Conference room E-275C, Lowrise
John F. Kennedy
Federal Building
100 Cambridge Street
Boston, Massachusetts
DR. IEZZONI: Okay, we are going to start again for the afternoon session. Elliot, do you want a minute to kind of digest your food? You can start.
MR. STONE: I'll start any time you want.
DR. IEZZONI: Why don't we start with Rob because I thought poor Elliot has had lunch. Let's give him a chance to just kind of digest.
There has been a little feedback during the morning with the microphone, so if people feel that is a problem, let us know, and we'll try to ask the sound folks to fix it. I would also mention that I hear that people had trouble hearing the committee or the transcriptionist had trouble hearing the committee this morning, so when people make comments, if you could speak into the microphone so the new transcriptionist can understand. I guess you have to put the microphone right in front of your mouth to be heard, so, Rob, that's for you too. Thanks Rob.
MR. RESTUCCIA: I'd like to thank you for inviting me here today. My name is Rob Restuccia. I'm Executive Director of Health Care For All, a consumer advocacy agency that represents consumers in Massachusetts around issues related to health care.
We do this in many ways. First, we do it by answering people's questions about access to health care and assisting them in getting the needed care. Last year we got 6,000 phone calls from people to the health line. Most of the people were uninsured and were looking for health care.
We also do legislative and policy advocacy. We led a campaign For Chapter 203 in 1996 which expands health access for children, low-income people and seniors. And we're currently leading a campaign in Massachusetts for a report on the managed care system. We represent consumers on various commissions that are set up to address issues, such as, free care, managed care and dental health.
Health Care For All believes that managed care can result in higher quality care and improved access for Medicaid beneficiaries. If managed care programs are implemented only to cut costs and without careful planning and strong state oversight, there is obviously ahuge risk to beneficiaries.
As populations that are previously exempt from managed care, they're covered under managed care programs, these programs face new challenges. We know in Massachusetts we have expanded the Medicaid program dramatically here. It now includes the homeless population. We have a waiver to include the duly eligible population, both Medicare and Medicaid. So Massachusetts is pushing the frontiers of managed care.
We know that outcome studies and consumer surveys have shown that low-income, disabled and chronically-ill Americans fare less well in managed care than in traditional plans, and I think our organization believes that the focus of managed care quality ought to be on the high-risk populations.
How do we protect consumers and improve quality of care for Medicaid beneficiaries?
Data obviously is a very important component of this. Some have suggested that competition is the key to high quality, and that we need to develop information that will assist consumers in choosing their best plans. I don't know if people have ever experienced helping someone choose a plan. Medicaid recipients will receive in the mail a large folder with seven choices, and there will be a long description of each plan. I've had the experience of people coming up to me and saying, "Well, I have these seven choices. Why should I choose one versus another?"
Thinking about how to do that, what is the data needs that someone actually could -- what would the information that people would have to actually assist them in better choosing a plan? It is quite challenging to think about. Particularly, in Massachusetts where all providers basically are in most of the plans.
I think -- we think that the view that data measuring quality of care will make the market work more effectively is misguided. Note data has proven to be very helpful to beneficiaries in choosing between plans because of factors such as the market risk adjusters, the nature of Medicaid enrollment and differences in methodology.
Comparative indicators that come out of HEDIS, for example, and other types of measurement systems that are developed are really not geared to the individual; they're really geared to the benefit manager who has many more resources to think about, and the whole approach to choosing a health plan is quite different. For example, thinking about the relevance of birth weight is lost on most consumers.
I think that the data must be viewed not as an end to itself, but as a component in an overall quality control system. Data can help raise questions about what accounts for differences, but in and of itself is not too sufficient to provide answers. Data will be helpful in assisting the state in writing and enforcing contracts, helping plans in their quality and improvement efforts. The department here in Massachusetts needs to develop a stronger program that will assist in explaining patterns that the data suggests.
As I had mentioned, in Massachusetts we're looking at sort of expanding Medicaid to include the homeless population and duly eligible, people who are qualifying for both Medicare and Medicaid, and that presents particular challenges.
Readily available data, such as reports by purchaser groups, tend to lump seriously and chronically-ill together with people with minor illnesses, making it hard to tease out experience in the sicker population. This speaks to the need to oversample certain populations.
Key areas: Home care, dental care, mental health, where Medicaid provides more comprehensive benefits than the typical insurance package, there is very little data available, the availability and the use of those services, and tying it back into the Medicaid population.
I think when you look at the homeless population, for example, you think about how will managed care work with someone who has no permanent address, no phone, and will people actually be able to choose a health plan? Will this choosing of health plan be based on what basis? Will they be going to the network that they choose? What will providers end up -- will the disproportionate share providers who have traditionally served these people still continue to serve these people? Although maybe they have already chosen Harvard or some other plan.
Looking at preventible hospitalizations within this category may be a way to evaluate whether people are getting appropriate access of primary care.
I think the future is now. We have entered the brave new world of health care. We're seeing a shift from for-profit to non-profit -- from non-profit to for-profit health care across the country and in Massachusetts. I think it presents a challenge for all of us who in the past relied on the good-will of providers and health plans to provide high-quality care to anyone. We're seeing much more of a profit-motive-driving health care system. And I think for those folks who are looking at issues relating to quality, it presents new challenges for you and for us.
Thank you.
DR. IEZZONI: Thank you. Why don't we hold questions. And, Elliot.
MR. STONE: Ms. Iezzoni, Members of the Committee, I did bring some copies of my remarks which will be available to the committee members. So, for the record, my name is Elliot Stone, and I'm the Executive Director and the CEO of Massachusetts Health Data Consortium.
With me today is Sandy Harrington who is our chief information officer and the director of our affiliated health and commission network project.
The consortium is a private sector and not-for-profit regional data repository, and I appreciate the opportunity to speak with you today about data collection issues in managed care and, actually, non-managed care, for Medicaid and for non-Medicaid populations.
Our testimony today is not so much about policy as it is about process, the process of how to collect the data, but the goal of creating standardized data repositories to assist research and analysis. My testimony is not only about Medicaid and managed care, but about the need for adoption of data standards across insurers, across purchasers and health plans and providers.
I think just as this committee is seeking advice today about how Medicaid wants to research the data issues, there are activities going on around the country in all 50 states: Medicaid trying to determine what they want to collect in every state, as well as Harvard, Pilgrim Health Care trying to determine how they want to manage health care, and Tufts Health Plan and Blue Cross trying to determine how they want to manage Medicaid. And all of them in their own way, some publicly, as this committee is doing, and some privately are trying to determine what are the data sets they want to have in their public warehouses or private warehouses. And, so, I think we need to keep that in mind as we have this conversation about data sets.
As Medicaid recipients are increasingly enrolled in managed care plans, the data collection efforts will only be eased if data are collected and reported on in a standardized format by all the users and all the holders of the data. I don't have to tell this committee, but, for the record, the Administrative Simplification Provisions of HIPAA support the adoption of standardized administrative transactions for a wide range of activities: Enrollment, eligibility, referrals, benefits, claims, claims attachments, premium billing, coordination of benefits. The point of mentioning those is to say that all of those aspects can be rolled up into Medicaid managed care data sets.
Again, we, at the Health Data Consortium, believe these requirements must be adopted across the sectors; that we should be careful about saying that this is just a Medicaid managed care data set. It may apply just as well to private sectors managing patients as well as just managing the Medicaid population.
So what I wanted to do today is reiterate some testimony I gave a while ago to the National Committee, back in December 1995, looking at core data set issues, and through this testimony to urge the National Committee and the National Library of Medicine to rapidly develop categories of data elements into what I'm calling a meta-dictionary to be used across the data dictionaries that are held in public agencies, in private agencies, and to begin to look at these categories rather than the specific elements themselves.
And I would direct the committee's attention to a couple of publications. Actually, Sandy, in my briefcase back there, somewhere we have got the IOM report. So this is a report that the Institute of Medicine did a few years ago that I would direct the committee's attention to for the record, and in that Institute of Medicine report there is a list of categories that I put at the back of my testimony, and I draw your attention to that, to those seven categories in the back of my piece that are in the IOM report.
When our consortium was created 20 years ago, one of the first things that the chairman of our board talked about for years was making sure we're clear about what the question is, and when the National Committee on Vital and Health Statistics looked at this topic a few years ago about core data elements, just coincidentally I had gotten a book from Dr. Densen that gives us some sense of humility. And it was written in 1972, and it was guidelines for producing core data sets for health care plans.
So the National Committee has been at this for a while. I think that is really part of the major theme that, in my testimony, it takes a long time to get adoption of these things. NCHS was thinking about this 26 years ago: How to get health plans to think about Medicaid as a health plan, thinking about what the minimum data sets should be. And here we are in 1998 continuing to talk about those guidelines. So we continue to have a debt to Paul Densen.
And my way of looking at this, to answer his favorite opener, "What is the question," is to frame the question the following way: Can we develop a set of core health data elements on persons and encounters or events that can serve multiple purposes and would benefit from standardization?
My answer, of course, is yes, but only if the National Committee is faithful to the other key concepts that we discussed previously. I agree that the National Committee can provide significant value-added to HIPAA implementation and to what is going on in the states as long as the National Committee's approach acknowledges these three concepts:
First, that the core data elements that we've been talking about since the 70s will have multiple and unanticipated uses; second, that the core data elements will have definitions which will be refined and changed continuously; and, third, that the core data elements that might come out of this investigation and your hearings, these core data elements will take a long time to evolve into general practice.
I wanted you to know about some of the planning that we're doing locally, and to go beyond the traditional inpatient and outpatient data sets that we're most familiar with.
And, again, to reiterate, these categories, because I think they're very helpful to keep thinking and going back to these seven categories of demographic data, administrative data, health risks, health status, patient medical history, current management of health conditions and outcomes data, all of which are going to be essential to the Department of Health and Human Service's responsibilities.
So we collect a lot of inpatient data, we collect millions of outpatient records, but we've been also looking at this phenomenon of CHINs and CHMISs around the country, those states that had Hartford (phonetic) foundation grants. And we've been looking at what they've been doing for eight years. And any time they went through processes like this in their own states to think about how to share core data sets and centralized repositories, they struggled, and these states tried to anticipate the needs of all users, but, in fact, many of these projects, these state-wide projects, had big problems and may have failed because of this idea of centralized data requirements, putting all the data into one big repository.
On our own we could very well tell you about all kinds of gaps that we uncovered over the 20 years that we've been collecting data. There is no question that there are gaps in many areas that we'd like to look at, but, again, they fall within those categories. We know that we don't have data elements in each of those seven categories, particularly in the areas of outcomes and the areas of health status and current conditions.
And, in fact, what we're finding more and more, as we work with the health plans and the hospitals and our technology partners, is that these gaps are not just in our centralized data warehouses; they're in all of the holders, the hospitals and the health plans.
One of the vendors that we talked to, one of the information technology vendors, told us that they have a data dictionary with 14,000 data elements in it. I don't know if the National Committee is planning to do anything like that on the list, but they are very proud of the fact that they have 14,000 -- you know, they could point to a data model that has 14,000 elements in it. You know, what do you do with that? How do you categorize that? You know, to them, they were very proud of saying we have 14,000 elements. And I haven't the slightest idea what to do with that, other than to try to encourage them to begin to categorize it, to assist others.
So we have - I'm going to skip through some of the things that we've done locally here, but, you know, essentially we have gone through the data wars about what kind of information should be stored, what kind of information should be shared. But having gone through that, we've come up with a new approach since 1994, and we call this approach the Affiliated Health Networks Project. And what it talks to is that instead of trying to anticipate all future uses of one big centralized community database, the preferred approach is to focus on commonality of the elements, the commonality of terms and of definitions. And what we're trying to do is facilitate the connectivity between and among competing organizations: Harvard, Pilgrim and Tufts Health Plan, the Tufts Health Plan and Blue Cross. The patients move around from place to place, and we want to make sure there is a longitudinal record that follows that patient asthey move from health plan to health plan.
One average we heard was that people might have seven health plans over their lifetime. That seems pretty reasonable. As I sort of checked that off in my own head, you know, from birth to retirement, that's pretty reasonable. But do your records follow you over time? It's unlikely that the separate islands of information that we have between all the different health plans don't allow that to happen.
