The Subcommittee on Population-Specific Issues of the National Committee on Vital and Health Statistics was convened on Monday, June 23 in the Hubert H. Humphrey Building in Washington, D.C. The meeting was open to the public. Present:
Subcommittee members
Lisa I. Iezzoni, M.D., M.S., Chair
Hortensia Amaro, Ph.D.
Vincent Mor, Ph.D.
M. Elizabeth Ward
George Van Amburg
Staff
Carolyn Rimes, HCFA
Ronald Manderscheid, Ph.D., Center for Mental Health Services
Marjorie Greenberg, Acting Executive Secretary (NCHS)
James Scanlon, Executive Staff Director (DHHS)
Olivia Carter-Pokras, Ph.D., Office of Minority Health
Patricia Golden, NCHS
Others
Janet Spinks, AQUAS, Inc.
Michael Tate, American Dental Association
Larry Crecy, National Caucus and Center on Black Aged, Inc.
Gwen Moulton, BNA, Inc.
Leigh Ann Albers, The MEDSTAT Group
Lynnette, National Center for Health Statistics
Kim Jones, VHI
Jack Emery, American Medical Association
Clair Rodriguez, Fordham University
Lydia Buki, National Hispanic Council on Aging
Claudette Bennett, U.S. Bureau of the Census
Gwendolyn Clark, Office of Minority Health
Clyde Tucker, Bureau of Labor Statistics
Manuel De La Puente, U.S. Bureau of the Census
Roderick Harrison, U.S. Bureau of the Census
The Subcommittee on Population-Specific Issues met on June 23 to develop its work plan and hear reports on matters related to race and ethnicity reporting. They began by reflecting on the many merits of the hearings held in San Francisco in early June.
The Subcommittee then was briefed by Ms. Greenberg, Dr. Manderscheid, and Ms. Rimes (respectively) on the Committee's core date elements project, the supplementary recommendations on mental health/substance abuse core data, and the report on long-term care/disability data sets. Ms. Greenberg stressed that the Committee's research agenda has so far gone unaddressed. The Planning and Implementation group will review the core data elements recommendations in the light of unfinished NCVHS business. An effort will be made to coordinate SAMHSA's work on core data elements for mental health and substance abuse with the Committee's follow up on its core data elements recommendations.
The findings on race and ethnicity reporting are summarized below (see Dr. Harrison's summary). The work plan, a framework for which was proposed by Dr. Iezzoni at the outset, is reflected in the following action items.
Dr. Iezzoni called the meeting to order. Following introductions, she commented on the June 3-4 hearings in San Francisco on privacy, confidentiality, standards and coding. They were a vehicle for thoughtful input from 44 presenters representing a wide variety of perspectives. She congratulated the staff for the success of their efforts, adding that the tone felt different from Washington, D.C. meetings -- more concrete and more candid. Mr. Scanlon noted the benefits of addressing privacy, standards and coding issues in the same meeting.
In response to a question from Dr. Amaro about the impact of administrative simplification on resource-strapped agencies, Dr. Iezzoni said the representative from Los Angeles County testified eloquently on that subject from her agency's perspective. Pursuing this issue will be considered as a possible part of the work plan (see below). Mr. Scanlon added that L.A. County favors standardization, but is concerned about timing and resources.
Dr. Detmer observed that the hearings showed the real tension between privacy and various uses of health data, something the Committee will have to struggle with.
Turning to the Subcommittee's work plan, Dr. Iezzoni proposed the following categories for its forthcoming work:
I. Unfinished business
A. Recommendations on State and Community statistics
B. OMB Directive 15
C. Core data elements project
D. Spanish translation of surveys
II. New Business
A. Discrete, focused issues needing immediate attention, e.g.:
B. A broader health data issue the Subcommittee would take on as an extended project resulting in a report, e.g.:
The Subcommittee returned to a discussion of its work plan at the end of the meeting.
