she appreciates the importance of data from the perspective of a family member, knowing that the system cannot be improved without meaningful data. The complexity of data collection is an issue for families, which must deal with different eligibility requirements, languages, and other issues.
Her remarks focused on children with special health care needs, and more of whom are being served by Medicaid managed care. She briefly touched upon several issues: the collection of data at the family level, data used by HCFA and AHCCCS, and information dissemination. She raised two overarching questions: “Why is the information being collected? and, Is it really going to improve services or make a difference to the people served?”
She stressed the need to improve coordination in data collection, as the same information is gathered many times for the same child, even by the same agency. Based on her conversations with other parents in Family Voices, she described as disruptive and intrusive the number of occasions on which people regularly come into homes to collect information. These data are connected to eligibility determination. This appears to be an enormous duplication of effort and waste of money.
Another issue of concern is the large number of people in Arizona for whom English is not their primary language, and the limited number of providers who can help with this language problem. "Filling out all of these forms is hard enough when you are college educated and speak English," she added.
An important concern is related to balancing the issues of privacy and confidentiality with the needs for data, since parents whose children have special health care needs are often fearful of losing benefits. "It is terrifying how hard it is to get services, and if that [access] is lost, and your child's not Title 19 eligible, you're up a creek." Employment and insurance coverage may be at risk and families wonder what will be done with the information.
Ms. Trombino said that from her perspective, the important questions are how many children with special health care needs are served by Medicaid managed care, and whether they are getting the care and services they need. It would be useful to have data on frequency of service denial, number of appeals, and their outcome. Other interesting information to analyze is: providers payments for caring for children with special health care needs, and degree of appropriateness of these payments. She noted that some pediatric specialists choose not to work with Medicaid because it is too complicated and the system does not make it worth the trouble.
Family Voices, in collaboration with Brandeis University, is currently conducting a nationwide survey of families whose children have special health care needs, to examine their experiences in managed care. Arizona is one of the first states being surveyed. The early findings may be of interest to the Subcommittee.
Ms. Trombino turned to dissemination, stressing that pertinent data should made available "first and foremost" to families and to organizations like Family Voices. They can assist in interpreting the data and by providing a family perspective. She also recommended that agencies and providers establish family advisory groups, to advise as to needed changes based on data from satisfaction surveys and other information from families.
She suggested that any information collected be reported in practical terms, "user- friendly language" so that people can utilize it in dealing with health plans and in their advocacy efforts. Furthermore, she encouraged use of data collection and analysis to examine the impact of the system on children with special needs.
Discussion
Several Subcommittee members expressed appreciation for Ms. Trombino's testimony. She was asked how she would balance confidentiality issues involved in providers' sharing data with families' discomfort with redundant data collection, given that there is a tradeoff between the two. She agreed that this is a complex issue and she suggested talking with a group of families about what works for them that would not jeopardize their privacy and control over who sees their children's records. Family Voices could help the Committee look at that issue.
Dr. Iezzoni and Ms. Trombino discussed the practice of repeated needs assessments for children whose conditions (e.g., cerebral palsy) are not likely to change. Ms. Temm noted the requirements for continual client assessment and tracking, adding that these data are not being fully utilized because they are entered into a one-way system. Program contractors who enter the information and who could utilize it in their case management are unable to pull the data out for their own reporting. Ms. Trombino estimated that families are asked for data on average every 30 to 90 days. New forms are required whenever there is an interface with a new program or provider.
Dr. Amaro indicated the need to develop tracking systems to learn if the health care system is working for people in different categories, instead of just looking at overall data. Ms. Trombino cited satisfaction surveys as a related issue and indicated those who are the most overwhelmed by the system may not have the energy to respond to satisfaction surveys.
Asked to explore further what solution might overcome concerns about privacy and confidentiality, Ms. Trombino said that some of the tension might be resolved simply by the way the questions are asked and how clearly it is explained what will happen to the information. It may also be a matter of who is asking. She repeated her recommendation that family advisors be consulted in the development of questions.
Asked about meaningful outcome measures, she cited "The ABC's of Managed Care for Children with Special Health Care Needs," which lists that information and was developed by families. She added that continuity and communication are important factors for families, as well as knowing what to do when they are not satisfied with services. She strongly stressed the importance of learning how to work within the system of managed care.
Asked about behavioral health carve-outs, Ms. Temm was said the state contracts use essentially the same standards as the AHCCCS program for encounter reporting requirements.
To clarify an earlier discussion topic, Donnie Keaton Hall, the principal and CEO for a health care consulting firm in Phoenix, noted that eligibility data are collected by the Department of Economic Security. These data systems are often incompatible with those of AHCCCS and of health plans.
Dr. Carol Lockhart
Dr. Lockhart is president of Lockhart Associates, a health systems consulting firm. She helped plan Arizona's Medicaid managed care program (AHCCCS) and managed it for its first five months. She is an advocate for public health and populations. The focus of her remarks was policy, design, and context, rather than specific data issues. She addressed three basic issues: the effect on beneficiaries in the system; important issues of Medicaid managed care; and what sanctions, enforcement and monitoring are needed. After quoting Peter Drucker that "All effective innovations are breathtakingly simple," she noted that Medicaid managed care systems are not simple and therefore will experience difficulties.
Although problems in the system have permitted some patients to fall through the cracks, she stressed that this has been minimized in Arizona because of providers' commitment to their patients. A major source of the problems is the unwillingness of legislatures to adequately fund the resources needed for implementation and data collection. People must be convinced about the importance of data. A barrier to getting data is the unwillingness of agencies to share and cooperate. Also, information from previous innovations is frequently ignored. States often re-create the very problems that were experienced earlier elsewhere. Limited insurance and capital requirements, minimal oversight in general, grievance processes, and marketing approaches have continued to place patients and providers at risk in the system. She added that legislators will not deal with a lot of complex data, so the information must be presented simply. Indeed, she noted a lack of legislative involvement and understanding in many states.
She commented that Medicaid managed care programs too often act as fiduciary agents, when they should actually use their purchasing power to guide the system and the outcomes, both for individuals and for populations. She then touched on several other areas where changes and improvements are needed:
· greater transparency in plan certification and patient and provider grievance procedures;
· parts of the system will be at risk and a mechanism to study these should be set up. These include, e.g., traditional and minority providers, medical education, research, community health centers, rural and inner-city patients and providers;
· linkages in managed care with the public health services and systems;
· sanctions and enforcement are too often seen as a "black box";
· need for a reward structure for plans that respond to and creatively move patients toward good health;
· a new set of regulators: people with program and policy expertise; and
· insurance and capitalization requirements should be addressed.
