The Subcommittee on Population-Specific Issues of the National Committee on Vital and Health Statistics was convened on Monday, September 8 at the Sheraton Hotel in Washington, D.C. The meeting was open to the public. Present:
Lisa Iezzoni, M.D., Chair
Hortensia Amaro, Ph.D.
Vincent Mor, Ph.D.
George Van Amburg
M. Elizabeth Ward
Staff
Carolyn Rimes, HCFA
Olivia Carter Pokras, Ph.D.
Patricia Golden, NCHS
Paul Placek, National Center for Health Statistics (NCHS)
Michele Adler, Office of the Assistant Secretary for Planning and Evaluation
Howard Bradley, Social Security Administration
Neil Russell, NCHS
Gerry Adler, Health Care Financing Administration
The Subcommittee plans to follow up on these items:
The NCVHS Subcommittee on Population Specific Issues met on September 8 during a break-out of the full Committee meeting, to hear presentations about five disability-related activities of the federal government:
· Revision of the International Classification of Impairments, Disabilities and Handicaps (ICIDH)
· Development of disability questions for the Year 2000 Census
· The Disability Evaluation Study (DES) of the Social Security Administration
· The Disability Supplement of the National Health Interview Survey
· The Medicare Current Beneficiary Survey
Dr. Iezzoni explained that the Subcommittee is in search of the best way to collect information on patients' functional status, as part of the Committee's work on core data elements for health care encounters.
After Dr. Placek described the ICIDH revision process, the Subcommittee talked with him about some of its limitations, the reasons for the classification's slow adoption in the U.S., and the recent growth of interest in it here. He stressed the importance of replacing the myriad of politically determined disability definitions in existence today with a standard classification.
On the proposed disability questions for the Census, the Subcommittee had several concerns about their content, the way they are being developed, and the means of testing and evaluation. They plan to comment on the questions when they are published in the Federal Register.
With the DES, the Subcommittee focused particularly on sampling issues. They made several suggestions and offered to be an ongoing resource on this survey, which is still under development and which was reviewed by IOM, under contract to SSA for review of research related to the Disability Process Redesign.
At the conclusion of the presentations, Dr. Iezzoni remarked on the challenge presented by the diverse ways that people think about disability. Asked to comment on how the Subcommittee might be constructive, Dr. Placek and Ms. Adler stressed the great progress that has been made in disability data in just a few years and the Committee's key contributions to that progress. Ms. Adler highlighted the need for the scientific development of questions on disability for children and adults. She also observed that as the work environment changes and assistive devices evolve, notions of disability need to be rethought. Dr. Placek reemphasized the need for a gold standard that would standardize data not only among U.S. disability programs but internationally.
The Subcommittee decided to ask the Subcommittee on Data Needs, Standards and Security to work with them on functional status coding. They also discussed the need to monitor access to assistive devices across economic groups and geographic areas, especially in respect to the impact of Medicaid managed care.
Following introductions, Dr. Iezzoni explained that the purpose of this Subcommittee meeting was to learn about a few of the federal government's 45 definitions of disability. She welcomed the first speaker, Dr. Paul Placek of NCHS.
The North American Collaborating Center (NACC) is housed at the National Center for Health Statistics, which is working closely with the Canadian Institute for Health Information on the revision of ICIDH. ICIDH was published in 1980 and has been under revision since 1992. The current revision title is the International Classification of Impairments, Activities and Participation. Information about the classification and the protocols for beta testing can be downloaded from the World Health Organization Website-- http://www.who.org/programmes/mnh/mnh/ems/icidh/icidh.htm.
The revision is being translated into a dozen languages, and testing includes linguistic analysis and three other approaches. The National Center for Environmental Health, part of CDC, is supporting additional research using ICIDH-2. Field trials are beginning to test the beta draft, and will end in December 1997. Then WHO consultants will produce another draft in time for the March 1998 meeting. In 1999, a refined version will go to the World Health Assembly for approval.
Applications of ICIDH and ICIDH-2 include RADIUS, DISTAT (a database of the U.N. Statistical Division), and Bay Area Managed Care (on which Gretchen Swanson reported in the June 1997 San Francisco NCVHS hearing).