Marjorie Greenberg suggested viewing the core data elements project in the context of the broader NCVHS mission, in which core data sets and electronic medical records are complementary parts of a continuum. The core data project grew out of the health reform discussions that began in 1992. A major recent influence is HIPAA legislation, which resulted in six DHHS implementation teams, all but one of which (systems security) deal with issues addressed in the core data elements project. HIPAA also gives NCVHS a key role in advising on core data and electronic medical records. These recent activities and responsibilities follow naturally on the Committee's long history of involvement in standardization and uniform core data sets. The HIPAA legislation picked up on many things the Committee had been advocating for many years. The Core Health Data Elements Project recommendations were released in August 1996.
Ms. Greenberg proposed a hierarchy of needs that starts with a privacy framework and includes classification systems, agreement on core data elements and their definitions, and a data dictionary. In this context, she stated the major findings of the core data elements project, noting that they predated but are consistent with HIPAA:
· The health care information field solidly supports standardization.
· The field is seeking a recognized leader and group to forge consensus and is receptive to the Committee's playing that role.
· A consensus exists for many data elements, but there is less agreement on data definitions.
· Several other data elements (e.g., functional status) are critical but need further work.
The Committee recommended not just a set of elements but a process, asking the Data Council to circulate its report within the Departments, to place it on the home page, and to refer it to the National Uniform Claims Committee. All of these have been done. It also asked for stable resources; the Council has none of its own to share, but it has invested in continuous mutual involvement with NCVHS.
The report categorized the core data elements as 1) ready for prime time, 2) close but not quite ready, and 3) needing considerable additional work. The Committee asked the Data Council to pursue standardization of the items in the first category, and HIPAA has become a vehicle for some of that. The Executive Subcommittee recently agreed to review the recommendations in terms of what is and is not being standardized. The recommendation that still needs Departmental attention is the additional work on the highlighted critical elements.
The data were put in two categories, personal enrollment data and encounter data. The former belong in a database but would not be collected at each encounter. They do not appear on the standardized claim. The personal unique identifier was considered critical, but difficult. Ms. Greenberg briefly reviewed the Committee's history on the personal identifier. It recommended the Social Security Number (SSN) in the UHDDS and UACDS, but the core data elements recommendations suggested using not just a number but confirmatory information and basic demographics. (This is similar to an identifier being developed and piloted in California.) She then highlighted some of the elements, noting that the Committee recommends mother's years of schooling as a proxy for socioeconomic status that is highly predictive of health status and health care use. She called attention to the fact that the Census Bureau and NCHS use credentials instead of years of schooling for the same information. Another element of note was patient's reason for visit.
Looking ahead, Ms. Greenberg predicted that legislation will be necessary to protect privacy and confidentiality, as well as for a unique identifier. She reiterated that the Committee's research agenda has so far gone unaddressed, adding that the Committee's Planning and Implementation group would review the core data elements recommendations in the light of unfinished NCVHS business.
Dr. Iezzoni then asked Dr. Manderscheid to brief the Subcommittee on the core data elements project in the area of mental health/substance abuse. He explained that SAMHSA, and ADAMHA before it, have been developing minimum data sets in this area for years. SAMHSA is now updating and adding outcome information to the most recent (1989) iteration. It assessed what is going on in the field and released a report on that, and is now completing its recommendations for enrollment and encounter data, due out in August. A second report will cover what managed care firms are doing.
Regarding the interface with the National Committee's work, the Subcommittee on Mental Health Statistics made two recommendations on enrollment and encounter data: 1) that any effort to look at the question of functional status should include social functioning, and 2) that relevant items be embedded in the medical record rather than in enrollment data, in view of the need to prevent the red-lining of these clients.
SAMHSA also has started a work group on the electronic patient record for behavioral health care, looking in part of the question of how to bring together the diverse types of data needed. Dr. Manderscheid offered to share its report with the Subcommittee.
