[This Transcript is Unedited]
Renaissance Hotel
999 9th Street, NW
Washington, DC 20001
Call to Order, Welcome and Introductions, Review of Agenda - Dr. John Lumpkin, MD, MPH, Director, Illinois Department of Public Health
Update from the Department - Data Council - James Scanlon, ASPE
Update from the Department - Data Standards
HHS Implementation of PL 104-191 Health Insurance Portability and Accountability Act of 1996 - Dr. William Braithwaite, ASPE
Data Standards Quality Panel: Racial and Ethnic Minorities
Dr. Carolyn Clancy, AHRQ
Dr. Olivia Carter-Pokras, OMB
Dr. David Nerenz, Michigan State University
Review Draft Functional Status Report - Action Item 6/28 - Dr. Lisa Iezzoni
Briefing on the Use of Race and Ethnicity in Health and Social Statistics
Dr. Jennifer Madans, NCHS
Dr. Jorge del Pinal, Census
Roderick Harrison, Joint Center for Political and Economic Studies
Review Early Draft Letter to Secretary on HIPAA Standards - Action 6/28 - Dr. Simon Cohn
DR. LUMPKIN: Good morning. My name is John Lumpkin, and for those of you who are here in the room or recognize my voice, you may be wondering why I tell you who I am. And it's not just for the people who are, unfortunately, behind us. So I apologize for my back but we are, as usual, going out over the Internet and for people who are not able to be here, it's important for all of us, of course, have our voices recognized as we will do during the introductions.
We have quite a full agenda for this meeting. We have a lot of materials to cover, so we will probably need to get started. I just have one bit of good news to report and that is that in Illinois, the last two months -- as you know in April in May, I tend to not be able to travel much out of state, but this year we were actually successful in that our budget for tobacco has grown in the last few years from about $1.6 million to $48 million for anti-tobacco activities. So we're really looking forward to an opportunity to have a big impact on smoking in our state as has occurred in other states. We're pleased with the process.
We're going to start off with introductions and then we will proceed to go through the agenda. So we will start on my right.
MS. GREENBERG: Good morning. I'm Marjorie Greenberg from the National Center for Health Statistics, Centers for Disease Control and Prevention and Executive Secretary to the Committee.
DR. LUMPKIN: And your ICF code is?
MS. GREENBERG: I know my ICD code.
DR. LUMPKIN: Marjorie broke her foot on a trip. It's metatarsal. It's a transient disability.
MS. COLTIN: I'm Kathryn Coltin, Harvard Pilgrim Health Care. I'm a member of the Committee.
DR. FRIEDMAN: I'm Dan Friedman with the Massachusetts Department of Public Health and a member of the Committee and, John, I'm looking forward to when the VA can institute video streaming along with our Internet.
MS. FYFFE: Kathleen Fyffe. I work for the Health Insurance Association of America and I'm a member of the Committee. I'm also obligated to tell you that I represent the HIPAA on the National Uniform Billing Committee, the National Uniform Claims Committee, the National Council for Prescription Drug programs.
DR. LENGERICH: Gene Lengerich, Penn State University and a member of the Committee.
DR. ROTHSTEIN: I'm Mark Rothstein, University of Louisville, School of Medicine, member of the Committee.
DR. SHORTLIFFE: I'm Ted Shortliffe from Columbia University. I'm a member of the Committee.
DR. ZUBELDIA: I'm Kep Zubeldia with Claredi Corporation, a member of the Committee.
MS. FROHBOESE: Good morning. I'm Robinsue Frohboese, the principal deputy and acting director of the Office for Civil Rights at HHS.
DR. BRAITHWAITE: Bill Braithwaite from HHS, staff to the Committee.
MR. ALTERESQUE: I'm Louis Alteresque with the General Counsel's Office at HHS working on privacy.
MR. BLAIR: Jeff Blair with the Medical Records Institute and member of the Committee.
DR. NEWACHECK: Paul Newacheck, University of California-San Francisco, member of the Committee.
DR. MAYS: Vickie Mays, University of California-Los Angeles, member of the Committee.
DR. COHN: I'm Simon Cohn and I'm a practicing physician and the National Director for Health Information Policy for Kaiser Permanente.
DR. STARFIELD: I'm Barbara Starfield from the Johns Hopkins University and member of the Committee.
MR. SCANLON: I'm Jim Scanlon from HHS, ASPE, and I'm Executive Staff Director for the Committee.
DR. LUMPKIN: Okay. We're going to do introductions of those in the audience.
MS. ADLER: Jackie Adler, and I'm with NCHS.
MR. WILDER: Good morning. I'm Tom Wilder with the American Association of Health Plans.
MS. CURRY: Good morning. I'm Ann Curry with the Academy of Managed Care Pharmacy.
MS. CANON: Susan Canon, writer for the Committee.
DR. ZIZONI: Lee Zizoni, Harvard Medical School, Emeritus Committee member.
MS. BERRY: Erica Berry, Staff Attorney for the National Health Law Program.
MS. WEIL: Violet Weil, Office of Minority Health.
MR. STEINDEL: Steve Steindel, Centers for Disease Control and Prevention, staff to the NHII workgroup.
MS. HORLICK: Gail Horlick, Centers for Disease Control and Prevention, staff to the Privacy and Confidentiality Subcommittee.
MR. NEARING: David Nearing from Michigan State University.
MS. JACKSON: Debbie Jackson, NCHS, staff for the Committee.
MS. WILLIAMSON: Michelle Williamson, CEC, National Center for Health Statistics.
MS. WOLFER: Heather Wolfer from the College of American Pathologists.
MR. MONGELLO: David Mongello, College of American Pathologists.
MR. PETTICORD: Doug Petticord for the Coalition for Health Information Policy.
MS. CANTOR: Julie Cantor, Weinberg-Guidant Corporation.
MR. FANNING: John Fanning, HHS.
MR. STOSS: Eric Stoss, Multinational Business Services.
MR. BELTMAN: Warren Beltman, Center for Regulatory Effectiveness.
DR. DUFFY: Katherine Duffy, American Medical Association.
MS. HAMMOND: Allison Hammond representing the American Association for Occupational Health Nurses.
MR. TAYLOR: Shaw Taylor, Public Health Foundation.
MS. PARNECK: Andrea Parneck from the Office of the Assistant Secretary for Planning and Evaluation at HHS.
MS. BISHOP: Susan Bishop, American Pharmaceutical Association.
MR. ENGER: Stan Enger, Agency of Research and Quality and staff to the Quality workgroup.
MR. CLAIRE: Jean Claire, contractor.
MS. FRISCH: Linda Frisch, HHS.
MS. NORTH: Marsha North, Intern from the University of Alabama-Birmingham. I'm completing my internship in ASPE.
MR. HITCHCOCK: Bill Hitchcock HHS, staff to the Committee.
MS. ROLLISON: Marietta Rollison, CDC, National Center for Health Statistics.
MS. WHITE: Gracie White, NCHS.
MS. JONES: Katherine Jones, NCHS.
DR. LUMPKIN: Welcome. Are there any other conflict statements from members of the Committee other than mine? I have a teenage son.
(Laughter.)
MS. FYFFE: Let's not go there.
DR. LUMPKIN: Okay. Review of the agenda. Any changes? Hearing none, we'll proceed to the update from the Department and I'll start with Jim.
MR. SCANLON: Thank you, John. Let me report on a few developments since our last meeting in February, and Bill Braithwaite will report on where we are with the HIPAA data standards and Robinsue will talk about where we are with the privacy regulation.
Let me focus on a couple of other things, population-based health and other issues. I believe I gave you just a little description at the February meeting of a study that HHS has been asked to do for the National Academy of Sciences relating to the adequacy of race and ethnicity information in health systems administered by HHS or by any of the partners with whom HHS interacts.
The Public Health Service Act amendments last year called for a study at the National Academy of Sciences. Let me tell you a little bit about it because I think the Committee may want to interact on a more formal basis or certainly on a more interactive basis with the Academy on this study.
Again, this is required by law and it directs HHS to fund the Academy of Sciences to conduct a comprehensive study of HHS's data collection systems and practices and any data collection or reporting systems required under any of our programs relating to the collection of data on race or ethnicity including other federal data collection systems such as the Social Security Administration with which the Department interacts to collect such data.
And then not later than one year after the date of enactment of the law itself, the Academies is to prepare a report to HHS and several of the committees in Congress that does the following and I'll pass out this precis for you to look at.
First of all, identifies the data needed to support efforts to evaluate the effects of socio-economic status, race, and ethnicity on access, healthcare, and other services and on disparities in health and other social outcomes. The data needed to enforce existing projections for equal access to healthcare. So it's quite a task and quite a scope.
Secondly, examine the effectiveness of the systems and practices of HHS and the effectiveness of these systems, including tribal agencies in the private sector in collecting and analyzing such data and to prepare a report that contains recommendations for ensuring that HHS, in administering its entire programs and activities, collects or causes to be collected, reliable and complete information relating to race and ethnicity.
So it's quite a scope, quite a task. It would also include projections about the cost burdens and so on that might be associated with full implementation of the recommendations.
Within HHS, ASPE is putting together a funding consortium of civil rights as part of that to fund the study at the Academy. We're probably within a week or two of actually beginning the effort with the National Academy of Sciences. This would probably be a 12-member expert panel at the Academy, and we'll work closely with them in getting the panel set up.
It's the usual Academy approach, probably 5-6 meetings. We're proposing that the Academy consider a workshop to deal with these issues in the private sector where they can bring in folks from plans, insurance companies, and providers as well to look at what the obstacles may be, what the practical issues may be and to try to get us a report on that aspect as well. So we're about to start, and I think the Committee itself may want to interact on a more formal basis.
DR. LUMPKIN: All of that in one year?
MR. SCANLON: Yes. One year from the date of enactment which was probably last December.
DR. MAYS: Can we get a copy?
MR. SCANLON: Let me pass out this precis. This is actually the language of the statute itself.
MS. GREENBERG: This is separate apparently from the panel that's looking at the problem of racial and ethnic disparities in health care and potential threats, potential means to eliminate those disparities which Dr. Mays has already testified to.
MR. SCANLON: Yes, that's right. Now, we may want to think as we go along that the Committee may, and obviously the Academy does these things independently, but we can normally work out arrangements in terms of liaison and even possibly membership between the Committee here and that study as well.
MS. FYFFE: Jim, are there going to be public notices in the FEDERAL REGISTER so that people will know about this in order to get industry input?
MR. SCANLON: The Academy now is subject to something like a FACA(?) requirement. They're not exactly like the Committee, but they announced the formation of Committees and they ask for nominations. So we'll notify the Committee as soon as that is done, but there should be room for recommendations.
DR. LUMPKIN: But a lot of their work is done in executive session. So they'll have public sessions and closed sessions.
DR. STARFIELD: Do you have any feel for how this is going to play as the first idea talks about socio-economics, race, and ethnicity and the third only addresses race and ethnicity and the rationale behind that or how they're going to deal with that?
MR. SCANLON: No, I think, Barbara, I'm remembering how this requirement came about and it was my recollection that there was a fair amount of discussion about how can the private sector particularly and how can the healthcare data, for example the Medicaid and Medicare data, how can race ethnicity information be improved in those systems specifically.
You'll recall that there was a fair amount of confusion. I don't think there's any confusion in the law or the regulations, but there clearly was some confusion among plans and others about their ability to collect this information and I think, in fact, HCFA issued some clarifying letter about collecting such information, that it was both legal and obviously careful how it's used.
At any rate, I think the Senate, at that time, was interested in moving along a little further and how could they improve the situation. They clearly didn't want to absolutely mandate reporting this kind of information particularly in activities not related to federal programs.
I think they wanted the Academy to take a look at how you bring about this situation as to the barriers. I think it's less so in the federal reporting systems. We're pretty much collecting race and ethnicity information in all of our major data collections.
There are some wrinkles here and there. Our problem there tends to be more the size of the sample and the size of the cells. In all of the administrative data systems now we're at least aiming to collect this. It's not perfect obviously and there are still some problems.
In terms of getting private sector data, apparently there was a real mix and there was a fair amount of confusion about what could be collected and how it could be reported. I think this was an attempt to clarify that and give it to the Academy which is viewed as an objective, independent body, and get the best thinking and give them some recommendations.
DR. STARFIELD: I was referring specifically to this discrepancy between them, the FCS.
MR. SCANLON: I don't know, Barbara. I think the Academy will have to sort this out. My guess is that it will focus primarily on healthcare systems and probably public health systems. How they will deal with the whole FCS issue, I 'm not sure.
Let me just quickly bring you up-to-date on two other things. In terms of the data collections plans within HHS, the Data Council is looking at what is probably will be the next stage of data and survey integration effort. You'll recall that we started several years back a process to look at our major surveys and how to integrate them and coordinate them a little better.
The first stage of that effort was the HHS Survey Integration Plan which I think you've been briefed on and we're actually well along. A second focus was on integrating employer-based surveys. And this involved HHS plus the Labor Department and Census and others.
Then we were to look at health systems data and now for a variety of reasons, it looks like the long term care data area seems to be shaping up as an area that provides some opportunities in integrating data across several of our systems in HHS.
Under the Data Council's strategy group, we've put together an interagency workgroup to pull together some of the major needs for health services research and data in the area of longterm care, put together a plan to address those needs in a coordinated way and we have some initial plans under development.
I think we'll probably offer to brief population subcommittee and the full committee is that's desirable on where we are. It's in the initial stages, but we had reached a stage where, for a variety of reasons including periodicity of surveys, everyone wanted to take a look at our longterm care area.
This involves nursing home surveys, community based populations with IADLs and ADLs and so on, our Medicare current beneficiary survey and our MEPS survey as well. So we have a group looking at this and there are enhancements being thought about in all of these surveys. I won't go into the six areas. I won't go into this today. Maybe I'll talk about it at the population committee this afternoon.
DR. SCHWARTZ: Could I just ask a questions about surveys of the institutionalized population? We've done nursing home surveys for a number of years and there have not been really comprehensive surveys of the institutionalized population particularly those under 65 and children. Is there any consideration of going below age 65 in these surveys or going beyond just nursing homes?
MR. SCANLON: There are, Paul. I think this is one of the very reasons why people wanted to take a step back here and say this is an area that's changing a fair amount. Before we automatically conduct another nationalistic survey, let's think about this and I think this will begin with an effort to develop a comprehensive inventory or master facility inventory of longterm care places.
This would include not just the nursing homes, but probably assisted living types of communities as well as the younger population as well. In fact, I think in MEPS there is a vision that MEPS would include an oversample to get at adults with ADLs as well as children with one or more limitations as well as the caregiver folks as well.
DR. LUMPKIN: Let me just make a suggestion when this does come up in case I'm not available to be in the population when it's discussed. Some thought still needs to be given to the fact that, for instance, in our agency, we have about 250 surveyors who are in longterm care facilities who carry CMS purchased laptops to carry out certification surveys and with some sort of linkage between the survey process that's going on and the actual on-site validation. I think that that connection really hasn't been made and is something that you may want to be looked at as we're reviewing this concept of nursing home surveys.
MR. SCANLON: I think, in fact, that the survey certification system is being looked at as a source of some of the basic facility inventory information as well, as well as the administrative data systems that are coming online, the dataset for nursing homes and the OASIS dataset for home health agencies. There will be a look, John, and I think if the Committee can help us with the, how can we improve this from the administrative side?
DR. LUMPKIN: Yes, because the important thing is, and from many years of experience, what's in the system or written on paper isn't necessarily what's there.
MR. SCANLON: I have one final item to report quickly. Dr. Lumpkin and Dr. Friedman were at our Data Council meeting earlier this month when we talked about an effort to integrate the HHS Health and Human Services data on not so much on the data collection end, but on the dissemination end.
And what we have in mind here, and actually we started this under the Data Council's auspices, is a Web-based one stop portal or gateway to virtually all of the HHS statistics, holdings in HHS. Publicly available statistics I'm talking about.
So this would be perhaps one or two clicks off the HHS home page would basically be a gateway to all of our program evaluations, published research in progress. The Web pages of virtually all of our statistics programs as well as some other links to state data resources, for example, among other things.
This would include state data that HHS agencies have developed as well as links to the state Web sites as well. It would be links to data policy Web sites such as the Committee here to our Data Council and to OMB and other places as well.
We have actually moved fairly well along on this and the only hold up now is that we've asked someone to develop a search tool to make it a little easier to find the topics that you might be looking for in data. Actually, we were able to get someone who has done some work for National Library of Medicine and Library of Congress.
So we're probably not more than a month or so away from having at least a test available. This would be an attempt to pull together from a user-friendly perspective, clearly our agencies have tremendous resources and holdings in statistics. Some of them are unique in the world and highly regarded.
In many cases, you almost have to be an aficionado to know exactly where to look. If you don't know about MEPS or about expenditure data or health insurance data and so on, it's not that easy to navigate unless you already know the layout. This would be a way of pulling it all together at the front end of HHS.
You'd be able to search, you'd be able to look from our major data systems. In many cases, the Web sites we take you to would not only have data, but they allow you to actually construct a table or to request a table through a protected confidential research data center. So John and Dan were there for the discussion and it looks like something. Certain portal-type pages are becoming more popular on the Web and this provides a way for us to pull all our stuff together.
DR. FRIEDMAN: It's a very exciting endeavor and I think if you want Beta testers, you may be able to find some people on the Committee who would volunteer.
MR. SCANLON: Actually, a lot of you have had a lot of experience particularly in the states with Web sites like this where you've actually had interactive Web sites as well. I think you've found as well that the user isn't interested in the agency sponsoring the data or the name that we call it or whether we call it evaluation or statistics or survey. They're interested in the data and the easier we can make it, the better.
Just one final word, we, as you know with federal agencies, are usually involved in three years of budget planning in any given year. This is FY01 in the federal government so agencies are progressing on that basis.
Before the Congress is the request for the FY02 budget, President's request. We've had hearings, but no final action on those yet. But that will be coming in the months ahead. And we've begun planning for the FY03 budget and, here again, the Data Council will be pulling together a summary and a cross-cutting analysis that the agency will be coming forward with in their FY03 budget submissions.
Again, we'll be looking for interagency initiatives particularly such as a long term care data initiative. Patient safety, I think all of you know is a big focus and you'll be hearing about that tomorrow. Quality measurement issues, quality reporting issues are also fairly big inter-agency initiatives that we'll be looking at coordinating. Let me stop there.