In terms of data collecting and monitoring, she suggested considering whether the goal is survival or efficiency of the program. Both sanctions and rewards should be looked at in light of the answer to that question. She concluded with the advice that market signals are not enough to ensure patient welfare, and that regulatory control is a necessary part of pro- competitive indigent health care policies.
Dr. Diane Greeneich
Dr. Greeneich is an associate of Fox Systems of Scottsdale, Arizona, a national Medicaid consulting firm. She was the research administrator for AHCCCS from 1995 to 1997. She asserted that the most important data issue is the association of managed care data with other data sets. This linkage is complicated by confidentiality issues.
Another issue is collecting encounter data is in the appraisal of health care quality and cost effectiveness. Refinement of encounter data allows state agencies to demonstrate the efficiency of managed care contracted plans, leading to a more realistic capitation rate structure. In addition, provider data furnish information on primary care specialists, referrals, PCP affiliations, and provider capacity. She also highlighted the need for standardization of enrollment file requirements, including demographic information (now considered optional when ascertaining eligibility) and linkages between families found within different state and Medicaid databases.
Dr. Greeneich emphasized that full evaluation of Medicaid managed care will depend on in-depth outcome studies. Although NCQA measures present a beginning approach to performance and outcome measurement, she hopes the Subcommittee will have recommendations regarding outcome data in Medicaid managed care and comparisons with the state population findings.
Impediments to data collection are most sensitive in the collection of encounter data. She noted impediments are caused by defined objectives, costs, and technology. In Arizona, HEDIS measures have proved useful in the tightening of collection and use of encounter data. Standardization of reporting parameters would facilitate realistic cost projections of needed encounter data collection. The technological impediments include: failure to define interface requirements, electronic formats and data exchange frequencies, the maximization of electronic data transfer, and Web-enabled technology. As health care moves into these new stages, one of the big stumbling blocks is meeting new technological requirements.
On the issue of association of data with other existing data sets at both the national and state levels, she cited her work as a member of the data sharing committee of the Arizona Asthma Coalition and said that "it's only through such coalitions that we have an opportunity to really speak about some of the issues that we face, a dialogue that will bring about a partnership." She added that Medicaid managed care currently exists in a condition of "DRIP: data rich, information poor." She offered the following recommendations for the Subcommittee:
· Plan methods of validation of electronic data (a point she emphasized strongly).
· Consider cost when planning standardization of reporting. Data sharing efforts can facilitate cost-effective ways to handle health data.
· HCFA should maintain a data base for Medicaid managed care.
· Adopt a minimum encounter data set as an industry standard.
Dr. Greeneich noted that a description of the referenced minimum encounter data set was included in their handout.
Dr. Sundin Applegate
Dr. Applegate is medical director for Community and Family Health Services, the organizational unit in the Department of Health Services that is responsible for maternal and child health. He noted that the AHCCCS was created as an alternative to traditional Medicaid after the state experimented with "one of those runaway government programs."
Dr. Applegate's special interest is the interface between existing public health programs and Medicaid, especially in the program for children with special health care needs. He focused on the Children's Rehabilitative Services, one of several programs in the Health Department that interface with AHCCCS. It is a carve-out providing specialized medical services while clients' primary care services continue to be provided by the general AHCCCS programs, creating the familiar interface problem. The answer to that problem, in his opinion, is moving towards an era of "specialized managed care." He cited pilot programs in California and Washington, D.C. which "carve out the entire child" -- adding that he would like to do away with the term "carve out" and instead use the phrase "specialized managed care."
He identified three interface-related issues of relevance to families and physicians:
· definition of a child with special health care needs (a national issue for which there is finally a momentum toward a solution -- e.g., a project by the National Association of Children's Hospitals and Related Institutions),
· need for functional assessment (which his agency is addressing through a pilot project that will enable them to track outcome), and
· the need for socioeconomic data (for which Medicaid enrollment is used as a surrogate).
He discussed the need for Medicaid managed care information from the public health perspective, noting that data gathered about a public health concern provide a point of entry for intervention. He indicated that the Arizona Partnership for Infant Immunization, has gained the cooperation of all the major commercial managed care plans in assessing the immunization status of two-year-olds. This joint effort has significantly increased the immunization rate. A similar initiative is the Asthma Coalition, mentioned earlier. Since asthma calls for education of both patient/family and providers, again, managed care organizations are good venues for reaching both.
He noted that with the maternal and child health block grant states are moving toward agreement on a national set of performance indicators. Medicaid data will be a source for performance indicators. As an aside, he suggested that HIPAA's great value is in promoting national provider codes and sharing data across state lines. A commonly agreed-upon provider identification will make life much simpler.
Turning to newborn screening, Dr. Applegate noted that access to the Medicaid AHCCCS program enables them to identify beneficiaries' primary care physicians -- an example of what he considers the advantages of managed care for continuity of care.
Discussion
Asked to comment further on linkages and interface with public health, Dr. Applegate asserted that some areas of managed care are becoming more stable and less subject to member turnover or plan collapse. He advises consumers to choose plans with roots in the community, and cautions against plans that are created with minimal financial and staff resources. Dr. Greeneich agreed.
Asked whether current data systems are able to provide information on the extent to which prevention is being integrated, the panelists said that some do. Dr. Lockhart noted that a local hospice was unable to identify categories of patients with a particular condition and to know what was happening to them as they were dying.
Asked for more comments about functional status, Dr. Applegate identified domains where an impairment is graded on an objective severity scale. Dr. Iezzoni noted that HCFA is now debating the value of the functional independence measure for adults, and Dr. Applegate responded that it would be nice if there was national agreement on this.
Returning to the issue of information-gathering from families, Dr. Applegate said his program has a paid cadre of parents on contract who sit on all policy committees. They have settled on a basic demographic profile that fits everybody's basic input. Program-specific eligibility requirements may be put in an addendum.
Asked if the presenters had any solutions to the difficulties in getting access to data for analysis, Dr. Greeneich suggested the development of a national standard and a national data base. She also called for better ways of working with financial data, as well as better dialogues about evaluation. Dr. Applegate added that if people are really serious about getting data, they can do it, "but if they don't want to do it, they can find a thousand reasons why they cannot." He questioned the need for a federal mandate to resolve this problem.