Dr. Placek described the structure of the classification, which is hierarchical and can be collapsed. He reviewed the conceptual scheme and the three classification systems that comprise it: organs that don't work (impairment), something one can or cannot do (disability/activity), and a societal disadvantage one experiences because of the disability (handicap/participation). There are no automatic interrelationships among the categories. ICIDH has 1,200 codes and ICIDH-2 has close to 3,000, grouped into ten chapters.
In response to questions from the Subcommittee, Dr. Placek said the ICIDH-2 impairment codes are similar but not identical to ICD-10 diagnostic codes. The two use similar groupings and chapter headings. Unlike ICD, ICIDH identifies the organ system that does not work but does not identify why. The system contains considerable information on psychiatric diseases and cognitive impairments. Bridges can be, but have not been, built to DSM-4. Inter-rater reliability testing was done on ICIDH-1 and is in process for ICIDH-2.
Dr. Mor commented that for ICD-9 and DSM-4, reliability deteriorates outside of the major classes and for ICD-9 it is very erratic for the last two digits. He speculated that the same would apply to ICIDH. Dr. Placek agreed that this is a problem for "the whole ICD family." The exception is Chapter 5 of ICD-10, which was extensively tested by NIMH.
Continuing his presentation, he said training tools have been developed for ICIDH, including a computerized one called Code IDH that is on DOS and Windows and is inexpensive. He predicted that a contract will be in place within six months to develop standardized coding procedures for ICIDH-2.
Dr. Iezzoni explained that the Subcommittee is in search of what it should recommend as the best way to collect information on how patients are doing, in conjunction with the standardization of information on the health care encounter. She asked why ICIDH is not more widely used in the U.S.
Dr. Placek pointed out, first, that there are 45 disability definitions in the U.S. because legislative aides have created definitions based on political realities for various Congressionally funded programs. None of the definitions meet any U.S. or world standard. He speculated that the ICIDH has not been more widely accepted in the U.S. because it was not developed by Americans and "we're very ethnocentric." He added that federal agencies are beginning to pay more attention, possibly because Americans are involved in the revision. He cited the National Center for Environmental Health, which has three projects to develop assessment tools, and NCEH, which is building a new BRFSS disability system using ICIDH. In response to another question, he said that any disability survey can be back-coded to ICIDH. Surveys will probably be back coded only to a one- or two-digit level. Four areas of disability are being targeted in the 25 BRFSS questions.
Asked how the questions will be structured on the test to get consistent and reliable data, he said they have been tested in the NCHS cognitive lab.
On this topic, the Subcommittee heard from Michele Adler of ASPE, who noted that the decennial census is very important from a research perspective. Briefly reviewing the history of the Census in this regard, she noted that the first one (1790) was only for apportionment and then social variables were gradually added. Health and disability variables were added in 1830, but disability questions were not continued.
Because of limited space on the Census, a statutory basis or program need is required for all components, including disability data. There are many statutory uses of disability data, but nothing says it should be gotten at local levels from the decennial census. Therefore, disability is included as a "program need."
Ms. Adler showed sample questions from the last three censuses, and noted that considerable progress has been made. Disability was only viewed in terms of ability to work in 1970. A new question in 1980 on transportation accessibility proved impossible to understand. In 1990, two questions were added as a proxy for IADLs and for ADLs, and generated estimates were not inconsistent with other surveys. Unfortunately, a recent national content pretest showed a high index of reliability for the ADL question, as a result of which the Census Bureau initially displaced them. This decision was taken despite expert arguments that the poor results were a function of the methodology, relating to the sensitivity of ADL and IADL questions to change over time and a change in mode, as well as to the respondent.
The Census Bureau then designed a series of questions on disability which actually expanded the space on the decennial census dedicated to disability, which Ms. Adler called "an incredible victory." Still, the space is very limited and she noted the difficulty of deciding what few questions will pick up as many aspects as possible and make sense of the complex phenomenon of disability.
At the request of federal agencies that wanted a review of the disability questions, OMB convened an interagency work group which worked intensively over the summer. It generated a version that is being tested in a cognitive lab along with the Census version. There may be a dress rehearsal in the spring, in which the final version will be included. The work group wanted questions that were relevant to children as well as adults. The version they developed includes a broad mental impairment question, a sensory question, one on physical aspects, and one on roles. Because the Department of Justice wants to identify geographic areas where people with disabilities are subject to job discrimination, another question will target need for work accommodations and accessibility.