DR. BRAITHWAITE: My report is very brief. With respect to all of the HIPAA transaction rules except for privacy, nothing has happened. As you know, we expect the final security rule, the final provider identifier rule, and the final employer identifier rule to be out some time this year once we get our skids greased and the rulemaking process working again.
We have NPRMs in the works for the claims attachment transaction standard and for fixes or updates to the transaction standards that have already been issued in final. One of which has been preconveyed by response to the letter that the Committee sent to the Secretary about the recommendation that the code set for drugs or pharmaceutical reporting on the institutional and professional guides be removed.
The letter response from the Secretary says that he plans to issue an NPRM which would withdraw the NDC standard for all of the transactions except the retail pharmacy transactions. Then, of course, the industry has just about completed the process of recommending changes to the implementation guides for the adopted standards. And as soon as that process is finished, then HHS will begin its rulemaking process to adopt those changes. So there's a lot of stuff that's about to happen, but not a whole lot to report on what has happened since our last meeting.
MR. SCANLON: Thank you. Any questions?
MS. FROHBOESE: Good morning again. As Bill has said, I, on the other hand, have a lot to report to all of you because a lot has happened since I last met with you at the end of February. Unfortunately, I need to leave around 10:00. When I looked through the agenda, there are just so many issues that are really important to the Office for Civil Rights. Clearly, Jim's discussion of data collection and the panel that will be on right after I make my remarks on racial and ethnic disparities are critical issues to the Office.
I, unfortunately, will not be able to meet with the privacy subcommittee tomorrow, but Louis Alteresque who is sitting next to Bill and is in our Office of General Counsel within the Office for Civil Rights as well as a couple of other members of our privacy team will be able to join you. So I'll look forward to hearing about what comes out of that subcommittee meeting tomorrow morning.
When I met with you at the end of February, I told you at that time that we found out that there was a little glitch in the Privacy Rule and that was the Congressional Review Act. The papers that needed to get to Congress transmitting a copy of the Privacy Rule, didn't get there as anticipated when it was published in late December.
So I brought you the news that the rule would not go into effect at the end of February and that, in fact, we had to start counting the 60-day implementation point from mid-February when the rule was actually transmitted to Congress.
Shortly after I met with you, the Secretary decided that given the fact that we had this built-in period. It was really important to hear from healthcare providers, hear from the industry, hear from consumers about the final rule that was published in late December.
And he had two major issues that he particularly was looking for in these comments. The first was that given the fact that there were a number of changes in the final rule that was published in late December from its earlier version in November 1999, he wanted to get comments, feedback on the major changes that were made in between the initial notice and the promulgation of the final rule.
The second area in which the Department was particularly interested in getting comments was about the workability of the rule. You know that we've been very concerned and interested about any unintended consequences of the rule in terms of actually being able to implement it and make it workable in the healthcare context.
So the rule was reopened for comments for a 30-day period and we didn't get quite the volume that we received in the first go 'round where there were more than 40,000 comments, but we did receive over 11,000 comments in that very short period of time and what we did to be able to review, analyze, synthesize the comments in a way where we could give meaningful information to the Secretary for decision-making purposes was to reconvene the privacy team that was in place, that actually drafted the regulation.
Thanks to the wonderful people at ASPE who then headed up this project and brought back in the contractors to deal with the volume that was coming in of the mail, we did this amazing 30-day intense period of reviewing the comments. And I'll tell you that most of the 11,000 comments came in on day number 29 or night number 29.
But we did seriously review and take into consideration all of the comments and met with the Secretary, and as you know, the Secretary and the President made the decision to go forward with the rule as written on April 14th.
So the Privacy Rule did take effect in mid-April, and we're now into month two of its implementation and what I would like to focus on is what the activities of the Office for Civil Rights will be during this two-year interim period where healthcare providers, clearing houses, health plans have the opportunity to come into compliance with the rule.
Before I do that, I'd like to go back to the comments just for a moment because it was very interesting. We found that as a result of all of the comments that we were receiving that there really are a lot of misconceptions out there about the Privacy Rule and we talked about some of those misconceptions when I was here in late February, but it really became apparent through the comments that we received and some outreach that we were doing that there was a significant level of misunderstanding and that the Department needed to do something very quickly to address these major misconceptions.
The other thing that we found out, and think this was no surprise certainly to the privacy team who worked so very hard over a concentrated period of time to grapple with all of the complex issues of privacy, is that there continues to be a difference of opinion out there about some pretty major areas, everything from written consent and how that works, contractual arrangements with business associates, and even the fact that the rule covers oral communications as well as electronic transmissions.
To respond to the misconceptions, the Secretary decided that it really was critical to get something out very quickly that would help the healthcare industry debunk the myths that are out there. And so, the privacy team after finishing opening up the last envelope in e-mail, went back to work to develop a very comprehensive set of guidance that should be published in the very near future. I would say within the next few weeks.
It will cover all of the major areas in which we are aware that there are major kinds of misconceptions. It will be in a format of some brief factual information and then have a lot of questions and answers. We will be posting that guidance on our Web site which is WWW.HHS.gov\ocr\HIPAA and you'll be able to get into that guidance.
We'll be using mailing lists and other ways of disseminating the guidance. It will cover the major areas of consent. What is meant by minimum necessary information? World communications, there will be a segment on research where there are a lot of questions an concerns and I know that certainly this Committee is very interested in those issues.
Business associates, marketing. The public and consumers have a lot of concerns in this area. Government access to health information another big top as well as access by parents and parents of more children to health information.
The Secretary, in addition to this guidance, the Secretary is also taking the comments into account to determine whether or not any changes need to be made to the rule. We know that there are some basic workability issues that do need to be addressed. I'm sure you all heard of the major snafu that a pharmacist, as the rule is currently written, a pharmacist cannot fill a prescription before getting written consent from the individual.
Clearly, those kinds of glitches in the rule, were not intended by the Department, but the drafters of the rule and we wan to make sure the that the rule allows the normal course of health operations to get forward and ,does not cause a major disruptive influence as we also carry out the major responsibility to insuring and protecting privacy of the information.
So as we consider the kinds of workability and perhaps some policy changes that may need to be made in the rule, we will be discussing these proposed changes with the Secretary, and we anticipate that any modifications to the rule would take place within the first year of implement of the rule so that there will be sufficient advanced notice and any changes that do take place, we anticipate should not really disrupt the kinds of actions that are already being put in place to comply with the rule.
From the Office for Civil Rights' perspective, we are going to be spending the next two years on a couple of major activities. The first activity is building our capacity from within. As I mentioned earlier, we are really relying on the expertise of individuals who are o he intra-agency privacy team and still drawing from that collaborative effort from ASPE and the Healthcare Financing Administration now CMS.
NIH and other important components of the Department who have been a significant part of building the privacy rule and we're building our capacity by hiring individuals from within the Department as well as outside the Department.
During the past month, we posted a number of advertisements for key positions, key senior positions and really hope to be able to fill those positions within the next month. We're looking forward to the next time that I meet with all of you to be able to introduce our privacy team that will be in place in the Office for Civil Rights.
We're going to be spending our efforts during this two-year period on public education, outreach, and technical assistance. That really is going to the focal point of our efforts. We know that healthcare providers, the industry clearing houses, health plans are really yearning for a lot of assistance from the Department in understanding their obligations under the rule and we really, although we are enforcement agency.
Typically, OCR's approach to all civil rights enforcement, and it will be the same under the privacy rule, is to really work with providers to ensure compliance so that, in effect, our enforcement role really is minimized because the outcome, of course, that we're trying to achieve is compliance and to the extent that we can do that in a cooperative way, that really will be our focus.
We're going to be spending a lot of time developing technical assistance material. This set of guidance that I mentioned will really be the first step of a series of guidance that we anticipate issuing. We know that one size does not fit all and that the Privacy Rule covers everything from the smallest health plan to the largest healthcare provider.
So we really will be gearing technical assistance material to the particular needs of different kinds of covered entities under the rule. We also will be doing a lot of meetings with major groups associations and speeches at conferences.
We really have been doing that even during the period of time that the rule had been reopened for comments and certainly since the rule took effect in final form in mid-April, we really have been back on the speaking circuit and the privacy team has been out there meeting with major groups.
Last week I, unfortunately, could not make it out to San Francisco for the HIPAA West conference with the large audience. Bill was there alive in person and I had to appear via videotape. But we really are making every effort to get the word out and the really be accessible to groups and organizations and I think this is one way in which NCVHS could really be helpful to us in helping us identify groups that we really should be reaching out to make sure we are getting maximum exposure, coverage, and really spreading the word about the Privacy Rule in a way that is most effective and comprehensive.
That will be a major thrust of what we are doing with focusing in on the technical assistance. Simultaneously with that, we are putting together a team and Louis certainly will be instrumental in that from the General Counsel's side in actually developing an enforcement regulation. This regulation will be our roadmap of our enforcement activities and is intended to give consumers and covered entities more specific details about actually how the enforcement process will work.
Again, I think that you'll see in this enforcement regulation that while the Office for Civil Rights, the Department takes its enforcement role very seriously, that we also do want to focus on voluntary compliance wherever possible in getting the speediest possible resolution of problems where they exist.
That is an overview of the major highlights of what has happened since past February. And, as I said, members of my staff look forward to meeting with the privacy subcommittee tomorrow to discuss some issues in greater detail, but if there are any questions here, I would be happy to take them.
DR. COHN: Thank you, first of all, for the update. In relationship to the enforcement rule that you're beginning to ponder, is this specifically just for privacy or is it privacy and security or is it privacy, security and all the HIPAA regs?
MR. ALTERESQUE: The way the statute was written, the penalties, the civil monetary penalties, the criminal penalties are written in a broad way so they apply to any of the regulations that come under administrative simplification.
So the enforcement regulation would apply to all. It doesn't mean that as we develop that that we can't have differences between the two. In the privacy regulation now, we have a very formal complaint process. That complaint and compliance and enforcement section will be moved so it applies to all of the administrative simplification regulations. There may be some modifications that make sense.
There may be a different model that will be used for the transaction standard than there is for privacy in terms of dealing with complaints. The same things with dealing with penalties and how those are used. The statute provides certain restrictions, but there may be flexibility.
So we will look to see. There should be differences between how the different rules are handled under the enforcement regulation. That regulation, like the modifications that Robinsue referenced, would all be presented in proposed form, FEDERAL REGISTER notice, the whole process. There will be plenty of opportunity to get involved in that.
MS. FYFFE: Given that some of the other administrative simplification regs are not going to be out perhaps this year or maybe into next year, is there going to be flexibility in the enforcement reg so that the unknown will be covered or do you expect to have to re-issue the enforcement reg again?
MR. ALTERESQUE: If we find out we put out a new rule and that enforcement rule doesn't exactly apply as most appropriately to that rule, then there would be some modification in that. Just like we put out the compliance and enforcement section in the Privacy Rule and knowing that there may be some changes made as the other rules are developed further.
MS. FROHBOESE: The timing and the coordination of all the rules we recognize is a very tricky and difficult challenge and we really were cognizant of that and are trying to do our best to have parts of HIPAA move forward while being very conscious of the fact that it is a whole integrated picture that we need to keep in mind.
MR. ALTERESQUE: I must say that one of the questions that we frequently get or one of the concerns is in terms of the effective date versus the compliance date and we've got in a number of calls from researchers and from HMOs that say that providers and others are arguing that they cannot disclose certain information because it would violate the Privacy Rule, but they cannot disclose it now.
What we've tried to say as sincerely as we can is that you cannot use the Privacy Rule as a defense. The compliance date is 2003. Until such time, you cannot argue that the Privacy Rule prevents you from sharing information, that there is any limitation on what you do that's imposed because of the Privacy Rule. So you may be hearing that also.
DR. COHN: One other question about the guidelines that you're developing. Some have speculated about the volume that we should be expecting and that appears on the Web site. Is it going to be over or under 1,000 pages?
MS. FROHBOESE: I'm pleased to report significantly under.
DR. FITZMAURICE: Would it be possible for anything else that comes out to be a little bit easier at the beginning of the document so that we could find out where the actual rule is? It was kind of difficult last time.
MS. FROHBOESE: We are trying, in keeping with the fact that guidance will be under 1,000 pages, we really are trying to make this guidance user-friendly and there is always that tension between being legally accurate and being able to describe things in understandable terms.
And particularly when this guidance is intended for a very diverse audience, we want consumers to be able to pick this up and understand what their rights are. We also want covered entities to be able to use this guidance and understand their responsibilities.
We tried to strike that balance and we will include all sorts of things like a table of contents and putting it in a format that will be easy to use, but we do welcome any feedback since, as I said, we intend this to be the first in a series of a lot of guidance documents.
MR. BLAIR: I do want to thank you and Louis, Robin, for giving us the breadth of your description and being able to get into some of the detail. These briefings are a very valuable connection for the National Committee.
You're probably aware that we have a charter responsibility to advise the Secretary on Health Data Policy Issues as well as a HIPAA mandate to advise the Secretary on administrative simplification. On the transactions and codes side, we've held numerous hearing and people come in and talk to us about their implementation issues and they result in good letters that come out of the National Committee to the Secretary with specific recommendations.
I think that can also be helpful to you if you have particular issues that you want the national committee to get input from if there are particular groups that have a forum or talking about their implementation issues and then the wisdom of the national committee around here.
If you could identify those groups or those issues either directly to some of the staff, to John Lumpkin, or to the Privacy Committee, that would help us in giving advice to the Secretary.
MS. FROHBOESE: Absolutely. I know when I met with you in February, there was some consideration to holding hearings and getting information about workability and policy issues. I think that the comment period really did provide us, as I said, with breadth of feedback, but I also know that clearly as we need to implement the rule, there will be issues coming to the national committee for guidance and assistance would be very helpful.
MR. ALTERESQUE: And maybe we could talk about that some more tomorrow morning.
MS. GREENBERG: We actually have scheduled a hearing for August 21st with the Subcommittee on Standards and Security is meeting on the 20th and the 21st, half day on the 21st. Starting after lunch, we have scheduled a meeting of the Subcommittee on Privacy, a hearing on the 21st, 22nd, and I believe it's going off to a half day on the 23rd.
There was a conference call which your staff participated in. Tomorrow morning that's a major topic on the agenda of the subcommittee on privacy and confidentiality to talk about who should be invited. Of course, there will be an open invitation through the FEDERAL REGISTER notice of anyone who wants to provide testimony.
DR. SHORTLIFFE: Yes, one thing that would be very helpful to our subcommittee tomorrow is if your representatives could work with us in trying to figure out the areas in which public comment and testimony would be especially valuable.
One of the things that we want to stress at these hearings is that this is a very complicated and controversial regulation. This is not a 2.5 day session to retry the issue of whether there ought to be a regulation or statute I guess for that matter.
We want to the extent possible to alert the public and alert the people that we invite these are the issues of particular concern and so that that they can address them and so that we can develop recommendations that would be most helpful to the Secretary.
MS. FROHBOESE: That's good and the timing will be good because by then hopefully our guidance will be out and we'll have a little bit more information.
DR. LUMPKIN: I have a question then I know you have to leave and I'm going to give you the leeway not to necessarily answer right at this moment.
MS. FROHBOESE: This is where I can make my graceful exit.
DR. LUMPKIN: In Illinois, our agency, which we believe is not a covered entity - the Department of Public Health, and we're still doing that review - we do, however, believe that the privacy standard rule sets a standard for interactions with the public, a review of what we ask for and information that comes in data and also what we send out.
Many of the standard practices and organizational practices that would be the outcome of a Privacy Rule for a healthcare organization such as a Chief Privacy Officer and privacy review committee is part of our plan to respond. My question is whether or not it's your authority or others in HHS, to what extent is HHS applying this new standard for privacy to its operations?
MS. FROHBOESE: To the extent that we are covered providers and certainly in the Medicare program, that's one very large area, the Department is actively working on plans to come into compliance and the Privacy Rule does apply to the covered entity components within HHS.
John Fanning and Jim Scanlon have really been instrumental through the privacy of subcommittee of data in pulling together the components within HHS that are covered entities as well as other federal agencies.
So I neglected to mention that a major thrust of our activities with technical assistance, but clearly focusing federal on other federal agencies and internally within HHS will also be a major part of what we're doing in the next years. The newly named CMS has already put together a very extensive plan and timetable for coming into compliance. It really is a major operation. So we take that very seriously.
DR. LUMPKIN: I would add to Robinsue's remarks, all of our agencies in HHS have been asked to look at the implementation planning for the privacy regulation and not just those that are covered entities, those that interact with covered entities, those that are research agencies, regulatory agencies, oversight agencies, public health agencies and how they would interact with covered entities and other issues. They're actually coming together fairly nicely. n addition, they've been asked for resources in their FY03 budget in terms of actually implementing this. The Department's very serious about this.
MS. FROHBOESE: It's one of the Secretary's eight priority areas for cross-cutting initiatives is privacy.
DR. LUMPKIN: I think it would be important for us as a committee to report to the privacy subcommittee or the full committee to just have an opportunity to see the plan once it's together to give any advice that we may to HHS on the implementation for their internal operations.
Thank you for coming. Unless there are any other questions, we're going to move on to the next item on the agenda. If the panelists will introduce themselves.
DR. CLANCY: Good morning. I'm Carolyn Clancy. I direct the Center for Outcomes and Effectiveness Research at the Agency for Healthcare Research and Quality(AHRQ). I'm now struggling for that new technique one needs in public speaking which is to look poised while the technology is coming, but I do believe it's coming.
My job this morning is to quickly set the state for issues related to thinking about disparities in healthcare as a quality problem and specifically what are some of the data challenges. Then I'm going to be followed by Olivia and David who are going to discuss some of the more specific issues related to legal issues and public perceptions around collecting this data.
Just to again put this in context. Obviously, there has been a great deal of interest in the past several years in public reporting on clinical performance and notwithstanding concerns about perverse incentives. That is to say that this might cause providers or practitioners to avoid sicker patients so that they wouldn't look bad.
This movement has gathered steam so to speak. From the perspective of public policy since there is an ever growing number of publicly funded beneficiaries and managed care arrangements, this seemed like a very interesting opportunity to examine disparities as an important quality of care problem.
