Wyndham City Center Hotel
Washington, D.C.
The National Committee on Vital and Health Statistics (NCVHS) was convened on June 26-27, 2002 at the Wyndham City Center Hotel in Washington, D.C. The meeting was open to the public. Present:
UPDATE FROM THE DEPARTMENT: James Scanlon, ASPE
Mr. Scanlon briefed the group on HHS bioterrorism grants to states, the National Academy of Sciences study on race/ethnicity data, HHS responses to the “Shelby 2” Statute,” the HHS strategic plan, HHS responses to NCVHS reports and recommendations, and the HHS gateway to its datasets. He observed that while the bioterrorism grants could help create the market for the information systems and standards with which NCVHS concerns itself, they could also create havoc absent a framework to assure interoperability. CDC and HRSA have both issued guidances, and the Data Council will explore these issues in July. “Shelby 2” requires all federal agencies to publish guidelines for ensuring the quality of the information they disseminate to the public. This could have a big impact on data policy in all federal agencies. The HHS guidelines are posted on its Web site; the comment period ends in July and final guidelines go into effect in October.
CONSOLIDATED HEALTH INFORMATICS (CHI): Jared Adair, CMS
The Office of Management and Budget asked HHS to take the lead in pulling together work in the federal government on health care informatics and in developing the business case for consolidation. The resulting project, CHI, is in the office of the Assistant Secretary for Budget, Technology and Finance, and Ms. Adair is the project manager. The adoption of government-wide interoperability standards, including vocabulary, that would be built into each department’s future architecture is being proposed. Ms. Adair asked the Committee for its help on the CHI process, to which Dr. Lumpkin and others responded affirmatively.
Ms. Trudel briefed the Committee on the status of HIPAA standards. Since the February meeting, HHS has published two proposed rules (modifications to the initial transactions and code sets) and a final rule (the employer identifier). The Department is now fleshing out what the final rules will look like, and CMS is spending a lot of time on outreach and education.
In the discussion period, members expressed strong interest in cooperating on the CHI project and helping it succeed. Mr. Scanlon stressed that this is not a regulatory approach but one of leadership and example, as others have recommended for the federal government. The project leaders were urged to keep the NHII recommendations and perspective in mind. Dr. Lumpkin suggested that the next NCVHS meeting include time to talk about what is going on with the E-Health Initiative, the Markle Foundation Initiative, and AHA initiatives as well as legislative initiatives, and how they and the CHI project fit together.
UPDATE ON PRIVACY REGULATIONS: Stephanie Kaminsky, OCR
Ms. Kaminski reported that OCR staff members have been working very hard since the comment period closed to digest and analyze the more than 11,000 comments received in this round. They have a rigorous schedule to get the rule published as soon as possible so that the compliance date for the modifications converges with the April 14, 2003 compliance date of the final privacy rule.
DATA ISSUES IN MEASURING QUALITY OF MENTAL HEALTH AND SUBSTANCE ABUSE SERVICES
Ms. Coltin, Chair of the Workgroup on Quality, said this is the latest of a series of panels on data issues in measuring quality of care. Their testimonies will inform the report and recommendations on data gaps and data issues being developed by the Quality Workgroup.
Dr. Goplerud provided an overview of the topic. He described activities that include TOPS 1 and TOPS 2, a common performance measurement framework developed by the national organizations representing state mental health, substance abuse and Medicaid directors, a minimum data set developed by CSAT, and Carter Center Forum on Performance Measures. He stressed that prevention and mental health promotion are often-neglected but important aspects for which performance metrics are also necessary. In general, a strategic approach is needed to address these issues, because current approaches are fragmented and piecemeal.
Dr. Chalk described a project by the Washington Circle Group, a group of substance abuse experts, to develop measures for processes of care for alcohol and drug treatment that have been found to be closely related to outcomes. The goal is to use measures that capture the range of necessary treatment services based on a model of addiction as an often chronic, relapsing condition. After describing some of the major barriers to process measurement, Dr. Chalk concluded that evidence-based clinical guidelines are a central component of helping quality measurement move ahead.
Mr. Renner described NCQA’s evaluation programs with respect to behavioral health, which focus on HMOs and managed behavioral health organizations (MBHOs). NCQA has found that the major obstacles to measuring effectiveness in behavioral health relate to the fragmentation of the system, confidentiality, the inadequacy of coding, data availability, and cost issues. Fragmentation makes it difficult to design measures that span the process.
Dr. Horgan’s presentation concerned the performance measurement work of the Washington Circle Group and its experience with the challenges of using administrative data. The Group has identified seven measures covering the continuum of care starting with the population-based activities of prevention and education and moving on to recognition, treatment, and maintenance. The barriers to performance measurement stem from organizational structure and patient and provider behavior.
Dr. Eisen focused on the challenges and issues with quality measures derived from patient-reported survey data. She discussed the major protocols for collecting patient reported survey data and the limitations of each. Privacy concerns and resource constraints are major barriers to obtaining survey data from consumers in treatment settings.
Ms. Greenberg said that in preparing to implement HIPAA regulations, AMBHA members are identifying gaps in revenue codes and HCPCS codes related to behavioral health. Together with the national associations of state mental health and substance abuse directors and others, they have submitted recommendations for HCPCS codes and HIPAA regulations. AMBHA is concerned that providers will use the HIPAA privacy rules to defend a refusal to supply performance data, and that such difficulties will increase the animosity between providers and MCOs.
Ms. Lumpkin said that the behavioral health industry recognizes the need to measure the same areas measured by JCAHO, and she outlined some of the challenges faced in doing so. She raised a number of concerns of AMBHA and Magellan, and noted the enormous cost of changing systems.
Dr. Ganju reviewed the needs for data on quality and the status of quality initiatives in the public mental health sector. He pointed out that a significant proportion of spending on mental health is by states, and this together with spending by Medicaid and Medicare creates an ethos for quality initiatives that is different from that of other health sectors. While quality initiatives have come a long way in mental health in recent years, much progress is still needed to embed quality into data systems. One thing that is needed is standardized definitions, which can be addressed through
HIPAA coding. Data fragmentation and duplication across systems make it impossible to count the total number of people in treatment. It is also difficult to count the number needing services. The public mental health quality agenda has three prongs, each with its own strengths and weaknesses: performance measurement, implementation of evidence-based practices, and systems-level quality management and quality improvement. People are working on new models in these areas. In general, a strategic approach is needed.