So our new approach is to kind of facilitate this kind of connectivity. What we think is that the future is going to be less about these centralized data repositories than it is about instead networks of networks; not silos of data and not central data repositories, but more and more networking. That's not to say data won't be aggregated, we think the data will be aggregated. We hope to be an aggregator continuously for the next 20 years. We tend to be an aggregator, but on a need-to-know basis. We would like to go out and aggregate information from Harvard, Pilgrim and Tufts Health Plan and Blue Cross and Fallon and Medicaid, but do that on a need-to-know basis, knowing that every silo that we talk to has a common set of elements and common definitions.
You ask a question about funding in your preliminary questions, and just for the record, whether it's managed care or non-managed care, I don't know whether the committee is going to make recommendations on it, but what we have continuously talked about and given testimony about is the idea that there ought to be some administrative information systems feed that is recognized by every payer, that is built into the cost of the encounter and that enables the development of standardized data categories. And I'll keep getting back to these standardized categories. To me, it's sort of like a chart of accounts for data.
All accountants know what their chart of account looks like. But data managers, data warehouse managers, don't know that there should be a certain structure, a certain categorization of data no matter what facility you work for, and there should be that chart of accounts. I think it relates to the seven categories from the IOM, that every data warehouse, health plan hospital, ought to have this kind of chart of accounts to say what are we doing, what do we have in our outcomes data sets, what do we have in our health status data sets.
I'll just skip through some of the things that we're doing locally, except to say we have very active work groups trying to implement what you are recommending. You look to the standards groups, the National Committee makes their recommendations on standards, and what we try to do locally is implement your recommendations, and we are doing that with a clinical group and a business EDI group, and we have a provider identifier group that's trying to take all of your recommendations and roll them out for all the hospitals and all the health plans together in our region. And what we do is we have a big public meeting every year, and we talk about how we are implementing the National Committee's standards, and I'll leave this behind. Actually, I think we mailed one off to every member of the National Committee. I'll leave them behind for staff.
In that public meeting report we keep trying to raise the awareness of what the National Committee is doing, what HIPAA is about, and how we can implement these standards locally.
So let me get to my recommendations on some of the questions you've asked me, and how I think we might be able to try what we feel ought to happen with Medicaid managed care data sets.
First, we recommend that the holders of national databases, and this applies to the holders of local databases, but mostly the holders of national databases: Medicare, Medicaid, HCFA in particular, should take a leadership role in data set standards. We would like to recommend that databases of any significance should be expected to list themselves or register themselves in what we might call a national compendium or a meta-dictionary. And we would recommend that this be done on a scenario basis; that is, let's look at what's needed for acute care, what data elements are needed for chronic care, what data elements are needed for outcomes research or for maternal or child health or for mental health. And we think that these ought to be done in that sort of scenario basis.
Our own clinical work group did that for emergency rooms, and we collaborated very nicely with CDC and with their Data Elements for Emergency Room Department systems (DEEDS). And DEEDS, I think, is a nice example of how you can do this for one department. Peggy McManus has done a lot of work in chronic care and has a data set there as well.
We think that the holders and users of the data sets should be expected themselves to make refinements, and we'd like somebody to step forward to consider maintaining some kind of compendium, some kind of meta-dictionary on the internet to let everybody know what's in these data sets.
It's like the Genome project in some ways because as the compendium evolves and you know more about it, the National Committee could then reach consensus on minimum data sets and definitions and guidelines. We also would recommend that this committee recommend to ANSI that you establish a public health work group with the CDC so we could look for data sets on public health coming out of ANSI. I want to reiterate our recommendation that there be an accessible, on-line compendium or meta-dictionary of data elements. We think -- to me, what this means is anyone who has a data set, whether it is HCFA or Medicaid or Medicare or any health plan or any hospital were to register their data set, that each Medicaid agency would let the world know in this compendium what they're collecting. All of our energy tends to go to the National Committee to say let's narrow this down to 32 elements or something, but that is not a reflection of what's really happening out there.
We're working -- we think that if we had this kind of a compendium, if we had this kind of a meta-dictionary, where any one of us working on a project, like our emergency room project, we could see what each state is collecting on emergency room data. We know that's how things happen now. One Medicaid agency sends a memo around of their data dictionary to another state agency, and lo and behold we see that one state's data set looks like another state's data set regulations with the typos written right into it. You know, it kind of goes around the country and moves from place to place, and we know that that happens.
What we're recommending is if we had a compendium, perhaps done by the National Center, perhaps done by the National Library of Medicine, then it would be easier for those of us who need answers tomorrow, we can't wait for 26 years for minimum data sets to come out.
What we need is something interactive, we need something that is timely so we can quickly see what data are being collected on outcomes, what data are being collected in any one of these seven categories so that we can build our own data warehouses, and we can make recommendations to the hospitals in our region and health plans in our region to do that.
We're going to use all of the standards that come out of this committee, especially starting probably with the provider ID, and especially starting with the eligibility data sets, the 270 and 271 transactions. So you are going -- and those kinds of things. Eligibility wouldn't normally find their way into Medicaid managed care data set recommendations, you know? I mean, it's not usually the kind of thing that one gives testimony about. But, to us, it's -- you know, it's very important that these unique provider numbers get implemented immediately. It's very important that the eligibility gets implemented immediately.
To summarize, I would say that what I've said here, what I'm proposing, is not welcome advice for the Type A's among us who want kind of a neat list of 32 elements. It's going to -- Medicaid managed care data sets are going to be messy for a long long time. The hospitals are going to do it differently than the health plans, and each health plan is going to do it differently. What I would like to recommend is small incremental steps forward: Cataloging, cross-referencing and understanding the diversity of what every health plan is doing, what every hospital is doing, what every Medicaid agency is doing.
We would like to see some kind of a growing, on-line compendium, the meta-dictionary initiated by the Committee, perhaps in a collaborate venture with the National Library of Medicine. We think this would help evolve toward consensus on core data sets. We don't think we should spend a lot of time just trying to make a list of the top 32. We think you have to evolve toward that. Maybe we'll evolve toward all 14,000 data elements that this vendor is talking to us about. But we don't even know what data elements are out there right now, we can't be sure.
If a health plan here wants to know what to put into their data elements, it's not that easy to find out quickly what the best and the brightest have in their data warehouses and who uses the data elements and for what purpose they're used. And what we're trying to do is reduce the variation among all these competing formats, and reduce the variations about all these competing definitions out there.
And we would recommend that the National Committee should avoid the chaos of trying to choose one winning definition. That seems to be what the process has been like over the 20 years. We're going to come out with one definition, and all the discussions and all the meetings are about word-smithing those definitions. And we would strongly recommend that there be a different approach to that, and instead let the market prevail to some extent in data and, you know, in data collection, and see what is being collected by whom, and what their definitions are. And once we look across all these things and let everyone see what's being done, then I think there will be a lot more consensus among those who need to do it for research and those who need to do it for patient management and those who need to do it for analysis or other purposes.
So, in closing, I would say that HIPAA has created the environment for administrative simplification to standardize and simplify the ways in which data are collected, and I can't think of any better committee to take on this task. I think the National Committee has shown great patience and persistence and has always taken on the long-term view.
Our Health Data Consortium would be happy to help in any way they can to accelerate this idea of a compendium or a meta-dictionary of data elements. So thank you for this opportunity to present my view.
DR. IEZZONI: Thanks, Elliot. Neil Cronin was our last speaker on this panel. There doesn't seem to be somebody who would fit with the name Neil sitting in the audience right now. No, no, maybe he is with the taxpayers downstairs.
Before we open the floor for questions, you know, apropos to what Elliot just said, I cornered Brian Burwell in the hallway outside before lunch because I wanted to ask him a question, but we ran out of time. I wanted to ask him whether the MEDSTAT proposal for the data reporting requirements for Medicaid in Massachusetts, that he mentioned that MEDSTAT was contracted to do, whether those reporting requirements matched HIPAA recommendations. He said no. He said no. And he said, you know, "This is really funny because Rosanna Coffee, who some of you know, who worked at HCPR, now works at MEDSTAT, so we should have been aware of all this, and we really need to look into this." My jaw is about on the floor. I said, "Brian, you know, the HIPAA standards are going to apply to every health care transaction of the specified types regardless of the payer, and it really ought to be something that should be considered." And so I just found that fascinating, not because it's a criticism of MEDSTAT at all. I'm sure that MEDSTAT is doing a very credible job, as they do on all of their contracts. It's more kind of instruction to us on the Committee about how the HIPAA issues just still haven't permeated throughout the interstices of people who have to know about HIPAA. Just kind of interesting.
Elliot, do you know anything about what MEDSTAT is doing in terms of their contract with the state for data reporting mandates for Medicaid, specifically which is our interest here?
MR. STONE: No, because we have not -- no, we have not been asked to -- we reach out to them. We tell them what we're doing, and we have a lot of different forms and we have a lot of different work groups, so in general MEDSTAT knows, but not on this particular contract I haven't talked to Brian about it. But, yes, this relates to other testimony we have given.
There is huge awareness problems out there. We're trying, with our colleagues from Minnesota, Walter Squarez from Minnesota Health Data Institute, we are doing a lot of things.
We have been in Seattle. So I think there is pockets of it.
DR. IEZZONI: Right. I just guess around Medicaid specifically since right now there is a big opportunity for designing new information systems since there are so many new data kind of suggestion mandates.
MR. STONE: Medicaid in general, Lisa, is not just among our friends here. I don't think they have totally embraced HIPAA yet.
DR. IEZZONI: That's the point I was going to make. I think that there is still a lacuna there.
MR. STONE: Yeah. So it's not likely to be in their RFP. All it needed was a sentence from Medicaid saying make sure you're HIPAA compliant in your recommendations that you come back to us with. And that's all we're saying to every health plan and every hospital in our project, to any vendor they worked with: Whatever you come back to us with, are you HIPAA compliant period. It doesn't take a heck of a lot more than that, and I think we need to encourage every vendor to be aware of it, and I think we need to encourage every health plan and provider to expect it.
But Medicaid -- there's two concerns I think that Medicaid has about this: One, that it could be perceived as an unfunded mandate; and, second, the concern that if Medicaid asks them to be compliant, there may be additional costs that Medicaid -- that hospitals and health plans might come back to Medicaid and say, well, if you want me to be compliant, it's going to cost a lot more than if I wasn't.
DR. IEZZONI: But the point is, Elliot, that HIPPA doesn't just apply to Medicaid; it's Medicare, isn't that correct Marjorie, across the health system regardless of payer?
MS. GREENBERG: Well, I agree with everything that you've said, and as some of you know, I was talking to Elliot about this just before the meeting. NCHS sponsored a planning meeting for a workshop that is now scheduled for November 2nd and 3rd. The planning meeting was in January. Dan Friedman attended; Rachael Block was there; we had some state Medicaid people; we had public health people; Rosanna Coffee was there of Health Services Research.
The kind of theme was what are the implications of HIPAA for public health and health services research. Because right now the HIPAA standards really have had very little input from those constituencies. The standards that are going to be adopted, the first ones out of the gate, will not have had a lot of input from those constituencies, and I think in some ways Medicaid is like a -- is more like public health than maybe a commercial payer; in other ways it's more like an insurer or purchaser than public health. So, I mean, it sort of bridges both sides, I think. Clearly, it is taking care of populations and is dealing with some of the higher-risk populations, etc. which aligns Medicaid with public health, but at the same time it also is serving as a payer and as a purchaser of care.
So this is not an easy process, but we had quite a lot of agreement around the room at that planning meeting, that these different constituencies really needed to be talking to each other and understanding what they can contribute to standards development in the context of HIPAA because HIPAA really, if you are talking about encounter data or enrollment data or eligibility data, has to be the context. And I'm realistic about, you know, what we can accomplish in one workshop, but what we're looking for is actually developing an infrastructure to inform this process over time.
And I think that fits in very nicely with what Elliot was saying. This is not an activity for short-term members; this is really something that is a continual process.