Asked about languages, Ms. Adler said the questions will only be tested in English, with ten interviews per question. The Department of Justice has pushed for multiple languages, big print, and TVD or TYY for the Census, and all of these issues are being pursued by someone at the Census Bureau.
She agreed with Dr. Amaro that it is important to look at how many people need special equipment to carry out their duties in the home, but added that there is no room for it on the Census. She noted that a major focus of the ADA is employment discrimination.
Dr. Iezzoni said the Subcommittee may want to comment on the questions, and Ms. Adler said there would not be a public comment period. Dr. Mor cautioned her in the cognitive testing to watch for differing interpretations of the word "need." She added that the American Statistical Association has been asked to develop another test of reliability for the health disability questions.
Dr. Amaro called attention to a question that appears to refer only to physical disability and might be missed by someone with a psychiatric disorder or addiction.
Howard Bradley of the Social Security Administration reported that the RFP for the DES will be out in the next six months. The priorities for the study have changed since it was first conceived in 1983, stimulated by concern in SSA and Congress about the growth of the disability rolls. Several short-term projects were funded but the DES, the most extensive study, has been delayed.
The DES is a national sample of adults. The biggest questions are how many adults meet SSA's current eligibility criteria and what their characteristics are. Another question is how SSA could better evaluate adults in terms of their ability to work. In addition, what accounts for the fact that some people who could qualify for disability continue to work? The final question is what can be done in the future to continue to monitor progress.
In this context, disability has a work-related definition. The DES will collect work history, income level, and current work activity. The most important part of the study is the disability-related factors, which will be collected from self-reports and medical and mental status examinations by clinicians. There will be about 5,500 respondents, who will be paid. Subcommittee members noted that a bias will result from the fact that the economic value of people's time will affect their willingness to participate. Asked how to deal with this, they suggested paying well and with a sliding scale. They also suggested oversampling to ensure a sufficient number of events of interest.
Asked about the language(s) in which the survey will be conducted, Mr. Bradley said he did not think they had talked about this. Dr. Amaro pointed out that different language groups may answer the questions differently, and Dr. Iezzoni added that this may be the case as well for different geographic areas. Dr. Mor commented that there are big sampling issues related to the overlay of state disability policies.
In sum, Dr. Iezzoni expressed concern on behalf of the Subcommittee about sampling strategy and other matters. Mr. Bradley said they are aware of these issues and are trying to hire a project officer for the survey. Also, an IOM committee has reviewed the DES as part of IOM's review of all research related to SSA's Disability Process Redesign (http://www.nap.edu/readingroom/records/0309058899.htm/). Dr. Iezzoni offered the assistance of this Subcommittee as well, and Mr. Bradley invited them to send him comments, which he will feed into the process. He promised to supply the Subcommittee with written information on the study.
Continuing his presentation, he said that under the redesign, the SSA committed to research on changing the disability decision process and greater emphasis on direct evaluation of functions related to work. So the DES is intended to provide a national sample of adults for testing of functional tests. Another purpose is to provide data on what enables people to work in today's economy, with an eye to improving vocational assessments at the time of intake and job placement services. The final purpose involves linking the DES with other surveys and administrative data to monitor the levels of disability in the future.
Dr. Russell noted that the NHIS uses a nationally representative probability sample. He presented information on the Disability Supplement (NHIS-D) conducted in 1994 and 1995, which dealt with a wide range of issues. It does not use a single definition of disability, but rather can be used for a variety of definitions, including information that would allow the approximate definition used by the ADA. The Disability Supplement was coordinated by 18 federal agencies and one private foundation. He noted that it focuses on children and their special needs.
The NHIS-D Phase I was administered after the annual HIS core questionnaire. The Phase I questionnaire was the major screener for Phase II, the Disability Followback Survey (DFS). Both phases are face-to-face, and both permit the use of proxies. The DFS uses four questionnaires, three for different age groups and the fourth to collect information on the effects of post-polio syndrome.