During the discussion period, Committee members and staff commented on difficulties in measuring unmet need; the participation of mental health and substance abuse professionals in standards development; the relevance of the International Classification of Functioning, Disability and Health; and the importance of respecting patients’ concerns about privacy and confidentiality in designing quality measurement tools in mental health and substance abuse.
HEALTH STATISTICS FOR THE 21ST CENTURY
Dr. Friedman presented the draft report developed by the Workgroup on 21st century Health Statistics, NCHS, and the Data Council, and NCVHS recommendations drafted by the Workgroup. The report and recommendations are the culmination of a multi-year consultative process that has already produced an interim report and several presentations. The four core recommendations are for a reconstituted NCHS to perform “hub” functions, an autonomous health statistics planning board, a single state agency in each state to coordinate state activities, and expanded health statistics training. There will be an executive summary as well as a separate summary aimed at policy makers.
In their discussion, members focused on the proposed role of NCHS and how it relates to health statistics activities elsewhere in HHS. They also discussed strengthening the statements about confidentiality in a few places, clarifying the relationship to the NHII, the content of the recommendations about training, and the overall purpose of the report. Dr. Friedman explained that the purpose of the project is to provide a consensus framework for people engaging in health statistics and those who use health statistics, for use in more detailed planning and implementation. The report will be sent to the Secretary, and it is also intended for use by other levels and sectors of government and in private and research sectors.
Based on this discussion in the first day of the meeting, the Workgroup on 21st Century Health Statistics revised the report and recommendations and presented the revised version to the Committee on day two. The final decisions are outlined under “Actions” above, #3.
DRAFT RECOMMENDATIONS FOR CODE SETS
On day one of this meeting, Dr. Cohn reported that the Subcommittee on Standards and Security is drafting a letter in response to two notices of proposed rulemaking on electronic transactions. On day two, Dr. Cohn presented the letter, which he read aloud, on the NPRMs (CMS-0003P and CMS-0005P). A motion was passed unanimously approving the letter, which will be sent to the Secretary.
On day one, Dr. Zubeldia introduced a draft letter on code sets addressing the migration from ICD-9-CM to ICD-10-CM and ICD-10-PCS. It was developed on the basis of two hearings on the subject. Following discussion by the Committee, the letter was referred back to the Subcommittee, which decided to present a revised letter to the Committee in September.
GAO REPORT ON RECORD LINKAGE AND PRIVACY
Judy Droitcour and Nancy Donovan, GAO
The speakers presented a GAO report on record linkage and privacy that focuses on record linkage involving person-specific data conducted under federal auspices to generate research and statistical information. GAO initiated this R&D study because the issue is under discussion and a matter of concern in many arenas and a broad overview of the issue was needed to inform policy makers. The intent of the report is to inform and stimulate discussion by providing examples, not to provide a comprehensive survey of what is going on. The authors discussed the tradeoffs and combinations of information-gain and privacy priorities, and the goal of developing strategies that make it possible to achieve both. They then presented their schema for categorizing and analyzing various data sources and linkages, with examples in each category. The privacy issues related to record linkage are in the areas of consent, data sharing, re-identification, and sensitive information. The other major topics of the report are the privacy protection toolbox and data stewardship strategies.
NCVHS FIFTH ANNUAL REPORT TO CONGRESS
A motion was passed unanimously to extend the time period covered by the report to the end of June, to approve the current draft of report, and to authorize the Chair to approve the final version.
REPORTS FROM SUBCOMMITTEES AND WORKGROUPS
Most of the key elements covered in Subcommittee and Workgroup reports are reported in the Actions section and other parts of the executive summary, above. In addition, both the Subcommittee on Populations and the Subcommittee on Privacy and Confidentiality reported that they have hearings planned for the remainder of the year. Their findings will be conveyed in reports or letters.
CALL TO ORDER, WELCOME, INTRODUCTIONS, REVIEW OF AGENDA
Dr. Lumpkin called the meeting to order and asked all present to introduce themselves.
Pursuant to FACA regulations, Dr. Cohn informed the Committee that as a member of the AMA CPT Editorial Panel, he would recuse himself from voting on a letter related to code sets that would be coming before the Committee at this meeting.
Ms. Greenberg suggested that at a future meeting, the staff of CDC’s ethics office brief the Committee on conflict of interest policies.
UPDATE FROM THE DEPARTMENT
James Scanlon, ASPE
CONSOLIDATED HEALTH INFORMATICS (CHI)
Jared Adair, CMS
The Office of Management and Budget asked HHS to take the lead in pulling together work in the federal government on health care informatics and developing the business case. The resulting project, CHI, is in the office of the Assistant Secretary for Budget, Technology and Finance, and Ms. Adair is the project manager. The Department of Defense and the Veterans Administration are participating with HHS, starting with identifying standards. This is still the proposal period with the President’s Management Council; the proposal is for the adoption of government-wide interoperability standards, including vocabulary, that would be built into each department’s architecture. The idea is that the federal agencies that touch healthcare would lead the way toward standardization with the rest of the healthcare community in the U.S. by themselves adopting standards. Rather than following a date-certain set for adoption, they would gradually build toward a target architecture.
These activities would be compatible with HIPAA and other activities; CHI is depicted as one leg of a four-legged stool; the other legs are HIPAA transaction records and code-sets, security and privacy solutions, and significant changes in management processes in the federal enterprise and beyond. The planners want to learn from the HIPAA process about working collaboratively with industry, building on what others have done, and adopting practices that make sense to others. The entire process is intended to be “ever-greening,” with built-in mechanisms for moving to newer iterations over time.
The NCVHS recommendations on PMRI standards are being used as a starting place. Ms. Adair said her main reason for coming to NCVHS was to ask for the Committee’s help on the CHI process. To this, Dr. Lumpkin responded, “the easy answer to your question is, ‘Yes.’”
He asked Ms. Trudel to brief the Committee on the status of HIPAA standards. She reported that since the February meeting, HHS has published two proposed rules (modifications to the initial transactions and code sets) and a final rule (the employer identifier). One of the proposed rules contained modifications to the transactions recommended by industry groups to facilitate implementation; the other concerned changes to the pharmacy transaction and National Drug Codes as a standard and follows the recommendations of NCVHS. The comment period for these NPRMs ends July 1.