On this slide, I have results from one study that was published just about a year ago in the NEW ENGLAND JOURNAL and it's very reflective of many, many studies documenting two issues. I think there are essentially two messages on this slide.
The first is consistent with literally hundreds of studies you see that African American Medicare beneficiaries were less likely to receive evidenced-based treatment, in this case, reprofusion for patients who just had an acute heart attack, than their White counterparts.
The second take-home message, of course, is ample room for improvement for all beneficiaries since it's hard to believe that 59 percent of eligible patients receiving the treatment represents the Everest of our ambition.
It's no news that there are non-clinical determinants of health outcomes including patient and practitioner characteristics, characteristics of the setting, patient preferences, reimbursement. You can name many others. This has been a vigorous field of inquiry in health services research.
What's a little unclear even at this point it what are the pathways? Do these factors influence what happens before patients get to healthcare once they get into the healthcare system? Moreover, this comes from some data that Arlene Berman and I have analyzed from the Medicare current beneficiary survey and this is just looking at older women, Medicare beneficiaries and looking at the independent contributions of low education, low income, and raised to activities of daily living limitations or the likelihood of having three or more chronic conditions. The only point here is that these are independently predicted.
So, in theory then, disparities associated with socio-economic position, race, and ethnicity represent a critical quality improvement opportunity, but we can't possibly get there without relevant and reliable data.
It's our contention that performance measures should be stratified and that populationwide measures potentially should be adjusted if one were going to adjust payments as has begun to be done in the United Kingdom and some other countries. And, moreover, that one should ideally be able to account both for socio-economic position as well as race and ethnicity.
Again, conceptually, this is easy for me to tell you about the real challenges, where do you get the real data to be able to do this? And just to make this pictorially concrete, this graph is constructed from data that you can just get from NCQA's Web site where they report on the performance of various managed care plans.
What's shown here are the three overall averages of plans that publicly report in 1996-1998 and then a three-year average for an order, advice to quit smoking, treatment with Beta blockers for patients who've had a heart attack, breast cancer screening, cervical cancer screening, and lagging behind is eye exams for people with diabetes.
In theory, one ought to be able, if we had the data, to click on this and look at how individual plans or plans as a group are doing for different racial and ethnic groups. So, what are the challenges then? I've mentioned absence of data and leadership issues.
There are big concerns very, very familiar to this committee related to privacy and data collection and how exactly do we get these data? It's not such a well-kept secret that much of the data we have on race and ethnicity does not actually come from individuals reporting how they perceive themselves, but actually the interaction between a clerk of some sort and one who looks a the patient sand cheeks out when they believe to be the most likely ethnic background.
There are concerns about misuse of data as well as healthcare organization resistance and inertia for a number of reasons. There are a number controversies that are going to settled by data, but data could help us to understand more about the pathways in which was see these big differences in healthcare as well as overall health.
The big question is: How much of these observed differences in health associated with race, ethnicity, and socio-economic position are amenable to changes in healthcare. And how many of them exert their influence through other pathways that are pretty much independent or certainly beyond the control of healthcare systems.
That raises big questions about the boundaries of accountability. If a plan or a healthcare system has a substantial number of patients enrolled who are of different racial and ethnic backgrounds or lower socio-economic position, to what extent should they be held accountable?
Should they be paid more if, indeed, those are the kinds of patients enrolled in their plan? Then very, very importantly do we know enough to intervene? So, wither data. There is no clear inventory this moment about what currently exists within health plans of how relevant data should or could be collected which is really the purpose of this discussion today.
Almost two years ago, we held an exploratory meeting with managed care leaders, employers, and others, including Kathy Coltin thankfully, sponsored by the Office of Minority Health, the Commonwealth Fund, and AHRQ. So I wanted to just describe to you a little bit about what we learned at that meeting and also some test orders that we have in progress right now.
Some of the major themes that came out of this meeting is that first, there is enormous variability in the current capacity of plans with respect to the data that they collect on race, ethnicity. At this particular meeting, we had leaders from several plans that were either solely or largely Medicaid managed care. They were extremely enthusiastic and in many cases actually did have the kind of data that would be required.
The second thing that came up was, this is important. It's hard to ignore the results of multiple studies that have been published, but this only applied to managed care organizations or should preferred provider organizations and other fee-for-service arrangements be a part of this as well.
This is particularly acute as many managed care plans are feeling the remnants of a backlash so to speak. So why should they have to measured on their performance? Again, the point, do we know enough to intervene? How much evidence do we need on that point?
At this meeting, I think there was clear interest on the part of purchasers and employers depending on the ethnic background of their employers and how to collect the data both accurately and efficiently remains a very big challenge and both Olivia and David will be speaking to that a little bit.
Some of the other themes are concerns about the perceived misuse of data and the privacy of personal health information. Many concerns expressed about potential legal barriers both in terms of creating a perception that a plan might be perceived to be redlining and selectively enrolling patients as well as liability if performance differences were made public.
There was a great deal of interest in developing what might be called the business case for collecting the requisite data. The issue is something like this: many of the plans represented at this meeting had explored the feasibility in the past of collecting all kinds of data on patients once they enroll.
This could be health risk assessments; it could be a variety of types of information. And yet, being able to do that efficiently was something no one was able to come up with a very clear strategy for. Then finally, well known to this committee, is the data on income and education particularly income are much harder to obtain.
So with that I wanted to tell you about a couple of task orders that we have in progress right now that will help us begin to shed more light on this area. Last year, we started a new contract for an integrated delivery systems research network and there are nine systems.
They are not integrated delivery systems in the way that you might normally think of them. Our definition for these purposes was very, very broad. A consortium could apply that had to represent a wide array of healthcare facilities including hospitals, outpatient clinics, and they either had to have their own research team or subcontract with one.
So a task order was given to Denver Health to examine the capacity within three of these systems to conduct studies on the impact of race, ethnicity on access, use and outcomes of care. The second task order was given to United Care and RAND which, again, are a consortium under this network, to examine variations in cardiac performance measures associated both with race, ethnicity, and socio-economic status.
In this case, they're exploring the feasibility of using different proxy measures such as linking to zip codes, census data to actually use that as a proxy for SES. So to tell you a little more about examining the capacity within some of these systems, in this case, Denver Health is looking at the Intermountain Healthcare System and the New York Presbyterian Hospital System as well as their own healthcare system.
What you see here in yellow on this slide are healthcare systems that do not collect any data on race/ethnicity as a matter of policy. So that's sort of a first cut.
What they're trying to get to is to describe in detail the data that are collected and how and where. In some cases, for example, from New York State, this is required for hospital data. In other areas, there are different strategies.
Kathy Coltin in our meeting made us aware of some efforts by some plans in Massachusetts to link to vital statistic data for either birth weight or to other data sets. It's taken them a number of months to be able to make sure that they're actually understanding the questions in the same way so that they'll come up with commonly reported information.
They're then going to analyze the utility of each distinct dataset for research and identify strategies to enhance that capacity and then report back to us on what datasets exist now that could be used to conduct this kind of research.
Finally, they're going to present the results to the Chief Clinical Officers of these organizations to try to think about what would be the next steps and what would be some strategies that could be developed to collect this data. With that, I think I will close and turn it over to Olivia.
The Agency is going to be publishing at the same time, a report on disparities in healthcare and we will be going out within the next couple of months with a public call for input and suggestions for measures. I can assure you that all of you will be on the A-List to hear from us about that. We have not even had an opportunity to brief the Department formally about this, but I did want the people on this committee to know about that.
DR. CARTER-POKRAS: There is a very interesting study that we're going to be sharing some preliminary results from and we're continuing the conversation that Dr. Clancy had initiated on improving healthcare quality data for racial and ethnic minorities.
To bring you up-to-date from the presentation Dr. Coluree(?) had given during the past year in regards to why present to the National Committee on Vital and Health Statistics that there is an increasing interest by civil rights and healthcare communities in addressing and monitoring disparities in healthcare access and quality.
With the Minority Health and Health Disparities in Research Act of 2000, we have various components within that legislation that is hopefully going to move us further along. One of those is in calling for a National Academy of Sciences study of the Departments data system and its ability to look at race/ethnic data and disparities in healthcare, access, and quality.
That is a project that the Office of the Assistant Secretary for Planning and Evaluation is taking the lead on. We've just gotten word that Congress is asking for a briefing on how we're doing in implementing the various components of this particular legislation.
What are some of the concerns of managed care that we heard from this meeting two years ago and subsequent to that? We're continuing to hear about the need to confirm with the obligations of the anti-discrimination laws that are on the books including the 1964 Civil Rights Act as well as perceived legal barriers to collecting such data. Many still believe that it is illegal for them to collect these data.
We went to the academic medicine and managed care form in May and spoke to many of the academic health centers. At that point in time, they thought it was illegal. Dr. Jack Geiger was able to share with them the results of a couple of the reviews that I'm going to be sharing with you today, that bottom line it's basically not illegal at the national level as well as the state level for most states.
Also, privacy and confidentiality concerns. There is new legislation that will be penning in the State of California in regards to this. The cost of data collection in terms of time as well as money. In healthcare consumer perceptions regarding such data, how the health plans are going to be using this information.
In addition to these concerns that are specific to managed care and healthcare, access, and quality, what are other concerns regarding the race/ethnic data collection? As they have been well documented over the years, there are discrepancies between self-identification and observed identification.
It's not clear in the healthcare arena, how that information is collected. We have some changes in preference and categories used over time. We have not had the same set of categories used at every census and you're going to be hearing a little bit more from Dr. Pinal later today. We need information on subgroups, socio-economic position, risk behaviors to help explain why these disparities are occurring.
In regards to the barriers to examining quality of care by race/ethnicity, there are perceived and real legality concerns. Another concern is that collection of reporting is not always required. Although it's required to follow the anti-discrimination laws, we do not require collection of that information that's essential for monitoring compliance with the anti-discrimination laws that are on the books.
There are additional data quality concerns as I mentioned before in addition to missing information, managed care penetration and global billing. That makes it very difficult to look and see if the disparities are continuing within a managed care population or any health plan population.
I did mention the privacy initiative within the State of California. We are gathering currently signatures for a petition to put on the ballot for next year, next March a racial privacy initiative. This will prohibit the state from classifying individuals by racial, ethnicity, color, or national origin in operation of public education, contracts or employment.
It also prohibits the classifying of individuals by race, ethnicity, and color for other operations with the exception when there is legislation and it requires a two-thirds vote and the Governor's approval, identifying a compelling state interest in the area of medical research although that is not defined and as when it's needed to comply with federal law or eligibility and the particular concern was loss of funding is they don't comply with eligibility. So this ii a middle of August is when they need to get their 700,000+ signatures in to get it on the ballot for next year.
Addressing legality concerns. The Department of Health and Human Services in response to the concerns that were raised at this managed care data workshop, has really engaged in a lot of activities with various partners.
One of the things we did is work with the Health Care Finance Administration to issue an operation policy letter to Medicare Plus Choice plan clarifying at the federal level, it's not illegal to collect this information as long as the information is used to improve healthcare quality for all and not to discriminate.
The Office for Civil Rights and the Office of the Surgeon General joined together in issuing letters to AETNA first and then to over 30 managed care and healthcare professional organizations to let them know that it's not illegal and to encourage them to use the information to improve healthcare quality for racial minorities.
The Office of Minority Health has funded a project with the National Health Law Program to review state laws and regulations governing the collection and reporting of data by health plans and insurers. The Commonwealth Fund has funded the Shier and the National Health Project to review federal laws and regulations for Medicare in the state children's health insurance program. So we've got back to back projects that have been working together to look at this issue of legality.
Today, I'm going to have Violet Woo, the project officer, who is going to share now the findings.
MS. WOO: As Dr. Clancy and Dr. Pokras mentioned, several of the managed care representatives mentioned that there may be legal barriers in the state legislation which prevents them from collecting race/ethnic information from their enrollees.
The National Health Law Program (NHLP) is looking at the state legislation to see if there is any written policy which prohibits or requires the collection of race/ethnic data from managed care organizations and health insurers.
At this point, they have found that four states specifically California, Maryland, New Hampshire and New Jersey have a specific legislation or a specific law and regulation which prohibits either health insurers or manager care organizations from race/ethnic identity on application forms.
All 50 states plus the District of Columbia require application forms or other forms used by health insures and managed care organizations be submitted to the Department for approval before they're distributed to enrollees.
A few states have mentioned that they would disapprove any form that's submitted requesting race/ethnic information. In some areas though, the ones that would disapprove the submitting of applications is not written in their policy, but that's a matter of practice.
The first one specifically has that in the document, in written law. Those four states prohibit the collection of race/ethnic data. Ironically, South Carolina requires the collection of race/ethnic by health and HMOs.
Each enrollment form used by HMOs requires collection and reporting, but it's not told when they should collect that information either at application or enrollment or post-enrollment.
Regarding primary language, the State of Texas is the only one that requires the collection of primary language data by HMOs also. Each enrollment form used by HMOs much include a space where the enrollee lists their primary language. I'm not sure if that's optional or required. My guess is that is optional information.
Regarding discrimination, 41/51 jurisdictions provides some level of protection of discrimination on the basis of race or national origin for health insurance consumers and/or managed care plan enrollees.
This is in regard to insurance, unfair trade practice, HMO statutes or the general civil right section of the law. Seven states have a level of protection provided by health care insurers is uncertain.
Some states, Alabama, Georgia, and Mississippi only provide protections against discrimination on the basis of the person's disability and not in specific to race/ethnic minority identification. Several states have varying degrees of protection from managed care and those covered by health insurance. So it's certain as to the exact translation and interpretation of state was and that's beyond our contractor's level of responsibility to know that in depth.
Regarding Medicaid, now the centers of Medicaid and Medicare services have left the requirement to collect race/ethnic data for Medicaid to the state. Most states do report race/ethnic identify to SMC, but it's not required. As far as S-CHIP, the State Children's Health Insurance Program, 47 states have been found to collect race/ethnic data from applicants.
Some of the written law indicate that this is voluntary and others note that excluding such information will not affect the eligibility status and those that have been following the S-CHIP regulation know that the final rule on the requirement to collect racial and ethnic information is now on hold. It's pending.
DR. CARTER-POKRAS: There is an interim final rule that's open for public comment by July 25th and we encourage any of you who have a concern regarding any of these issues to participate in the public comment period. The interim final rule includes a federal requirement for the reporting of race/ethnic for S-CHIP.
MS. WOO: But it excludes the requirement for the collection of primary language for S-CHIP applicants. Do you have any questions on health at this time that the states are implementing.
MR. SCANLON: Will there be a report summarizing findings?
MS. WOO: Yes, there will be. At this point, they did all their data collection by database as well as telephone conversations with the state representatives. The next phase is visits to selective states. They're going to do hands on and more in depth analysis. So a report indeed will be produced and distributed.
DR. CARTER-POKRAS: Sorry for the switching around, but this appears to be the best place to have the computer at this point in time. Because of limited amount of time, we weren't able to have all the project officers present, but we're sharing information that Dr. Ruth Perot from Shier has said that we're welcome to do so.
This is a Commonwealth funded study of the review of federal policy and practices. And what they were looking at is the federal laws and regulations that govern collection of data within the Department of Health and Human Services.
What they found out was that health related data collection by race, ethnicity, and primary language is legal and fully authorized under Title VI of the 1964 Civil Rights Act. In fact, that is the communication that went out to AETNA as well as the academic medicine and managed care form and the various academic health centers that are part of that.
No federal statutes prohibit collection of racial, ethnic or primary language health data. First, we heard from NHLP that in general state laws do not prohibit the collection and reporting of race/ethnic data. In fact, one state requires the collection and one state requires the collection and reporting of primary language data.
Now we're seeing for the federal level that there is no federal law or regulation that prohibits it. In fact, what they found out was that three require collection and are reporting race/ethnic data: Maternal and Child Health Block Grant and the SAMSIS and Grants for Children of Substance abusers.
They found that two require collection of demographic or enrollee characteristics without describing what those characteristics are: S-CHIP and the Ryan White Cure Act. As we mentioned, there's interim final regulation in regards to S-CHIP which will impose a requirement for the collection of race/ethnic data, but not for primary language and public comments are requested through July 25th. The FEDERAL REGISTER notice was published early this week.
In regards to the review of federal regulation, they found two health service regulations require race/ethnicity data collection and/or reporting: end stage renal disease program and SAMS application for prevention activities. As I mentioned, there are potential requirements not only for the S-CHIP regulations, but also for Medicaid managed care that are pending implementation.
They also reviewed various data collection vehicles within the Health and Human Services. The Department has pulled together an inventory of it's data systems of almost 200 data systems where they are funded and maintained by the Department of Health and Human Services and it has contact information. It's available on the Data Council's Web site.
When they received the actual data collection vehicles, they reviewed over 100, all of them included fields for race/ethnicity and our review of the data systems included in the Department's inventory found over 90 percent included it.
The few that did not include race/ethnic categories that are either consistent with the Office of Management and Budget Studies and/or do not collect race/ethnic data at all were ones that we were very much dependent on other individuals for collecting that information that was in the grant program and mostly surveillance activities. Here is the Web site for the inventory that I mentioned.
The National Academy of Sciences' study that I mentioned earlier asked for a comprehensive study of the Department's data collection systems and practices and any data collection or reporting systems that are required under any of the programs or activities of the Department relating to the collection on race/ethnicity including other federal data collection systems such as the Social Security Administration with which the Department interacts to collect relevant data on race/ethnicity.
Although the study has not started, we're in the process of procurement and hopefully, we'll be able to report at the next or subsequent meeting about how the study is going. As I mentioned, the Office of the Assistant Secretary for Planning and Evaluation is serving as project officer for this project.
Other projects addressing the data quality concerns, those mostly were addressing the legality concerns. The Office of Minority Health funded a project with the Public Health Foundation which is ongoing to access eh quality of race/ethnic data from state Medicaid agencies and foster collaborative analyses with public health departments.
We're looking at a wide gamut form percentage of missing information on race/ethnicity on Medicaid, as well as whether to use the information infomrat policy when a report has come out, whether they share that with public health departments, a wide range interest.
The project officer from Public Health Foundation is here today if you have any questions. The states' more detailed analyses have not quite yet been identified, but we're hoping to be able to be completed with this project by this fall.