Dr. Lockhart agreed with Dr. Applegate, "maybe because we've both been regulators." Although it would be good to have some standardization, she cautioned against trying to collect everything that can be counted. The same, she suggested, is true for standardization: some is really necessary, but other things are better done within the plan or at the state level.
In response to a question, Dr. Applegate said his program gets very little back after feeding data on their program into the AHCCCS system. They have been more successful working directly with motivated individual plans. Asked by Dr. Detmer about incentives to help realize the potential in managed care for leveraging public health population improvements, he cited denominator assessments, using a set of systems, as "one of the richest potential areas for improvement in our health care system nationally." Dr. Greeneich suggested that more time be spent looking at a specific populations and how they can be more effectively and efficiently served.
Dr. Lockhart pointed out that Medicaid managed care programs were created to control costs, not to improve health care status; however, states with an executive and legislative commitment to positive health for the citizens are doing some very innovative things. She stressed that the Medicaid managed care programs can set clear agendas and specify them in their contracts, with rewards and penalties.
While agreeing on that goal, Dr. Amaro questioned how prepared and appropriate the personnel are, given that the public health model is different from medical care models. Her concern is how well data can help determine whether particular goals are being met with particular populations, e.g., prenatal care, AIDS, substance abuse, and mental health services. Dr. Applegate noted that one problem in behavioral health is that monies get diverted: the state monies were spent in matching for Title 19, so there is far less money available for non- Medicaid eligible clients.
Returning to the subject of public health and the contractual demands that the Medicaid agency can put into its agreements (e.g., culturally appropriate translation and related issues), Dr. Lockhart emphasized that the agency must specify contractually what it wants, just as it does with community providers. Another point she stressed is prevention: the more stable plans are beginning to look at prevention and asking public health providers to teach them how to do it. This suggests a clear role for the public health sector. She cited a project by Roz Lasker in New York, and the Subcommittee agreed to invite her to brief them on it.
Dr. Amaro pointed out that the Committee's recommendations for improving data systems have not adequately addressed the many issues that intersect in health care and ways for public health to play a bigger and more integrated role.
Asked about the need to develop a data system to track the services received by people from public health, and how to document the extent to which public health agencies contribute to changes in the health of a population, Dr. Applegate said there's a real opportunity and "we just have to learn how to do things differently."
Dr. Iezzoni asked if people in Arizona are aware of how they will be affected by HIPAA mandates. Dr. Applegate did not think there is widespread awareness. Dr. Iezzoni said the Committee is afraid that people are generally not aware of the federal transaction standards and their potential impact. Dr. Iezzoni invited participants to let the Committee know whether HIPAA is an issue on which they need help in educating their colleagues. Similar questions were discussed about the Balanced Budget Act.
On the need for partnerships to collect data, Mr. Hall pointed out that commercial managed care companies enter Medicaid to make money. If states put too many restrictions and regulations on health plans and physicians who participate in those programs, they will not do it.
The group discussed the Subcommittee's attempt to develop model language for contracts, which several panelists supported. Ms. Coltin noted that the lack of uniformity in contracts across the states is especially difficult for multi-state plans. The recommendation for a minimum and uniform data set across the states is sensible, since HCFA has not stepped into that void. Dr. Iezzoni said this suggestion might be part of their recommendations. She thanked the panelists for their participation.
Subcommittee Discussion
Dr. Iezzoni then invited Subcommittee members to share their thoughts about the themes emerging for their forthcoming report. Mary Moien will help write the final report. In addition, Sarah Rosenbaum and her colleagues at George Washington University are writing a report based on their investigation of contractual language, this will be synthesized into the Committee report. The Subcommittee needs to identify themes and areas they consider particularly important that might lead to recommendations.
Mr. Van Amburg observed that no one appears to be looking at population outcomes; rather, they're looking at plan outcomes. The former should be part of the focus--i.e., how to integrate data to get population-specific outcomes on such things as smoking prevention. No one seems to be looking at the Medicaid population as a whole. The issue of the analysis of data and feedback to both plans and to the population is critical, but little was said about that, either. As is typically the case, people put data in and it sits there, with nothing coming out. Asked if he thought plans or providers need to be given a quid pro quo if they are required to report data, he said they should at least be given access to the data so they can do their own analysis.
Expanding on that theme, Ms. Ward said the purpose of the data, still is not clear. The principle should be: “don't ask for data unless you know why,” --and delineate the purpose and standardization for these data. She said the report should encourage improvements, possibly outlining a set of principles underlying the desired changes. Citing Arizona's work on asthma data as a good example, she noted the plans' willingness to take time and energy to look at denominator population-based public health issues, may be something that will be dissipated if they are overwhelmed by requests for data.
Dr. Iezzoni raised the idea of a national standard for a minimal number of data elements, allowing for site-specific or state-specific overlays. Concurring, Mr. Van Amburg added that it should be oriented toward what the data collection system would allow the group to do, rather than specifying all of the individual elements. Dr. Detmer supported the idea of laying out principles and directions to aim toward, e.g., incentives for population health and special population needs.
Ms. Ward noted that the many demands for data are causing the fragmentation and disintegration of patient care. She recommended standardizing all government reimbursed care, so the same standards could be used for Medicaid and Medicare. The group then discussed the implications and specifics of the Balanced Budget Amendment.
Mr. Van Amburg noted that Medicaid managed care is going to have to come up with population measures because they are responsible for knowing about immunization status and youth tobacco use. This implies standards and principles of uniform data from all the different plans, as well as the Health Department and other safety net providers.
Dr. Amaro recommended that the clear articulation of questions should be a first step in any data collection process. Ms. Coltin suggested a framework for key questions about the Medicaid population as a whole, not just those in managed care. Then, asking about the managed care plan including the traditional sources of information for states, and how those sources are affected by managed care.
Members agreed on the need for linkage to public health programs and HCFA. The areas to cover include:
· what data are needed to answer key questions about the plans,
· which are needed at the aggregate level and which at the individual level,
· what the states are getting from the managed care plans,
· how the data vary from state to state,
· whether states are getting data through contractual provisions,
· the quality of the data, including completeness and accuracy,
· how the data are being used to answer key questions, and
· the impediments to getting, improving, and using the data.
Ms. Ward observed that this enormous effort will be a total waste of time if HCFA sees it as an audit by the Committee. It would be best to find ways to collaborate with HCFA.