A team is fleshing out what the final rules will look like, and the Department is working hard on the remaining four rules ¾ final rules on security and the provider identifier, and proposed rules on claims attachments and the plan identifier. The Department has received more than 20,000 compliance extension requests under the Administrative Simplification Compliance Act (ASCA). Some Medicare contractors are already in production with the 837 claim, which Ms. Trudel called “a tangible success point.”
CMS is spending a lot of time on outreach and has just aired its second HIPAA video by satellite broadcast and Webcast, reaching some 7,000 people. It recently held its second roundtable session, with 700 connections on an audio call to give feedback and identify issues. These have been well received, as has the CMS administrative simplification Web site. Ms. Adair said Ms. Trudel did an excellent job in the roundtable, answering questions for more than an hour.
Discussion
Dr. Cohn thanked Ms. Adair for her work and expressed the Standards and Security Subcommittee’s interest in working on the CHI project. He noted that she may want to participate in the Subcommittee’s forthcoming work on clinical terminologies. He asked about the timeframe for the CHI project and was told, with the caveat that it is still in the business case mode, that it might be a two-year period.
Dr. Lumpkin reiterated the Committee’s commitment to a strong partnership with the Department. He encouraged Ms. Adair to identify any issues that the Committee might bring out at hearings and on which the CHI project would like recommendations from NCVHS.
Noting that CDC, VA and DoD all have architectures and there are others as well, Dr. Fitzmaurice asked how CHI would pull them together into a single workable architecture. Dr. Adair said every unit is required by Clinger-Cohen to have an architecture; the intent is to have each party build interoperability standards and vocabularies into their respective architectures, resulting in what Dr. Fitzmaurice described as “different buildings with the same building blocks.” Asked if the CHI plans are built into the FY04 budget, Ms. Adair said the idea is that the partners would work together on the activities they are already planning and budgeting for.
Dr. Shortliffe hailed the intention “to enable the sharing of health information in a secure environment to improve health” (the CHI vision) and noted that the characteristics of federal healthcare systems should make the envisioned federal changes possible. He asked what the thinking is about using and analyzing the standardized data this interoperable system could generate ¾ i.e., “what will we be able to do that we can’t do today?” He added that “gripping examples” of the possible applications are needed in order to capture people’s imaginations. Moreover, it will be critical for bringing along the private sector, which is far more fragmented than the federal sector. He emphasized the Committee’s desire to help in the effort to “bring along the private sector.”
Dr. Lumpkin said the Committee is thrilled that the government is taking “exactly the approach that we recommended in our most recent communications.” Mr. Scanlon observed that this effort builds on the successes of HIPAA and on the NCVHS PMRI recommendations; federal agencies are converging where the private sector is already heading. He stressed that this is not a regulatory approach but one of leadership and example, as others have recommended for the federal government. He suggested that the Committee serve as “the FACA arm” of the CHI workgroup when hearings or expert testimony is needed in a public setting.
Dr. Zubeldia applauded this initiative as “the first step on the road to the National Health Information Infrastructure.” He urged that the NHII recommendations be kept in mind, along with efforts to avoid a payer-centric approach and to work toward interoperability with “the outside world” in the context of the NHII vision. He asked if the project would also look at communications interoperability, an omission from
HIPAA that has caused problems. Ms. Adair promised to return another time with an answer to that question.
In view of all the activity in this area, Dr. Lumpkin suggested that the next NCVHS meeting include time to talk about what is going on with the E-Health Initiative, the Markle Initiative, and AHA initiatives, as well as legislative initiatives, and how they and the CHI project fit together. In conclusion, he said that the Standards and Security Subcommittee would follow up soon and the full Committee will want a briefing sometime in the future.
UPDATE ON PRIVACY REGULATION
Stephanie Kaminsky, OCR
OCR staff members have been working very hard since the comment period closed to digest and analyze the more than 11,000 comments received in this round. More than 9,000 of these were on the topic of consent. They have a rigorous schedule to get the rule published as soon as possible so that the compliance date for the modifications converges with the compliance date of the final privacy rule (published in December 2000), which is April 14, 2003. (The aim is to have the modified rule take effect rather than the original “final” rule.) The Secretary is very interested and involved in the privacy rule and has been part of the process. The OCR is contracting for outside consultants to help develop technical assistance.
Dr. Lumpkin observed that the idea is not to trip people up but to be clear about what the rules are. The Department is trying to give as much guidance as possible to the industry as it implements HIPAA under the guidelines and timelines that have been set in motion.
DATA ISSUES IN MEASURING QUALITY OF MENTAL HEALTH AND SUBSTANCE ABUSE SERVICES
Ms. Coltin, Chair of the Workgroup on Quality, introduced the panel, saying that this is the latest of a series of panels that have spoken to the full Committee about data issues in measuring quality of care. Their testimonies will inform the report and recommendations on data gaps and data issues being developed by the Quality Workgroup. The panelists were given a broad set of issues to address with respect to their field, concerning such matters as the adequacy of existing federal data systems for identifying key topics, the adequacy of public and private sector data systems for developing and implementing quality measures, and privacy issues involved in collecting the information.
Eric Goplerud, SAMHSA
Dr. Goplerud provided an overview of the topic and presentations. SAMHSA and the behavioral health field have been concerned for a long time about the use of non-financial measures of quality of care for mental health and substance abuse. SAMHSA commissioned an Institute of Medicine study on the subject in the mid-1990s, resulting in the report, “Managing Managed Care,” recommending responsibilities for producing common performance metrics. While much of the field disregarded the recommendations, the national organizations representing state mental health, substance abuse and Medicaid directors developed a common framework for performance measurement that has provided the grounding for subsequent work. Panelist Vijay Ganju will describe that.
Two pilot studies, TOPS1 and TOPS 2, a project of the Washington Circle Group, are looking at the feasibility of selecting common quality measures of performance in mental health and in substance abuse treatment. Panelists Mady Chalk and Connie Horgan will talk about them. Dr. Goplerud observed that prevention and mental health promotion, which are often neglected and are not represented on the panel, are important aspects for which performance metrics are also needed. He called attention to the minimum data set developed by the Center for Substance Abuse Prevention.