There does seem to be a disconnect here though, and I don't know if I can articulate it well. I agree with what -- Elliot, with what you're saying about the compendium, of the meta-dictionary, meta-data dictionary. In fact, Bob Maze (phonetic) at HCFA is very much involved in trying to sell this concept to the department. It has generally a positive feedback and positive response from the Health Data Standards Committee of the Data Consul. He is working very closely with the Department of Defense, which if they heard about your venture with 14,000 elements, they'd say why so few. So, you know, talk about a mammoth system, and many of which are crossover to health care of any population, and then a number of which relate to when you are in battle. So, you know, they have quite an extensive data dictionary. EPA is also working on this, and they're using as a model an Australian system that is available on the Web which I can provide you with later or anybody who is interested in which there is actually an international standard, an ISO standard, for registering data sets, and they're trying to actually -- I think they are starting a pilot project with some considerable support from the Department of Defense, and the VA as well is participating, to begin to register data elements and to -- they're going to register the core data elements. When the National Committee is recommending, they're going to register the HIPAA elements. But given that and the fact you are arguing over one definition, I agree, is so difficult, how does this fit with trying to get people to implement the HIPAA standards? Which really is onedefinition for every data element, given the variability that's out there. And I think this is really the challenge. I think, yes, at least people overtly have to say we want to be HIPAA compliant, but there's a lot between saying that and getting there, and I think that is the major challenge.
MR. STONE: There is a talk-show guy that always talks about BST, big scary trucks, on the road. There is a BSM, big scary manuals. If you have ever looked at the manuals for the HIPAA transactions, the implementation guides are big scary manuals, but something that's not as scary is the data dictionary that is coming out. We have seen some of the preliminary work that is coming out of the National Committee, and there are some nice, simple data dictionaries that you put out that summarize all of the elements across all the transactions that you recommended, and, people, if you register with them, that's great. People need to look at those and say, you know, well, I guess there is a lot of different names. There's patient. I can choose patient names, subscriber names. There is a lot of ways of coming up with the name field. There's lots of different variations on lots of other fields.
And, again, Lisa, I think the best thing is for those who put out the RFPs, to be at least aware that they should be asking to be HIPPA compliant.
DR. IEZZONI: Okay, let's broaden it. Do the committee members have questions for either Rob or Elliot? Hortensia.
DR. AMARO: I kind of wanted to go back to our initial questions which were, you know, as a subcommittee looking at this, are there current data systems in place that help us if somebody wanted to look at the issues of access and quality of care and how it's affecting populations as they switch to managed care? Are those in place? And, you know, I couldn't help but remember a recent incident in a community hearing that we had as the Boston Public Health Commission in a community in Boston I won't mention, and where a woman came up and said, you know, when I call my health care agency, I'm told that they will call me back because they don't speak my language, and they've yet to ever call me back, you know, and I've waited three or four hours by the phone. When I do finally manage to get through to get an appointment, it takes three months to get an appointment. When I get to the appointment, it takes several hours, and I'm kept waiting, and then when the doctor finally sees me, he always tells me to take a Tylenol.
I thought that was a very good description of the kind of issues from the patient's perspective that somehow we need to get to. I do think we need, clearly, data sets on patient's specific information. And what you have outlined, Elliot, is very helpful. What I was wondering, especially, Rob, have you given any thought to what kind of system factors we need to be tracing and tracking over time, and, you know, how if you have been involved in trying to do that?
MR. RESTUCCIA: Well, I actually haven't spent a lot of time on issues related to data; that is not sort of the focus of my organization, but we certainly have looked at issues that are facing the Medicaid population, and I think your experience, your description, of a person in a community in Boston is a typical one in a lot of ways.
In terms of my organization, the issues that constantly come up for the Medicaid population is lack of interpreter services. There are a number of HMOs that are providing health care to population groups that are not prepared to do that, and I think there is lack of interpreter services across the system, although I think we have done much better in Boston than New York, for example. Problems around dental care are enormous. We know, from surveys that have occurred within the schools, that as many as 40 percent of the kids in Boston don't have screening and have not had sealants put on their teeth. Anecdotally, I haven't seen any health plan, any data, related to that coming from a health plan or coming from the Medicaid office.
Although, I know Medicaid is very interested in this and have set up a task force, and I am on the Governor's Commission on Dental Care.
Issues related to home care are -- we're constantly getting concerns about what is happening with home care and failure of -- you know, I think what we see often happening is HMOs that are not normally used to providing care to Medicaid populations. Anything in that benefit package is somewhat at risk. So transportation, dental care and home care often times people are not getting. So I think, you know, as we move forward and as Medicaid expands into these new population groups, there are clearly challenges in thinking about how you actually measure this, how do you inform people around the choices that they're making.
Frankly, I feel like the data is probably more useful to the agencies that are enforcing the contracts than it is to the consumer making the choice. And, you know, my point on some levels, let's not -- let's look at data for that purpose and not think that we are going to translate it down for the consumer. So when the pregnant woman gets that brochure that, you know, goes this long (indicating) with all the seven plans, she says, well, should I choose HMO Blue, Tufts, should I choose the PCP plans. I think on some level what I am asking for is less choice because I think we need to raise the bar and perhaps eliminate some HMOs that aren't providing those services and not allow them to participate in the Medicaid program if they can't fulfill their program. I don't know if that answers your question.
MR. STONE: Let me tell you something we have done recently. We have looked across all of the on-line directories of the health plans who work with us and the hospitals who work with us, and we ask them what kind of data elements do you have about the physicians; what do you tell patients who are looking for a physician about the doctors. And we have a long list of data elements, and we've made a compendium of what's in the various physician directories that, you know, some of which are in the phone books, some of which are on-line, and only two of them list the language spoken by the physician.
And so what we're trying to do is bring prominence to that, and we'd say -- we intend to do a study this summer to ask patients to look at a long list of elements and then say to them what do you want to choose. Well, one of those might be a photograph. Well, of the 13 that are on-line directories, only one has a photograph of the doctor. I don't think any of them have the race of the doctor, ethnicity of the doctor, but they do have a photograph of the doctor at least in one of those directories. Some of them eventually may have the physician doing surgery for all I know, you know, you go to the directory and see them operate.
But we want to give prominence to what questions do Medicaid beneficiaries ask when they choose their doctor from a list of data elements, and feed that back to everyone to say, you know, most of them wanted to know about language or most of them wanted to see a photograph, and you don't have that in your...
DR. IEZZONI: Can I follow up on that specifically? A lot of the hospitals are creating websites, Elliot, where a lot of this information is stored. For example, there are photographs in our website about our primary care doctors.
But, Rob, how many of your pregnant women are logging on to the internet to choose, you know, to use the webs to choose their doctors? I mean, is that something that...
MR. RESTUCCIA: Realistically, people choose their physician or choose their health plan mainly through networks. You know, they ask their friends, they get experience. It isn't often that you have people sort of -- you know, this doesn't drop from the sky.
You know, in general, I guess I feel like we need to really think about this, and I think it is important to really identify -- to really provide information to people about cultural competence, and I think -- our experience, a lot of places that have interpreter services, for example, overstate what they have so, you say, well, they have Vietnamese. Well, they may have a housekeeper that speaks Vietnamese. But I think that the Latino Health Institute Study indicates that, you know, we need to create standards for things like cultural competence and services.
But I think we could and I think it would be really useful to create new organizational forms that really relate to these populations, and you see it, for instance, among parents of disabled children who have created an organization of Family Voices, and that's a national organization --
DR. IEZZONI: We heard from them this morning.
MR. RESTUCCIA: Right. They have a lot of trade, a lot of information around provider issues, around issues related to quality of care. And in some ways I think, you know, it may not be as scientifically valid or actually may be more scientifically valid because, you know, they're looking at it from the point of view of personal experience, and I think you see consumers gravitating towards surveys as opposed to hard data which, you know, presents some issues in terms of how you explain this.
But I think investing in that kind of network, investing in, you know, communities that are different cultures and creating some information from those organizations, that's really how people are going to end up choosing their Medicaid provider. They're going to turn to the local agency that works with physicians in the area. They're not likely going to go to a website and look at the Beth Israel physician list and say, oh, this guy looks good. You know, is he Hispanic or was he in Florida for a long time or something. You know, it's not really an effective way of communicating that information.
DR. IEZZONI: Paul, you had a question, and then Marjorie.
DR. NEWACHECK: This is a question for Robert. In your initial remarks, you pointed out the need to look beyond just simply our data systems, but the whole Medicaid population, lumping in the whole population, and that we really do need to have some ability to look at some populations that are vulnerable one way or the other. You pointed out, for example, the chronically-ill population. Are there other subpopulations that you think should be identified for purposes of monitoring and tracking managed care in Medicaid?
MR. RESTUCCIA: Well, certainly -- I think, certainly, people who are homeless which is now going to be enlarged upon in the Medicaid population in Massachusetts. I think people from different cultures, it will be interesting to see sort of that type, and I think will be very useful to sort of understand how they experience health care. Certainly, we know there is -- you know, people approach health care very differently. Issues related, for instance, to the compliance are very different based on cultural norms.
I think that as we look at the expansion of Medicaid and to see the kind of diversity that there is occurring in the Medicaid population now, I think we'll come up with other ways of dividing it up that may be useful.
But, you know, when we move into the brave new world of dealing with the duly eligible population, I think that is really going to present tremendous issues for us. I think that is a population that is not well served by either Medicaid or Medicare right now. We could do a much better job. And I think, you know, Carol and Bob spoke earlier or will speak, have shown that, but there are going to be some real challenges for us in figuring out how to break up those populations. I mean, Carol is proposing a very very -- you know, looking at risk adjustment and breaking it down very finely in terms of different stages. AIDS, for example.
DR. IEZZONI: Marjorie.
MS. GREENBERG: I just wanted to say, and this may be preaching to the converted around this room, but I think it is something to think about, as when the HIPAA standards you get published for public comment, the kind of feedback, some of the issues that we're discussing here, it's very important that those go into the comments as well as the more traditional comments that will be received from say payers, etc., because with the national provider file or system which will be used to identify providers, I can tell you that is very much of an uphill battle to get race and ethnicity as an element in that file, and language didn't even have a chance, and the data elements will be part of what we publish. You know, these huge rules, and people may not, you know, get to choose.
But I've been asked on several occasions to justify why having race and ethnicity of providers is an important element, it's not needed to enumerate the provider. And, of course, the focus is on data elements needed to enumerate the provider. Maybe it isn't needed at the national level, but, certainly, on the other hand, to look at access, it may well be. Language as well.
So I think it's very important these voices get heard in the review process because I think one of the big arguments against collecting that is that the plans or the payers or whoever it is enumerating people don't know it. So if they don't know it, this does tell us something. I mean, if they don't know it to include it, you know, as one of the data elements, they don't know it as a mechanism of cultural competence in meeting the needs and access needs of those publications.
DR. IEZZONI: Can I ask Rob a question that's kind of related to that? Rob, this morning we heard that some state organizations do not collect race and ethnicity on patients because they view it as potentially usable by people to discriminate; whereas, other states feel it is important to have it so you can look at patterns of access and care for people by race and ethnicity.
How would your group come down on that? Would you support in Massachusetts the collection of race and ethnicity?
MR. RESTUCCIA: I don't think we would have a problem. We have always argued for the importance of looking at that. So we don't have an official position. I mean, I think, actually, consumer organizations, frankly, there is still some time, you know, particularly involved in this area, and I think what your --
DR. IEZZONI: They should be Hortensia and I are saying.
MR. RESTUCCIA: You know, to some extent it's a result of a focus on other issues, and only recently do you see -- I think it is more -- it's becoming more of an issue among organizations that represent chronically-ill folks and disabled folks, but it's clearly not -- I mean, I represent a very large consumer organization comparatively, and it's not high an origin. So, you know, it has always been something we put on our agenda, partially because I have a familial connection to health quality, but it's not easy to get funded, and it's not easy to sustain. You know, provider organizations have tremendous resources; HMOs have tremendous resources. You know, Family Voices is really a great example of how, you know, a group of moms got together and sort of pushed through some important changes, and, you know, they're now working with Brandeis on national studies. You know, it's a big leap for those consumer organizations.
DR. IEZZONI: Vince, it looked like you had a comment.
DR. MOR: I'd like to go back to your vision of the compendium of data sets and the registration. When you ordered this menu of 14,000 data elements, not all that many. Actually, when you think about data elements, are those data elements that are aggregated definitions as well or actually only the raw collected bits of data? Because at any given level, there may be many many different data elements which are essentially composites of sort of microdata elements, but that, of course, makes it even more important they be very precisely defined because they could be readily manipulated.
So when you think of registering a data set, for example, are you just registering the definitions of the collective microdata elements or of the distributed and aggregate data elements?
MR. STONE: I think we need definitions of all of those things.
DR. MOR: Your vision, have you thought about it?