Dr. Amaro mentioned the need for standardization of enrollee satisfaction data, some way of looking at denials for service, and ways of tracking complaints about health plans. Ms. Coltin suggested that Sarah Rosenbaum look at these issues. Dr. Iezzoni suggested also including patient-reported functioning and patient-reported experiences with their care. Dr. Amaro noted another theme, the need for support structures to fund the massive system changes necessary to implement the changes. Dr. Iezzoni saw as a theme the linkage of feasibility and cost to the utilization of the information. Mr. Van Amburg added another important category, race and ethnicity data of subpopulation groups unique to the plans.
Dr. Iezzoni suggested that Ms. Moien begin with a very broad outline of where the Subcommittee is going, using the camera metaphor to frame the levels of questions. The meeting then recessed for the day.
Nelda McCall, Laguna Research Associates
Ms. McCall has been involved in evaluating the State Medicaid Managed Care programs in several states since the early 1980s, concentrating on management information systems and their importance in a Medicaid managed care model from the standpoint of the evaluator.
She gave a brief overview of the AHCCCS program, which is composed of separate programs for long-term and acute care. The AHCCCS's management information approach is based on a prepaid management information system implemented in 1991. She enumerated the management information needs (MIS) needs for a plan like AHCCCS, with functions including procurement, monitoring, enrollment, capitation payments and data on service utilization. The prepaid MIS is composed of 11 subsystems. The information system (PMMIS) was designed as a relational data base, with much of the information available online. One of the best and most tangible benefits of the PMMIS is the ready access it provides AHCCCS staff to information about many aspects of their program.
Their evaluation looked at the collection and recording of utilization data. By the late 1980s, AHCCCS's senior management had also become convinced of its importance. AHCCCS has surpassed other states in its effort to collect and use utilization data. By the end of the evaluation in 1995, AHCCCS was using their data internally to support a number of functions including capitation rate setting, program budgeting, monitoring utilization and quality issues.
Ms. McCall identified several things any prepaid managed care information system needs to have in place:
· hardware and software to develop the special systems are important for the program in terms of quality assurance,
· capability with respect to utilization and AHCCCS monitoring,
· support for financial analysis and rate setting, and
· future planning for cost estimates of program modification requires data to satisfy external information requests.
She identified as critical the ability to monitor the contracting entities in managed care for financial solvency, thus providing an early warning of serious problems. The ability to secure competitive rates depends on the availability of data that can track costs and enable risk assessment and adjustments. She also stressed the importance of standardization across states, programs, and providers. She ended with the idea that giving up records of individual encounters seems to be a very dear price to pay for implementing capitated managed care; AHCCCS has demonstrated that it is not necessary to choose between the two.
John Murphy, Flinn Foundation
Mr. Murphy is the Executive Director of the Flinn Foundation, a private charitable trust or philanthropy in Phoenix. Previously, he was with the Robert Wood Johnson Foundation. Having moved to the Flinn Foundation just as the legislature adopted the AHCCCS program, he thought it would be useful to apply techniques developed by the Johnson Foundation (RWJ) to the Arizona low-income population prior to introduction of AHCCCS, and then to follow up with similar studies after its introduction.
They used the same survey methodologies in 1982, 1984, 1989, and 1995, with interviews of a random sample of 4200 persons. Telephone interviews were conducted in English and in Spanish, and low-income people without telephones were interviewed in person. In the initial survey, they looked at the positive and negative effects of county variations might be on each county, and at how the source of provider alters the frequency of utilization and ease of access.
The first results from the comparison of the 1982 and 1984 samples were very positive in terms of the AHCCCS program. Among the highlights of the 1989 findings (validated in the 1995 study) was that AHCCCS is serving its intended population, and that most eligible persons were enrolled. Most AHCCCS enrollees obtained their health care from a doctor's office, or a free-standing clinic. Fewer used the hospital emergency room as their primary care source. Many individuals who previously had used public health facilities were seen by mainstream physicians. And, finally, the satisfaction of AHCCCS enrollees with their health and the care was comparable to persons interviewed who had fee-for-service or HMO coverage. In closing, Mr. Murphy indicated that the foundation plans to continue such studies periodically because of their value in providing helpful information.
Rep. Susan Gerard, Arizona Legislature
Rep. Gerard has been in the state legislature 10 years, and on its Health Committee for nine of those years. She praised the good job that AHCCCS has done in developing, through a lot of trial and error, very good management information systems and methods for tracking billing and costs. They have also become very good at managing costs. She indicated that they are not yet doing a good job on data collection for managed care. She sees this as the next “evolutionary step." She faulted satisfaction surveys for not asking if the patient got better or was able to return to work.
She indicated that they have not integrated their public health goals with their Title 19 health delivery systems. She wondered whether anybody at the Medicaid office ever examined the Healthy People 2000 goals, and she recommended that these be integrated into health plan contract requirements to get the plans to report data on immunizations, prenatal care, and low birth weight babies. Dr. Iezzoni expressed agreement with those remarks.
Rep. Gerard then said that managed care organizations need to integrate their different departments so savings in one area does not mean increased costs in another. She thought it might be valuable to compare private and public systems. She noted that, in a public system, there is an opportunity to collect the kind of data that can't be obtained from private companies. Behavioral health data collection lags far behind acute care. The Department of Health oversees the Title 19 behavioral health, not the AHCCCS program. She pointed out that even when evaluation and outcome information are required of the providers or the regional behavioral health authorities (RBHA), the Department of Health enforced compliance with their contracts. A current crisis in Maricopa County has created an opportunity to move all the stakeholders toward needed changes.
Discussion
Dr. Amaro also expressed agreement with Ms. Gerard's comments and noted the possible role of the legislature to inform the public about allocations of resources. She endorsed the idea of integrating data "across the boxes," and especially the behavioral health care box, and asked for Ms. Gerard's thoughts on how legislators can be brought to better understanding of the need for allocating dollars for information systems. Ms. Gerard replied that the problem is that they (legislators) have had so many major failures with information systems. Moreover, state government staff have allowed contractors get away with things that the private sector would not tolerate. She suggested that the government might want to look at what the private, for-profits are doing since they are way ahead of the public sector. Although the chances of getting money for computerization is better now in a good economy than it was in the late 1980s, it is still tough, considering the track record, to give five or ten million dollars to do something that seems never to get it right.
Mr. Van Amburg commented on the somewhat contradictory testimony on the AHCCCS data system, particularly concerning the ability to use the system to extract useful information. Asked for her perspective on this question, she said that "when we don't get the data we want, we put it in a statute" or "turn the heat up" on ACCCS, which in turn puts pressure on the health plans.