The Health Care Financing Administration funded also a project to analyze state and Medicaid, maternal and child health data by race/ethnicity and they are going to look at data quality in this project. AHRQ is analyzing he CAPS data to look at because there was concerned for the managed care workshop that we aren't just focusing on those in managed care plans, that we are also looking at those who are in fee-for-service and looking at also those programs that the Department of Health and Human Services funds as well as the federal government.
In regards to HIPAA that the Commonwealth funded study with Shier and NHLP looked at. As you know, HIPAA requires the Department to adopt code sets for electronic transmission of health data. Each code set is designated as required, situational or not used.
I should mention that Dr. Braithwaite who just made a presentation has been very much key to this process and CDC and ASP have been taking the lead in developing the business case within our department to make sure that requesting this information is possible.
For healthcare encounters, currently, the race/ethnicity code set is not used which makes it more difficult for those who are interested in collecting that information at the point of encounter.
For enrollment, race/ethnicity code set is designated as situational and we're trying to change the not used which is the guide for the code set to at least situational. As I mentioned the business case has been developed and prevents, but has not been implemented for the next version, the revision of the guides.
There is another project that was funded by the Office of the Assistant Secretary for Planning and Evaluation that I just wanted to bring to your attention because it certainly impacts on the ability to monitor and enforce the anti-discrimination laws on the books as well as improve healthcare quality.
That is, a review of the literature conducted by Dr. Tom LeBees(?) of Johns Hopkins University with funding from the Office of the Assistant Secretary for Planning and EVALUATION. This is our measuring race/ethnic discrimination in healthcare settings. The final report is to be posted on the Data Council's Web site.
Just to summarize the findings, Dr. LeBees noted that we need routine monitoring of administrative records to screen for patterns of racial disparities in the use of medical procedures and that a second phase that could be considered is to send trained testers to facilities to determine the process whereby race/ethnic disparities are being produced.
Since we just received the report, the Department hasn't yet decided which approach they're going to be going with or how best to address these recommendations, but at least I'm sharing them with you today. We invite your thoughts about it.
Finally, I would like to end with a quote that was submitted to the Office of Management and Budget during the review of the federal standards for race/ethnic data that was completed in October of '97. That very eloquent statement was that we do need data, that measurable gains in advancing the civil rights agenda to bring all Americans into the economic, political, and social mainstream would have been extremely difficult if not impossible if we did not have adequate information on racial groups.
DR. MAYS: Can you tell me a little bit about where we are with S-CHIP because it's a little confusing as to say something about, one, what the proposed policy might be and, two, to get a sense of where we are?
DR. CARTER-POKRAS: Certainly. As you know, we've had a change of administration and that's required a review of all pending regulations and that's included in the one for the State Children's Health Insurance Program as well as Medicaid managed care.
And our department has worked with other partners to see what is the best way to address the concerns that were raised during the public comment period last Fall and what has happened subsequent to that is there was a FEDERAL REGISTER notice that was posted a couple of weeks ago, June 11, around there saying that they were still waiting a little bit longer to review the comments.
Earlier this week, they published an interim final regulation. That interim final regulation for the State Children's Health Insurance Program includes the January version of the requirement for racial data reporting.
We found that virtually all the states are already doing so. So this would not impose a burden to the states and that was a concern to some is that, is this going to be an additional burden of reporting. But the primary language information they decided to not include the interim final reg and the rationale that I read in the FEDERAL REGISTER earlier is because some states may decide that they want to include primary language. Some states may decide they want to use first language, but public comments are requested on this interim final regulation which includes all of the reading of the public comments and what they're doing in regards to this. That included social security number, cost sharing, other issues which you may want to take a look at. So I encourage any of you to participate in the public comment period.
DR. NERENZ: Good morning. I'm here to talk to you about a specific project that we're doing with a number of health plans around the country addressing the issue of whether health plans can actually obtain and use data on race/ethnicity of their members for purposes of stratified quality reports and, ultimately, for quality improvement.
DR. LUMPKIN: Could you identify yourself?
DR. NERENZ: I'm sorry. I'm Dave Nerenz from Michigan State University. The starting point for this, this project has actually been going in two phases starting about two years ago supported by the Commonwealth Fund and we've been connected over these two years with the Office of Minority Health and AHRQ and doing a workshop we talked about earlier.
The premise is that in a number of domains in healthcare, we're used to the idea of report cards that produce comparative reports of quality. By and large, these report cards do not contain data on race/ethnicity. So the reports are not stratified by race/ethnicity.
So the consumers who might be interested in choosing a health plan or hospital on the basis of quality of care for their particular group, would not be able to do so and also that organizations wishing to work on disparity issues or to improve quality of care for a specific target group, would also not be able to do so.
So that's really the gap that we're trying to step in. We have, as I said, been supported by the Commonwealth Fund. There has been a planning phase for the project that's already been completed. I'll share a little bit about that.
We've been working with quality of care in a number of different domains, but all at the health plan level and this summer, we're in the middle of phase two of the project which is a multi-plan demonstration project. In phase one, our objective was to identify and recommend a set of concepts, a set of quality domains, specific measures, high priority clinical conditions to be included in the report card and also to make some recommendations about how information can be collected and how reports would be generated.
This was a collaborative effort of three organizations: Henry Ford Health System and that was a time when I was at Henry Ford; Herman Healthcare System in Houston; and Loveless in Albuquerque.
The project was a little unique in that right at the beginning we sought the input of community leaders specifically from African-American and Hispanic communities. Many of the report card projects are built on technical expert opinion or technical expert built on scientific evidence. We also had that, but we wanted to get the views of community leaders about these questions of what measures, what concepts, what clinical conditions, how information should be collected.
We spent a total of three days with these folks spread out over a year's period of time addressing these questions and those inputs have shaped the project as it's gone forward. We had some very open-ended brainstorming meetings, but also put the groups through some formal priority setting exercises that led eventually to the choosing of measures to be worked on in the second phase of the project.
There is also a phase of pilot testing on a very small scale. We don't have time this morning for me to share the results of that, but perhaps if we have time for questions, I could get into that if you like.
The three broad policy recommendations of these groups first of all, was that when they thought about quality at the health plan level, they thought about something bigger than HEDIS. So you'll see as we go along that we do have some things that add to the HEDIS dataset.
They did quite strongly feel that it was appropriate for health plans to have data on race/ethnicity and to use that data for purposes of quality improvement and they also indicated that they were interested in two kinds of comparisons.
They wanted plans to be able to look at disparities in the quality of care within their own membership. So you're comparing within plan, across groups. But also from the consumer point of view, they were interested in the opportunity for members of a specific group to compare quality of care for that group across plans serving the same area. So again, as we've carried the project forward, we've tried to support both kinds of comparisons.
As I said, we're in the middle of phase two right now. What we're doing is actually working with nine plans to access the feasibility of various methods of obtaining data on race/ethnicity and using those data to generate stratified quality reports.
We started on this last summer. Right now, we're in the middle of data collection in a number of domains and we expect to have results out by the end of this calendar year. The questions are essentially very practical feasibility questions. Can plans actually get data according to certain methods that I'll indicate in just a minute?
Once you have that data, can you actually link it to other datasets at the plan level in order to produce stratified reports on quality of care and if you do that, what do you see when you look at variability within the plan and what do you see when you look at differences across plans for care of the same groups?
The content of this project as we're going along includes some of the HEDIS effectiveness of care measures, but not all of them. These are the ones that were ranked as the highest priority by the community leader groups. It includes analysis of the CAPS survey. These are basically the measurable domains of the CAPS survey.
We have three surveys and this is where we expand on HEDIS a bit. The community groups were particularly interested in concepts of a patient involvement and decision making, adequacy of information, adequacy of information for self-management. So we've designed some survey items about that and we have one survey that's being administered to parents of children with asthma.
We have a similar survey that includes a few different items designed to go to adult plan members who have diabetes. In prostrate cancer, this is interesting because this is of very great significance to the African-American community leader group.
It's difficult to find evidence-based guidelines for care either in the diagnosis or treatment of prostrate cancer, but the groups agree that there are certain concepts about adequacy of information, involvement in decision making that one might be able to obtain through surveys. So we are testing a survey on those topics administered to a sample of plan members with newly diagnosed prostrate cancer.
We also have a plan level survey about some dimensions of cultural linguistic competence. This survey is done by staff of the plan. It's not administered to plan members. It gets at a number of structural process aspects of cultural and linguistic confidence that were identified as high priority by our community leader groups. There's a survey that can done either in paper, pencil, or telephone.
These are the plans that are actively participating. There are nine plans. Somewhat diverse in representing at least different geographic areas of the country. It's interesting that we have two plans in Hawaii that are focusing on 4-5 groups that represent a breakdown of the Asian Pacific Islander group.
The other plans are working either on Hispanic or African-American plan members or perhaps both, but in Hawaii, the plans are trying to break down their reports by a finer grain analysis and they're using some self-report survey data to try and do that.
Very recently, we had an opportunity to expand the project. We have a HERSA contract to add six plans that are affiliated with HERSA. These plans are plans that are formally affiliated with community health centers. Care Oregon is one of the plans that's committed.
We're in active discussion with 7-8 other plans and don't have formal commitments yet. So I probably should not name them, but they'll essentially be doing the same things. They'll just be coming at it from a slightly different perspective.
The interesting thing for the group this morning is how are these plans doing this? Where are they getting the information? Without exception, the plans do not normally have race/ethnicity data on their members. So there are 4-5 different methods that these plans are testing.
Each one of them is doing 1-2 things. They're not all doing all of them. The plans that serve Medicaid enrollees are getting information from the state Medicaid agency and that information is obtained through the eligibility or enrollment process, and then once the plans have that, they're merging it with their other records to generate HEDIS reports.
For plans that are doing the HEDIS hybrid method, I presume most people are familiar with that, and it involves eventually identifying a sample and reviewing actual medical records. At that point, the plans are looking in the medical record, trying to find data on race/ethnicity.
The key question for us is simply whether it's there. We're not asking the question of whether it's accurate or perfect, but we're asking the question simply whether you can find it.
In some plans, they have a strong affiliation with a provider group that has its own clinical record system which includes race/ethnicity. So even though you do not have it at the plan level, you do have it at the provider level and as long as you can match, we think that's a feasible way.
We've done some pilot testing of that already. A couple of plans are using surveys and they're done on a broad cross-section of their members that includes a self-report item on race/ethnicity. The question there is can you match that with your clinical data that you need to generate the quality reports.
Also, we're trying a couple of things in the proxy area either based on geo-coding or at Loveless in New Mexico, they have a discerning recognition software that they've been testing for many years, that they can use to identify Hispanic plan members. That's worked pretty well for them.
So we're working oh that and we'll have data out. I can give you some anecdotal things if you would like on how things are going, but things are happening, people are merging these datasets, they're generating reports.
The question is long term. We know in a legal sense that plans may do this, but we still have the practical questions of can they do it? Are there feasible ways of getting this information?
In this project, we're trying to test several of them. We know that they can work in an absolute sense in the sense these plans are doing it, but we want to learn something about how much work is it? How many missteps do you do when you're trying to match a Medicaid data file to your member file?
We'll try to report some of that out. A lot of questions of should this be done? We've covered a lot of ground this morning on the general theme that it should be done, but there is still some concerns about privacy. There will probably be some concerns about whether there are HIPAA regulations effects on any of this that we need to work through.
Then we get into some more detailed questions. When, ideally, do you get this information? How do you do it? What are the legitimate sources? Which ones work? Which ones don't work? Must you always do this on the basis of self-report or are there other methods that work acceptably well?
Ultimately, what can you do with the information? If you have these stratified reports, can you use them effectively either for consumer choice, for internal quality improvement? I acknowledge support from HERSA. I apologize. In your handout version, we got the resources and services backwards. I did correct it here, but my apologies to HERSA and you about that.
DR. SHORTLIFFE: I might be correct by now.
DR. NERENZ: I don't think they've changed.
DR. LUMPKIN: Do we have any questions for the panel?
MS. COLTIN: I'd actually be interested, I know that in the process of trying to recruit plans to participate in your project, there were plans because I know ours was one of them, that just looked into could we do this in any way? What were the opportunities to participate and to try to do this and found that it was really very limited. So, I was curious whether those that are participating represent a significant fraction of those that you approached and whether those that were not able to participate had common barriers that they were encountering in terms of their ability to do the kinds of things that the nine plans are trying to do?
DR. NERENZ: That's a great question. I don't have the specific numbers, I'll give you ballpark figures. I think we mailed invitations to 30-40 plans. So in that sense, we're dealing with a 1/3-1/4 of those who were contacted.
I think we had serious discussions with perhaps 12-15 plans, meaning that they were interested and at least wanted to go to the point of asking questions, getting more information and then we have the nine who are involved. It's a little sense of what the flow is.
Clearly, this is a select group. I don't want to create the impression that this is a representative sample of all plans nationally. These are plans who have stepped forward. These are plans who have some unique capabilities, who were interested in doing this. So it is a select volunteer group.
But I think for this stage of the game, that's okay. If we see that there's a group that can do it. Then some of those lessons might be expanded. Also, I might mention that in the HERSA part of this project now we have talked with a couple of plans who have special barriers.
For example, a plan in New York was interested in doing this, but didn't feel ultimately capable because the reports that they provide to the Medicaid program are not standard HEDIS reports. So even if technically they could do this, they would not be producing standard things that would be comparable to the other plans in the project.
Likewise, I think there's a CAPS-like survey that is not exactly CAPS. Finally, they weren't concerned that they could actually get accurate complete data from the states. So it was a bit of a three strikes you're out.
So, clearly, there are plans who, in spirit, would like to do this, but simply cannot. They can't get the data. It's just not available through any of these mechanisms. What fraction that is nationally? We haven't tried to explore that. I can't say.
DR. LUMPKIN: Danny, you had a question?
DR. FRIEDMAN: Yes, just a quick question for Dr. Clancy. The data that you presented indicated that both race/ethnicity as well as economic position independently contribute to provision of healthcare services and outcomes as well. Yet, you also indicated with that little asterisk that data on economic position would be much more difficult to obtain. I was just wondering if you could really quickly explain that assertion.
DR. CLANCY: I'm not sure I have a clear, cogent explanation of why that is. I think the best example that I heard about not too long ago is National Cancer Institute has sponsored this prostate cancer outcomes study where a cohort of men are asked quite intimate questions about genital, urinary, gastro-intestinal, and sexual function and response rates for those items are way up there.
Response rates for income and education level, particularly income, are well below 5 percent. So I don't know if this has something to do with an American sense of that might be mutable or what the story is. That Medicare health outcome survey which will be publicly reported in the Fall, which is going to report on a 2-year change score in SF36 adjusted for comorbidity and so forth, has a lot of data on demographic characteristics, including race/ethnicity. There, again, the response rates on race/ethnicity have been much, much higher. Education is not bad, but income is terrible.
DR. STARFIELD: This is self-report, not when people are asked; is that what that is?
DR. CLANCY: Yes, and these are male surveys for the most part with telephone call.
DR. CARTER-POKRAS: Census found very similar findings too. It was actually a follow up of those who filled out the long form and they found over half the people who filled out the long form, found the questions to be invasive on income, but much less percentage on race/ethnicity.
DR. LENGERICH: This is a question for David. Just to follow up on the plans. What's the intention of the plans to use the data or the commitment that they made when they got into that study?
DR. NERENZ: The commitment is really within the scope of the project. We really crafted it was a researcher demonstration project. So there's a committment to the information that was sharing with the project staff and with each other. There is no commitment to public reporting of any of this.
I think the plans understand that that is the longterm direction of this, but at least under the current structure, there will be no public reporting of them by name. If they choose voluntarily to do something with it, we've not bound them in any way. But their commitment is to obtain the race/ethnicity data, produce the stratified reports, share data among other participants in the project who have a comparable population, say, the Medicaid enrollees and then have a report come out collectively out of the project.
DR. LENGERICH: Do you have a sense of how they will voluntarily do that? Have they indicated that sort of thing to you, how each plan voluntarily will use those data that are now collected and available? Are they going to modify their administrative plans, their marketing plans or anything of that nature with this data?
DR. NERENZ: I want to make sure I understand the question. In terms of using the data for the specific activities of the project, they've already made that commitment, they're doing it. If the question is specifically, how are they doing it or is the question, how are they actually going to use it to improve quality down the line?
I can't give you a definite answer. It really depends on what the data show. If there are essentially no disparities within plan, I think that will provide a certain comfort. It may make plans more willing to make the information public, but may serve as a less catalyst for going to work on internal issues. It would be the opposite if there were really enormous disparities seen, but since we haven't done the analysis yet, I don't know which way it's going to be.
In general, the mechanisms that they use for quality improvement across a number of domains would certainly be available here which means you can identify a target population. You can try to understand the processes that lead to the disparities you observe and then go through whatever you have in terms of incentives of CQI processes or policy changes to work on that, but it's a wide open range of what might conceivably be done. It all starts by what they see in the first place. What's the shape and scope of the problem?
DR. SHORTLIFFE: Yes. Vicky, please stop me at any point if anything I say is going to be dealt with in the next panel. It seems to me that our ability to get meaningful data on race/ethnicity will be made even more difficult over the coming years by two rather diametrically opposed political forces.
On the one hand, there's a group of people who say that the G-nome is sequenced and we know that it's 99.9 percent similar, that it's clear that race is a social construct, it has no scientific value and we shouldn't use race at all for anything and we're merely reinforcing the racial fallacy. On the other extreme are individuals who don't want to use race/ethnicity for the political endpoints that they see in terms of affirmative action, et cetera.
So on the political spectrum, you could argue that they're on the opposite sides, but what they're arguing for has the same affect on the ability to generate race/ethnicity data for health research. My question is: To what extent are your research plans taking those social/political currents into account?
DR. CARTER-POKRAS: I'm sorry. Are you addressing that to anybody in particular? If I could take a stab because I've been working in this area of racially active data for 20 years. There were many for a long time who said it was illegal to collect this information.
Now that we're somewhat addressing that legality concern, we are getting additional concerns that are being raised. That doesn't take away from the fact that we have segments of the population with differential access to services, differential access to education, live in poor housing, are more likely to live near toxic waste dumpsites.
That doesn't take that fact away. So there are many who are concerned that it's very similar to what's happened in various states where they have not allowed the reporting of abortions. Not allowing the state to report abortions, doesn't mean that the abortions are not occurring.