Regarding the NCVHS response to the Mental Health Subcommittee recommendations, Ms. Greenberg said they were referenced in the main report, which included the recommendation on social function. The non-inclusion of mental health status in the core enrollment elements was in part a result of the Subcommittee's recommendation. The Committee's recommendations were reviewed with SAMHSA and found acceptable. Mr. Scanlon noted that the Committee's work on the electronic medical record will provide a context for continuing joint efforts. He suggested that SAMHSA's work be shared with the relevant HHS work groups. Dr. Detmer, Dr. Manderscheid and Mr. Scanlon will follow up on that.
Dr. Iezzoni asked how the Subcommittee on Population-Specific Issues could support SAMHSA's efforts. Dr. Manderscheid said they would like the work of the two groups to be consistent and coordinated. He suggested that his group report to the Subcommittee periodically, starting with a report on the recommended enrollment/encounter MDS. The Subcommittee agreed that this report should be to a joint meeting with the Data Standards Subcommittee.
Ms. Rimes then briefed the Subcommittee on a report on long-term care and disability data elements that ties together and cross-walks several data sets in that area. Unlike the aforementioned reports, it does not include recommendations but is merely descriptive. Dr. Amaro noted that the report's significance is in promoting the idea of getting away from categorical types of measurement. Asked about the report's recommendation that a functional status measure be included, Ms. Rimes said in her view there are no functional measures structured to specifically address the long-term care population, and this is one of the areas that needs methodological work before anything is recommended. The Subcommittee agreed to authorize payment for the report on Long Term Care and Disability: Data Elements.
Mr. Scanlon suggested that the Subcommittee be briefed about a joint project of HHS and Census to develop disability questions for the census long form. They will ask for such a briefing. Dr. Iezzoni noted that the Social Security Administration also is starting a big disability survey, and the Subcommittee might also be briefed on this. Keeping track of how the field is looking at functional status would be a way of staying involved in disability issues. She remarked on the contrasting views on the ICIDH expressed at the San Francisco hearing, and Ms. Greenberg said NCHS would like to brief the Subcommittee on the ICIDH revision process. The NHIS disability survey also was mentioned, and the group agreed that these briefings on surveys should be held in conjunction with the Data Subcommittee.
Mr. Van Amburg then discussed a summary and set of recommendations on community assessment and state/community statistics that grew out of the work of the former Subcommittee on State and Community Health Statistics. The members of the Subcommittee on Population- Specific Issues expressed support for forwarding the report to the full Committee, for transmission to the Secretary with a cover letter from Dr. Detmer. Mr. Van Amburg noted that the current timing is good because state surveys are under consideration and there is awareness of the critical need for community assessment. At Ms. Ward's suggestion, the report will also be sent to the state centers and to NAHDO and ASTHO, so that states know of the Committee's recommendations.
Regarding Spanish translations, it was noted that the HIS is the last major need for translation and that the decision to translate it will improve the survey's data. Mr. Van Amburg pointed out that the issue of bilingual interviewers is not addressed in Dr. Sondik's letter on the subject to Dr. Detmer. Ms. Golden said the NHIS people are working with the Library of Congress's translation service. She offered to arrange an update on that effort, and Dr. Iezzoni asked Dr. Detmer to ask for this in his response letter to Dr. Sondik. The Subcommittee will continue to monitor this area.
The OMB began reviewing its 1977 Directive 15 in 1993. It convened an Interagency Committee which created a research working group to investigate the effects of proposed changes on data quality and collection procedures for various agencies. The first empirical study was the Current Population Survey Supplement on Race and Ethnic Origin, which examined the effects of 1) adding a multiracial category to the list of races, 2) including Hispanic in the list of races, and 3) alternative terminology. A working group developed the supplement in consultation with experts on race and ethnicity and other experts. The overall design divided the survey sample into four panels on whom different versions were tested. The CPS is administered repeatedly, so they were able to compare responses at the beginning and end of the survey period. The overall nonresponse rate to the supplement was 17 percent.
Dr. Tucker reviewed the findings for each of the variables for the four panels. The net results are as follows:
· The multiracial category has no statistically significant effect on the proportions identifying as White and Black, but it causes a 20 percent drop in the American Indian count and a drop in the Other category.