Asked what kind of information she considered most useful, she talked of the need to monitor the health plans in terms of changes in costs and/or enrollment. Regarding the outcomes she was specifically interested in, Ms. Gerard mentioned visits to emergency rooms, hospitalizations, and restrictive formularies and how these affects outcomes.
Ms. Coltin noted that pharmacy costs are the fastest growing cost component in medical care, and asked whether the legislature, or any legislators, plan to try to improve data collection on this important cost item. In response to a question, Ms. McCall said AHCCCS does not collect pharmacy data for the acute care plans, a decision that goes back to the early 1980s when AHCCCS was having a lot of trouble collecting data. Dr. Iezzoni noted that many other Medicaid programs do collect that kind of data.
Dr. Detmer expressed delight with Ms. Gerard's ideas about moving toward a value- based health care system, and asked how well she thought the state legislature would receive preemptive federal laws on confidentiality and privacy. She said, in effect, that federal measures that "[give] the state the impression of a little bit of wiggle room" are more likely to be well received.
Joe Gaudio, Arizona Physicians IPA (APIPA)
Mr. Gaudio is chief financial officer of one of the AHCCCS health plans, Arizona Physicians IPA, covering over 135,000 lives. Their focus is young mothers and their children. His remarks focused on data utilization in Medicaid managed care related to reporting on budget, finance, encounter, bid preparation, and claims liability. He also touched on the data they collect on quality of care and risk management, primarily on young or pregnant women and young mothers.
Speaking about claims liability and bid preparations, he mentioned that several plans filed for bankruptcy in the early 1980s because of their inability to identify and estimate their total claims liability reserve, partly due to an inadequate information system. He highlighted inpatient costs and pharmacy costs as of particular concern.
APIPA uses a lag methodology to estimate their liability. Their information system requirements are tremendous and they use data from their prior authorization system. On deliveries, they track every pregnant mother who comes onto their system, from their first prenatal care visit through delivery outcome. The "lifeblood" of their work is bid preparation, which also places a great demand on information systems. From the quality of care perspective, he said their plan is "100 percent compliant with HEDIS 3.0 reporting requirements." Their quality of care and risk management system tracks provider and member grievances, and those data are reported back to the state on a quarterly basis. Additionally, they manage an extensive physician profiling system. From their complaint tracking system, compilations are made on a weekly basis and reported to the service improvement committee, and to the quality management committee when warranted. In closing, he mentioned the IPA's extensive credentialing program which includes visiting physician officers for chart reviews.
Brian Lensch, Medicaid long-term-care program
Mr. Lensch has worked in the area of developmental disabilities with the Medicaid program, providing services both to the Medicaid and non-Medicaid population. They are capitated program contractors with a variety of relationships with different plans, providing both acute and long-term care. Most of his experience has been in the program side of the information system.
Seconding Ms. Trombino's emphasis on knowing the purpose for data collection and use; he noted the need to balance data gathering with Medicaid's mission. Even good intentions for gathering data raise barriers to doing a good job. For example, the way the provider information system is set up for AHCCCS tends to limit the network they can use.
In respect to the problem of multiple requests for the same information, Mr. Lensch said his office is trying to implement a program called Targeted Case Management. One cause of problems is the multiple sources for demographic data. The problem is compounded because clients are asked for information by many SSI or welfare case workers. Families are constantly asking why the government can't just ask for this information once. Another problem is that much of the data they're required to gather and update annually, particularly demographics, change often and people forget to mention those changes. Thus, data systems are not always accurate after the initial entry.
Echoing other speakers, Mr. Lensch cited the need for a link between the public health systems, the Arizona Physicians's IPA, and their system, e.g. immunization. In the absence of these links, the best intentions in data collection and analysis are "not going to get us exactly where we want to be."
Discussion
Responding to Mr. Van Amburg's questions, Mr. Gaudio said they submit encounter data to AHCCCS within 30 days of paying the claim. It’s bids include program information as evidence of quality of care -- for example, the Young Women's Case Management Program. In addition, they have created a complaint tracking system. He reiterated that the complaints are summarized weekly for their service improvement committee, which determines whether they warrant further processing into the quality management committee.
Dr. Iezzoni praised Mr. Gaudio's company for the marvelous job it seems to be doing. Referring to the many small plans in Arizona so crucially important to serving primarily rural communities, she wondered how these small plans manage to monitor quality. She also praised Mr. Murphy's foundation for their public service and for doing their own survey, and asked how that work related to the plans' own satisfaction surveys. He replied that the foundation's latest survey asked questions that are more specific than those asked by the plans.
The issue of consistency in reporting was noted, and Ms. McCall recalled that her evaluation criticized AHCCCS about nonstandardized reporting with respect to grievances and appeals. She added that AHCCCS has done their own satisfaction studies. Ms. Ward questioned the value of data on grievances and complaints if the sates don't know what to do with the data. Mr. Gaudio pointed out the importance of the type of complaint and the broader question of whether the population is being mainstreamed. He said his plan reports the status and action plan on every grievance, which he affirmed is well worth their trouble.
Dr. Amaro wondered if the panelists had a way of tracking the quality of their services to mothers as a way of looking at outcomes for children or factors that might contribute to those outcomes. Mr. Gaudio said they do track members referred to a behavioral health organization or support group, and they have protocols to ensure that a referral gets to the appropriate counseling center. He noted that incidence of low birth weight has dropped since they developed interventions. They have attributed the significant positive outcomes directly to the Young Women's Case Management Program.
Returning to yesterday's concern about multiple layers of reporting, Mr. Gaudio said that he recommends that reporting come through the plans. Plans can capture the data and present it to the state in a useful form. He added that he'd rather have billing data costs funded submitted to the plan rather than the state, letting the state focus on other areas.
Asked about the lessons from the initial experience with 50 percent cost overruns on information systems, Ms. McCall recommended that some level of cost overrun be budgeted and that people simply expect things to "cost more than you think."
Dr. Iezzoni asked Mr. Gaudio whether and how his company is planning to implement HIPAA standards. His response was that he could not comment on the specific details of the standards, however, his plan would follow the lead of AHCCCS and do whatever is necessary because they understand the importance of data reporting and standardization. He added that in their last RFP, AHCCCS made reference to the fact that they will be developing standards for electronic reporting.
The group discussed the new state immunization registry and the problems stemming from the unwillingness of illegal aliens to be registered for fear of being deported. Mr. Gaudio speculated that low immunization rates are not a utilization issue, but a reporting problem. Mr. Lensch thought the registry would not resolve the issues of access by families who are in the country illegally.