That is a real concern in the civil rights community. They want to make sure that the data are there. The best way to address these concerns is to take them one by one, and see where there are common areas of concern that you can work together. Certainly, those who are working in the genetic area with the human g-nome project are absolutely correct that race is not a biological construct, it's not a genetic construct.
It is a social/cultural construct. We are totally in agreement. Then we move to the next step. How best to measure that social cultural construct? How best to address the reasons behind the disparities that are occurring? We need to see where those common areas of interest end, have them join us in working together to address that.
Utopia, right? This is what we're working towards, but we do have an eliminating disparities goal for the Department for the year 2010 that's been incorporated in healthy people 2010 and we recognize we have to address these concerns and we have to reach out to folks we normally haven't been working to be able to achieve that goal.
DR. SHORTLIFFE: Your answer addressed the merits and I agree with you on the merits. The question really goes to the strategy of how you deal with these criticisms coming from both sides and it's a very difficult thing to do.
DR. CARTER-POKRAS: Again, we have to find out where the common areas of interest are and work from that because instead of saying they're completely in a different planet than we are, we are sharing some common concerns. We've got to find out what those are. Its not always easy, but that's what we've been doing in regards to the S-CHIP regulations as well as other areas where these discussions are occurring. They're not occurring just out there in the press. They're occurring in the Department of Human and Health Services today.
DR. CLANCY: I guess just to build on that. I think to some extent how the problem or challenge gets framed is very important. I think in some ways this recent forum sponsored by AETNA that Olivia mentioned was quite representative of this. I think that disparities have come to be seen as an issue of fairness rather than one of affirmative action.
I think fairness has a much different political slant than the concept of redressing past wrongs regardless of how one feels about affirmative action specifically which I think is very important. I also think that if one sees this as part of an over-arching goal to improve quality of care for everybody which is why I said there are two take-home messages from the slides showing disparities, one being disparities, the other being we're really doing pretty badly overall.
That becomes another strategy in terms for framing it. I'm not sure if that gets at your issue. I think the first argument that you presented is going to be a little bit more tough in scientific community rather than in the political.
DR. LUMPKIN: This has been a good discussion. I know that other people want to ask questions, but we have an extremely tight agenda today. So I'm going to have to move on. We're going to be taking a short break at this time. If you do have lingering questions, you can catch our speakers as they're wrapping up here.
(Short break.)
DR. LUMPKIN: As you know, for a number of years, we have been struggling with the issue of functional status. It's an area that we think if extremely important as a committee and we've ben pushing the envelope in our ability to classify and me recommendations. We've had some very hard work by the population subcommittee and Lisa, who thought she could get away form this project by resigning, she learned better and she is back. We appreciate your hard work over the years in pulling this report together and keeping this report alive. I'll turn it over to you.
DR. IEZZONI: Yes, I am back for a very brief appearance we hope. John told me during the break he hopes it's very brief and you can help make it brief, but I see a number of new faces around the Committee and what I'd like to do is just give you a teeny weeny history. The rest of you have heard this before many times. So I'll try to be brief and hopefully we can spend most of our half hour or so trying to wrap this up.
The official start date of this project was probably the Fall of 1999, but the concept dates back to the mid-90s with the core data elements project that the National Committee on Vital and Health Statistics was the leader in. In the core data elements is a place holder for functional status with a statement saying we don't know how to do this.
What we though we would do is try and answer the question, how should you do it? Our original goal in thinking about this was to actually do a broad survey of how do you measure functional status. We quickly had a reality check. That was simply not feasible given the subcommittee's time and budget and not necessarily on budget for what we should e doing giving the natural of this committee and its mandate because there are a lot of people who measure functional status in very many different ways to care for parties.
So oriented to focus on HIPAA amend the fact the HIPAA mandates our committee to look at code sets. And so actually thought that this was timely because we were going to do this around the same time that the WHO was wrapping up a 10-year revision of a code set called the (ICIDH) International Classification of Impairments Disabilities and Handicaps which resulted in May 2000 in a final revision, the International Classification of Functional Disability in Health(ICF) which is a code set complimentary to ICC() The way that we went about doing this by holding hearings and we did three of those.
During the hearings we asked speakers to talk about three thing. Number one was the utility of functional status reporting in administrative records, number two was the appropriateness of the ICF for this purpose, and number three was what steps would be required to think about going ahead to consider this a feasible thing to do.
We were ably assisted by a number of wonderful staff, most notably Susan Canon who really pulled this off for us. Susan is the author of the report you've seen, but there were help of other HHS staff, Susan Queen, Ceil Kennedy, Jerry Hendershot, Paul Plasik, and Marjorie Greenberg was extremely helpful as the WHO liaison through the NCHS.
So we came up with some recommendations that are of a couple of types. The first is a statement of principle. This is echoing what we said in the core data elements. Functional status information is important to include on medical records and maybe to report administrative data.
We also feel that the ICF is the only viable classification system at this time for classifying functional status information. The third recommendation is this is not ready for prime time yet. If we can agree on these principles, there is still a lot of work to do to prove that coating ICF in administrative records is logistically feasible and worth the cost in terms of the information value added and the business case for doing it.
Our global overall final recommendation might be that HHS spearhead research to investigate the practical issues and that the NCVHS monitor that investigation. With that brief introduction, you all have had an opportunity to review the report and, John, do we need a motion before this is an action item or should we go straight to people's questions?
DR. LUMPKIN: It's on the agenda not as an action item, but as a discussion. However, tomorrow it's an action item. Depending upon how the discussion goes, I may entertain a motion today because our agenda s very full.
DR. COHN: Actually, as a discussion item rather than a motion; is that okay?
DR. LUMPKIN: Yes, that's appropriate.
DR. COHN: Lisa, first of all, it's wonderful to have you back and I've missed you. I'm overall very supportive of the report and I think it's very good. I did find it an odd mixture of things which I strongly support which is that we do need a standard coding set for functional status and this other piece which is we ought to be putting functional status in administrative records which I found myself scratching my head a little more about because I don't exactly know where or under what circumstances it's appropriate, but I'm not going to start wordsmithing that.
However, I do have a significant question and concern about recommendation two which to my view, I like all of the recommendations. I think I would much more comfortable if it were removed and I think things would stand fine without.
And recommendation two has to do with basically the functionally status information should be reported at appropriate intervals in standardized code sets including claims and encounter records as well as computerized medical records, the issue being that it just seems a little premature to be making that recommendation at this point since we're still investigating the utility and appropriateness of all of this. It may come out later on that that's what we recommend.
DR. IEZZONI: You're right. Recommendation four says that we should be investigating what might be the appropriate time frames at which to do this and what would be the logistical implications of doing it.
I think that we had hoped in this report to make a compelling case that collecting this information "routinely" is a useful thing to for a variety of different reasons. That's what I think that this recommendation is. It's not really recommendation, it's more a statement of principle if you will view it that way. But then we do go on in number four to say we need to work out all of these things about the implications of statement of principle number two. How do other people feel about this?
DR. NERENZ: Could we add that phrase to it as a statement of principle?
DR. IEZZONI: Yes, I think that would be fine.
DR. COHN: That's already in number one. I'm still not satisfied on that, but let's hear other comments.
DR. IEZZONI: Other comments about that point?
DR. ZUBELDIA: I think it's a problem to recommend that this should be part of the claim or encounter. The claim, especially a HIPAA claim is encumbered with a number of things that have not been on the paper claim for years. It's a kitchen sink and to add one more data collection requirement that would be a burden to collect that is not being collected now because they have to report it in the claim, I couldn't support that.
DR. IEZZONI: We're not suggesting proposing that right now. I think we're suggesting that we study it and we need to remember that there are healthcare providers for whom what they do is supply functional status related services like the occupational therapist, the physical therapist, a lot of nurses who do home healthcare.
It's not reported on standard physician claim or the diagnostically driven claim. Nevertheless, what that means is that we're not representing a lot of healthcare that's given to chronically ill and disabled and functionally impaired people.
So I think at this point that what we're saying is that there are costs to this. We recognize it. Once we study it, we might say the costs are too great, but what we're recommending in number four is that we should study it and see. Again, study it throughout the different aspects of the healthcare system not just the diagnostically driven aspects of the healthcare system.
DR. LUMPKIN: I think what the subcommittee is recommending here is a study of the feasibility of using ICF in administrative records, not that we do it tomorrow or any specific date.
DR. IEZZONI: No, not at all.
DR. ZUBELDIA: That's not what point number two is. It says functional status information should be reported including claims and encounter criteria.
DR. IEZZONI: Maybe what we should do is eliminate recommendation two.
MS. GREENBERG: I just wanted to clarify that as I understand it, recommendation one is really about patient medical records or clinical records. It's not about two. It's really on patient medical or clinical records where two was on standardized datasets including claims and encounter records.
On the other hand, hearing what everyone is saying and given that recommendation four qualifies this considerably, one could, I would just propose one possibility is that number two could be retained with the addition of functional status. Instead of saying functional status information should be reported, should be considered for reporting.
DR. ZUBELDIA: Maybe if two becomes a bullet under four. That may be more appropriate.
DR. LUMPKIN: My trouble with this is I can think of a number of instances particularly in dealing with psychiatric illness, head injuries, other injuries where rehabilitation is a component of that. There is a lot of business that's going in claims attachments that might be eliminated by the addition of 3-4 spaces in the encounter forms or those transactions for an assessment of functionality.
The question that comes up repeatedly is, is this person making progress or have they progressed to the point where we're no longer going to pay the claim or they've exhausted their benefits. There is some assessment that goes in there. I don't think we can resolve that now.
We do think we need to raise the question in a placeholder fashion that this information has business use which is what the claims forms are really for. Over and above the fact that we, as a society, need to monitor the state of disability.
DR. IEZZONI: Proposals for changing the wording, Simon?
DR. COHN: I guess I agree with what John is saying except that I think you make an equally strong case for having the code set that's used in the claims attachment. That's the conclusion I come to that this is may not be a general use item, but for those people that need functional status measurement, it may be absolutely the right code set for that.
DR. IEZZONI: Or for claims for physical therapy and occupational therapy. Maybe we could add claims, claims attachments because I think that in our discussion of this, we talked about the claims attachment being a place where this could be considered. Again, remember that we're not recommending that this happen now. We're recommending that it be studied.
DR. LUMPKIN: What if we were delete the clause between the commas so that it should say functional status information should be reported at appropriate intervals in standardized data sets as well as in computerized patient records?
DR. COHN: That probably would work. I was going to comment that number nine also was a statement that says much of the same thing, but I think with the change you made, probably that's fine I guess.
DR. IEZZONI: Are there other comments from people? How long should we wait to see if people raise their hands, John?
DR. LUMPKIN: If someone could raise their hand to move that we adopt this. Paul just did that. It's been moved and seconded. The motion is that we would document with the modification in number two that the clause "including claims and encounter records" would be deleted. All those in favor, signify by saying aye. Opposed say nay. Any abstentions? Motion passed.
(Applause.)
DR. IEZZONI: Thanks to everybody who hung in on this project especially Susan.
DR. LUMPKIN: Thanks to you, Lisa, for shepherding it through. I am willing to entertain a motion to reconsider our next meeting so Lisa has to come back.
(Laughter.)
DR. LUMPKIN: At this point, we are slightly ahead of an agenda that was already slightly behind. We are only adding to our lunch half hour and making it a lunch 45 minutes.
DR. MAYS: I'm probably going to put you back on time with this request. After the last panel, I was actually supposed to do an update and can either do it after. I just want to make sure we do it the most efficient way of time, an update on the IOM meeting and the measurement of health disparities. We were either going to do it at the end of this last panel or we kind of quickly caucus with the staff. I can do it at the end of the next panel. Since you have time, it's up to you.
DR. LUMPKIN: I think the best thing for us to do is to break for lunch because we have such a short time frame, we need to be back at 12:30. Now, we are in a hotel, but if you take that corridor off there to the left and follow the signs to the food court, there is a food court. If you have conversations that you want to make during lunchtime, that you get your food first and have your conversations later.
(The session recessed for lunch at 11:45am.)
DR. LUMPKIN: Our next item is a briefing on the use of race and ethnicity in health and social statistics. Our first presenter is Dr. Madans. I have a major mile post coming in July and I've already passed the one where my glasses only have one refractory and doesn't always work all the time. I apologize for not being able to read that.
DR. MADANS: First, let me thank you for switching your schedule to meet my schedule. I need to catch a plane. So I'm going to talk very fast and leave before you can ask any questions, but I have brought Jim Reid with me. He's in the back there and he's been working on this general topic and can answer all the questions and he's agreed to stay this afternoon.
I just want to touch on a couple of things and I know for the subcommittee or for the full committee if you want more information, we'd be happy to come back and provide that and I hope to write some of this up at some point, but I didn't quite make it for today.
I want to just do three things. Since I'm the first speaker, normally the OMB person does this. I'm going to do what the directives are and why they're there. My interpretation of it, I'll try and say what I think OMB would say, but don't quote me.
I'll tell you a little bit about what has been happening at NCHS and some of the issues we're confronting and then outline a little bit of the research agenda that we've been developing and undertaking in the last couple of years in relation to this.
First, it all starts with Directive 15 which was the original OMB directive on the collection of information on race/ethnicity which is back in 1977. The statistical agencies, and NCHS in particular, have been collecting information on race for much much longer than this.
But in '77 we were all in agreement that we needed to do this in a consistent way across statistical agencies and this really came from the civil rights legislation and the need to monitor progress towards a social variable. So usually there's a whole lot of conversation about whether it is or isn't and for our purposes, it's purely social. I think that is the way OMB is treating it and that's what it says on the directive.
In '77, there were four basic reporting categories, collecting and reporting categories, but agencies were free to do more than that as long as they could collapse back. The other thing that OMB likes to say is once we think about revising the directive which is now called standards, lots of issues were raised about what it would mean for how we collect and analyze and interpret and do bridging. Very insightful questions were raised about the measurement of race/ethnicity. That's all very true.
All those questions were true in '77. They were true before '77. There aren't a lot of new problems. There are some, some are exacerbated, but a lot of the issues that we spent a lot of time that we spent discussing and debating have been there all along. This is not all of a sudden we realized we have problems in the way we do measurements.
In consultation with the statistical agencies, this was not something we just plopped down, we didn't know about it and one day we get a set of directives or standards. Is all measurement basically a compromise? It doesn't optimize anything. It's a little good for a lot of things, but it's not best for any one thing.
I think after several years, we're willing to live with the realization that for some things you have to go beyond and maybe in a totally different direction than these directives take you. You also sometimes get the question of why are we doing this at all?
I think from a public health standpoint this chart says it. These are infant mortality rates back to 1920, way beyond the directive, way before the directive and these differentials have been there, been there for a long time and they're still there.
From the monitoring of civil rights from the OMB perspective, however they got that way, we want them to go away. There is a whole lot behind this. Why is there a differential? How is SES involved and whole bunch of other variables involved.
No one would dispute that. This is very shorthand, but at the end of the day, if we still see these differentials, something needs to be done. At the point when they all come together, maybe we can say fine, we don't need to collect this anymore. It's a monitoring activity and testing ourselves, are we doing the right job as a public health agency in terms of intervening to make this go away.
So what happened in '97? There was an understanding that there were some demographic changes going on in the country. We were finding some resistance to answering the question and it's interesting that the HIS since 1982 has been asking the question on race the way the directive for '97 was written.
Why did that happen? I really can't find anyone to fess up to this. Why did we change? My assumption is that people started giving us multiple races and being good survey researchers, we don't argue with our respondents when they tell us this is the answer to the question.
Dutifully, we wrote it down, realized we were out of compliance with the '77 directive, figured we had to get back into compliance and hence the follow up question of if you had to choose, which one do you relate most closely to and that's how we tabulated all these years.
Most people forgot that we actually ever asked the question this way. Users were unaware of it. They never really went back to the source questionnaire. It was always there. So you have some resistance, you have some demographic changes going on and I think what it came down to was we'd rather do the change when the percent reporting multiple rates were smaller rather than larger in that it would be less of an impact on the system.
If you waited another 10 years until the next census, what we're going through right now would actually be harder. You wanted to get ahead of the curve. If we're doing this, this is going on, we want to standardize as early as possible.
Even then, this is not a quick process. This has been going on and being talked about probably in the early '90s and gets implemented early '97 and we're still figuring out how to implement it. This is a 10-20 year transition. I'm sure most of you know this. Let me just throw it on real quickly.
Revised categories. There are now five rather than four, some change in the labeling. The revised category are White, Block or African-American, American Indian or Alaskan native, Native Hawaiian or other Pacific Islander, and Asian.
Previously the Native Hawaiian or other Pacific Islander or Asian were one category. Asian Pacific Islander, but that's been broken up into two. The Hispanic origin question is now asked before and the labeling has been changed to Hispanic Latino.
I'm not going to talk much about this, but for bridging and for data, what we show in our tables, the fact that we went from four to five in the way we did it, is somewhat problematic. As you can imagine, the Native Hawaiian other Pacific Islander is very, very small. Most national surveys will not pick up enough to show separately, but by pulling them out of the API, you no longer can do trend analysis on the API. So it has some implementation issues.
The main change that everyone has been talking about is allowing people to put in more than one race. The guidelines on the tabulation report, not the actual standards it says that agencies much collect, tabulate, and report information on multiple race persons provided that the data meets agency standards because the surveys we run are very different sizes, very different designs.
It's very hard to put out one set of guidelines that will meet everyone's needs. But there are questions asked about why you do what you do in terms of collecting and disseminating. So you have to say we're doing it this way because of sample size of because of precision.
I also want to mention that what the census does is estimate the number of people in these categories, but what those of us in the other agencies are interested on the characteristics of the people in the categories, adding a whole 'nother level of complexity. You may have a large number of people, but if you're looking at a very rare event, it won't be enough. That's why this subject to the agency requirements.
What are we doing at NCHS? I've divided this up into the three basic data systems. The population based surveys, the provider based surveys and vital statistics. Population based surveys are in compliance and have been for quite some time especially the HIS which has been collecting data this way since 1982.
That's the data set that provided the basis for a lot of the analysis that's in the guidelines for tabulations report. We have done a lot of analysis and doing more analysis of that information. In terms of bridging, that's pretty easy.