The most interesting current activity is by the Carter Center Forum on Performance Measures. Leaders of the prevention and treatment fields are developing common, well-specified performance measures that can be used across payers and provider platforms. The content groups are adult mental health treatment, adult substance abuse treatment, child-adolescent mental health treatment with a subgroup on adolescent substance abuse treatment, and prevention. There is good agreement in the adult mental health and substance abuse treatment areas about how data should be collected, thus providing the foundation for common specified performance metrics. SAMHSA’s performance partnership grants now promote common data reporting by states on mental health and substance abuse services; this will reinforce the Carter Center Forum’s common measures.
Dr. Goplerud stated that behavioral health will make progress only when it is integrated into the more general health enterprise. IOM’s Crossing the Quality Chasm is used as the bible of the quality improvement effort in behavioral health. Turning to specific issues, he noted that publicly available data sets are disappearing and in managed care, data for benchmarking are not available. In administrative data sets, one general concern is how to integrate care between primary care and behavioral health, notably with respect to depression. Also, HIPAA transaction codes do not allow for the information needed to implement evidence-based practice. He added that Medicare doesn’t pay for ambulatory case management, which is critical to the effective treatment of depression. And “if you can’t pay for it and you can’t bill for it, good quality evidence-based care won’t happen.” In conclusion, he urged the Committee not to “wait another five years to have another panel” on these topics.
Mady Chalk, Center for Substance Abuse Treatment (CSAT)
Dr. Chalk directs the Office of Quality Improvement and Financing at CSAT. The Washington Circle Group decided to focus on measuring processes of care rather than outcomes for alcohol and drug treatment. A lot is known about the processes that are central to appropriate treatment and positive outcomes. The goal is to use measures that capture the range of necessary treatment services based on model of addiction as an often chronic, relapsing condition. Related services also need to be measured, such as those to improve health and social functioning.
The thinking was that the process measures, viewed as an intermediate set, would be measurable at program and systems levels, but there are significant barriers. They include pervasive stigma and problems of will (e.g., providers’ unwillingness to screen), the lack of a needed data infrastructure (only half of all substance abuse providers have computers) and the cost of data infrastructure development. CSAT is very involved in trying to improve the data infrastructure. Besides the lack of procedure codes, another major issue is data fragmentation across systems, related to the rising use of medications, treatment for substance abuse disorders in primary care, the rising prevalence of dual diagnoses, and the treatment of criminal justice populations. Finally, measuring movement across levels of care within alcohol and drug treatment is virtually impossible. CSAT and CMHS worked successfully with three states to integrate mental health, substance abuse and Medicaid databases. There were 32 requests from states, but “this would break the bank.”
Dr. Chalk concluded that evidence-based clinical guidelines are a central component of helping quality measurement move ahead.
Phillip Renner, National Center for Quality Assurance
Mr. Renner directs measures development for NCQA, which measures and reports on health care quality using accreditation surveys, the Health Plan Employer Data and Information Set (HEDIS), and a series of surveys including CAHPS 2.0H. The key evaluation programs with respect to behavioral health are of HMOs and managed behavioral health organizations (MBHOs). One issue deals with bridging the gap between these two types of organizations.
NCQA now has four HEDIS measures in place for MCOs relating to mental health and substance abuse, and four others under development. It has found that the major obstacles to measuring effectiveness in behavioral health relate to confidentiality, the adequacy of coding, and data availability. There are cost elements as well. Confidentiality constraints drive toward measurements based on administrative data rather than taking anything out of patient charts; and as noted, administrative coding is inadequate and imprecise. For example, there is no administrative coding for most of the relevant forms of screening and assessment.
Data availability issues arise in part from the fact that HMOs carve out MBHOs, creating data silos that degrade the ability to measure performance. Pharmacy is another data silo. It is thus very difficult to design measures that span the process covering initiation of treatment, referral, and treatment. Mr. Renner recommended a continued focus on ensuring that enactment of HIPAA does not present an obstacle to quality measurement as a part of health plan accreditation activities. He also recommended development and implementation of coding systems that support collection of quality of care information that is only available in charts.
Constance Horgan, Brandeis University
Dr. Horgan is with the Heller School for Social Policy and Management. Her presentation concerned the performance measurement work of the Washington Circle Group (WCG) and its experience with the challenges of using administrative data. The WCG is composed of substance abuse experts who held their first meeting at the Washington Circle Hotel.
The Group’s continuum of care framework is key. It describes the process of care starting with the population-based activities of prevention and education and moving on to recognition, treatment, and maintenance. Using this framework, the WCG has identified seven measures ¾ four for treatment and one each for the other domains. Dr. Horgan described three measures, all of which are derived from administrative data: for identification (recognition), initiation and engagement (both aspects of treatment). The three measures did work to discriminate between the three domains, but there were major challenges.
The first concerned the organizational structure of service provision. Many workplaces have employee assistance programs that provide services and are often outside the medical plan. This reduces the identification rates and could distort utilization patterns. Another organizational challenge relates to benefit design: limits, deductibles, and cost-sharing are common for mental health and substance abuse services and they lead to underreporting. Carve-outs are another challenge. Dr. Horgan reviewed the multiple pathways to behavioral healthcare, which has major implications for where organizational responsibility lies. Decisions about service provision are made both by employers and by payers. She noted that one-third of Fortune 500 companies carve out behavioral services.
Another set of challenges relate to patient and provider behavior with respect to privacy concerns and misreporting as a response to stigma, benefits, and financial incentives. Dr. Horgan noted that substance abuse is illegal behavior. The final set relate to coding issues with respect to both diagnosis coding and procedure coding.
Sue Eisen, Boston University
Dr. Eisen is with the Center for Quality Outcomes and Economic Research, a VA health services research and development center. She focused on the challenges and issues with quality measures derived from patient-reported survey data. Research has shown that well-designed consumer self reports can be reliable and valid measures of quality of care. However, there are several limitations. Response rates are never 100%; consumers are not the best sources of information about every aspect of quality of care (e.g., technical aspects); and in some areas (e.g., mental illness) the information provided may not be accurate. The patient represents one perspective; others include clinicians and family members. Other limitations are that differences in patient samples may make comparisons difficult, and the lack of standardization on consumer-reported quality indicators limits the ability to profile mental health services.