The morning session adjourned for lunch and reconvened at the Arizona Health Care Cost Containment System (AHCCCS). For the afternoon session, all presenters were members of AHCCCS staff.
Kari Price and Jan Hart
Ms. Price is the health plan financial manager in the Office of Managed Care, AHCCCS. Ms. Hart is the ALTCS (long-term care) manager in the same office. Ms. Price started by describing briefly the kind of data they require from health plan contractors. All divisions of AHCCCS collect information throughout the year from health plans and program contractors. She gave a brief overview of their financial reporting requirements as well as the collected encounter information. They also receive quarterly grievance reports; collect quality indicators, as well as information about fraud and abuse. They have claims information for their small fee-for-service population, and collect encounter information on everybody assigned to a managed care plan. In response to a question, Ms. Price said there are differences in the quality of data they get from contractors.
Noting that AHCCCS does not perform eligibility reviews for the acute care program, Dr. Iezzoni asked if they collect information independently on race, ethnicity, and other sociodemographic variables. Ms. Price said they rely on the agency that takes the information, and collects no additional demographic information. Ms. Hart explained that AHCCCS determines eligibility for ALTCS. She did not think they gathered information about primary language.
Ms. Ward asked what they would recommend to other states from what they have learned, and what they like and don't like about their information system. Ms. Price felt that their system is excellent and they have technical people who can pull the information for them. Ms. Hart acknowledged one weakness: that individuals can't go in and query the system. It can be a lengthy process to put in a systems request to the IS department, i.e. may take months for response. She thought it would be useful to have a system capable of downloading data so people could run queries independently of the IS department.
Asked about evaluating pharmacy costs, Ms. Price said that they haven't figured out a way to handle that much information. Dr. Iezzoni inquired if they are thinking about developing a risk adjustment method for setting capitation rates. Ms. Price indicated they already use age and gender-specific. She did not think the health plans would want to be more specific because of the SSI rating category.
Ms. Price responded negatively when asked if they use the data to calculate specific rates for pregnant women who might need intensive services, e.g. substance abuse treatment or mental health. She indicated that there is one rate, regardless of risk profile. There was a brief discussion of how they set capitation rates. They develop an actuarially sound rate range, and bids have to fall in a range that which was developed from encounters and financial information. Ms. Hart said the ALTCS program is very similar. For their developmentally disabled population, they have one contractor who covers the entire state with a statewide capitation rate.
The long-term care program has a capitation rate specific to each county, because of population variation on county-to-county basis. For example, Cochise County has a high Hispanic population with extended families that "take care of their own." Many people at very high risk stay in their own homes. Dr. Amaro asked if they could use their data to develop different approaches to capitation in order to make taking care of those populations doable and attractive to providers. Ms. Hart replied "no"; that the capitation rate reflects these variations because encounter and financial data is used to develop capitation rates. Often health plans will actually alter their capitation rate to physicians with more severely ill patients. Ms. Price pointed out that they are assigned their populations. Ms. Hart added that several of the health plans have maternal child health case managers to help address high-risk pregnant members. AHCCCS has a re-insurance program for acute care and long-term care; its stop-loss program.
Ms. Coltin asked about incentives for managed care plans to become involved with the AHCCCS program, and why some choose not to do so. Ms. Price said a couple of commercial plans became involved several years ago and struggled for a couple of years, but one got out last year. "They didn't have an understanding of what the AHCCCS population was about." Ms. Hart thought the problems stem from several factors: the population and types of services they needed, as well as the state's administrative and reporting requirements. She noted that commercial plans have to satisfy their stockholders, "and the bottom line that's acceptable to the State may not be acceptable to the stockholder."
In response to a question about the impact of the welfare reform law, they suggested contacting Diane Ross, assistant director for the Division of Member Services, who has been assigned to study that subject.
Referring to the printed questions, Ms. Hart identified a gap in data collection related to dually eligible enrolled members. Their office is unable to get data about the Medicare population, and especially those in the Medicare risk plan. When dual members are on a fee- for-service basis through Medicare, nothing is known about what services are provided or what costs are incurred.
Asked for more information about how they use data to derive capitated rates for behavioral care, Ms. Price explained that they had calculated it on a per-user basis. Now they base it on the entire population, so the rate went down significantly. Contractors are paid based on the utilization information reported by Regional Behavioral Health Authorities (RBHA). Ms. Price promised to send more information on the behavioral health carve-out from their policy manual and/or contracts. For most age groups, persons enrolled in the acute care plan who need mental health services are referred to the RBHA, which assesses them and provides the services. Patients can also self-refer. Long-term care members and families have case managers, who would do assessment and referral.
On the subject of quality, Ms. Hart commented that they do thorough reviews of each health plan. These provide opportunities to know programs and help them judge the validity of complaints. She considers this amount of oversight unique. The use of case managers in long-term care provides administrative oversight that lends to the safety and quality of that program.
Dr. Iezzoni speculated based on the morning testimony that some of the big plans might be trying to crowd out some of the small plans, and she asked if that might be happening. She noted that the small plans would be unable to spend what the large ones do on information ($3 per member per month in the case of APIPA). Ms. Hart acknowledged this reality, adding that they don't want that to happen, since they want members to have a choice. Ms. Price noted that the state tries to limit administrative requirements so they're manageable. Asked further how they are trying to make sure that plans can remain competitive, Ms. Price said they have no specific policy but they want to have at least two plans in the majority of their counties. In their large urban counties, they now have four or six. She reminded the Subcommittee of the reinsurance (stop-loss) provision. Ms. Hart added that in the long-term care program, five counties now are contractors (service providers) due to legislative action. The contract for years 2000/2001 will be a competitive bid for long-term care. She said they always ask the plans for recommendations for the RFP.
Dr. Iezzoni asked about the involvement of AHCCCS with the Native American population in Arizona. Ms. Hart said the state had case management agreements with 8 of the state's 21 tribes. The tribes have a choice to join in AHCCCS or the Indian Health Service, both for acute and long-term care. Neither knew whether tribal members were choosing AHCCCS or IHS.
Asked to whom they give information from their audits and reviews, Ms. Hart said reviews for both acute and long-term care go back to the health plan or program contractor. In both programs, they send a draft report and the plan has an opportunity to comment. In the event of deficiencies, they require a corrective action plan. They also forward the information to HCFA.