We just re-tabulate what we had before to what the current standards are and we can go either way. Any survey that's based on the HIS, for example, the NSFD at some point is also in compliance. Haines has switched its questions. Our population based surveys are in pretty good shape.
The question with OMB is will they allow us to keep the follow-up question of which one do you most closely relate to? We have been successful in doing that basically for bridging because other people are using that information for their bridging work. Also, we feel that it actually tells you something about race and it in itself is a piece of information.
Even if we don't tabulate according to that variable, it's a useful variable to have, an analytic variable to have on the file. So we're making a distinction between the variables that are used to tabulate according to the guidelines that will be consistent with what BLS puts out, what transportation puts out, and so on as opposed to what we would want to use analytically to understand why we see those parallel lines and the lines aren't on top of each other.
There is some kind of stuff going on with how we do the waiting to the census and the groups aren't exactly the same, but that's the least of our problems actually. The statisticians are working that out.
Provider based surveys. They are currently all in compliance. They allow on their forms and this is the hospital discharge survey, the ambulatory care surveys, the long term care surveys, you can check more than one race.
The problem is, as you might imagine in provider based surveys, not a lot of people are checking more than one race. I was talking to someone in the last year of the national nursing home survey, there were 10 multiracial people.
So in this case because we do not have any control over the primary data collection. That's being done at the hospital for purposes totally unrelated to what we're using the data for. It's an administrative record, medical record, with some idea that it may be used for research, but not primarily used for research.
We don't have any control. We're kind of stuck with what they give us. This is true for a lot of other surveys that are run by a lot of other parts of the department. We have denominators from the census that have multiracial people. Our numerators, while they have some, don't have enough. We will never be able to make estimates for that population.
This is similar to what happens in vital statistics. For some things, the multi-racial population will never die, they will never have children, and they will never get medical care because those boxes will not be checked on these provider based records and in some case, on vital statistics.
What do we do about that? When we have the denominators, as I say we'll never use, those people will go through total or some other category. They're almost treated as an unknown because they can't be analyzed by themselves. There aren't enough of them and we know there's bias there.
Some of those multiracial people from the census where we get the denominator are going to hospitals, they're just not being captured. So you can be in compliance, but still not really have anything useful.
Vital statistics. I think we've come before this group and talked about the revision to the standard certificate which does follow the census format. We had hoped to have that in place in 2003. So we knew we were going to have a real problem for 2000-2002.
It looks now like the revised certificates will not be adopted by all states in 2003. This is getting caught up into our desire to put more effort into getting electronic certificates. So states are not going to want to change three times twice or whatever.
So this is going to be a lot more complicated and take a lot longer to resolve than we had originally thought. We think in 2003, we will be able to get states to give us everything listed. The certificates now it's a box and you can write anything in there.
We would get that data and retabulate it. I don't know where we are on that. Eventually, we will have the standard certificates implemented. We hope to be able to go to a check box although there is still some discussion about that.
The real concern right now with vital statistics is for at least 2000-2002, the numerators and denominators are totally incompatible. Again, this is where you will have no multi racial people dying or having children and we either have to ignore that population totally, we either have to impute multi racial status to some vital events or we have to bridge back, we have to create a file from the census that is closer to what you get on vital statistics and we're wrestling with that.
That is something we would like some thought about. We've talked to statistical agencies. We all kind of agree that we're going to have to do this bridging file at least for a couple of years which is not what OMB would like us to do, but even they realize we're in a bind.
The concern for the future is more with the mortality data. Similar to what we found with the hospitals. Even if we have the option of reporting multiple race on a death certificate, we don't think it's going to happen.
We know we have underascertainment of death rates for American Indians, for Asians, for Hispanics. We'll probably be worse for the multi racial population. Even if we go to the electronic certificate with the census question, if the funeral director is not asking, it's going to be a problem. Even if they ask, it may not be what the person would have told you when they filled out the census.
This is really a major problem. It's not as much a problem on the birth certificate. It's a major problem on the death certificate. We did have a workshop last year around this time to look at this.
The main recommendation from that was to do a meg-match of the census to deaths the following year and births in the preceding year and try to get a handle for how this is working. I think that would be extremely interesting and we're trying to move ahead on that.
We've gotten some funding and we have to try to make a proposal to the states now. But if it looks really bad, I don't know what that tells us about doing death rates by race based only on the death certificate. You have to do some kind of correction. We may have to continue to do the matches to the census.
It's going to be a major methodologic issue and I don't know the answer right now, but we're ready to put out 2000 rates and we don't know what denominators to use. Just to give you a heads up, for rates for 2000, we will be putting reports out that are based on projected 1990 numbers and we all know how good those are going to be.
We will probably have to do that for the preliminary and the final reports. We're going to have do another report based on the 2000 estimates once we figure out how to do that. There will be lots of numbers floating around over the next couple of years and this may be a problem for more than one or two years.
To summarize what the implementation issues are, we are still interested in doing some research into best practices for data collection for various modes. We think we have the person-to-person household based survey pretty well underhand. What about phone surveys? What do we do about providers? What do we do about vital records? How do you ask this kind of stuff?
There are other parts of the federal statistical community particularly people who do education. Where they're getting their records from the schools are very concerned about that and that research is ongoing. We are very concerned about what we put out in our files and what we put out on our tables.
Because of the small numbers, even though we collect, for example, in HIS all this great detail, we can't release it because of confidentiality. So what do we release? Right now, we just release one multiple race category.
We may be able to break that down to two, but that would be it, two of the largest groups. There has been some suggestion that we really want to do an any mention, that that would be more useful for people. Unfortunately, you can only do one of those because once you put that together with all the other information, you get some very small sell.
So what we put on the public use file is an issue. It does affect what you can bridge. Right now, we could not, according to the guidelines, recreate our API category. We will only put out Asians. The Native Hawaiians get put into a total. So you will not be able to do trend analysis unless you come to our data center. I know that's problematic for folks.
What we do on the tables is also going to be a challenge. Basically, I can say our reports will be much thicker because what used to be half a page table is now going to be a three-page table. We put more and more into the printed material than we can have on the file.
So everything is just going to get very long. We'd like to put more of that stuff on the Internet rather than print it. We'll see how that goes, what people really want to see, but I think right now, we have single race, multiple race, three within multiple race broken out by Hispanic, non-Hispanic race within Hispanic race with a non-Hispanic and then any mention.
It's a very long stub for something like Health US. That may be the way they're going to go. In terms of analysis, we have to figure out how we're going to bridge back. Some of our major trends. We need to look for what's the best way for putting the data together to establish new trend lines into the future.
We are looking at cross sectional comparisons internally in any one year to learn more about differentials. The trend analysis is particularly important for us because of the healthy people. A major issue for healthy people is getting rid of disparities. If you can't measure disparities, we're not going to be able to meet any of the healthy people goals.
How we do this is is not straightforward and it can be quite important for the programmatic aspects of the Department. Then some basic research on what to collect. Here it's not so much the methodology, but what other information should we, as a health agency be collecting to help us explain the differentials.
The follow-up question is useful, but is there anything else that we want to be putting on as a matter of course or every couple of years? There's been some discussion about putting information on ancestry.
Real quick. I want to give you some examples of the research agenda. What we've done and what's in the pipeline is a lot a analysis of HIS. We've looked at the size and characteristics of the multi-racial population both over time and at a single point in time, the most recent point in time.
We have looked at folks who do select a primary racial identification, what do they select, who are they, are there any factors associated with that? Because the HIS gets race information for everyone in the household, we can look at the races of the parents and see what they are reporting for their children.
There seems to be either a major cohort effect or a major age effect in the reporting of multirace. They're all very young. Some of this follows the change in interracial marriage and childbearing. I can show you that not all of it.
Remember in our surveys, parents report for their kids until they reach the age of majority. There may be something going on there where there is more reporting by the parent. When the kid finally becomes an adult, they may report differently.
To me, it doesn't make it any less useful. It just adds some complexity into what's going on here. The question is: How do we best capture that? I don't really want to spend time on this because I'm going to have to go in about 5 minutes, but just a couple of slides from presentations that a bunch of people at NCHS have done.
This is looking at the percent in fair and poor health, comparing the multiracial group with the single race groups. The thing for this is the patterns are different whereas here the multiracial people are like that AIAN, as opposed, part of the multiracial group. You see a different pattern when you're looking at Black/White. So there's nothing consistent here in terms of how multiracial are bind g reported.
MS. FYFFE: Can you articulate that a little better?
DR. MADANS: Talking a little bit too fast I guess. The 25 percent of American Indian Alaskan Native White folks are in fair or poor health, 9 percent of Whites, if they report White single race are in fair or poor health, 14 percent of American Indian Alaskan Native, single race are in fair and poor health.
The multiracial group is s little more similarly race than the single race American Indian Alaskan Native. If you look at the White/Black, the percent in fair or poor health is higher for single race Black and intermediate for single race White than much lower of the combined group.
DR. STARFIELD: In other words, you expect them to be in between?
DR. MADANS: Right. And they're not always in between. These patterns vary by what health outcome you're looking at. When you have time, I can show you all the health outcomes we have looked at and they all vary.
The next one real quickly is just showing you what the primary identification is for the multiracial groups. So for Black/White about half identify as Black when you ask them what group they identify with most closely. For American Indian Alaskan Native and White, 80 percent identify with White. For the API and White, it's a little bi more spread out. Fifty percent identify with White, 35 percent with Asian Pacific Islander and about 18 percent still say no, they're multi racial.
Most of this stuff I'm showing will come out at a symposium that was held. One we do that, this is looking at employer sponsored health insurance by the multiracial group and also comparing them not the single race, but to folks who picked a primary race.
For example, 50 percent of American Indian Alaskan Native White have employer sponsored health insurance, 30 percent if they pick American Indian as their primary identification, 30 percent of them have health insurance. If they picked White as their primary identification, 54 percent have health insurance and if they pick one, about 50 percent have health insurance.
That is very different than if you look at the API White where it's pretty consistent. They're a multiracial group. The ones who pick API as their primary, the ones who pick White as their primary and the one who picks none as the primary just continue to be multiracial. All that's 60 percent.
Again, it varies by what combinations you're looking at and this will also vary by what health outcome you're looking at. The difference between this and the other slide is this is comparing multiracial with the primary identification. The other is looking at the single race folks. The primary here are subsets of the multi. In the other, they are not. It's a totally separate group.
This is how parents report for their children. If one parent is Black and the other is White, about 40 percent are reporting the child is multiracial, about 10 percent are saying White, 45-50 percent are saying Black and so on.
Parents might be reporting their children as multiracial are not reporting their children as multiracial. There is something else going on here. This is not surprising. This pretty much goes in line with how we think people think about race, but it's clear from these data that this is what's going on.
What are we doing with vital statistics? Some trend analysis of births to parents of different races. We have gotten data from four states, California, Hawaii, Alaska, and Washington, to give us everything on their birth and death certificates, all races listed.
These are states that have high multiracial populations. So we can do an analysis of that. This is in preparation of when we hope to get this from all states. Then we're trying to follow up on the recommendations of that workshop I told you about.
Interracial births by mother, this is just showing you the big increase from '71-'95 in the number of interracial births. So we would expect to see an increase in the multiracial population. So this is just from the vital statistics.
The have some cross cutting analysis where a paper has just been submitted where we did an expectation based on the birth data of what you should see in HIS going back and following cohorts.
There is a cross over in time, but in modern, more recent years, HIS is reporting less than what you would expect for vital records and it varies by the nature of the races of the parents. Hopefully, that will be coming out. I think I'm going to stop. I had some other issues, but I'll stop and I'm going over. Thank you very much. Jim is here.
DR. LUMPKIN: Our next speaker is Dr. Jorge del Pinal.
DR. DEL PINAL: You won't mind if I speak from here.
MR. BLAIR: We're broadcasting on the Internet so you have to speak near a microphone.
DR. DEL PINAL: This is what our experience has been in the Census Bureau collecting information on the population by race/ethnicity. The modern chapter of collecting race/ethnicity at the federal level begins really on October 30, 1997 when the Office of Management and Budget really issued new federal regulations called the revisions of standards for classification of federal data on race/ethnicity.
Essentially here are what the new standards are called. They used to be OMB Directive 15 which is a nice catchy title. They didn't like that directive stuff. So they gave us this long title. Essentially, these standards for maintaining, collecting and presenting federal data on race/ethnicity did several things, four of which have some impact on what we do.
The first is that they asked us to sequence Hispanic origin ahead of race. The idea was, all the research showed that people are not answering the Hispanic origin question because it followed race and they figured once they had answered one of the race groups, that that took care of the Hispanic origin question.
So we have very high non-response for that. The cure for that was seen to move the Hispanic question ahead. There was some terminology changes. People were saying some of the terms you're using are ancient, outdated, and offensive.
One of the key changes was that they allowed people to select more than one race category. For us, this was new, but as Jennifer has already related to you, they've been looking at this issue for some time now.
To complicate things further, OMB decided to create a new racial category. These last two things have an enormous effect on the data product and what we're going to show in the Census Bureau. The first is to select more than one racial category.
The category that was split up was the old Asian and Pacific Islander category. Basically, Pacific Islanders made the case to OMB that they're really quite distinct form other Asian groups.
Our pulling them out, I'm not sure that they did a great service either, but it's distinctive and people are reporting in this group if there is enough observation to say something about it.
Basically, the new OMB categories are American Indian or Alaskan native, Asian. Black or African American is a new title there. The Pacific Islander group was called Native Hawaiian and Other Pacific Islanders. We have the White category.
And because people could select more than one of these racial categories, OMB has said, you should really show how many people reported more than one race. Show it in as much detail as you can so that you don't violate confidentiality or statistical reliability of the data.
Also the Census Bureau asked OMB and we were granted the waiver to have a category called some other race. Generally, this is s a catch all category and people put all kinds of stuff in there.
We tried to do our best to interpret what they write-in and either identify a racial category that appropriately fits that write-in or leave them in the other race category when we have no basis upon which to do something with that information.
Now, just to clarify what we mean by two or more races in the OMB context, it's really combinations of two of these major race categories. If somebody selects category White and selects the category Asian, then they are two races, White and Asian.
However, in our forms we also have break out of subgroups within racial categories. For example, a lot of different Asian and Pacific Islander groupings are in there. So somebody could check two of those boxes. Like Japanese and Chinese which come from the same Asian category, for the purpose of OMB, we treat that as one racial category.
So you have the six basic response categories in the Census, but there are 57 combinations of these six categories that are possible. These include combinations of the two of these, three, four, five or six. There's only one combination of that possible.
Basically, here on this handout I've given, I'm showing you what the Hispanic origin question is and it has five check out blocks plus a write-in area where people can tell us what other type of Hispanic they are besides Mexican, Puerto Rican, and Cuban.
It's followed by the race question which has 15 check off blocks all of which can be reduced to six categories. We have three write-in areas, one under the American Indian category. We captured up to two write-ins in that area.
We have a write-in area for other Asian and other Pacific Islander share a block. We can capture up to two responses in there. Then we captured up to two responses in some other race. It turns out that the possible combinations of 15 check off blocks gives you over 32,000 possible combinations. If we add in the write-in areas, I'm sure everybody in the world could have their own unique racial identification.
That's how big it potentially is. In our first try at this was in the dress rehearsal in 1998 and it was really conducted in three locations: one in Sacramento, California and several counties in South Carolina and one county, and American Indian reservation, Monamine, Wisconsin.
Basically, what we found there was the question everybody wanted to know is what's the extent of the reporting of two or more races? We found that about 5.4 percent of the population in Sacramento reported that they had two or more races, but less than one percent in South Carolina and a little over 1 percent in Monamine.
So Sacramento was a surprise on the upside, but a lot of people were worried about the reservation, but a lot of American Indians will tell you they're mixed Americans. The one thing we found on all three sites is that 9/10 people who selected 2 or more races, selected only two.
Even though there are 57 combinations, most people are dealing with errors of races and sometimes triples. We also found that Hispanics are likely to say they are two or more races than the general population. Often we found that they're selecting a race category like White and then selecting another race and telling us they're Mexican or something like that as their other race. We found that.
The next shot was the American Communities Survey which is conducted in 31 sites in 1999. I've selected the five largest sites to talk about: Fort Bend in Harris County, Texas which is Houston; Broward County, Florida; Bronx, New York; Franklin County, Ohio; and San Francisco, California.
There the percentage of two or more races varied between about 1.6 percent for Franklin, Ohio to about 3.8 percent for the Bronx, New York, a little bit lower than what we found in dress rehearsal.
Again, in most sites, we fond that 9/10 times the people selecting two or more races are selecting pairs of two races and that's over 90 percent of the time except in Franklin, Ohio which is almost 90 percent.
When we look at this by whether they reported that they were Hispanic or not, we found that much higher proportions of Hispanics reported two or more races. You can see it there. For the non-Hispanic population, it ranges about 1.3 - 2.1. Among Hispanics, it ranges from 3.6 in Harris County, Texas to 9.5 in San Francisco.
Quite a big difference. What we examined because it was so much higher for Hispanics, we wanted to see what they were reporting.
MR. BLAIR: When you say Harris County, is that 3.6 percent of the total?
DR. DEL PINAL: Of the total, yes. In Harris, it was 2 percent overall. Just among Hispanics, it was 3.6 in Harris.
MR. BLAIR: Just to make sure I do understand this. If in Houston, you had 20 percent of the population was Black and 35 percent was Hispanic, I'm just guessing, then the 3.6 percent was in addition?
DR. DEL PINAL: No, the 3.6 percent represents the portion of Hispanics that identified as two or more races.
MR. BLAIR: It's of Hispanics, not of the total?
DR. DEL PINAL: No, of the respondents who said they were Hispanic, what proportion reported two or more races.
MR. BLAIR: Thank you.
DR. DEL PINAL: Now, the key thing is we know from dress rehearsal that Hispanics are reporting some other race as one of their races. If we eliminate pairs of races in which one is a standard race category and another is some other race, then we get the next figure I show.
Now, the reporting of two or more races among Hispanics if we exclude those pairs with some other race, is very similar to what it is for non-Hispanics. So often, what Hispanics are really trying to do is they select a race category, but they also want to tell you again in the race question that they are Hispanic. That seems to be what was happening.