Two widely-used protocols for collecting patient reported survey data are an annual survey of HMO or MBHO enrollees regarding all services received during the year, and contact with individuals during or soon after the treatment period. NCQA’s ECHO Survey, a mental health module of the CAHPS survey, is an example of the first protocol. One limitation is that this protocol excludes uninsured people and those who are institutionalized. There are other limitations related to the use of mail and/or telephone for the contacts. This is one reason for low response rates on such surveys. Also, patients are sometimes frustrated at having to generalize about all the care they received over a year.
The main limitation with obtaining survey data from consumers in treatment settings is that resources are often inadequate. Also, there are privacy concerns. Dr. Eisen concluded with five recommendations, most concerning standardization and confidentiality protections.
Pamela Greenberg, American Managed Behavioral Healthcare Association (AMBHA)
The members of AMBHA collectively manage mental health and substance abuse services for more than 110 million people in the U.S. Ms. Greenberg (no relation to the Executive Secretary) said that in preparing to implement HIPAA regulations, AMBHA members are identifying gaps in revenue codes and HCPCS codes related to behavioral health. Together with the national associations of state mental health and substance abuse directors and others, they have submitted suggested codes to the HCPCS panel and will do so to the National Uniform Billing Committee regarding revenue codes. The lack of specificity of the codes for behavioral health is one barrier. AMBHA advocates designation of the DSM as a descriptor code set for diagnosis and reporting of mental health disorders under HIPAA. The organization is working with the X12 steering committee and others to address its EAP transaction concerns.
Practitioners’ concerns about patient privacy impede the use of medical records to measure quality. AMBHA is concerned that providers will use the HIPAA privacy rules to defend their position, and that such difficulties will increase the animosity between providers and MCOs. Ms. Greenberg noted that the term “minimum necessary” is open to a range of interpretations. She said AMBHA supports the proposed modification to the privacy NPRM that offers the option of obtaining prior consent from patients.
Suzan Lumpkin, Magellan Behavioral Health and AMBHA
Magellan, an AMBHA member, is the largest MBHO in the U.S., with some 70 million clients. Ms. Lumpkin (no relation to the Chair) said that the behavioral health industry has determined that they need to measure the same types of areas that are measured by JCAHO. She outlined some of the challenges faced in doing so. As claims generally go to the primary payer, this information is not available to MBHOs; nor is information on follow-up and on timelines for care. Coordination with health plans to capture information is difficult. Surveys are the best source of information on outcomes; it is much more difficult to get clinical outcomes information, for the reasons mentioned by other panelists, including stigma. It is difficult to identify the treatment received through surveys, and some patients may have trouble understanding the rating scales or concepts in the instruments. Getting timely information is another challenge. Ms. Lumpkin raised a number of concerns of AMBHA and Magellan and noted the enormous cost of changing systems. AMBHA members support the use of a standardized set of measurements.
Vijay Ganju, NASMHPD Research Institute
Dr. Ganju said his presentation would try to link the needs for data on quality to quality initiatives going on primarily in the public mental health sector. He has been involved in several federal and state performance measurement initiatives. He pointed out that a significant proportion of spending on mental health is by states, and this together with spending by Medicaid and Medicare creates an ethos for quality initiatives that is different from that of other health sectors. He noted, too, that the Surgeon General’s Report on Mental Health identified huge gaps between what we know and what is going on in mental health practice. That represents an opportunity with data implications. He noted the fragmentation that exists in the public mental health sector, encompassing the criminal justice, child welfare, state hospital systems among others.
Quality initiatives have come a long way since creation of the MHSIP Report Card in 1996. For example, 48 states implement parts of the MHSIP report card, using similar measures. And what began as a five-state study in 1997 has expanded through data infrastructure grants so that 47 states report the same data elements and do the same consumer survey. Still, much needs to happen to embed quality in our data systems. Much of the focus of the public mental health initiative is on outcomes such as reduced symptoms, increased functioning, and reduced hospitalization. Mr. Ganju showed a report card on data available for mental health quality, which shows movement toward standardized definitions but still considerable variation in the meanings associated with nomenclature. This is an issue for HIPAA coding.
Many people who have been recognized as having a mental health problem are not captured by existing data, and in some cases there is duplication between systems. So using data that cuts across funding streams and sectors, it is not possible to count the number of people receiving services. For planning purposes, it is impossible to determine how many people need services. Mr. Ganju enumerated other data limitations related to contracting, cultural competence, implementation of evidence-based practices, and data integration. He noted the number of references by other panelists to data fragmentation.
The public mental health quality agenda has three prongs: performance measurement, implementation of evidence-based practices, and systems-level quality management and quality improvement. People are working on new models in these areas. Mr. Ganju reviewed the status of mental health performance measures with respect to standardization, implementation, and quality management with respect to utilization rates, process measures, outcome measures, and consumer surveys. The greatest issues related to standardization for process measures of evidence-based services, standardization and quality management for outcome measures, and quality management for consumer surveys.
Turning to data limitations, he mentioned duplication across systems and the need for standardized definitions with administrative data, response rates and methodology with consumer surveys, and standardized instruments and benchmarks for different populations for outcomes measures. Several of the organizations represented on the panel have tried to get new mental health procedure codes into HIPAA. Another priority is performance measures related to implementing evidence-based practices. In general, a strategic approach is needed to address these issues, because current approaches are fragmented and piecemeal.
Discussion
Dr. Newacheck noted that NHIS estimates of unmet needs for mental health care among children are very low and have limited face validity, suggesting incorrectly that children’s mental health needs are being well met. The same is probably the case with adults. He asked if research is underway to identify better measurements of unmet needs, and whether there is any point in collecting data with low reliability and validity. Mr. Ganju agreed that there is a large unmet need, which is reflected in national prevalence studies. He added that another issue is the people who establish contact with the mental health system but don’t come back. Ms. Eisen noted the need for consensus and a standardized definition of unmet need.