Noting that the Subcommittee had come to Arizona to learn first-hand what to recommend to the rest of the country, Dr. Iezzoni asked the speakers what they thought should be replicated elsewhere. Ms. Hart's advice was to be very specific at the outset about what should be reported and the methodology. Ms. Price stressed the importance of defining the requirements, ensuring consistent reporting, and doing insuring compliance.
The speakers noted that other states might find their "readiness assessment tool" (RAT) useful: After a plan has bid and been accepted pending a review, they go on-site to see if the plan has a claims system and other essentials in place. State people are also available to give technical assistance as needed. Asked about sharing data among plans for the benefit of the less developed ones, Ms. Price said they do not do this, but during bid years they do produce a data book that has utilization information by county.
Ms. Moien asked about the health plan that opted to drop out after three years, and whether the health care of the enrollees suffered. Ms. Price said that quite the opposite happened: the plan was not controlling its utilization, and its members got more care than the average. If members had not been getting adequate care, it would have been brought to their attention fairly quickly through their grievance process.
Terri Keagle
Ms. Keagle is research administrator of the Office of Managed Care. She focused her remarks on encounter data collection and validation and the use of this information. They have collected encounters since the beginning as managed care under an 1115 waiver. In addition to reporting to HCFA, they provide feedback to the health plans, each of which is measured against quality indicator and utilization reports. (They have three programs: acute, long-term care, and behavioral health. She uses the term "health plan" for all of them.)
Ms. Keagle said HCFA has permitted them not to collect pharmacy encounters from their acute care plans. Dr. Iezzoni noted that they may be losing a lot of valuable information by not having the pharmacy claims, and she asked if this is an issue they are re-exploring. Ms. Keagle said the next group of panelists are lobbying hard to have them collect that information. Such a change would require internal and external support as well as funding from the legislature. They do receive data on pharmacy claims from some of the plans and have a system for processing the information. It would take time to make it consistent and storage costs are a major issue. The group discussed this issue at some length.
Moving to the HCFA requirement that they provide technical assistance to their health plans, Ms. Keagle described the staffing in detail. They have six staff positions dedicated to facilitating the processing of encounters and providing technical assistance. She then described the various sanctions imposed by the Office of Managed Care in regard to both processing and validation. Requirements are outlined in the request for proposal and contracts. There also are annual reviews of the operational and financial aspects of the plans.
Asked about their requirement that data be submitted within 240 days, Ms. Keagle said that about 90 percent of the encounters they receive come within the first 6 months, but they allowed more time so that plans can submit "clean encounters." She agreed with Dr. Iezzoni that for those using the data for oversight and quality management, this can be like "chasing a train that's way down the track."
Dr. Iezzoni asked if Arizona requires hospitals to submit discharge abstract data to a state health data repository or central data authority, but the panelists were uncertain whether the Office of Vital Statistics or some other entity gets those reports. Dr. Iezzoni expressed surprise that AHCCCS was not tapping into that data source She noted that Massachusetts mandates that it be reported to the state, and it is done very quickly. Ms. Coltin noted that one problem is that they cannot identify the Medicaid members through the health plan. Ms. Keagle added that the hospitals in Arizona do submit data to the state Department of Health Services, but not in a timely manner.
Ms. Keagle then described in detail the procedures for collection and processing of encounters. They define what they want from the plans and how they want the data presented, as set forth in a manual they provide. The health plans submit the data on cartridge or reel tape, but AHCCCS wants it submitted electronically. They now have EDI for fee-for-service claims, and hope to automate the rest later this year. Asked if they are taking into consideration the HIPAA standard, Mr. Ratterree said they're planning to follow a national standard format.
Continuing her discussion, Ms. Keagle explained that once each record passes through edits and audits, it is finally approved, adjudicated, and put into their data base; and once that cycle is run, they process tapes to go back to the health plans, one with the encounters approved and placed in the AHCCCS data base, and the other with encounters they didn't pass ("pending"). They process about a million encounters a month, and approximately 4 percent of the encounters they process are "pended", of which about 8 percent take more than 100 days to resolve and thus accrue sanctions. Mr. Ratterree estimated that it takes five to seven days to get the information back to the managed care organizations.
Brent Ratterree
Mr. Ratterree described the encounter validation methodologies. They include comparison of medical records (used primarily for professional services), file comparisons, and comparison of independent report on newborn births. For procedure codes, they use CPT primarily. Hospital-based procedures are ICD-9-CM-coded. They have some state-only codes for behavioral health. Asked about enrolling newborns, he described a complicated process involving phone calls immediately upon birth, as well as other forms of reporting.
For the validation studies, they use a random sample using an error-based sampling methodology. The sampling frame is the eligibility file, so they sample people who may have had no encounter through the system. Their sanction policy allows a 5 percent error rate, all the errors are weighted the same, and the plan is responsible for the provider's errors (which is where most of the errors are found). Asked if any of the patients ever question confidentiality of records, Mr. Ratterree said that has not been an issue. They do some real-time data quality evaluations to check whether they are getting the data they should be getting.
As for recommendations, the panelists said state programs should make sure when they start collecting information that they know what it should look like, how to monitor it, and that they're getting complete and accurate data as soon as possible. Mr. Ratterree pointed out that managed care organizations appreciate learning early on about errors which later would have subjected them to sanctions. Dr. Iezzoni suggested that the panelists look at HCFA's evaluation and management coding guidelines.
Ms. Coltin asked if there are changes in the error rates as a result of feeding information back. In addition, she asked if the plans are passing along the information to providers. Mr. Ratterree said that some of the plans pass on sanctions to the providers; if a provider is consistently high in errors, it becomes a contract issue for the health plan. Generally, the error rates do decline over time.
Mr. Ratterree said they plan to expand their medical record methodology as they move towards an EDI process for encounters. He gave several examples of how they use encounter data:
· as a base for their capitation fee-for-service rates,
· in calculations for their reinsurance program,
·for regulatory oversight,
· for utilization reports,
· to monitor fraud and abuse, and
· as a basis for validation.
They've learned to keep the communication channels open with contractors, plans, and managed care organizations.
Asked if they believed that their sanctioning policies have had a beneficial effect, Mr. Ratterree said some cases require both the carrot and the stick: their carrots include providing better information that contractors can use in the next round of bids. The Subcommittee was supplied with AHCCCS's manuals for data management and encounters.