MR. STONE: We've thought about it, you know, in general. I think about it in terms of things like Dr. Amaro mentioned. For example, we work -- you know, what we find is that hospitals keep very detailed track of operating room time. Probably Dr. Amaro's consternation, we don't keep track of when patient's arrive in a waiting room and then are seen by the doctor for this type of visit, and then walk out of the room. But in an operating room, the data elements are, you know, you get rolled into the operating room so you are -- admit time to the operating room, skin time, when the scalpel touches the skin, the scalpel -- the physician --
DR. IEZZONI: It's called skin-to-skin time.
MR. STONE: Skin-to-skin time. The physician finishes. Then there's even places that keep track of the physician walking into the room; what time the surgeon walked into the room; what time the surgeon left the room; and then what time the patient left the room. All right, so I don't know what you call that in your definition, you know, the aggregation. But there is a sweep of data elements about what happened for O.R. time. I wish we had this for medical visits, but it is in great detail, and the vendors who offer those systems say to us we'll give you anything you want. You want skin-to-skin time, you want patient roll-in time, patient roll-out time, we've got operating room times coming out of our ears. And that causes some confusion when we try to recommend a standard. So we don't try to recommend a standard. We say here is what you can collect, here is what most people are doing, and here is what we probably really want to know. We really want to know for resource planning how long the patient was in the operating room. Because the biggest use of that information is the physician's efficiency and the turnover of the operating room. So to do the analysis on those two areas, which of all of these operating room times could you use? And there are -- there could be -- you know, as I've just said, there is quite a long list. So we want both. I mean, I think we want each of those items, the micro items, you know, skin-to-skin time and then, you know, the macro items.
I would say, though, that some of the data elements that you're looking for on the definition of race and ethnicity and the definition of access times, how long someone is waiting, many different providers do collect this information. I think the problem comes when the National Committee and others try to squeeze it into one short list and everyone starts getting worried about what's going to be required; what's going to be mandated; what will people use. Instead we take sort of a market -- I can't believe I'm saying this, but we take sort of a market-driven approach to it where we sort of lay out what people use, and that's how we did the operating room time. We just asked everybody. What does every hospital in our project do, we have about 20 hospitals, how do you keep track of operating room times. And that's when we were amazed when we saw this multiplicity of different times for one different event, and I think you would see that for everything.
If you said out there register your data set with us and tell us how you define ethnicity, register with us a category, tell us how you keep track of access or how long patients are in your facility, and, you know, waiting times, how do you measure, a lot of facilities absolutely do measure waiting times for patients. For marketing purposes, they do measure that.
If we had broad categories and allowed people to tell us and inform us how do you do this, then I think we slowly could inform the whole marketplace to say here's how the world -- here's how some of the best and the brightest are collecting ethnicity or collecting information about wait times.
DR. MOR: Recognizing that we Americans have a negative reaction to anyone ever telling us what to do, imagine this great marketplace of web-based ideas, how do you decide the relative utility of one form of a measure of data element versus another? For instance, you get your 20 hospitals together and they say, well, really, you know, while we use it for this little waiting time, O.R. time for this, and we actually do blood coagulation studies and all these other kinds of fun things, and it's useful for that, but really what we need is X.
MR. STONE: We went to the end use, we would go to people who do the end-use analysis. We would start at the end and say, well, we have people collecting all of this, but we would talk to the outcomes researcher or the resource planner and say what is the purpose of the end use of this study, and then go back and talk to the data collectors, and they will say, well, if the end use is resource planning, then you probably want these two operating room times.
If the end use is efficiency of the doctor, the surgeon, in the O.R., you probably want these elements. And it really depends on the end use.
DR. MOR: So it's end use definition.
MR. STONE: It's end use derived.
MR. RESTUCCIA: I think looking at the subpopulations that traditionally have been underserved and unrepresented in the health care system and need to be part of this because I think what we're seeing is, and I actually think opportunity -- Medicaid managed care has provided the user opportunity to change the system for some of those subgroups who previously were just not wanted by anybody.
When I started this Health Care For All, we were fighting for physicians to take Medicaid. You know, times have changed. And I think what we're seeing is the importance of sort of, you know, looking at populations that have historically been underserved, that have had problems with the health care system, and figuring out how can we collect data and use data in special ways, and I think what we're beginning to see is, you know, instead of homogenizing this, the average person may not need the same kind of data that a person who has a severe disability, a person who doesn't speak English, a person who is not connected to -- doesn't have a shelter. You know, so I think there are really ways to be very creative about this and to think about it from the consumer standpoint, and, you know, put the consumer first on this. How would we collect data to make it a much more responsive health care system for that person who came into the clinic and, you know, talked to a physician without an interpreter, you know, and the physician doesn't have a clue; or even worse, the physician thinks they speak Spanish, and they really don't, which, you know, some of the studies are showing now. It's quite remarkable.
So, you know, I think sort of thinking about it from the end user consumer point of view, we could be very creative in really developing a health care system that is much more responsive to these populations, and managed care could be a very important part of that.
MR. STONE: I would say that race and ethnicity is another good example of, you know, my testimony. We would not look to the National Committee to tell us how to code race and ethnicity if we were working in a community with a multiplicity of cultures because you're going to end up -- you know, the National Committee probably has to say all right, we'll have seven codes, you know, and everybody is going to fall into other. And in our own little clinic, in a community health center, we might have 30 different codes for a wide variety of constituencies that we have, and I would look to those folks to advise us.
If I was trying to advise another community health center, I would like to the compendium of data elements for people who serve populations that I'm interested in. I wouldn't look to the National Committee because all my patients would be under other because they're just not broken out the way that I would need it.
DR. IEZZONI: Elliot, I think I need to clarify something for you. Our subcommittee is not a core data elements promulgator nor is the entire committee right now focusing specifically on core data elements. However, when we did a number of years ago, specifically generated or prompted by the effort towards health care reform and thinking about the data and instruction that was needed to support health care reform should it have happened, the first thing we did was kind of do a market survey and look around to all the different people who were collecting these data and asking them how they were doing it, the problem was a lot of them were proprietary organizations who refused to tell us.
So I think that, you know, we also have kind of gone beyond, a step or two beyond, where we were back a few years ago. And so hopefully we’re all going to be very forward thinking at this point, and we enjoy learning from you.
I think, though, that it looks like there aren't anymore questions around from the subcommittee. Thank you Elliot and Rob. I understand that there is a minor emergency in the lobby of the building. In Faneuil Hall, okay, there has been a fire in Faneuil Hall, a small fire in Faneuil Hall, and so because of that there is a lot of fire trucks around, and it's making it harder for people to get here, and so that might be why Neil Cronin was unable to get here.
What I'd like to do is are any of our people for the 3:00 session here? One person, two people. Okay, great. So what we'll do is we'll reconvene at 3:00. We'll take a little break here. Reconvene at 3:00, and see how things go. If we break early, we'll break early.
Thank you Rob and Elliot. (Recess taken)
DR. IEZZONI: Okay, can we get started again. Carolyn is just passing around Phyllis Freeman's testimony. And, Phyllis, do you want to tell us a little bit about who you are, and let's hear from you.
MS. FREEMAN: Sure. Thanks very much for inviting me. I'm Phyllis Freeman from the University of Massachusetts, Professor and Chair of the Law Center and a senior fellow at the McCormick Institute of Public Affairs. And the two topics that are near to my heart that I am pleased to have a chance to talk with you about today are, first, I'm always, when I look at HIPAA and get in conversations with people who are concerned with data sets, always arguing for some semblance of equal time for health in the discussion along with financial and administrative concerns, and I want to make some remarks about the Medicaid managed care and how we might be using data to help improve the health of the population with regard to medical care and with regard to things that are broader than medical care. Many of the people have touched on that today, but not quite so directly. So I want to say a few things about that, and then talk about the privacy issues within the context of what are we trying to do or what could we at best be doing for this population in addition to worrying about the expense: who pays it, is it costing more than it needs to, and all these sorts of things that generally occupy people a great deal when they talk about Medicaid.
I'm also a fan of the committee's website, and have been a great consumer of a lot of your materials which I've used in teaching and a lot of other ways, so I want to appreciate you for posting as much as you do, and I have done written testimony for you today in that spirit in case you're going to do that again. And I don't need to say all of that to you, but I did it for your readership with the sites and all of those bells and whistles because I think it has been such an incredibly useful tool in my work, so I just wanted to support you in that way.
The first point I want to talk about, and I will shorten what I've given you considerably, I don't need to do all that for those of you who know this so well, is that there are a lot of reasons to collect Medicaid data and use it for improving health, not only with how we use and direct medical services, but also how we could be directing community intervention, school interventions, work site interventions to prevent people from exposures and risks that may be all the more, unfortunately, prevalent for Medicaid population for all the reasons that you know. And, yet, I don't hear that often in the data conversations. Although, it certainly came up several different times today, and Marjorie Greenberg was talking to me about it a little while ago how it is we could be combining public health and Medicaid and managed care data in ways that help us understand and do better with the health outcomes of this population. First, understanding health status, and then intervening and doing something about it. Our society, as you know, expects much more from medical care than medical care has ever been able to deliver, and much less from public health than it can accomplish, and that's my concern that my colleagues and spouse and I have written about a great deal, as sort of the Tony Robins, the centerpiece, of the remarks before I get to the privacy part.
And I'm afraid Massachusetts is not particularly an example of a state that has taken that to heart and had the Medicaid program operating in a full capacity as a public health organization, even though it's not the public health organization. It seems to me that some states have done better combining the two capacities of the public sector organizations to improve the health of the whole population and of the lower income, including Medicaid population.
So what I'd like to do is talk about for a minute the reasons why I think that data is important, what sort of data am I talking about, and then give you a few examples in states that have done something interesting. And as I was talking earlier with my colleague, Dr. Ackerman, I mean, she has some examples and experience where sometimes not all states asked for the information that may actually help improve health status the most because we may not know enough how to do that. And so I hope some of the examples I give you are contrast to that, but I don't think probably we know enough to be sure yet, but it's certainly in that spirit that I offer them, rather than, you know, the harassment of managed care organization which might be another other spirit which is not the one I'm reading.
I think the data that I'm talking about in order to accomplish health improvements in a very general sense, the day I'd like to have would be to understand and help us as a society make certain kinds of decisions, and I think the fundamental decisions sort of worth restating when and where to employ public health interventions that reach into the community to prevent disease and product health; how to make the best use of medical interventions to treat the ill and injured; how to direct research and development resources to the efforts that are most likely to improve the effectiveness and efficiency of all health enterprises, including public health programs; and how to allocate our resources intelligently among those activities. So I'm talking about data in a very broad sense because it seems to me that that overall is, you know, an important mission that Medicaid programs could be playing an even better role in.
I think Medicaid managed care is a special opportunity for a number of reasons, not the least of which being Medicaid is a public set of programs that don't live with the proprietary restrictions on data and have the ability to work with private provider organizations in a different way than when it's in the commercial sector, just, you know, as is true in some instances with Medicare as well.
And I've been particularly concerned when I look at data issues and in privacy issues, the whole issue of what's in the public sector and the public domain and what is not. I think it's terribly important for Medicaid programs to use their leverage via statutory/regulatory or the most recent of the mechanisms which are the contract provisions in Medicaid managed care contracts to elicit the data that really does help us understand the health status and population and to improve it.
Now, a lot of people who talk about Medicaid managed care in the world that I travel in immediately lead to the concern that by moving into managed care, that health for vulnerable populations will be sacrificed and that there was something magical that the providers, often Department of Public Health providers, or clinics funded through those mechanisms, have provided in the past, and some things could be lost for this patient population if they moved into the more commercially-based managed care organizations, and I'm sure you've collected lots of testimony about that in your travels.
It's a real question, but for me it gives rise to a broader question, and I want to just give an example from Florida, and I've cited in my testimony a study that was done, I think HCFA was involved, March of Dimes and RAND did the actual work, with prenatal care comparing a population both before and after the advent of managed care. And what the study suggests, without going into all the details about it, is that there was an unintended unfavorable effect on birth outcomes with the shift of the population into managed care. And what I think that that suggests that is important in this inquiry is that the State of Florida and any other state that is concerned with the same issues really needs to understand what was it that the other providers were doing and how were they doing it that it yielded better results. For me, it's not the magic of who, but what was being done, and that seems terribly important to understand. And I think one mechanism that Medicaid programs could use, if they use it wisely, it's very helpful, is to figure out how do we learn about those things and how do we use data in the managed care situation when you have whole populations to much better understand that. And that study didn't tell us much about that. It told us that the outcomes were different and not as favorable.