· The preferred terms for the respective groups are "Hispanic," "Black" and "American Indian."
· The size of the Hispanic population will be larger if a separate question is used.
· The size of the White population will be reduced if Hispanic is included as a racial category.
· Mexican-Americans would be a larger proportion of the Hispanic population if race and ethnicity were combined into a single question with only one response allowed.
· A large proportion of the Other race category move to Hispanic when that category is available on the race question. Under the same conditions, 4-5 percent of the White category moves to Hispanic.
· Hispanics prefer to be in a racial category (in the absence of knowing how this would affect the counts).
· The American Indian/Alaskan Native population count will probably be reduced if a multiracial category is included.
· American Indian self-identification is less stable than other races and changes over time more than the others.
This survey and the 1996 Race and Ethnic Targeted Test (RAETT) are part of the Bureau of the Census 2000 research and testing program. It used a national sample of more than 90,000 households representing 95 percent of the country. Because it occurred in mail-out/mail-back areas and used telephones for re-interviews, it is less representative of the American Indian and Alaska Native populations. The three objectives were to determine the effects of:
· adding a multiracial response category to the race question
· placing a Hispanic origin question immediately before the race question
· the combined changes
The survey found the following:
· The multiracial category did not affect reporting for the White, Black, American Indian, or as Asian and Pacific Islander populations at the 90 percent confidence levels.
· While there were no statistically significant effects on the Asian and Pacific Islander population, there was some evidence to suggest that they were affected.
· The combination of placing the Hispanic origin question immediately before the race question led to a reduction in reporting in the Other race categories, but did not eliminate the non-response to the race question. It also lead to a reduction in the non-response rate to the Hispanic origin question.
The RAETT, unlike the other two, was targeted to get enough African Americans, American Indians, Alaskan Natives, Asian-Pacific Islanders, Hispanics, and White ethnics, using Census tract-level data. The sample included 112,000 households but is not a national sample, so the results cannot be generalized outside the areas in which the samples were taken. The survey had four purposes:
· to determine the effect of allowing the respondents to report more than one race
· to determine the effect of collecting information about race, Hispanic origin and ancestry in a combined question
· to determine the effects of placing the Hispanic origin question immediately before the race question
· to test alternative terminology
The major findings included the following:
· The multiracial category had no effect on the Black, American Indian, White ethnic and Hispanic targeted samples.
· The "mark one or more" instruction had an effect on the Asian and Pacific Islander sample, unlike the other groups.
· Combining race and Hispanic origin reduced the nonresponse rate to the race question.
· For both Asian and Pacific Islanders and Hispanics, the information garnered from ancestry write-ins with a combined question was not comparable quantitatively to that obtained in the race question alone (i.e., the distribution could not be recreated and a separate question generated better data).
· As with the CPS, placing the Hispanic origin question before the race question reduced the nonresponse to Hispanic origin by non-Hispanics and reduced this group's reporting in the Other race category.
· Findings on terminology were consistent with the CPS.
Dr. Harrison highlighted the following conclusions from the three studies:
· Less than 2 percent of the population will exercise options to report more than one race, create a category or check more than one.
· The options for reporting more than one race do not have effects on estimates of White or Black populations, but do affect Asian-Pacific Islanders and American Indians. (Dr. Harrison explained many nuances of American Indian reporting.)
· The Asian-Pacific Islander population seems most likely to be affected by an option to report more than one race because of the high degree of intermarriage in this group. American Indians also are affected, for the same reason.
· Sequencing the Hispanic origin question from the race question produces [reduces?] a nonresponse to the Hispanic origin question and leads to lower reporting by Hispanics as Other race and higher reporting as White.
· Reporting as Hispanic alone would drop substantially in a combined question.
· The combined question works well for capturing the total Hispanic population but less well for capturing specific Hispanic origins. The same is true for Asian/Pacific-Islanders. There is a tradeoff between combined rates and details.