Susan Cypert and Alan Schafer
Ms. Cypert distributed a copy of the Quality Management Program quarterly report prepared for HCFA, which contains all of their performance indicators. She described how AHCCCS has used encounter data to define performance measures, initially as part of their waiver continuation as a demonstration project that began in 1994. The first step was to identify indicators for which valid encounter data were available, initially using indicators modeled after HEDIS. They produced baseline indicators for such factors as well-child visits, low birth weight, and mammography screening.
The provision allowing plans to submit their encounter data in up to 240 days works against getting quality indicators, as it doesn't allow the health plans to make timely interventions to improve their performance levels. As a result, chart audits have been used. A major problem for AHCCCS is getting access to demographic and other data collected by the state Department of Health Services, because of federal confidentiality restrictions. Sister agencies are very reluctant to share information, even without the names of individual members. For example, getting aggregate lab data would require having the health plan ask individual providers to authorize the release of that data, and then re-aggregating the data -- a process they find unworkable. Plans are to seek state legislation to resolve this.
For pharmacy encounters, they are forced to have the health plans report that information rather than take it from their encounters. Another difficulty is that there are no regional or national benchmarks for them to measure themselves against. In regard to the stability of the denominator, Mr. Van Amburg asked how frequently they recertify eligibility. Mr. Schafer indicated some are recertified in six months, but mostly on an annual basis. One of the advantages in defining indicators is that AHCCCS has locked-in enrollments.
Responding to a question, Ms. Cypert indicated that ultimately they would use the immunization registry for certain data. They have some concerns about how slow it is in getting off the ground, but DHS is going to add incentives to have providers report immunization information to DHS. She doubted that there would be confidentiality issues with that information on members.
One of the things that AHCCCS does differently from HEDIS is they validate their own data. On immunizations, they report the state aggregate, and then the health plan's specific data are given to the legislature and included in their RFP and contract requirements, along with an indication of needed improvements. In their last RFP cycle, they did not require some of the health plans that had high immunization rates to meet some of the requirements in the RFP, as a carrot.
Ms. Cypert mentioned another limitation that pertains to prenatal care. Because the encounter systems were originally structured for payment, they do not provide information on the number of visits a woman was getting or when she initiated prenatal care. Now, after three years, they are developing prenatal indicators within the context of the encounter file format.
To a question about grievance reports and consumer satisfaction, Mr. Schafer said that both their acute care and long-term care units have their own tracking system for showing trends in quality of care complaints. In addition, AHCCCS has done a member satisfaction survey, and the grievance and appeals unit does operational reviews on a monthly or quarterly basis that are handled at the AHCCCS administration level.
Ms. Coltin asked why AHCCCS has been unable to get benchmarks on some of the indicators, and Ms. Cypert cited Arizona's long history of dealing with encounters and the fact that other states may not have locked-in enrollments. Thus it is not always possible to compare Arizona with other states. Asked what kind of benchmarks they'd like to have, Ms. Cypert replied that the Healthy People 2000 goals seem to be the most commonly accepted and Arizona has used some of these goals. Immunization rates from Healthy People are the nearest they get to having a benchmark.
Mr. Schafer described their varied efforts to get data for quality assessment. They're just starting the second year of collecting data on long-term indicators, and don't know if they are valid.
Mr. Van Amburg asked if they have given thought to comparing to state data and risk- adjusting it. Ms. Cypert indicated although they initially shied away from risk adjusting, they are finding that they may need to address this. However, they have not figured out how to do that, given the limitations of data sharing and the inability to get demographic information. They are doing better on the ALTCS side in getting demographic data because everyone has a pre-admission screening.
Dr. Iezzoni asked if they rely on the minimum data sets that HCFA promulgates for nursing home residents. Mr. Schafer said that when they are automated in June the first thing they want is standardized data for home health agencies, but it will be limited because of the rate of growth of home-based care. Asked about new government activities around home-based health for the Medicare program, Mr. Schafer implied they have not seen enough of OASIS to know whether it would be useful. He does not think the SF36 is helpful for an at-risk population, point on which Dr. Iezzoni concurred.
On the first member satisfaction survey, Ms. Cypert said they first focused on satisfaction with prenatal care, to test the level of effort needed for a member satisfaction survey. Questions were tested on focus groups. The second, larger member survey proved to be exceedingly expensive. Though the results were quite good, they did run into some significant issues in gathering data.
Mr. Schafer said that on the ALTCS side, the plans' member satisfaction surveys are somewhat "simplistic." In contrast, AHCCCS does annual member surveys during these case management service reviews, using random samples on each of the program contractors, and asks some basic questions. They are finding satisfaction levels above 90 percent. Ms. Cypert mentioned they are now embarking on a provider satisfaction survey, and added that the MDS will help with the long-term care population.
Ms. Coltin mentioned the Consumer Assessment of Health Plan Survey developed by AHCPR for Medicaid, which might be helpful in allowing them to compare themselves with other states, even though there are many differences in how states use that standardized survey. She believed the version for the Medicaid population is in Spanish as well as English.
Asked if the legislature is giving adequate funding, Mr. Schafer said that his unit has adequate staffing, and the state Medicaid agencies work very closely with the managed care organizations. These close relationships to effect change when necessary are the key to the whole process. He questioned how other states can effectively implement Medicaid managed care without adequate staff. The panelists discussed the size and location of the state staff. The Information Services has roughly 200 staff, all in Phoenix. Staff in the Office of the Medical Director do a lot of traveling to perform audits, assess quality of care, and provide technical assistance. The Division of Member Services has 500 in administration, as well as in 13 to 15 local offices throughout the state. Mr. Van Amburg commented that other states are considering going from fee-for-service to managed care to reduce not only medical costs but personnel; but it appears that reductions in staffing will not occur if they're doing it right.
Ms. Cypert said that they use their PRO organization to do concurrent review and utilization management and for specific audits. They also use them for quality of care issues where they have peer review concerns.
In response to a question, Mr. Schafer said they are building a longitudinal record for each of their clients. He was uncertain about the process for allowing patients to access and correct their records. To another question, Ms. Cypert said they have a specific AHCCCS identifier, which is not the individual's Social Security number. Mr. Schafer added that the ALTCS program has a more specific tracking system than the acute care side.
Dr. Iezzoni thanked the panelists for their extremely helpful information and for sharing their time with the Subcommittee. She concluded the proceedings at 3:47 p.m.
I hereby certify that, to the best of my knowledge, the
foregoing summary of minutes is accurate and complete.
/s/ Lisa I. Iezzoni May 27, 1998
Lisa I. Iezzoni, Chair Date