But the George Washington Center for health policy has done an interesting report called "Negotiating the New Health System" which you're probably familiar with. There looks like a little shine of familiarity there, in Lisa's face at least, where they have looked at the provisions in lots and lots of states in the Medicaid managed care contracts. I picked out our examples where I think collecting and reporting requirements actually look as though they may yield some very important, useful information over time. I mean, it hasn't been completed; it hasn't been evaluated. So this is really a hopeful list rather than a tried and true list of where the Medicaid program and the Public Health Departments did work together with the managed care organizations in ways that could be quite useful for actually improving the health of the Medicaid population.
The first example I want to mention is in Iowa where the interesting collaboration seemed to be between the managed care organizations under contract with the Medicaid agency and the WIC program. Not only did they make referrals which is, I think, fairly common, but they also -- the state maintained pediatric and pregnancy nutrition surveillance systems, and they insisted that the managed care plan, when it referred patients, use their format, send along the medical information so that tests didn't have to be repeated, and then they established an accountability system back and forth between the WIC sites and medical care providers to push towards better outcomes in the nutrition area. And I think that that's a little bit unusual, both that they save some medical costs, perhaps, and they also may have a better idea why they're getting what outcomes they get and some hints about how to plan better and more efficient nutrition programs, avoid the duplication, and establish this greater habit of accountability between the two public pieces of the system and the providers that they're paying for.
The other examples I don't want to go into in as much detail. You have them in front of you. But one of them is in Oklahoma, has to do with the Immunization Registry, that all the Medicaid providers participate in the state-wide registry so that they knew entirely what was going on with coverage levels and match it with the surveillance and other important things.
In South Carolina, I gather you have a South Carolina member of your subcommittee.
DR. IEZZONI: He was unable to attend.
MS. FREEMAN: Okay. Well, he probably could tell you about this one better than I, but there is a pregnancy risk assessment that I understand is very well-known and appreciated nationally in South Carolina. And the South Carolina Medicaid Department and Public Health Department worked together such that the data was maintained in the public sector which was terribly important to do in case management and monitoring and, again, working on pregnancy and birth outcomes, and I think also South Carolina stood out as having some very serious sanctions when the results weren't reached which is useful to remember, though not the cheeriest part of the conversation.
And the fourth one that I go into a little bit more in the written testimony is in Missouri where there was a series of internal quality indicators that the department was to keep -- I mean, that the Missouri managed care contractors were to report on HIV, STDs, tuberculosis and lead poisoning. But what seemed like the interesting twist on this one, that some of the accountability actually came by treating the health department services as services and reimbursing them, paying them, under the contract on the same basis they would have paid other providers to which referrals were made. So it treated it a little bit more like a grown-up part of the medical care system. In fact, the state was relying on those services or have them somewhere else, but to have it appear in their accountability and sort of feedback system of care, billing and outcomes.
All of this for me is terribly related to the ongoing debate about privacy which I fear.
I think your committee has been a wonderful exception in trying to really reach out and engage in a much broader public in these questions, but I think my experience locally has been that even though there is a great deal of discussion going on, it's not a well-developed discussion in the sense that people rarely are talking about the same thing at the same time. And I don't think that is a very hopeful way for leading to good policy, and I'm personally very worried that we're going to have terrible privacy law. And my colleague and I wrote a paper recently called "The Health Privacy Debate Can We Achieve Comprehension Before Closure". And by closure, I meant the fact that there are, you know, the HIPAA deadlines, and the Congress is likely to do something. But I'm very worried about the comprehension piece, and I want to relate it back to the public health and health outcomes conversation in a particular way. I think several of the important weaknesses in how we as a society talk about privacy of personal health data has to do with our incredible lack of clarity about the purposes for which data are going to be used and who understands what about that and what expectations we have about who ought to understand what about that.
I was sort of amused this morning when, was it Brian Burwell, talked about one of the problems that the plans in the Medicaid agencies haven't sorted out was that they didn't understand the purposes of a lot of the data they had. And I was sitting thinking myself imagine the Medicaid recipient population if you were to put them on that scale, and nobody dared mention that this morning, perhaps. But I think, you know, we all know that what any of us understands has been hard to come by, and probably the Medicaid population is among those that find it particularly hard to come by information that they might consider very relevant.
My primary concern and motivation has been that I think the data we have available to try to improve the health of the Medicaid population and any other has to do with the threshold issues about how we end up handling privacy and identification of personal information over the coming years and over the last 20 or 30 years that this debate has been going on, and that the current political situation leads me to believe that the data to improve health is likely to be permissible or impermissible to have and use in certain ways that might really improve health, depending on how it is that we resolve the struggle between two other sets of interests. And those are individuals who are totally appropriately concerned about the privacy of their personally identified information, as represented often by consumer or privacy advocates in the debates that go on nationally. And the commercial enterprises on the other side who are wisely, from their point of view, exploiting the commercial value, the economic value, of health data for many many purposes that don't have much to do with improving the health of the population.
I think what's left over in the middle are, you know, the public health and health research and licensure and the other sorts of quality issues where data really has to do with improving the health of people, the status of the society, how well we deal with one another, all sorts of very important things, and they're simply lost in most of the privacy debate.
The paper that Tony Robbins and I did that I left you some copies of and is also reprinted in the National Meeting Summary that Elliot Stone mentioned and said he sent to all of you from the Mass Health Data Consortium, it lays out a way of carrying on the debate that I have a little bit of optimism could engage more people, and more people in the comprehension of what the issues are and being able to have a voice about, and before we're done, and set law and get ourselves in potentially some worse trouble.
Without trying to take you all the way through the analysis, let me just say that it really has three parts. The first is to do something that I think as a society we've been very sloppy about, and that is to describe in a way that can be reproducible, without regard to the policy preferences of the people doing the describing; what data exactly we're talking about in any transaction; exactly what are the elements; to whom are they going to be given; in what form are they going to be given vis-a-vis identity; can it be traced back; who can trace it back; to whom is it identifiable; and what is the purpose for which the data set or the data element is going to be used, whatever it is that's in a transaction. So we lay out a way in which we could as a group of experts or across society talk about the same things at the same time which seems to me an important step in any analysis which we haven't reached in the privacy conversation.
The second step which I think matters enormously to all the data subjects of the world is what purpose is the data going to be used for and who is going to get the benefit of that. I think real people care a lot if they're in a medical care situation and the data is actually being used to benefit their care; or if there are a whole set of other purposes for which are being used but don't have much to do with them that may take over and have a life of their own. That is a great concern to people.
All of us in this room today, I'm sure, are terribly concerned with all the public purposes for which data can be used that have to do with improving health and a variety of other things, up to and including law enforcement, which isn't necessarily health people's favorite use, public purpose of health data, but, nonetheless, it is on the list, and we have to address it in the privacy debate or we're tricking ourselves.
And then there are economic benefits for economic enterprises, and I think often or, in fact, something Tony and I have laughed about is we have rarely seen any enterprise describe its use of data without attributing the most altruistic motives to the use of data. I mean, that's just a standard way people interact in the world, but I think if you subject that to a simple kind of analysis, it's pretty ease to sort out where we can agree or disagree on who the primary beneficiaries are of any data transaction; and where we disagree about that, we can figure out why, and then we're getting close to be able to be smarter about policy, but we don't do it that carefully. And this paper lays out a way in which we could be, perhaps, a little more direct about that. I mean, I've been frustrated by the studies over the last 25 years since we invented computers and electronic transfer of data because even the most recent one from the National Research Council last year carefully skirted the issue of economic advantage of the use of data and doesn't really take it on, and I don't think that's a good way to run a democracy. So we suggest a way that we might get at it, and hopefully civilly.
And the third part of this analysis is really to suggest how people, who may not be familiar and comfortable in the world of formulating policy, go about understanding the trade-offs. When we talk about privacy, we're talking about a trade-off between privacy and something else, and I think people react very differently about what is being traded off. If you're trading off improved care for yourself, it may feel very differently than if you're trading off the profit for some commercial enterprise to which you have no relationship, but happen to have access to your pharmacy records and the CVS example or others. So it's sort of a guide for people to understand the trade-offs, understand how to articulate what they do and don't like about what we've been doing as a society, to the extent that we know; it's not very clearly laid out anywhere, and how to lay out for, you know, the legislators in Washington or any other group that's interested, what it is we think we ought to be doing as individuals participating in a democracy to end up with a way where we don't undermine our uses of data for good public purposes, where we don't overstate them, where we're actually, you know,engaging and intriguing a lot of researchers, but not doing much that's good for the public purposes, but how we understand the differences and get some more attention to the values of research and public health.
And with regard to the Medicaid population, I think the important thing to say is while through various legal mechanisms, we have more leverage over the data; for people who are paid for through public programs, I think we ought not as a democracy to treat their privacy differently. And what that means is we really have to take on this debate across society and work this out in the broadest framework. And then if we're talking about Medicaid patients and Medicaid setting of getting their care paid for or getting care, that we ought to use the same privacy principles that we do for anybody in any other system. And at the same time, I think we ought to use the public leverage very much with managed care organizations to get the data that allows us to improve everybody's health and be able to compare outcomes of the
Medicaid population in medical care as well as with the other kinds of broader public health interventions.
And I think I really don't want to say more about the privacy because I think it's much too complicated. If you have a few questions later, I am happy to talk about it.
DR. IEZZONI: Okay, thank you, Phyllis. Let me just clarify. We do have a copy of the paper that we can distribute to the committee later. Some of us may not have realized why exactly we were getting the volume from Elliot and may have filed it in different places other than the place where...
MS. FREEMAN: I suspect he had many other reasons for giving it to. I just realized what he had, that I ought to mention that.
DR. IEZZONI: All right. Why don't we move on to Kate Ackerman. Dr. Ackerman is next.
DR. ACKERMAN: This is a test, but it's accurate. I wanted to step back to Data 101, if you will. My perspective is as a medical director for Boston Medical Center's Medicaid Managed Care Plan, I wanted to tell you that I'm saturated in managed care, but not from the Medicaid perspective. That's part of my background and bias. I'm saturated in managed care having grown up in the northwest where group health, Kaiser, was the provider for my family. This was a philosophical choice by many members of my family. The plan served us well, and I think many of you know them as sort of the grand-daddies as the best of the group staff model HMOs.
I went into medicine as an adult learner, after having done some other things, and was pleased to go to school where there was a strong public health influence, although I didn't get a chance to exercise that further until I actually had been in practice a number of years.
I chose to practice in a group staff model HMO in Upstate New York which is where I also had my residency training. Much to my surprise, this was in '82, and you're welcome to guess my age from there, but in '82, in Upstate New York, a group staff model HMO was still a tremendous threat to a provider community, and I was literally told I was a communist which was remarkable in my mind.
I worked as a primary care internist, a primary care provider, in that setting for a handful of years, and then I got sucked into the "wouldn't you like to work on this project for us" which turned out to be quality management. HMOs were being barraged around the issues of quality, and I was asked to begin to develop a quality program for this HMO.
From there, I went on to other projects, including a Medicare demonstration project which later became a risk contract and later than that a Medicaid demonstration project for the state of New York which I had a brief opportunity to be involved in before some other things happened in my life.
I came to Boston about three and a half years ago because of a career change in the other part of the family, and that gave me a chance to decide whether I wanted to continue to work in managed care in a big commercial environment or not, and at the time I decided not. I took the opportunity to revisit some clinical arenas that I hadn't had the opportunity to visit in Rochester, and that was with the home care program at Boston University. The old Medicare demonstration project had taught me that primary care internal medicine was fine, but geriatrics was where my heart aches. Being at B.U. with the health care program also gave me the opportunity to get back to the public health issues and work on public health degree.
The experience with what was then quality assurance, quality assessment, now quality improvement/quality management world, meant that I was also given the opportunity to become a NCQA reviewer, and I have continued to do that since coming to Boston, and that's been about a tenure experience for me. So there is a bit of the background and, as I said, the bias which will invest the comments that I am about to make to you.
I joined the managed Medicaid program at Boston Medical Center in January of this year, and we immediately were tasked or challenged with responding to the 1998 RFR. We have a contract based in 1997, but we were told that the expectations would now be changed to match the 1998 contracting process. And this has been a wonderful educational opportunity for me to learn quickly about issues and to plan, and it has also introduced me to many of the data issues which I hope to move on now to talk with you specifically about.