So now, in Census 2000, we reported on the national level, 2.4 percent of the population reported two or more races. We were expecting about 2 percent so this is a little bit higher again. We found that among non-Hispanics, the portion reporting two races is 1.8 percent compared to about 6.3 percent among Hispanics. So much higher reporting again. We found that in excesses of 40 percent of the two or more responses had to do with some other race. Again, we found that.
When we looked at what number of races were being reported again, 9/10 times people who reported two or more races reported only two combinations, much, much simpler. People also wanted to know what were the largest combinations. There was some interest to see if the multiple race reporting of a particular category exceeded 1 percent of the population.
It never did, but the largest combination was White and some other race and a lot of these people are Hispanics that are reporting that. The next largest category was about 16 percent of the two or more races were those who reported White and American Indian as their race combination.
The about 13 percent reported White and Asian and about 12 percent reported White and Black. The next largest combination was Black and some other race followed by Asian and some other race and finally Black and American Indian with about 3 percent.
Some people aren't as interested in that some other race combination because we're not sure really what it means at this point. The question is how are we going to portray the information. Here's where seeing these tables in front of you is going to help and it's probably going to baffle listeners over the Internet.
How are we going to deal with this data? One of the ways is we're going to show a complete table that shows the total population and we will eliminate one group. For example, the total population in the United States was about 281 million.
We're going to show the White population broken into two groups, those who selected White by itself which we call White alone and it turns out in the US, that's about 212 million people that reported that.
However, another 5.5 million that reported White in combination with one or more other races. That's another almost 6 million people there. The total of that is what we call White alone or in combination which is about 217 million.
We really had these different measures of the White population here. We wish the story was as simple as it is for the White population, but it gets a little more complex.
Another way to look at this is of everybody who selected that they were Black as one of their races, how many selected more than one race? That's another way to ask it. I was asked that several times particularly by young people. They were curious about that.
To answer that question, we have to assemble the pieces. The population who selected Black alone was about 35 million and then there is about 785,000 people who said they were Black and White. Another 182,000 that said they were Black and American Indian.
Another 112,000 said they were Black, White, and American Indian. We had 107,000 who said they were Black and Asian. There were about 600,000 other people who had Black in other combinations. So if you add all that up, it gives you the Black alone or in combination number of 36 million.
In the press, when we release this data, they were saying the Black population is between 35-36 million is the story they gave. Then I break it out assuming the Black alone or in combination is 100 percent of the people who said they were Black, what proportion gave us one race and what proportion gave us some of the others?
When you look at that, you see that just slightly under 5 percent of the Black population selected more than one race as a category. Remember, the national average is about 2.4, but for the Black population, it represents about 5 percent of Blacks.
Remember that some of those Blacks are also counted in other race categories. The groups in combination overlap with some of the other groups. If you're interested in the Black population, this might be an interesting way to organize it.
If you were going to bridge the data, that is, decide that some of these combination really belong with one category of their own, you're going to hit a number somewhere in between these. You're going to decide some groups go to Black, some groups go to the other area and there are lots of people that have different methods to do that.
There is a federal working group that's trying to decide and look at different methods and report back to us what difference it makes if you select several of these. The picture for African Americans doesn't appear so bad.
If we do the same thing for Asians, we can see that the Asian alone is about 10 million and if you add to that Asian/White which is about 900,000, you get to the Asian alone or in combination of about 12 million. It ranges from 10-12 million essentially.
If you ask again what percent of Asians select more than one race category, now you see it as 86 percent of those who selected Asian, selected only Asian meaning the balance of about 14 percent reported Asian in conjunction with another race and so forth.
We were talking about 95 percent for Blacks and now it's 86 percent for ASsians. Let's look at American Indians, what happens to American Indians? American Indians alone, there are about 2.5 million people that said that.
Over a million said they were White and American Indian and if you add up all the combination, it gets to a little over 4 million. Now you have it varying between 2.5 million and 4 million depending on which measure you use.
If you put it in percent terms, you find out that 40 percent of the American Indians also selected another race category. Look at those who reported Pacific Islanders by themselves, that's about 400,000, but if you add all the pieces together, the largest combinations are Pacific Islanders/Asians which shouldn't be surprising, Pacific Islanders/White and then Pacific Islanders/Asian/White, those are the three that are big.
The top number if you add all the pieces together is 874,000, quite a variation. In percent terms, 46 percent of the Pacific Islanders selected that by itself and the balance about 54 percent have it in combination with other groups.
The next slide says what would be the population change depending on whether you took the minimum or the maximum of those two numbers. You can see for once it isn't too bad. The White population either grew by 6 percent or by 8.5 percent. The Black population either grew by 16 percent or 22 percent.
The Asian population either grew by 48 percent or by 72 percent. The Pacific Islander either grew by 9 percent of 140 percent. That's the kind of variation we have here to deal with. As you can see, the other story we found here is that the smaller groups are much more dramatically affected by this than the overall population or the major race categories.
The other thing is it would be nice if the 2.4 percent of the population everywhere reported that there were two or more races. We didn't have that luck. It turns out that certain states are reporting much higher combinations. A few people weren't surprised. I was surprised at how large this was, but a lot of people weren't.
Hawaii came in with 21 percent of its population reporting two or more races. That is one place where there are combinations that some of these groups that are in excess of 1 percent of the locality at the state level. At lower levels, it might even be greater than that.
The other thing we were able to discern from the data we released to date is that reporting two or more races is much higher among children than it is among adults. So the next graph shows among the percent of children, under 18, what percent reported two or more races.
I used the same states and they should be ten largest. Hawaii, 37 percent of the kids are reporting that they are two or more races. It's a bigger issue for the kids.
DR. LUMPKIN: Were they reporting or their parents?
DR. DEL PINAL: Their parents were reporting for them, but in my household, my son got to report for himself. Essentially, that's what we found. In terms of the Hispanics, they are more likely to report two or more races, but it's because they're using the race again to tell us that they're Hispanic and that somehow that modifies the race that they've given.
That's essentially what I brought. We can discuss other issues that may be of interest to the group here.
DR. LUMPKIN: We're going to have to move on to the next speaker. We have one speaker and no more time.
MR. HARRISON: Roderick Harrison from the Center for Political and Economic Studies. Thank you for the opportunity. I will try to be very brief given the time and allow time for questions.
I wish there was some way to convince you and others, hopefully, some of the numbers Jorge just cited would do so that the federal statistical system and the part that you're concerned with, the vital statistic system, is in great jeopardy of becoming virtually very difficult and very meaningless for conducting the kind of analyses by race that you've conducted in the past.
I would point to several problems. Jorge gave you the statistics. If the population of Blacks is 5 percent higher if you count those who reported more than one race, than if you count only single race Blacks which you might be able to deal with.
It's 60 percent higher for American Indians. It's about 16 percent higher for Asians. Thais is at the national level. As Jorge pointed out, when you get into specific localities, these ratios are going to be much greater. These ranges of possible population counts and these are the denominators you will be using through much of the health surveillance system.
So how to count the multiple race responses would become very important to the quality of the statistics and the meaningfulness of the statistics collected particularly for American Indians, but certainly also for Asians and Blacks in some localities.
I'd like to point out, for example, some Blacks at 5 percent nationally might not be that great, or might be within the margins of other sorts of errors and the number might be close enough. But if you get into California which does have one of the higher rates of multiple reporting, there are 14 counties in California where the range is from 5-10 percent the ratio of all including Blacks.
Those include Fresno, Riverside, San Diego, and Mercer Counties, some places that are not unimportant although clearly not the concentration of Blacks that Los Angeles or San Francisco has. These are not unimportant places for the Black population or health statistics on Blacks in California.
It would be 10-20 percent in 16 more counties including Merrin, Orange, Ventura, and Santa Barbara counties. When you get down to specific localities, you're going to see in some cases wide ranges, very wide ranges in the estimates in the size of the population much less it's characteristics depending upon how one counts the multiple responses.
I wold like to argue that what is needed in this situation is adequate bridge statistics that would tell us what the population distribution and characteristics of these populations would have looked like, the education, income, et cetera, absent the revision to the guidelines and the Census Bureau will be conducting a bridge study.
But I'm afraid I'm not sure what the sampling scheme finally came down to. I really think you need bridge statistics at the county level given the wide variations in multiple race reporting of counties and given the degree to which the federal statistical system, certainly the health statistic system, but even civil rights reporting is important at the county level or aggregations of counties into metro area.
I'm not sure at all that the Census Bureau is planning to provide county level estimates or that it has even discussed the issue with the kind of seriousness and with the kinds of audiences, users, that it needs to.
That's one problem. That's the problem of how are you going to repair trend statistics, how are you going to construct meaningful trend statistics from the populations? The size of multiple race responses particularly for American Indians, Asians, but certainly even for Blacks in some areas is going to make it difficult to determine whether changes that you observe as you gather statistics on the new system are attributable to real world changes in health conditions or changes in these reporting systems.
It's going to be very complex to tease out, virtually impossible without good bridge statistics. That's one set of issues which is the bridge that is past. How do you construct bridges that enable you to use trend statistics? Even more important and less recognized is the problem with the bridge to the future.
The telling statistic which Jennifer Madans cited, but people tend to pass over. If you take one thing away, less than half of the people that we can identify as being in households where the parents are of two different races report the child as multiple race, less than half.
Some of these households, the parents might not be the biological parents of the children. It might be that in some of those cases, the child was the monoracial child of two White parents and this is now a Black/White intermarriage or something of this sort. Even allowing for that, what we know about intermarriage statistics would suggest that a large proportion of children of parents of different races are not reporting more than one race.
Why is that a problem? The problem is that when you start producing and the federal guidance is full of page after page of alternative ways of presenting data that might well be meaningless. When you start presenting data on a birth rate, a death rate, a cancer rate for a Black and White population, what are you talking about? What is that statistic supposed to mean? Are you to compare that statistic with a cancer rate, a birth rate, a death rate, a heart disease rate for the Black population, the White population as you have in the past?
If this population, the people reporting as Black and White in the census, people who will be reported in various registries in the health system as Black and White, if these people do not represent a racial population based on parentage can be easily identified. I think many people will think that's what these numbers represent.
These are the people who have one Black and one White parent of maybe one Black or one White grandparent somewhere in the mix. You're not getting that population. You're getting somewhere in the order of half of that population. Until you know how the characteristics of people who have parents of different races or grandparents of different races, how those who choose to report more than one race differ from those who report only a single race, until you have that, I don't see how you can make any meaningful use of the statistic collected on a Black and White population from the Census or your registries.
I think it would be professionally irresponsible to present a thing as here is the cancer rate for the Black and White population. Compare it with the cancer rate for the Black population. Even the graph that Jennifer Madans showed you about that.
It's interesting and important to see that sometimes it's between and sometimes it's not, but who those people are who are reporting more than one race. The referent in the real world has yet to be defined in any meaningful way.
You're reporting what people reported, but whether it's a statistic that's meaningful in a house surveillance system or not or in an education system concerned with poverty, et cetera, is the question I don' t think has yet been posed and I don't think people have yet begun to answer. I'm afraid there are no good answers at that point.
I don't think you can make very good sense of those statistics. I wish as much time has been spent in the guidance addressing that issue. I see in your letter to Katherine Wohlman that you are raising this issue with her.
But as much time needs to be devoted to that as how alternative ways of presenting statistics that refer to a group that nobody can define independently. It's the group who reported more than one race. Do we have any other way of identifying this population? No.
The third problem is the consistency reporting of people reporting more than one race was on the order of 50 percent. That means that half of those who reported more than one race in the dress rehearsal, did not do so when re-interviewed. In the national content survey that we conducted as part of informing the OMB decision, half of the people in the re-interview who were asked if they ever reported more than one race before or reported as multi racial category, about half of them had not.
They had not done so in reporting to school systems, in reporting to employers and probably in reporting to hospitals or health documents. Anywhere else, this was the first time they had done so. Maybe some of them will do it again the next time they go to apply for a job, but the consistency rate suggested many of them are not. The third problem is going to be horrendous numerator, denominator matching problems.
There is no guarantee that there will be sufficient stability in the reporting of people who reported more than one race in the denominator when they come to the context where they're going to report numerators.
Those are three problems. Many of them, the problems suggest that OMB, the Census Bureau., the statistical system needs to be much more professionally responsible than they have been and take much more seriously the issue of providing adequate bridge statistics as they did when they changed the unemployment questions in the CPS.
I think frankly that some of the political sensitivities about asking people who reported more than one race to do the kind of NHIS follow up has locked OMB and the Census Bureau has made them shy about doing what professionally needs to be done. If you're going to salvage any kind of racial analysis out of the vital statistic system, somebody is going to have to convince them to do it.
DR. LUMPKIN: Thank you. We are significantly overtime and I don't want to minimize the importance of this particular discussion. I think it's complicated because we're in the midst of a social change process. That social change process is dramatic and ongoing.
In my father's time, he had no choice about what race he was. He was never asked. He grew up knowing what race he was even though if you do a genealogical study, that would not be the race he would be under the current definition.
I personally am also even more complicated than he is, but in my childhood, I had no choice in what race I was. As our society is evolving, it is becoming more and more complex. We're dealing with a term for which there is no scientific definition that holds up through any sort of external analysis.
We're dealing with an incredibly complex social structure that is unique to this country and the social structure of this country. The meaning of race changes as the social context changes from nation to nation. Given that there are so many questions and so many issues, obviously we're not going to have a full discussion here.
Hopefully, our subcommittee on populations will be up to the task of beginning to make some sense of what kind of recommendations we can make. Clearly, disparities exist and clearly we need to be able to document them if we're going to reach our objective of eliminating disparities.
The transition of our health statistic system to be able to fully be sensitive to the existence and change in those disparities has to be a major priority of the committee. So we have time for a few questions.
DR. ZUBELDIA: I would like to make a couple of observations. These are personal observations on the process by which this information is collected. First, I would like to say that I like very much how the Census form was structured because it was very clear.
However, a year and two days ago, I had a baby boy and a year ago at about this time, we were leaving the hospital. The social worker came with one of those new forms that the State of Utah was using and had been recently trained on the use of this new form and she happened to run into me.
Going through the form, I felt like I had to check multiple boxes for my boy, maybe four or five of them. She was adamant. I could not check more than one box and it was part of the training that she had received from the State of Utah. I could not check more than one box. After some discussion, I checked the boxes I felt appropriate and she took the form and it was her problem.
About a month ago, I received from the State of Utah, a driver's license extension application. This is a form that every five years you fill out. They give you a sticker and you put it in the back of your license and you don't have to go in person.
It has a list of race/ethnicity listings and it says will appear in the future as a numeric code on your driver's license, one code. It starts correctly with all the Hispanic classifications, Spanish, Hispanic, Latino, Mexican. I checked all the above.
Then it goes into other things like Puerto Rican, Cuban, White, Black, African American, Negro. I'm not sure that most people would know how to fill this out. If they do check more than one, which is the code that will appear on my driver's license?
The process is clearly flawed in the collection of the information. Somehow, there has to be additional instructions, there has to be education of these people. This needs to be made easier for the normal population to answer.
DR. LUMPKIN: Kepa, I'm sorry, but why do they need it on the driver's license?
DR. ZUBELDIA: I don't know. The form talks about racial profiling. There have been complaints about racial profiling so they want to put it in driver's license, but which code are they going to put on my driver's license?
DR. LUMPKIN: It's in order to identify those police agencies that are engaged in racial profiling.
DR. ZUBELDIA: Here it says due to recent national concerns of alleged racial profiling by law enforcement agencies, we're going to do this.
MS. FYFFE: Jorge, does the Census Bureau count people who are citizens, non-citizens? How would the Asian/Indians who were here as high-tech workers be counted? Are they included in the census?
DR. DEL PINAL: Sure. It's a household based survey. So if you live somewhere and we have that address, a form goes to you and if you don't answer the census, then you may get it through an interview or something. It tends to count everybody whether you're documents or whether you error legally.
There are some people here in town who are diplomats and things like that that live in their own soil. There are a few people like that, but they're not counted. If you're there, we want your information.
DR. FRIEDMAN: I'm looking at your slide 18, Dr. Pinal, on the Hispanic population. This is based on the census question and not the OMB standards. According to this slide, almost 94 percent of Hispanic population said they were one race, of that 94 percent, 45 percent selected some other race. I'm wondering if not in the census standard, but in the regular OMB standard that does not include some other race as one of the allowable race categories what do you think is going to happen to those 40-45 percent of Hispanics who say they're some other race?
DR. DEL PINAL: In other words, what race would they normally report. They're saying that my race is Hispanic or Mexican or Argentinean. You have all those responses that they put in there. Now if some other race would disappear, I think a lot of them wouldn't answer or they might reluctantly select White.
Ones that have been here longer say, especially in Texas they're very sensitive about this. Once they've identified they're Hispanic and they know they categorized as White because a lot of the school districts in Texas were integrated that way, half Hispanic and half Black. A lot of the Hispanics already know.
DR. FRIEDMAN: I might add the percentage of people largely Hispanic on the CPS and other surveys who are refusing to answer even after being prodded in several ways to do so is increasing and it's probably going to be a large program, not having that option will present a problem to increasing percentages of respondents.
DR. LUMPKIN: I would like to thank the speakers for a very thought provoking presentation. I have a ton of questions I still would like to ask, but if you are available, we would appreciate if you could stay over. We are having a subcommittee on populations.
The first item on their agenda after we take care of one more item of business will be discussion of implementation of race/ethnicity. If you are available, we would appreciate it if you could stay over for the subcommittee meeting.
The next item on the agenda, Simon, it's a letter that was at our place?
DR. COHN: I don't think so. There are actually two letters for consideration which will be moved for action tomorrow. Should we take the easier one first? Do we have copies of the first letter?
DR. COHN: I apologize to the subcommittee. The second letter we need to go through with some discussion.
MS. GREENBERG: It was an earlier version in the agenda.
DR. COHN: We're not even handling that one. This is a transmittal letter we're passing out to everybody. John, would you like me to read this transmittal letter and explain what it is we're doing?
DR. LUMPKIN: Yes.
DR. COHN: The good news is that this should not be particularly contentious.
"Dear Secretary Thompson, as part of its responsibility under HIPAA, the NCVHS provides advice to you regarding the standard electronic transactions and code sets adopted under HIPAA. NCVHS has recently reviewed the first recommended set of changes to these standards as recommended by the DSMOs.