The 1994 Phase I data have been released. The response rate was 87 percent, with no bias from the drop-outs. Phase II is expected in early 1998. Phase I of 1995 data are due in December 1997, and Phase II in spring 1998. All the HIS supplements fielded in 1994 -- e.g., family resources, access to care, and health insurance -- are linkable by person, using a unique number (not the SSN). The data can be analyzed by age groups, policy groups, and so on.
Gerry Adler of HCFA reported that his agency has had an ongoing survey of Medicare beneficiaries since 1991. The MCBS was stimulated by the need to estimate the sources and objects of health care spending, and to enable the Office of the Actuary to model the consequences of alternative program policies. The survey is now housed in the Office of Strategic Planning. The survey uses a nationally representative random sample of from 12,000 to 16,000 persons. It is longitudinal, giving an ongoing record of services, coverage and payments. The interview is conducted in the household or long-term care facility in person, using CAPI. In addition to a broad array of standard questions, others can be added on timely topics in supplementary sections. The data are linked to respondents' Medicare claims files.
The sampling uses six age strata, oversampling for the under-65 and over-85 groups. Every year the sample is augmented with new enrollees, and people are rotated out after four years. Proxy respondents may answer. After collecting utilization and payment data they get a full accounting of charges and payments from all sources. Mr. Adler said managed care patients are asked for the same information, but because they do not have good claims HCFA has to rely on interview data. He acknowledged that people may not remember correctly, but said amalgamated data on utilization can be used. He then gave some examples of questions and findings, including health status in different age groups. Finally, he called attention to a report Westat, the contractor, will produce annually, "The Health and Health Care of the Medicare Population."
Reflecting on all five presentations, Dr. Iezzoni remarked on the challenge presented by the diverse ways that people think about disability. She asked Dr. Placek and Ms. Adler to comment on how the Committee might be constructive.
Ms. Adler observed that when she presented five or six years ago to two NCVHS Subcommittees, there was virtually no information on disability. The Committee was very helpful at that time in supporting the notion of a Disability Supplement, for which it helped develop the mental health questions. Today, there are disability data from several sources, and the question is what to do with them. The Disability Survey is the first and only comprehensive survey of disability in children in the U.S. It raises many research issues and methodological issues. A core set of questions on children's disability and on working age and elderly people needs to be developed more scientifically than is being done for the Year 2000 Census, with recommendations on data elements that work.
Ms. Adler also called attention to the fact that changes in the work environment are changing the meaning of disability. People in the special education field are now saying that anyone, regardless of disability, can use a computer, and thus they can function. In addition, assistive devices are taking off, and notions of work environment and work accommodation are changing. Finally, she mentioned some methodological issues that the Subcommittee could help with.
Dr. Placek echoed Ms. Adler in his acknowledgment of the Committee's role. He noted that several years ago NCVHS hosted hearings on ICIDH. It was the largest gathering in the U.S. on that topic, and served as a jumping-off point for involvement in the revision process. He pointed out the plethora of non-comparable definitions and ways of assessing disability being used in the U.S. today, making it impossible for programs to talk to each other statistically and for the U.S. to talk to other countries. He said his work with ICIDH has convinced him of the need to look for a gold standard to use in framing disability questions and analysis.
Dr. Iezzoni suggested that the Subcommittee on Data Needs, Standards and Security might want to join this Subcommittee in thinking about how to code disability information. She thanked the panelists for their presentations.
Ms. Ward commented on the need to know whether employers and schools are responding to demands of ADA, and Ms. Adler agreed that measures of the environment are needed in respect to disability, to get at questions of accessibility.
Dr. Amaro asked if there is information on the availability of the technology that improves functioning and participation that compares different socio-economic groups and geographic areas, to determine whose lives "are really being improved by all this technology." A related question is the impact of managed care on access to these devices. Dr. Iezzoni observed that this is especially important prospectively, in terms of whether the data systems exist to monitor whether persons in Medicaid managed care are getting the assistive technologies they need.
In conclusion, she highlighted these issues for follow-up:
Dr. Iezzoni then adjourned the meeting.
I hereby certify that, to the best of my knowledge, the foregoing
summary of minutes is accurate and complete.
/s/ Lisa I. Iezzoni 12-08-97
Lisa I. Iezzoni, Chair Date