The first issue is that much of the data that we are asked to collect and report, and this is in no way a criticism of the State of Massachusetts, this also comes out of my prior experience with commercial populations in HMOs and the previous Medicare and managed Medicaid experience, much of the data that we're asked tocollect does not necessarily tie itself together in ways that allow us to tell a story about what our plans are doing, about what our programs are doing for the enrollees or the members. And in thinking about data overall, the three big areas that I, as a medical director, need to know about and want to know about on behalf of these enrollees, is I want to know what our current status is, and our current status of health indicators, our current status of satisfaction, our current status unmet needs, whatever our current status is. I want to know also about -- I want to know there is data that will allow me to manage resources, both money and people in the short-term. I want to know about data that helped me to plan for the longer term; program development; changes in the practice, so-called medical management changes, how do I influence my colleagues to do things in a different way because we believe the outcomes will be better in this new way.
Also, I want data that will teach and motivate on its own without a lot of investment on my part in terms of being the coach and being the cheerleader. I want the data to be clear and usable because it relates to goals that we have already agreed upon or set for ourselves as a practice environment and because we have comparisons. We have got benchmarks, we have got threshold, we have got best practices. That the data ultimately will speak for itself, teaching and motivating.
The data needs to be economical to produce. Many plans and programs have been tasked with adding new personnel in order to collect the data or process the data or whatever to the data. Many are being faced with major upgrades in terms of hardware and software and other resources, and many have been faced with changes year after year in data collection tools or various other support kind of tools for their data framework. I will get to some better examples in a moment. Data also needs to be real time data. By that, it may mean literally to the hour or the day, whichever. And, again, it needs to relate back to the goals. So if it's appropriate to have your data on a quarterly basis because that tells you enough about how you are doing in reaching your goals, fine. But if it's data that needs to be used on a weekly basis in terms of your access and availability or on a daily basis in terms of telephone access, those sorts of things need to be thought about in the context of the data.
So what are some of the observations and concerns that I have coming back to this Medicaid world? There is less data obviously available. I find this really interesting when I think about the commercial populations I worked with and the Medicare populations that I've worked with. There are fewer official reports just showing up on my desk. There are fewer published studies that just may find their way. There are fewer trade groups that are manufacturing newsletters and brochures and handbooks and glossy circulars. There are fewer advocacy groups giving me input about what is happening with Medicaid and what should be happening with Medicaid populations. There are fewer thematic journals, journals that are dedicated to issues related to Medicaid. Now, this doesn't mean none of this exists. I can go to the library and find health affairs. I can go to the library and I can find official published reports. I can go to the library, but this stuff is not coming across my desk. I'm not being found, okay, in ways that occurred when I was dealing with more Medicare and commercial population. So there is this lack of sort of an environment around Medicaid which I find very interesting.
Now, some other observations and concerns, the data acquisition process in my three short months with the plan at BMC, the data acquisition process is already clearly a more difficult issue. Telephone, okay, mailing addresses. So surveys that we might in other populations be able to accomplish using those, what are often considered basic matter-of-fact resources, have already compromised a small study that we're doing of the first 100 patients hospitalized, that we're attempting to determine whether those patients left the hospital knowing ho their PCP was, actually got an appointment made with their PCP and actually followed through.
Of the first 20 charts that we collected and gotten our background data on half of them, half of them we cannot any longer access through the telephone number or the address that was given as part of their initial hospitalization.
So data acquisition is problematic, and I know that case-based acquisition, every time an individual comes in updating that information, I know that's a tremendous burden having been in a situation of working in an ambulatory care setting where we attempted to do that, and yet that may be what it takes in order to stay linked with patients for the purpose of key data acquisition.
Outreach, the census model where literally, you know, you had outreach workers going or will have outreach workers going in 2000, is that the next one? Door-to-door and canvasing neighborhoods. We may need to move back to those approaches.
Another observation, the excess of process measures, proxies. I realize that outcome data is hard to come by. As an NCQA reviewer, I'm really pleased to see the news that NCQA has made for more standard data collection and a clearer articulation of what the expectations are during a review, and then it's companion activity that the HEDIS measures. But even the HEDIS measures, as many of you know, are still more focused around process than outcome. And I credit the HEDIS initiatives with a lot of time and a lot of energy and a lot of input from many people trying to define and be very precise about what is included.
But encounters are not outcomes, and there's this tremendous focus on describing an encounter, describing which provider saw which patient, how many times, and what they did each and every time they saw them, but that tells us nothing about the patient's satisfaction, about the actual functional outcomes, whether that person is getting to school, whether they're getting to work, whether they are physically and mentally as able as they're going to be.
Let's see, some final observations and concerns. In our Medicaid population, the definition and stability of denominators and numerators is difficult as patients move in and out of the plan on a frequent basis. They lose their eligibility for Medicaid. They don't like the provider selection we have and can move to another plan on a monthly basis. Patients moving in and out of the plan make it very hard to track and keep track of data. This is not to say I want to refuse any patient the freedom and the flexibility to make their choices, but it would be delightful if there were ways without creating confidentiality and privacy issues to somehow track patients through the greater system.
Patients move geographically and doctor shop providers move and aren't always available. Twenty-four availability means that patients end up seeing an OB other than the one who was their primary OB, a primary care provider other than the one who is their assigned primary care provider. So trying to keep all these issues of assignment and identification clean and clear when we develop ratios and indicators is an issue.
Lack of standard formats, the experiences with Medicare taught me great respect for the MDS activities in nursing homes and their own data sets and a great respect for the RUGS in New York, Resource Utilization Groups. Here in Massachusetts there are Massachusetts medical minutes or something like that, the actual resource source utilization of nursing homes. I realize they're not perfect, but I'm always impressed, when I do nursing home care, with how much more information I can find in a nursing home chart on behalf of dictation, than I can find in even the best of the group staff model HMOs where medical records tend to be much more voluminous than they are in network and IPA model HMOs.
Nursing home report card initiative of Massachusetts just put on the web, you know, again, probably not a perfect process, but a structured set of information available for use by consumers and providers, and I know a lot of work went into that.
The recommendations that I would have in a perfect world would be to clearly define and set the goals of data collection and to focus the measurement around that. Is it to achieve healthy people 2000, public health kinds of goals? Is it to manage to a budget? Is it both of those? But be very clear about why we are collecting data and where we hope to go with the data. Is it a monitoring activity or is it an activity that is meant to work its way towardsome sort of improvement?
The data related to measurement needs to be categorized. The data needs to be clear, and the elements need to be clearly defined. They need to be ranked. If we can't collect ten pieces of data this year and can only collect five, which are the most important five? I have a family who are involved in a census in planning for the year 2000 and the census collection, and they tell me about the great battles that are raging around which of the federal departments are going to get to put their questions into the census list because there is really only one reason, the major reason: The mandate for census collection is the one man one vote, one woman one vote, and distribution of representatives in the House. But other than that, everything else has to get bought, okay?
So on behalf of states of Medicaid, what are the goals, what is going to be measured, define it, rank it and rate it. Not all data is equal. I think the U.S. Preventive Health Services Task Force did a wonderful thing when it created that model for us that rated the evidence as A, B, C, D. It can be rated high, moderate, low. But I think it's important to recognize that claims-based data is not the same as referral-based data is not the same as narrative in the RFR process that I've just been through. What is the strength of this data?How valid, reliable, useful is it? That would be lovely.
User-friendly formats, identification of best practices, best thresholds and benchmarks, and flexibility, allowing and encouraging local adjustments to reflect the reality of the service delivery unit. Case mix, resource intent, delivery models are generally samples of this.
Let me give you one example from our own plan. We are enrolling standard members and basic members. In Massachusetts that's the AFDC population, as well as a group of mostly males who qualified because they are unemployed or fall below the poverty threshold even if they are employed. Our population mix is almost half and half, but even in the AFDC population, we have a much lower than expected number of children.
When I look at the admissions, when I look at just those demographics that I gave you briefly, it's substance abuse, substance abuse, substance abuse that I would be concerned about if I had my druthers, but I'm being driven. That's not -- you know, no doubt that it's important we have child health initiatives, no doubt that asthma has a large impact in our population and community as a whole, but the population I'm responsible for, that's not actually what we are saying. So local adjustments, okay?
In terms of intensity of resource use, the diagnoses themselves and how care is delivered, one of the things that I'm very interested in and appreciative of are the wonderful folks who have expanded the ability to care for patients, so it's not just the model I was trained in which was the one-to-one doctor/patient. As an internist, it was one-to-one. I learned from family medicine colleagues, I learned from geriatric colleagues. The multi-disciplinary team truly does live. It's extremely important. Not every patient necessarily needs or deserves -- deserves isn't the right word, but is appropriate to include in a case or care management model. So the ability to direct patients is very important.
Ultimately who uses the data, what are its purposes. I've given you my bias as a medical director. Here is a group that is concerned that if they lower their admission standards any further, they're going to put their accreditation at risk. But it is the key question: What are the goals; why are we collecting this data; who is going to be impacted and affected by it.
Thank you for your attention.
DR. IEZZONI: Thank you. That was extremely helpful. It's wonderful to hear information from different lines. Before we take questions, Dr. Castellano Turner, could we hear from you, and introduce yourself first.
DR. TURNER: I'm Castellano Turner. I'm on the faculty at University Massachusetts Boston. I have been in Massachusetts since 1968. I got my Ph.D. in psychology at the University of Chicago in '66. I'm primarily a mental health person, and I was at the University of Massachusetts Amherst for 21 years until 1989, on the faculty of the Psychology Department. And the faculty of the Psychology Department of UMASS Boston was trying to start up a new clinical psychology program and found me over there at Amherst and seduced me over here to Boston to be the first director of that program which I was for a while, for seven years.
Now I am a faculty member in the Pyschology Department and more or less a regular faculty member, but I have one other role which I think is relevant here, and that is there are several ethnically-designated or identified institutes at the University of Massachusetts. The most recent has been the Asian-American Institute which I think is about three years old. But included is the William Trotter Institute for the study of African-American life, and the Gastone Institute for Latino research and development. And there is several institutes that have been approached as long ago, I suppose, as two years ago. Primarily, it's through the McCormack Institute which is yet another institute which is a public policy-oriented institute rather than a ethnically-oriented institute. But this long story, I'm getting to the point soon, if you'll hold on. The point is that in the process of negotiating a contract with the McCormack Institute which is a policy-oriented institute which has its own Ph.D. program in public policy, the Department of Medical Assistance, the Administration of the Medicaid for the State of Massachusetts negotiated a contract with the McCormack Institute around the issue of providing some consultation having to do with the large issue of access to medical services in the State of Massachusetts. And, in particular, given that they are institutes dedicated to the notion of we need to know something more about what's happening with minority populations, essentially they were contracting with us to do a study and to educate them at various levels in terms of the observation; that, in fact, there is some discrepancies in access to medical services within the state, as there is around the country, but Massachusetts is not different in that regard. But, in fact, they wanted to know what the impact of race and ethnicity was and along with that language problems in defining the general issue of access.
And the whole process, what I'm about to describe to you, is basically I'm privy, since I have no prepared text, having been brought on fairly recently, I'm reading from a report that we wrote for the Department of Medical Assistance which does address basically a summary with some recommendations having to do with health issues for blacks and other minorities in the State of Massachusetts.
Essentially, what we have been doing over about a year and a half has been, in this educational process, making, I would say, conference-like presentations colloquia to the staff, to the administration, of the Department of Medical Assistance here. And the scheme has been essentially the following:
That, as a group of consultants, people who know something about minority populations, said that we might be able to begin in about four steps, let's say, ending up presumably at the same level as the previous presenters, having to do with best practice issues or supposedly our bottom line. But to begin, we had to really make some extended presentations about who the population of Massachusetts is; that is, are there African-Americans here, Latinos and Asians around Massachusetts, and how many and what is the growth patterns and things of that kind. So we made one large presentation having to do with the issue of where the ethnic minority populations are.
Now, it won't be a surprise to you to discover that one of the things that we've discovered in this process, basically getting available information from various institutes within the University of Massachusetts and from the Massachusetts Institute for survey research which basically gathers in all the information at Amherst on the population of the state, that the minority population is not spread evenly throughout the state, and clearly the notion of urban/rural is a major dimension that we wanted to point out.
But, also, among the problems that I identified almost immediately trying to discuss with Connie Chan and Edwin Melendez, along with Carol Upchur (phonetic) and, in fact, I'm sorry that I'm the first one from our group to be making this presentation because, in fact, Carol has a written document, testimony, that she has allowed me to look at and probably would clear up a lot of what I'm saying.