"The DSMOs are a set of organizations that have been designated by the Department to modify standard based on in industry input. The first set of changes was produced by then DSMOs under a fast track process that the NCVHS requested to expedite changes that are necessary to facilitate success implementation.
"The process required the DSMOs to do their review and analysis work in a compressed time frame so that changes could be made to the standards during the first year after adoption and in time to assist with timely industry implementation.
"The DSMO representatives responded to the NCVHS request in extraordinary fashion devoting substantial time and effort to meeting our deadlines and achieving industry consensus. They are to highly commended. This goes well for the ability of the HIPAA management process to be responsive to future health industry needs.
"The NCVHS has reviewed the DSMO recommendations or changes to the standards. We recommend that you accept all of the changes subject to technical corrections and begin the regulatory process to adopt these changes as part of the standards.
"In addition, we urge you to expedite the process so that health plans, providers, and clearing houses can adopt these necessary changes as soon as possible. We recommend that the final rule implementing these changes be published no later than January 21, 2002.
"It is clear that the current HIPAA code sets do not cover all procedures and diagnoses. This will always be a problem due to the evolving nature of healthcare. Where gaps have been identified, we endorse the DSMO's position that requesters work with maintainers of the currently designated code sets to address these gaps before new code sets are considered.
"We are planning to address these and broader issues related to code sets and terminology gaps at hearing some time in the late Fall or winter. One topic we are likely to address is the extent to which this process is responsive to requester's needs. We appreciate the opportunity to provide these recommendations to you."
This is a transmittal letter and we have not made copies of all of the changes which are mostly technical changes. There are 67 of them. We spent about four hours at the subcommittee yesterday going through them.
If any of you have any questions about them, we're happy to address them, but we're hoping that based on your confidence in the subcommittee and the DSMO process, and they were really exemplary in their process. We'll be happy to make copies for any and all of you who would like copies here or if you'd rather, we can mail them out afterwards.
DR. LUMPKIN: How many DSMOs are there?
DR. COHN: ANSI X 12, HL7, NCPDP, NUCC, NUBC, the ADA's Dental Content Committee.
DR. LUMPKIN: Since there is a relatively number and we're appreciative of their efforts, should we not maybe list them in the body of the letter.
DR. COHN: We accept that as a friendly amendment.
MR. SCANLON: Simon, this is a recommendation for the Secretary to issue a proposed rule that incorporates these in total as modification?
DR. COHN: Yes, exactly.
MR. SCANLON: Are they referring to it at earlier meetings? How would these fit in with the implementation guides, for example?
DR. COHN: The implementation guides would be modified, updated, to incorporate thee changes and require new NPRM in final rule.
MS. GREENBERG: That friendly amendment to John's suggestion to name the organizations and maybe the co-chairs because they spent hours with the subcommittee, including their vacation time, their birthdays. It was incredible.
DR. COHN: We'll get them some glory. Another friendly amendment.
DR. LUMPKIN: Is there a desire to do any additional study or can we vote on this today?
DR. ZUBELDIA: Let me clarify something for Jim's benefit. It's not only proposed for one NPRM. There may be several NPRMs that come as a result of this. There are at least one on the NDC codes and there is one on the implementation guides, but then there is also some other policy-type issues in the back of the document that requests for the Department to either clarify something or address some new policy issues.
DR. LUMPKIN: Thank you. I think what we want is for something to be done quickly. If we ask for clarification or a policy issue, that's going to be a lot slower than some of the other things that we want. So the point is we want there to be a rapid response and there may need to be a triage as you suggest so that there may be a rapid NPRM and then a standard track NPRM.
MS. COLTIN: I would like to see a copy of the implementation guides than I am in being able to read some of the information about the code sets and the recommendations.
DR. LUMPKIN: Are there any other members who would also like to see the entire?
DR. COHN: We'll make as many copies as people would like to see.
DR. LUMPKIN: You're the only one who's holding this up. But because every member is important. We are more than happy to do this.
DR. COHN: So my understanding is that we'll have two copies made before the end of the session for you both to review. The next letter, you all have copies of an earlier draft. The subcommittee spent a number of hours rewriting this yesterday and it's significantly different. I think the conclusions are pretty much the same.
MS. FYFFE: The revised letter has what data on it?
DR. COHN: 6/26.
MS. GREENBERG: Just can the subcommittee assume, I'm thinking of what's going to transpire now in the next 24 hours. The subcommittee is going to be meeting this afternoon and this will come back tomorrow, but can the subcommittee assume that at this point there aren't any other really major comments on this letter?
DR. COHN: The first letter.
MS. GREENBERG: The first letter because there would be an opportunity to revise that letter after reviewing details. If you have any comments on the letter itself, tell the subcommittee so they can address them in their next session.
DR. COHN: Okay. And there were some further modifications that you had on the second letter. Okay. I'll read the letter, but I don't know what else you had.
DR. LUMPKIN: If I can suggest a way to go through this rather than to read the letter. I think what we need to do is get some issues on the table and what I would like to urge given the time and the sequence that people who are not on the subcommittee should be the ones trying to get their issues on the table for.
If there are some controversies on the points, then it might be useful for you to suggest what the Committee is going to be trying to tackle on coming up with a final. Or is this a consensus of the full subcommittee? This is the consensus of the full subcommittee. So, to that extent, maybe you can walk through the individual points rather than reading the entire letter and highlight where you think there may be controversy.
DR. COHN: I don't think so, but I will read it in sections.
MS. FYFFE: Paraphrase it.
DR. COHN: I will do my best to paraphrase, John. But think since people are on the Internet, they might appreciate having it read.
On May 31, 2001 the NCVHS held hearings on the subject of industry readiness related to HIPAA administrative simplification. Testimony was heard from a number of industry representatives and advisory bodies and there are a list of advisory bodies and people for your review.
From these testimonies as well as written statements and letters, the Committee concluded the following:
One, there was overall agreement that HIPAA standards are necessary and viable. In all the testimony and comments heard by the Committee, there was strong support for the administrative simplification provisions of HIPAA and for the efforts of HHS to promulgate the implementing regulations. In all instances, no negative feedback was received. The only concerns voiced were the speed and the timing of the regulations and the timeframes for compliance.
Two, there was general agreement that all the final rules especially those for security, claims attachments and provider identifiers should be published as soon as possible. The reason for this position was different for each entity and included, one, the final security rule as necessary to properly implement the Privacy Rule. Two, the final provider identifier rule as necessary to allow payers to design changes to or replacement of claims processing systems. Three, the final healthcare claims attachment rule is needed to increase the percentage of claims that are processed electronically introduce paperwork.
Three, there was testimony from a number of industry sources for a delay in HIPAA standards compliance states as well as for a number of others that oppose a delay. The NCVHS has not heard a delay proposal that we can support.
Thee is a new sentence right here which says "a delay that is not coupled with a strategy to productively utilize the additional time is unlikely to contribute to a successful implementation. Supportable proposals for delay much include a rationale and a sequence of measurable events that would lead to successful implementation for the whole healthcare industry."
We've added one sentence there that you don't have, but it doesn't really change the meaning of that item three. Would you like me to pause for a moment to see.
DR. LUMPKIN: Any comments on the introductory section before the recommendations? I have one. The goal of all of this is to achieve administrative simplification not to adopt code sets and standards. I don't think we say that, that concern about the delay is that this is a first important step in moving forward towards administrative simplification and rationalization of the health information process. Perhaps some introductory statement like that would speak to the Secretary in a way that this doesn't.
MS. FYFFE: So we need to get out of the weeds a little bit here?
DR. LUMPKIN: To put into context.
DR. COHN: Good idea. Other comments?
MS. GREENBERG: Can you read the new sentence again?
DR. COHN: Oh, sure. This is right after, under 3 where the word support is which is the second sentence. It says "a delay that is not coupled with a strategy to productively utilize the additional time is unlikely to contribute to a successful implementation." The recommendations obviously fall from these observations. Recommendations:
1. To achieve the desired rapid publication of final rules, the NCVHS recommends that you take a leadership role in providing for orderly transition and implementation. This role could include, one, providing early guidance. HHS could provide early guidance to the industry through letters or white papers revealing the directions being taken by HHS.
For example, a letter or white paper illustrating the size, format, and rules for assignment for the national provider identifier could help covered entities make progress on system changes before the final rules were issued.
2. Opposing delays. The NCVHS opposes any of the proposed delays in the compliance states for the HIPAA standards. The Committee particularly opposes open ended delays of compliance dates such as that found in Senate Bill 836 as being detrimental to the industry and the possibility of a successful implementation process.
It is imperative that we not undermine current implementation efforts and continue to promote the urgency of working on the implementation of standards.
The NCVHS heard that a fixed delay up to one year would not result in significant stoppage and would allow the industry to continue on an orderly implementation schedule. Any delay greater than one year would incur the risk of significant stoppage and disruption because most entities in the healthcare system must comply with the standards to provide the benefits for all. This is a critical point. Forget the last sentence.
DR. LUMPKIN: So we're scratching the last sentence?
DR. COHN: Yes, the last sentence doesn't make any sense at this point. Comments?
DR. LUMPKIN: This says we want you to proceed quickly with the other rules for which you have not been final, that for the rules that aren't final, it would behoove the nation as people have a hint about which way it's going.
DR. COHN: Especially for those that have been out for the NPRM and where the final rule is known internally to HHS. It's obviously inappropriate to start giving the industry guidance on things that haven't yet been subjected to an NPRM process.
DR. LUMPKIN: Then we're opposed to delay.
DR. COHN: Opposed to delay. Exactly.
DR. LUMPKIN: That delay greater than a year is particularly detrimental.
DR. COHN: Exactly. It's the open-ended that's especially detrimental.
DR. LUMPKIN: Any questions or comments on that? Okay. Next one.
DR. COHN: If a delay is considered, we recommend that it be accompanied by a schedule of intermediate accomplishments so that the industry continues to move forward towards an orderly and successful implementation. For example, these are really meant as examples, healthcare claims and remittance, advice transactions could continue on the current timetable and other transactions could be delayed to support a phased implementation.
Or all transactions could be required to be ready for implementation by the compliance date, but additional time could be given for implementation and testing with trading partners. Or all transactions could be required for implementation by the compliance date, but the requirements for implementing adopted code sets or identifiers could be delayed.
DR. LUMPKIN: Any comments on that section?
MS. COLTIN: Among the examples that are provided, the last one seems to be very responsive to the concerns that were raised about why delay and the final rules for the identifiers haven't been published yet and there hasn't been adequate time to really get changes to code sets made by the transactions. They all have the same problem with identifiers or with coding. I'm not sure what it helps there at least in terms of the concerns.
DR. COHN: The third one is a concern that was heard. I think the subcommittee has heard a variety of different issues going all the way from local code sets which is the issue you're bringing up to not enough time, too much to do by the compliance date, et cetera.
There is not a single issue that by doing "x" you can solve the problem. We thought that it made sense to give the Secretary some examples as opposed to dictating exactly what ought to be happening. That was the rationale here.
MS. TRUDEL: Another one of the issues that we heard in the testimony was that testing was going to be a significant issue. It was going to require a long time allotted resources. The alternative that speaks to providing more time for testing is responding to that particular concern.
DR. COHN: The first one, at one point, we recommended a phased implementation even within the regular timeline and this adopts that and stretches it out a little bit.
DR. LUMPKIN: Let me toss an idea and ask if the Committee has looked at that issue. If we go into a hospital or other healthcare facility and we identify major problems with the physical plant, there is a process that under certification that begins which is to find them out of compliance and require if it's a significant problem that it be corrected within a certain period of time.
There are provisions under the CMS rules that allow for what's an extended plan of correction so that a facility, if they haven't been in compliance by a certain date, they would have a plan of correction that would be a year whereby they would file a plan and they would be held to implementing that plan and coming within that time frame. Has that concept at all been explored?
DR. COHN: I think we're including that in the next bullet which is flexibility and enforcement. Maybe we should look at that one and see if it meets what you're describing. You have a very good idea there and it doesn't exactly say that, but should I read the next bullet?
Flexibility and enforcement. Realizing that the administrative simplification subtitle requires compliance within 24 months, the NCVHS recommends that the HHSD explore the flexibility available for enforcement under Section 1176 of HIPAA and publish the intent to utilize this flexibility and enforcement to ameliorate industry concerns.
HHS may do this by monitoring the implementation situation in the industry and granting criteria based leniency in enforcement as being necessary. This doesn't quite address what you're saying, but close maybe.
DR. LUMPKIN: I like that point, but this would be an alternative if a delay is decided. Entities are required to make changes, that while we might delay the enforcement date for a year, at the end of that 24-month period of time, they would be required to file a plan saying that this is what they're going to do.
The danger, and I think what everybody is concerned, is that if you delay it a year, everybody says okay, I don't need to do anything for a year and we lose a year of implementation. Filing a plan at the end of the 24 month time frame would require that they actually have an implementation plan, that they have begun to make progress and they're showing that since it takes longer than the 24 months to do i, that they're actually making progress during that period of time.
It makes it different than enforcement. Enforcement flexibility says you screwed up, you're under enforcement, but we're not going to penalize you for a period of time. This says we're not going to hold you to the standard as long as you have an adequate plan to come into compliance within a certain period of time. It actually has a different feel to it.
DR. COHN: This would be one of those examples.
DR. LUMPKIN: As an example.
MR. SCANLON: That cold be one of the criteria that you suggest here for forbearance.
DR. COHN: That's's the question. Is it a criteria related for flexibility and enforcement or related to an example about how delay might be handled?
DR. LUMPKIN: The problem with flexibility of enforcement, once you get into an enforcement environment, your costs are going up because now you've got to hire lawyers. Just having run a regulatory agency, there's a big difference between delaying or being flexible in enforcement and delaying implementation. If you file a plan, then you just have your IT folks work on the plan, put it in and that's what you file with CMS or whoever the enforcement agency is.
If you now, make your 24 months, if it's enforcement, first of all, you've got to be caught. Once you're caught, then you've got to have lawyers to file your plan for you because you're afraid of being penalized. It just changes the environment a lot.
MR. ALTERESQUE: This would be the fourth example that you're proposing.
DR. LUMPKIN: Right.
DR. ZUBELDIA: So you're proposing as one of the intermediate accomplishments, the creation and filing of a plan?
DR. LUMPKIN: That's correct.
MS. FYFFE: On the one hand, you have the lawyers and on the other hand, you have HHS staff that's already in the hole not being able to review plans. There is no staffing for that. Six in one, half dozen in the other.
DR. LUMPKIN: But if you have a plan filed, you don't need to worry about enforcement, whether or not to adhere by HHS. I think the concern of people despite the fact hat I haven't heard that there is a large enforcement force that's been hired by HHS is that they don't want to be cited.
If they filed a plan, and there's some untoward event occurred, they may consider themselves if they're in compliance, one is that HHS may determine whether or not they're in compliance, but the other is that a court of law may say you're required to have this in place, and you didn't. So and so got harmed, now you're liable.
DR. FRIEDMAN: The only concern that we might have about this is that if we put that in this letter at this time, there is going to be other people besides the Secretary who will be reading this letter, including people probably on Capitol Hill. I'm wondering whether some folks will feel as if that is very open-ended. The other examples tend to be a little tighter.
They're basically saying you must go forward. The option that you've just mentioned almost seems to communicate that if you don't make it and you just have a plan at the end of the two years, you're okay.
DR. LUMPKIN: A plan to come into compliance within 12 months. It's looking at that same time frame. I don't think we need to spend a lot of time.
DR. ZUBELDIA: The only concern I have with the plan is that even though it's a good mechanism to avoid fines and avoid a compliance deadline, it doesn't contribute much to administration simplification. It needs to contribute to administration simplification I think. And the plan is actually a delay mechanism.
DR. LUMPKIN: You were looking at it in good faith. Maybe we can spend time in subcommittee.
DR. COHN: Let me move through the remainder of this. The Committee recognizes that to achieve rapid publication of the final rules, it is essential that you apply the necessary resources to reaching this goal.
This would include devoting the required means to promulgating the rules and implementing the identifiers. Prolonging the process of assigning identifiers will cause increased expenses to those who must run parallel systems to accommodate the old and new numbering systems simultaneously.
The NCVHS recommends that once a system for assigning provider identifiers is ready, identifiers should be issued very rapidly to all providers so that transition time is limited.
Three, the Committee believes that HHHS should expedite the HIPAA change process so that changes and standards that are voted on by industry and recommended by NCVHS are promulgated in final rules as soon as possible. The example like we just discussed previously. To assure this process, the NCVHS recommends that you provide the necessary leadership and resources to have these changes advanced as soon as industry consensus is reached.
Four, because the data requirements of the HIPAA standards for electronic claims are not identical to those used on the current paper claims form, the NCVHS recommends that the HHS asks that the NUCC and NUBC explore ways such as issuing in implementation guide for each paper form to make the data used on paper forms consistent with the HIPAA standards. Sincerely yours, John Lumpkin, MD.
(Laughter.)
DR. LUMPKIN: Any other questions or comments, concerns for the subcommittee.
MS. COLTIN: One issue that isn't spoken to in the letter and I don't know if it's because no one brought it up is federal agencies requiring implementation deadlines for some of these standards prior to the dates that are laid out for HIPAA.
The example is in this submission on encounter data to HCFA for Medicare Plus Choice plans, they're basically saying that if those claims or encounters do not contain the standard codes, they're going to be rejected and sent back by July as opposed to the October timeframe, July 2002.
That's raised an issue because essentially at least for that population in the health plans, it means that you have three months less, almost four months less to come into compliance and the code sets are one of the more problematic areas for doing that, getting rid of the local codes that quickly. I don't know if that was brought up at all, but it's something that has been addressed.
DR. COHN: The subcommittee will discuss that issue. I'm a little concerned about putting too much in this letter and it may turn out to be immobilizing for the Secretary if you don't watch out, but I certainly share your concern. If we may all the changes, is the Committee comfortable with the letter as it's been laid out basically?
MR. SCANLON: Basically, no delay. No delay.
DR. LUMPKIN: You feel duly guided? At this point, I think we have completed our business for this afternoon's session. We'll be breaking into subcommittees. The subcommittee on standards and security is right here. And populations is Room 2. NHII will be back here.
(The meeting of the full Committee was concluded at 2:30pm.)