Dr. Mays commented on the complexity of measuring mental health and the difficulty of including it in surveys like the NHIS. She noted that the health care system and the dysjunctions between subsystems is a major source of measurement problems, both at federal and state levels. She expressed concern about impacting patients negatively in the push for quality data, and noted that patients’ concerns about stigma lead to behaviors that distort the data. She concluded that there are some good reasons why mental health has been kept separate that need to be considered in the move toward integration.
With respect to the interest in functioning as an aspect of outcomes, Marjorie Greenberg pointed out the relevance of the International Classification on Functioning, Disability and Health (ICF). Noting that work is underway to map assessment tools to the classification, she referred the panelists to the NCVHS report on functional status.
Dr. Fitzmaurice asked whether SAMHSA is active with the SDOs in the development of codes and vocabularies for behavioral health uses. Ms. Wattenberg of CSAT replied that the Center sponsors a code set group and has been encouraging states to get onto all the relevant standards workgroups. She noted the lack of expertise in behavioral health and mental health, mirrored by a lack of clinical expertise on the standards committees. On the mental/behavioral health side, another constraint is a shortage of manpower.
HEALTH STATISTICS FOR THE 21ST CENTURY
Dr. Friedman presented the draft report developed by the Workgroup on 21st century Health Statistics, NCHS, and the Data Council, and the recommendations developed by the Workgroup and intended to represent only NCVHS. The report was written by Dr. Friedman along with Ed Hunter and Gib Parrish of CDC. It is the culmination of a four-year-long consultative process that has already produced an interim report and several presentations. The report has four chapters covering 1) the purposes and definition of health statistics and the health statistics enterprise, 2) a model of influences on the population’s health, 3) a model of the health statistics cycle and gaps in the process, and 4) a vision for health statistics that includes a mission, an overarching conceptual framework, core values, and guiding principles.
The chapters are followed by draft recommendations that are proposed for NCVHS approval. They flow from the ten guiding principles, which were originally presented in the interim report. They are also presented in a cross-cutting classification system based on their relation to issues that include data access, data set improvement, evaluation, and training. Four core recommendations provide the foundation for the others. The first relates to developing a reconstituted NCHS with a strong board of scientific counselors, to provide the essential “hub” functions of coordination, planning and standards. The second recommendation, which flowed from a recommendation in a National Academy of Sciences workshop, is for an autonomous health statistics planning board. It would set the agenda for the health statistics enterprise, periodically evaluate it in the light of the mission, engage in consensus planning, and so on. The third core recommendation is for coordination of state health statistics activities in each state by a single state agency, supported by state health statistics planning boards. The fourth calls for expanded training in health statistics.
Dr. Lumpkin added that an executive summary and another summary aimed at policy makers are forthcoming; these and minor text editing will not result in any substantive changes except for any recommended by the Committee in the forthcoming discussion.
Dr. Cohn congratulated the Workgroup and authors on the report. He raised a few questions about specific recommendations. Questions on geocoding and longitudinal data sets prompted a discussion of privacy and confidentiality implications. Dr. Friedman stressed that with any size data set, a basic principle is that confidentiality will be protected. He noted that the report may need to clarify this, and also the fact that the vision is not for “a big data base in the sky.” He added that there were many calls in the hearings for the capacity to do longitudinal analysis. Dr. Lumpkin cited the practice of linking death certificates and cancer registries, and infant birth and death records. Such information is very important but would be made accessible only for restricted uses. Mr. Scanlon commented that what is really being proposed is the capacity for data linkages, not the creation of data sets. Mr. Hunter pointed out that one of the recommendations (3.1) addresses privacy and a code of fair information practices.
Noting that funds are lacking to do what NCHS already wants to do, Dr. Mays questioned the cost of doing what is envisioned. Dr. Lumpkin pointed out that the cost of much data collection is now greatly reduced, thanks to automation. The 21st century health statistics vision, like the NHII report, is a blueprint for developing a new system in a new environment, identifying the capabilities this environment should have and what we should be building toward. Dr. McDonald pointed out that the functions being supported are those of “traditional public health,” and in his view the proposals could even go farther to expand data use for public health purposes.
In a discussion of the report’s definition of health statistics, Dr. Friedman said the report stresses that there can be multiple views and uses of the same data, rather than the generation of more and more data sets. To a question from Mr. Scanlon, he said the definition of health statistics does encompass public health surveillance. Several people asked that the authors clarify how the health statistics arena overlaps with, and is distinct from, the NHII and the world of medicine.
Dr. Danaher asked about the intended uses and impacts of this report, and what immediate need it is fulfilling. Dr. Friedman reviewed the genesis of the consultative process and the resulting document. Health statistics and population health in the U.S. have been not reflective and have been based on a narrow set of legacy data systems with a narrow set of outcome measures. Even the health statistics field has not clearly recognized the distinction between individual health and population health. The purpose of the effort and the document was to provide a consensus framework for people engaging in health statistics and those who use health statistics, for use in more detailed planning and implementation. The report will be sent to the Secretary, and it is also intended for use by other levels and sectors of government and in private and research sectors. Dr. Lumpkin added that it fits in the overall context of the NHII, fleshing out the population dimension.
Further discussion clarified the thinking about the intended role of NCHS, which is to advise on public health statistics within DHHS and to have responsibility for general purpose health statistics, not to take over all data collection within the Department.
Dr. Mays noted that there is a lot of activity now by schools of public health and medicine, working with state departments of health, to do the training recommended in the report. She suggested expanding the recommendation on training to talk about training in place as well as basic training.
Dr. Lumpkin said the Workgroup on a Vision for 21st c Health Statistics would meet following this meeting to process the foregoing comments, and a revised version of the report would be presented to the Committee on the following day for action.
DRAFT RECOMMENDATIONS FOR CODE SETS
Dr. Cohn reported that the Subcommittee on Standards and Security is drafting a letter in response to two notices of proposed rulemaking on electronic transactions. It will be presented on day two of this meeting.
He then introduced Dr. Zubeldia for presentation of a letter on code sets, developed on the basis of two hearings on the subject. Dr. Zubeldia said the letter addresses the migration from ICD-9-CM to ICD-10-CM and ICD-10-PCS, in response to testimony that ICD-9-CM is not capable of supporting current business needs for administrative, public health, and other uses. The letter says that the Subcommittee recommends that ICD-10-CM and ICD-10-PCS be implemented as soon as possible, but no sooner than October 2005, to allow a hiatus after HIPAA implementation prior to the transition to the new code sets. The Committee expects to continue deliberations on code set issues and to provide additional recommendations in the future.