But, in any case, among the things that I detected immediately was the issue of the different ways in which minority populations were presented with barriers to health care. It isn't all the same; that is, different populations have different problems. Among them is, for Latinos and for Asians, language. And I got up and I made the comments that, in fact, the issue for African-Americans around a barrier is not around language. We do share a language in large measure, and it really has to do with the history of race relations in this country rather than language barriers. And those barriers are rather long and complicated obviously, but they relationship, I think, an important piece to the whole issue of trying to understand differential access and so forth.
But promptly we had a presenter at that first conference who basically pointed out to me that there are language barriers. There are quite a few language barriers in terms of the use of language, the sophistication of language, the non-medical and medical overlap kinds of problems in minority populations, but African-Americans in particular, and giving lots of examples and bringing in the issue of Black English as perhaps a general barrier.
But, also, and here is my first substantive statement, I think, about this gathering of information that we were doing, is the issue of not imagining that even in the black population is somehow homogenous; that is, the idea that there are not African diaspore populations for whom language is not a problem, no, I'm not announcing anything new, but Massachusetts does, in fact, have quite a substantial Haitian population. What substantial means is itself a major problem because although I have been in touch with various people within the Haitian community, and we have an institute, a separate institute, for the Haitian population or student body at UMASS Boston, but nobody knows how many Haitians there are.
However, Carol came up with a statistic which I think may, in fact, help a little bit, and this may suggest something very important about this gathering of information about race in particular, and that is that only 35 -- or 35 percent of the black population is foreign born which was startling to me just as a beginning. The idea that we were not looking at, you know, what has been traditionally the African-American population, a population largely with historical roots in the south and immigration to the north and largely multi-century kind of contact with trying to make it in this country.
And I tell you, once that became clear to me, we both had lots of English, but we had Haitians and so forth, but in that 35 percent there is, in fact -- I'm sorry?
DR. IEZZONI: 35.
DR. TURNER: 35 percent in our foreign born of the group. That struck me as pretty large, but it isn't Haitian. It's not all Haitian. I didn't mean to imply that. My own family background, my father grew up in Jamaica, and, in fact, the Jamaican population is quite substantial in Massachusetts. In Boston in particular. There is, I think, some language barrier that is implied by the use of Jamaican, Creole, etc. But that is a major problem; that is, how to identify the needs of a population where there may be some discrepancies in definition, etc., beyond simply, you know, what the family origin was and that you're dealing with a heterogeneous population in the sense of foreign born as against migrated from the south or migrated from other parts of the United States or longstanding families within Massachusetts.
That began our process of uncovering really an awful lot of useful information. So our next stage really was to address the issue of health. Not only health, but the issue of what are the health problems of the several ethnic minority, the major ethnic minority populations. We had, as I said, representation from the Trotter Institute for African-American studies and the Gastone Institute for Latino studies and the Asian Institute. We essentially began a process of pulling in the numbers from various places to discover basically what it is that -- where the discrepancies are, how many of various kinds of individuals from different racial groups, etc. were having difficulty or had medical problems that were not being attended to.
Let me just give you just quickly a little bit of a summary of the definitional problem. In Massachusetts the population is estimated for African-American or the black population is estimated to be 5.6 percent of the population, and 45 percent of that population lives in the city of Boston, while nearly 85 percent live in eastern Massachusetts.
But in terms of the definition of what we are looking for, health outcomes, is there a difference in health outcome for these different populations? My responsibility was, as you probably figured out, the African-American population or the black population rather, and so what I'm going to talk about is only that population tomorrow. Tomorrow you'll hear about others.
We had initially the issue of trying to sort out in the literature the issue of mortality and morbidity and decided that we were going to essentially work with whatever came our way; that is, if people had done research that was based upon illnesses and populations as against death rates from different kinds of illnesses, etc., we would simply go with that.
Among them, the general issues that we wanted to get at, was the life expectancy discrepancy. Now, again, there is a couple of paragraphs here that address the issue of the origins of not only life expectancy differences between blacks and whites, but also the issue of whether or not life expectancy differences can be explained by access to medical services.
Our conclusion at the end of all of our various kind of presentations to the Department of Medical Assistance has been that the explanation for not only the access problems but any pattern that we see like shorter life expectancy essentially has multiple causes. It's not a simple issue of racism, but it's not also a simple issue of genetics, okay? Between genetics and the social cultural environment, there are an awful lot of complex issues that need to be taken into account.
But, in fact, the life expectancy differences that we discover in Massachusetts mirrors essentially the pattern in the United States. It is not a different pattern; that is, the black population of Massachusetts dies earlier. And you can look at the pattern of mortality associated with different illnesses, and it all adds up so to speak. But, essentially, it's important simply to know that Massachusetts' data is consistent with federal data, with national data, and sometimes we have used metropolitan data from Boston.
Among the problems that several people or the presenters alluded to was the issue of child, the infant mortality rate; that is, children who are dying before the age of one is much higher. Now, among the things that is problematic is obviously that of course is rather complicated as well, including the age of the mother and how much prenatal care and so forth and so on. So among the things that has been documented, not only a national problem, but a Massachusetts' problem, is all of the factors related to the higher rate of infant mortality among blacks show a discrepancy with the white population in the sense for those explanations; that is, younger mothers, less prenatal care, less medical follow-up and so on.
Now, again, all those statistics represent kind of the endpoint. They don't exactly tell you, you know, why that is taking place. I mean, those are factors. Among the specific conditions that we made some reference to, what we did was we determined the death rates from various kinds of illnesses in the state, in the population at large, and then broke it down into the death rates and basically the ranking of high to low death rates. And there are, of course, a lot of overlaps with the general populations that the minority populations have, but, essentially, the top of the list is cancer for the black population as the leading cause of death, heart disease next, HIV/AIDS which immediately finding it at the third level, HIV/AIDS as the third leading cause of death has to be in some sense rather shocking considering that, in fact, it represents a large part of the population that is dying from a disease process that is more or less now understood. But the next is stroke, okay, all of which -- another problem with some of the counting in these statistics, all of the things that we have kind of dug out from the literature, basically addresses death, but, unfortunately, it is sometimes difficult to tell whether or not a person had various other kinds of conditions, and the fact that they died of a heart disease may not tell the entire story.
For the black population which has basically chronic hypertension at a higher rate than the white population. The issue of stroke is a major problem. Likewise, I was myself, being from mental health basically, I don't have a long history in medical, general medical statistics, but I was, myself, shocked to find that diabetes was so much more prevalent among the black population and that, in fact, it was implicated in quite a few discrepancies between the -- between blacks and whites in terms of health outcomes.
And, finally, our recommendations had to do with other conditions that are particularly focused on the needs of the black population which includes sickle sell anemia and asthma.
And obviously since the former; that is, the sickle sell anemia which is the most common genetic blood disorder in the United States. It affects 1 in 500 black newborns each year. One of the major mortality risks for children with sickle sell anemia is stroke, and this is yet another example of the deception of both the conditions that people have and the causes of death, you know, did the child die from the sickle sell anemia or from the stroke, etc.
The asthma, of course, is somewhat more complicated because it is not necessarily only a matter of death rates, although the death rates from asthma among black children is substantially higher. It's also a matter -- also of the utilization of medical services that might be a major discrepancy.
Among the things that we have talked about in terms of gathering additional information, our new kind of initiatives, would require that there be good information. What we have discovered is that many of the data files that are available for the understanding of any discrepancies in service, access to medical services, the data are quite poor. And, yet, as I go through the literature, I find remarkable ability on the part of large numbers of researchers to get access to race data, and it seems to me that the data must be there, but somehow it is quite difficult to find databases that are, that do have substantial amounts of good data.
Now, what would be good data? Now, immediately I've already identified a couple of problems even within the black population about the heterogeneity of the population; that is, are we talking about the same people and the same problems when we talk about African-Americans and Haitians or Jamaicans, etc. Well, I think probably not. I mean, there are probably all sorts of cultural differences that might represent part of an explanation for the barriers to medical services that are different for those groups. It isn't only language, but maybe even very specific cultural differences in terms of the attitudes toward the medical services and access to medical services.
But, as you probably already know, there is also the whole issue of the population of blacks and the change in the census designation of mixed-race people, and it's going to be a challenge. I, myself, however, do not think that the issue of mixed race is going to be a major issue in the medical service kind of area, but, in fact, it complicates the issue somewhat.
Okay, I think I'll are stop.
DR. IEZZONI: Thank you, Dr. Turner. A series of very interesting and diverse presentations. Committee members, are there any questions or comments? The reins have thinned a little bit. Elizabeth? Hortensia?
DR. AMARO: Well, maybe a question for Kate. I'm wondering how, and maybe this is a little premature because the data may not be available yet, but I was wondering, you know, we continue to be interested and we have been asking in each hearing whether people are using race ethnicity data to look at issues of access and quality as well as outcome issues, of course.
I'm wondering whether you've yet, in the system you described at Boston Medical System Health Net, are doing this; if you have done it, how you've utilized race and ethnicity data or data about other subgroups that might have unique risks or other patterns in quality of care.
DR. ACKERMAN: The short answer is no; the longer answer is that it's not because of lack of interest around the various patterns of care and issues related to the care.
Currently, the data that we have is claims data, and claims data have no way of incorporating race. It's billing data basically. Likewise, referral data, okay, has no way of incorporating race. Since we're a member, a provider calls up for pre-authorization before a patient goes to a diagnostic study or goes to a specialist or whatever. That's not part of the exchange of information that occurs for a referral.
Ironically, despite the long application form that is filled out and provided to the state as part of the enrollment process, before the patient even selects or is assigned to the health plan, that data doesn't reach us on a timely basis. So we are very limited in terms of the descriptors of our population. Literally, the standard basic designation that I shared with you earlier: male, female and the age breakout are the descriptors that are available to me currently and to plan currently. I needed to stop my presentation, but clearly on my wish list is more descriptive data about the population and subpopulations that we will be serving.
DR. AMARO: So your data system doesn't link back to the enrollment data or to data that is gathered before clients or patients come to you that would have more background information?
DR. ACKERMAN: We're a very new health plan, so our experience, yes, may be different than older, larger plans, but we are literally having to build our own databases now, and we have received limited information through the broker that the state uses and limited information from the state itself about the enrollees who either select us for us or are assigned to us.
Building our own database is obstructed by the -- some of the issues I raised earlier, i.e., phone numbers and addresses that may no longer be current, if we had them in the first place. And that leads us then to -- those other potential data collection approaches which is literally every time somebody comes in for an office visit, to revisit with them, key data elements or doing outreach work. Both of them are time-consuming and expensive, but critical. Something that the plan should bear on its own, I don't think we can afford to do that. Something that we could do with others, possibly if it were felt important enough, if it ranked high enough on that list of priorities. Can I add to my wish list?
DR. IEZZONI: Of course.
DR. ACKERMAN: Staging data. If we knew more about individuals who have dual diagnoses, whether it's substance abuse plus other chronic illness or medical illness or just multiple medical illnesses, we tend to get a diagnosis, maybe two diagnoses. It makes it really hard to find the individuals who have the most complex set of overlapping needs. So definitely more descriptive data, dual diagnoses and the stage of disease. If we had some way of knowing that this is an insulin-dependent individual who has had their diabetes for 20 years versus 2 years, there is somewhat different strategies in terms of how we would approach them.
Also, more about the individual patient and what they want or what they wish. As committed as I am to the philosophy of population health and population issues, I highly respect the right of any individual person to say I don't want to be medicalized, I do not want to stop smoking right now, I do not want to get preventive care, and I do not want to go see the doctor on a regular basis. That is an individual's right, and that needs to be respected. I wish we knew more up front about what our enrollees want and expect from their care.
DR. AMARO: A follow-up question. You mentioned in your presentation that the problems most common in your population substance abuse, substance abuse and substance abuse.
DR. ACKERMAN: Given our current data.
DR. AMARO: Right. Well, this is an issue of interest to me that I keep bringing up to the committee, you know, behavioral health care, general mental health and substance abuse issues, and I was wondering from the perspective of the provider, what kind of information would be useful to you, to the plan, to the health net, to collect related to addictions and substance abuse?
DR. ACKERMAN: In general?
DR. AMARO: Based on, you know, your perspective and objectives.
DR. ACKERMAN: It is a big question, and I'm reminded of a conversation last night where someone, who's not known to me unfortunately, commented that he recently had seen in one of the British journals a study which described the validity of individuals who stated that -- validity of the data for individuals who filled out a survey regarding substance abuse, and this study apparently managed to check back in some way.