· The findings on terminology were consistent with those reported for the CPS.
The group discussed methodological issues and noted that self-identification is regarded as the gold standard for race and ethnicity.
Dr. Harrison said the survey findings have been made available to an OMB research subcommittee along with additional research, literature reviews and public comments. An interagency committee is reviewing everything and will issue a report to OMB the first week of July, to be followed by a 60-day comment period. The OMB will release its decision in October. The date for finalizing the Census dress rehearsal is October 15. Dr. Iezzoni noted that the Subcommittee will respond during the comment period.
Asked about reporting for people from Pakistan, India and the Middle East who have expressed dissatisfaction with the current categories, Dr. Harrison said OMB and Census are studying this issue and plan to release a detailed report within two months. He noted that the Census long form includes an ancestry question.
Dr. Carter-Pokras mentioned other surveys and statistical systems in which multiracial categories are an option, with the percentage using it ranging from two to eight percent. She also noted that a bill has been introduced in Congress to add the multiracial category; if it passed, it would supersede the OMB decision.
Dr. Golden pointed out that the OMB standards will be issued as a minimum and will not
preclude the collection of additional information. She expressed concern that the research reported on above is based on survey conditions, while the problems arise in respect to administrative records in practical settings.
Dr. Iezzoni thanked the panelists for their clear summaries.
Returning to its work plan, the Subcommittee accepted Dr. Amaro's offer to spearhead the response to the OMB recommendations, working with Dr. Carter-Pokras and Ms. Greenberg. Both the OMB notice and that of the Census Bureau will be addressed. The Subcommittee's response(s) will be forwarded to the full Committee for a final decision and recommendation to OMB and the Census Bureau. Dr. Amaro will be supplied with Dr. William's letter on this subject. She Carter-Pokras also will work on the linkage issue [?], in consultation with the Data Subcommittee.
The next issue for discussion was the Subcommittee's concerns about OASIS. Dr. Iezzoni asked the group to be thinking about possible actions as it hears testimony at the June 24 full NCVHS meeting. One possibility is a follow-up meeting to hear from patient advocates and people in the industry.
On concerns about HCFA's data release policies, the Subcommittee agreed that this is a general research issue that should be addressed by the full Committee. Dr. Iezzoni will consult with Dr. Starfield on this matter.
The group discussed in some detail concerns about the dearth of data on Puerto Ricans. Dr. Amaro noted that health issues for this population involve the island as well as the mainland and both have an impact on and are affected by the entire U.S. health system. Some Subcommittee members felt that this issue should not be removed from the context of the broader question of non-state entities, while others argued that Puerto Rico is a special case. It was noted that there is inconsistency within the Department regarding the treatment of Puerto Rico. Dr. Manderscheid described the inconsistency regarding the Pacific Territories and SAMHSA's efforts to address these issues at their most basic level, including assessing the communications structures that are preconditions of any kind of data system. SAMHSA also has been trying to get information on Puerto Rico, specifically.
Subcommittee members agreed to devote a meeting to these issues. Dr. Detmer suggested talking with Dr. Sondik about this and other Subcommittee concerns, in view of the strategic planning NCHS is embarking on.
Turning to item II-B of the framework she outlined for Subcommittee business, Dr. Iezzoni asked the group to consider what large, critical issue it would like to address in depth. She agreed with Ms. Ward that the most important action is taking place in regions and states, and the question is how a national-level Committee like NCVHS can make a difference. They agreed on focusing recommendations on things that can be done regionally. The Subcommittee agreed to focus on Medicaid managed care, which Mr. Van Amburg said is "politically hot in the states" and a focus of considerable concern. This relates closely to children's health and to the health of under- and uninsured people. The Subcommittee will revisit this plan in a July meeting that will focus on the OMB announcement.
I hereby certify that, to the best of my knowledge, the
foregoing summary of minutes is accurate and complete.
/s/ Lisa I. Iezzoni November 7, 1997
Lisa I. Iezzoni, M.D., M.S., Chair Date