Dr. Lumpkin asked if there were any concerns on the Committee about migrating from ICD-9-CM to ICD-10-CM. None were expressed. He then asked for discussion on replacing volume three of ICD-9-CM with ICD-10-PCS. Dr. McDonald expressed some concerns. Ms. Greenberg stressed that the industry has made it clear that volumes 1-3 of ICD-9-CM have to be replaced. The idea is for hospitals to make the transition to ICD-10-PCS and for physicians to continue to use CPT, at least in the short run. The long-term goal, as has long been recommended by NCVHS, is to have a single procedure classification system. Moving ahead with an NPRM would get the transition to ICD-10-CM and ICD-10-PCS out for comments and launch the process. Dr. Lumpkin asked the Subcommittee to discuss these issues further and bring a proposal to the Committee on day two.
The Committee recessed into Subcommittee and Workgroup meetings.
GAO REPORT ON RECORD LINKAGE AND PRIVACY
Judy Droitcour and Nancy Donovan, GAO
GAO published a report last year on record linkage and privacy, focusing on record linkage conducted under federal auspices, involving person-specific data, done to generate research and statistical information. Dr. Droitcour noted that GAO itself links records; it is not just observing this practice from the outside.
Record-linkage has flourished in recent years with technological advances and a growing recognition of its power as a research technique. Privacy issues are especially relevant because linkage can create new kinds of information on the individual and the data involved may be sensitive. GAO initiated this R&D study because the issue is under discussion and a matter of concern in many arenas and no one has created a broad overview of the issue to inform policy makers. The authors talked to the Senate Committee on Governmental Affairs before beginning the study and have conducted seven requested briefings on the findings on the Hill, as well as many other presentations in other settings. The intent of the report is to inform and stimulate discussion by providing examples, not to provide a comprehensive survey of what is going on.
Dr. Droitcour showed a graph that has “priority on information gains” on one axis and “priority on privacy issues” on the other, illustrating the multiple tradeoffs or combinations between the two variables. One goal of the GAO report was to challenge people to find strategies to assure both priorities. She acknowledged the help they received from Mr. Scanlon, Dr. Fitzmaurice and Dr. Fanning as well as many other officials in HHS and other agencies. Examples in the report pertain to either health or income data because these are sensitive areas and a good deal of linkage activity goes on in them.
Several types of data are involved in record linkage: sample survey data (generally voluntary and from individuals), archives (data on everyone in a target group, and generally mandatory) and contextual information. These are collected by statistical, research, and program agencies. Dr. Droitcour gave examples of each type of data set. She then gave examples of record linkage in the areas of health and income/wealth, within and among the various categories. Two examples are the linkage of SEER clinical data to Medicare insurance records, and SEER data and Medicare records linked to census-tract information. There are multi-person linkages, survey-to-archive linkages, multi-archive linkages, and linkages of survey or archive plus contextual data. The linkages provide an additional dimension of information and permit analysis that would not be possible otherwise.
Ms. Donovan then discussed some of the privacy issues related to record linkage, focusing on consent and data sharing. The report also addresses re-identification risks, potential sensitivity, and security issues such as the heightened risks of transfer through the Internet. Consent is related to a core concept of personal privacy¾that each individual should be able to control information about himself or herself. The authors found that consent is sometimes obtained and sometimes not before linkage occurs; in the latter case, people may not know that new information is being created about them. Consent may be obtained in an explicit or implicit form. Data sharing raises concerns over confidentiality and security because more than one agency or organization is involved. Some agencies observe the principle of “functional separation” by only permitting one-way data sharing (i.e., survey or research data are not transferred to program agencies).
The report’s third major topic is building a privacy protection toolbox. The tools include techniques for masked data sharing, measures for safer data and/or safer settings, techniques to reduce sensitivity, consent forms, and security measures. Dr. Droitcour noted that the techniques that create safer data tend to distort the data, making it preferable to create safer settings, as the Census Bureau and NCHS have done in recent years. She described the measures of using a trusted third party and of list inflation.
Finally, the report talks about data stewardship strategies. It emphasizes that mechanisms are important for weighing the risks and benefits of linkage. Such mechanisms are the IRB, getting input from data subjects, development of a government-wide checklist of items for use in linkage decisions, the use of technical expertise, and data disclosure boards. Accountability and organizational culture are important for data stewardship, as is leadership, as exemplified by groups such as NCVHS.
In conclusion, Dr. Droitcour said the authors will probably do a follow-up on the future of record linkage, as practices, technology and attitudes are ever-changing. She invited questions from the Committee.
Discussion
Members praised the presentation and the useful schema of the report. Asked about its intended uses, Dr. Droitcour said GAO’s intent was to raise awareness of both the issues and the strategies that could be used, and to encourage discussion. She noted that in certain other countries, when the general public learned about linkage activities, their strong negative reactions resulted in governments “throwing the baby out with the bathwater” and that it would be desirable to avoid this in the U.S. Ms. Donovan said they’ve gotten positive feedback from agencies that find the report helpful.
The authors agreed with Mr. Blair that the barriers to linkage are a very important topic. The report identifies this as a question for further study, using a cross-cutting approach. Asked if they had consulted the NCVHS NHII recommendations, Ms. Donovan said that because of their interest in the future of record linkage, they reviewed the interim and final reports to consider the implications of the GAO and NCVHS reports for each other’s areas. Issues related to NHII and the NCVHS work on NHII will be considered in any follow-on work by GAO.
Dr. Friedman noted that the table of types of record linkage does not highlight person-by-person record linkage within a single data set (i.e. longitudinal data linked across time) and he said he believes this is an important area that is problematic in terms of current oversight.
While acknowledging the importance of considering and minimizing the risks of the bad things that could happen, Dr. Shortliffe asked if the authors had found any evidence of actual harm to individuals or groups from record linkage. They said they had not, although Dr. Droitcour said that she was told informally of various instances where identifiable records were inadvertently released but retrieved before harm was done. She noted that in the Internet age, such errors would be impossible to reverse and could have widespread consequences. For example, as Ms. Donovan pointed out, even one inadvertent breach of confidentiality could, if publicized, result in more limited access to data by researchers in the future.
Dr. Lumpkin thanked the presenters and said the Committee looks forward to their future work, which coincides with and will be helpful to that of NCVHS.
NCVHS FIFTH ANNUAL REPORT TO CONGRESS
Mr. Scanlon reviewed the latest draft of the annual report to Congress on HIPAA implementation. In a move to optimize the timeliness of the report, he proposed extending the time period that the report covers from the end of March to the end of June, to encompass two significant HIPAA-related events that happened in that three-month period. A motion was passed unanimously to extend the time period, to approve the current draft of the report, and to authorize the Chair to approve the final version.
REPORTS FROM SUBCOMMITTEES AND WORKGROUPS
Executive Subcommittee: Dr. Lumpkin
The Subcommittee will hold a one-day meeting on July 26 in Chicago, to be preceded by two daylong hearings, one on the NHII and one on Quality.
Workgroup on Health Statistics for the 21st Century: Dr. Friedman
Dr. Friedman reviewed the comments on the report made the previous day and described the Workgroup’s responses to each one. To clarify the relationship between the health statistics enterprise and the NHII, a graphic from the NHII report will be included; in addition, Dr. Lumpkin will add a foreword tying the elements together. The foreword will also clarify the relationship to NEDSS.
Dr. Friedman offered revisions to several recommendations, in response to the Committee’s comments. Dr. Lumpkin suggested that after the Committee discusses the revisions, the recommendations go back to the Workgroup for final editing, with the Executive Subcommittee authorized to give final approval to the report and related documents. Accordingly, a motion was passed unanimously authorizing the Executive Subcommittee to approve 1) the foreword, 2) the final iteration of the recommendations, to be revised based on the forthcoming discussion, and 3) the executive summary and another summary for policy makers.
Dr. Friedman then read the revised language of the recommendation for a reconstituted NCHS with enhanced responsibility for planning, coordination and standards. This prompted further discussion of the implications of the recommendation for other HHS agencies such as AHRQ and NIH. Dr. Lumpkin noted the goal of greater standardization in the entire health statistics enterprise, both within and beyond the federal government. Dr. Friedman said that the most consistent message heard during the years of public consultation concerned the lack of coordination and collaboration in the broad health statistics agenda. He added that the recommendations are aimed at getting these issues on the table at HHS, rather than saying exactly how the problems should be solved. A motion was made and seconded to accept the revised version of recommendation 1.1. The motion passed, with one person opposed.
Another motion was then passed unanimously to accept all remaining recommendations, as revised.
Regarding finalizing the report, the Committee agreed on the following process: when the Data Council has reviewed the report (chapters 1-4, not the recommendations, which only represent NCVHS and are now final), any edits it recommends will be sent to the entire Committee for information. If anyone believes that any of the changes are substantive, the report will be brought back to the full Committee to discuss.
To bring the pieces together, a motion was passed unanimously to forward the report and recommendations, finalized in the manner outlined in the previous motions, to the Secretary.
NHII Workgroup: Dr. Lumpkin
The NHII Workgroup will conduct a hearing on July 24 in Chicago on three areas: the population health dimension of registries and surveillance and the overlaps between them, and personal health records, and the integration of email with clinical and public health information systems.
Subcommittee on Standards and Security: Dr. Cohn
The Subcommittee is drafting a letter on code sets, which will be brought to the Committee in September. Dr. Lumpkin asked for an accompanying factsheet of background material.
Dr. Cohn then presented a letter, which he read aloud, on the NPRM (CMS-0003P and CMS-0005P). A motion was passed unanimously, approving this letter, which will be sent to the Secretary.
Dr. Cohn reported that the Subcommittee has scheduled three sets of hearings for the remainder of the year. On August 28-29, the hearing will launch the Subcommittee’s next steps on PMRI with a discussion of clinical terminologies and code sets. An October 22-23 hearing will focus on the basic HIPAA administrative simplification transactions, with a report from the DSMOs on recommendations for changes and modifications to standards and a discussion with the DSMOs about improving the
HIPAA process. The topics for hearings on December 10-11 are not yet determined.
Subcommittee on Privacy and Confidentiality: Mr. Rothstein
This Subcommittee will hold three hearings this Fall, all outside Washington, D.C. in different sized communities. The aim is to get a sense of how covered entities are progressing toward compliance with forthcoming privacy requirements, where the issues are, and what kinds of outreach and education are needed.
Subcommittee on Populations: Dr. Mays
The Subcommittee will review and comment on a handbook under development by CDC on the measurement of disparities. Continuing its study of health disparities issues, it will hold a hearing in September in Denver on data needs of Native Americans and (time allowing) Asian-Pacific Islanders, and in November in Philadelphia on the barriers to collecting information on racial and ethnic minority groups. The September hearing may also have a presentation on language issues.
Workgroup on Quality: Ms. Coltin
Members are encouraged to attend the Workgroup’s hearing on data issues in quality improvement and to garner input to AHRQ for its National Healthcare Quality Report. Plans are underway for the Workgroup’s report that will summarize the issues and present recommendations regarding data problems and barriers to measuring quality and how they might be addressed. An early version of the report and recommendations will be circulated at the November meeting, and the final report will be presented at the February NCVHS meeting.
Ms. Coltin said she hopes to get input from each of the subcommittees on the draft report prior to the November meeting, if their schedules permit.
FUTURE AGENDAS FOR NCVHS MEETINGS
Dr. Lumpkin reviewed the dates for forthcoming meetings and asked members to put them in their calendars:
Dr. Shortliffe suggested continuing the discussion about Consolidated Health Informatics (CHI) at a future meeting. Dr. Lumpkin noted the flurry of disparate standardization efforts and the need to coordinate and leverage them, a role NCVHS could play. Some other initiatives besides CHI also deserve attention, including ones in the private sector and in Congress. Dr. Shortliffe agreed to help conceptualize a session with this focus, either in September or November 2002.
Other suggestions for the September agenda:
Dr. Lumpkin then adjourned the meeting.
I hereby certify that, to the best of my knowledge, the foregoing summary of minutes is accurate and complete.
/s/ John R. Lumpkin, M.D. 9/25/02
___________________________________________________________
Chair Date