Not surprisingly, people aren't very willing to admit that they use drugs. And so what I wish, of course, is to know whether they use drugs, what kinds of drugs they used, and the pattern of use: was it for alcohol, were they binge users versus chronic users. But I have to admit that's an area where I doubt that I'm going to be able to get that kind of information, and yet it's so important for program development, understanding how and where and when to do the outreach, the education, the -- let me stop there and ask you if I understood your question well enough to give you some perspective on that?
DR. AMARO: Well, I've always thought that since substance abuse is the major cause of hospitalization next to pregnancy-related hospitalizations, that, you know, managed care corporations would be very interested in reducing this and providing whatever treatment was necessary in order to reduce this because it's more expensive than providing treatment for these repeated hospitalizations in these populations. So I thought you might have a perspective on what kind of information would be useful to have, you know, for providers to -- both in terms of improving services to the population, containing cost, whatever perspective. Because I think that we don't have good systems in place right now for collecting that kind of information, especially for Medicaid populations, at least from what we've heard from our contractor about what is required. That information as well as lots of other information is not often collected. But since this is such an issue that affects, you know, this population as well as has high implications or a lot of implications for costs, and it's usually care that is not often available, it's not covered, and that you might have a perspective on what would be useful.
DR. ACKERMAN: Maybe one quick follow-up on that, and it relates to my new colleague's comments, Phyllis Freeman, and that is around not just substance abuse, but behavioral health in general, and making sure that the confidentiality issues are respected and preserved and that the information about diagnoses of substance abuse as well as other mental health issues can be shared appropriately with primary care providers so that the primary care provider can be part of this therapeutic intervention. That, I think, is probably the largest challenge we're facing right now.
DR. IEZZONI: You know, I wanted to hear Phyllis Freeman's views and perhaps also Dr. Castellano's, given that you are a psychologist, about whether data around mental health services, substance abuse, whether there is a different standard for protection of privacy or confidentiality that you would view as necessary or whether you view those data as data that should be treated the same way as data on diabetes treatment and other kind of standard medical paradigm physical ailment treatments.
MS. FREEMAN: I tend to think that we end up with a separate set of standards because we don't do too well in general, and if we really understood how to handle data with the kind of respect it deserves, that we wouldn't have to have so many grade decisions, and then change them weekly as we change our patterns of discrimination and inattention and many other foibles of human society. So I tend to go for the much more fundamental statistical issues.
DR. IEZZONI: Dr. Turner, do you have any thoughts on that?
DR. TURNER: Well, I'm not sure that I have particular differences that I would allow for mental health. I'm not sure whether you imagine that there is more need for privacy or less because, in fact, the community mental health movement, for instance, is basically on the idea that somehow instead of taking mental health problems as things you send off to the country, that somehow it becomes a community issue. And as far as I know, there is not a major concern about diagnostic processes, except maybe with insurance companies, the kind of finagling that goes on with diagnoses in order to get people paid or something like that, changes of that kind. And, of course, everybody is nervous, everybody is concerned about whether or not their records are going to be made available to just anybody who is a clerk or something like that because everybody has to be watched in terms of how much service they're using, mental health service that is being used, etc.
In a way, just trying to hook this up with the last statements that I was making before, is the idea of somehow trying to identify the vested interest that might be present in the availability of this information or of that information. And in terms of race and ethnicity, for instance, I'm obviously here to advocate for the idea that we can't understand our problems without that information. We can't understand what explains the death rates and relatively poor health outcomes unless we know the populations and can somehow track that. I don't think that there is any reason myself to believe that people in larger numbers are afraid to let people know they're black or white or from which country or what have you.
DR. IEZZONI: This morning we heard from Jack Fowler, who is also from the University Massachusetts Boston campus, so we have people springing all over the place from the University of Massachusetts, about surveys, and that one of the chief sources of information about how Medicaid enrollees now will come from the CAP service, a consumer for health plan survey, based on all of your own experiences, especially, Dr. Ackerman, based on your experience and not being able to find the addresses or telephone numbers of people who had been recently discharged, I think you said, from the hospital, what do you think; and, also, Dr. Turner, given your discussion about the cultural diversity of the black population, let alone, you know, the fact we have a heavy Portuguese population, we have a Hispanic population, what, from each of your perspectives, do you think are the cautions or things that we should be thinking about when the emphasis really shifts to doing surveys of patient's experience with health care as the metrics for evaluating how well Medicaid is doing? Dr. Ackerman, do you want to start?
DR. ACKERMAN: I'm not sure. In fact, can I ask you to repeat the last question?
DR. IEZZONI: Yes. I guess I just wondered, even based on your experience, what do you think are the cautionary details from your own experience about us beginning to vest a lot of our interest in evaluating Medicaid in surveys? You know, the fact that we have designed most of the surveys using white English, to use your phrasing, and we are mailing them out, most of them because it's too expensive to telephone, and even if we telephone, as you suggested we might not be able to find them, so I guess the kinds of things I am thinking are of biases. Why don't you talk about it, you know more about it than I do.
DR. ACKERMAN: It's based on Dr. Turner's point, if we don't know the substance, if we can't stratify the population for a number of -- in a number of directions, then we won't even know that we are biased, and it gets back to what are the goals, why collect the data, who is going to use it, what is being driven by this data. Yes, I'd be very concerned about a survey processes in English only.
DR. IEZZONI: Well, they have Spanish translations, but it's English and Spanish.
DR. ACKERMAN: Yes, that might be appropriate in this state, for instance. But certainly our population at BMC, not our planned population, but we do have statistics for the Boston Medical Center. Once you get past the English-speaking caucasians, it's actually Cape Verdian, after that it's Haitian-Creole, then we start down into the 10s and 7s and 5s and 2 percent of various Asian and Hispanic. We have 140 languages listed in our interpretive services directory. That many languages are represented at BMC in a year.
Now, again, those are not all going to be represented in our plan, but it gives you a sense of the tremendous diversity that we face in that environment, some of which is reflected in data.
DR. IEZZONI: Do you have any comments?
DR. TURNER: Well, I guess my overall assessment is that there is good reason to have more rather than less information. If, in fact, mail surveys are coming in at the rate of 20 percent, obviously there is biased information. I think we ought to certainly figure out the best way to get as close to the population universe as possible. Mail surveys are not, I think, going to get that kind of information. And, again, the important problem here is that the race and the ethnicity differences, the race and the ethnicity differences that we have data on, the most clearly is what? It's what? It's the outcome. People die, and all of a sudden we have statistics on the discrepancy in the death rates in the population. But if we were to design programs or try to figure out, you know, how to kind of narrow that gap, we would need to have to know something about the population earlier.
DR. IEZZONI: Well, what would you suggest then given that the CAP survey asked questions about access and waiting times, things that we need to know about for everybody, what would you suggest as a strategy for improving our ability to get that kind of information from the populations that you have studied?
DR. TURNER: Well, there are survey research techniques that will update the percentages somewhat. Even if you have as much as 60 percent of the population, it's a distortion of your population.
Now, how you overcome that with limited resources, I don't know. I mean, if I had all the money that I needed to do it, I would do, you know, what the Census Bureau does: I'd go house to house and basically survey people directly rather than sending them out surveys.
You know, is that at all plausible? Probably not.
DR. IEZZONI: I don't think so. Tony had a comment, and Kathy.
MS. COLTIN: I think one of the things that is interesting is the CAP survey is a member-based survey, but there are really very few questions in that survey that are pertinent to people who haven't used services. There are really only a couple of questions, and they actually screen, and people can screen themselves out at that point; and then the group that have tried, the only pertinent question really in there is whether they have had a problem in access for those who have used services, who have actually gotten in for an appointment or whatever. Then there are a whole range of questions that are pertinent to them.
So my point is if, in fact, you are finding poor address and telephone information among a group of patients or members who have recently used services, my expectation is that among the non-users, that probably is likely to be even worse because there is a good opportunity to at least update address and telephone information when patients come in and present for care. But if they're not coming in, what you have is the enrollment information which can get out of date very easily. So you have a bigger problem, I think, and the greater potential for non-response bias among people that haven't used services. And, in fact, in the study that CAPS did in Denver, when they did telephone follow-up and they looked at the 20 percent of patients who they were able to reach by phone, but who had not responded by mail, it was much more heavily young males and people who had not used services. And I think those people also didn't feel as vested in responding to the survey because they really didn't have much to tell not having used the services.
So I think that is one of the issues you get into as well for the sake of the couple of questions in the survey that get at whether people have tried to use services and not been able to, and you are trading off against a whole group of potential non-respondents because you have bad information about telephone and address. And are there other ways that one can use administrative data sets and enrollment files to identify the population that hasn't used services and, perhaps, look at other types of issues in that population?
The other point is when you are going to focus on people who are using services, there are other modes for distributing surveys, like on site, the time they come in for care, that they would either fill out and leave, you know, before they leave the site; or you give them an envelope, and they can mail it back. And there you have a whole lot of issues around bias and distribution of the surveys. So whatever way you go, you are going to have issues that you have to deal with.
I thought it was particularly interesting when Dr. Restuccia was mentioning the enrollment of the homeless population. I don't know how we are going to reach that group of service.
MS. FREEMAN: Massachusetts just set up a tracking system.
DR. IEZZONI: Tracking system of people without homes?
MS. FREEMAN: All homeless people. Actually, it's in Boston at this point, but as you can imagine, there is enormous issues with confidentiality because Baltimore is without any kind of privacy protection at all. In Massachusetts, we very carefully stripped all the data of its identifiers, so...
DR. IEZZONI: Interesting. Tony, do you have something?
MR. ROBBINS: I guess there is a public comment period maybe, just as a citizen though, I do work for the department, these are not department's views. I did give you, I think, enough copies of the issue that had the race article, which I think you will find interesting, on surveys versus the medical care system for gathering health data. I was also lucky enough to be in the UK late last fall just at the time that the new report on "Health in Equalities" came out. This latest issue of public health reports has a very brief piece that I wrote about this report and the history there, and I hope maybe your secretary can get you enough copies of this report on "Health in Equalities".
Finally, I wanted to say, with the experience of being an editor over the last three and a half years of public health reports, starting with a piece that we got from Richard Cooper about race and ethnicity hypotheses and epidemiologic and other studies, we have gotten very serious about demanding that authors who include data on race and ethnicity have some underlying hypothesis. It is far too easy in this country, where we are so conscious of race and ethnicity issues, to flow this data in when they may stand for education, they may stand for cultural differences, they may stand for income. And it really makes a difference because, to state it most boldy, we are participating in a very racist exercise if we use race without understanding why we are doing it and asking reasonable questions. Because what we are doing is, by the very use of it, we are implying that that is the underlying cause or difference.
I know we may not -- I know well-meaning researchers don't see it that way because I have spent a lot of time fighting with them over the last three years, but I do think that maybe if the U.S. had been more social class conscious and a little less race conscious, we might have described very different phenomena in this country than what it looks like when we read most epidemiologic studies in these days.
DR. IEZZONI: Well, the National Committee in 1992 actually recommended that we add to our core data set or discharges, I think Marjorie, wasn't it education, years of schooling? I don't think we had income on there. We wouldn't have had income. But I think education because I think that our committee, this was before I was on the committee, but I think that they recognize that we are using convenient proxies simply because that data is available for a more complex phenomenon.
MR. ROBBINS: The best single proxy in the U.S. for social class is the value of the real estate of the home values of which people live. It actually does very well in picking up both education and income, and obviously because people live in neighborhoods, it has a whole cultural aspect to it as well.
DR. IEZZONI: Good. Are there any other questions?
MR. NEWACHECK: I have one more question for the panel. In thinking about consumer surveys, because that may be the most important source of data on looking at plans, particularly Medicaid in terms of satisfaction outcomes, it is likely that the samples could be relatively small so we'll be able to look at the whole Medicaid population, maybe Spanish groups and such, but probably not very many of the subpopulations, I'm wondering if any of you of the panel have any particular thoughts about specific subpopulations that we should be sure to include in any kind of survey like this; that is, to oversample or make sure we have coverage of?
DR. TURNER: Well, I think you have already proposed it in the question. That is, oversampling is not an unusual procedure when you have subgroups within the population that you want represented, but you're not likely with the particular size of the sample that you're getting to include.
DR. IEZZONI: What would be the subgroups that you would want to hear about?
MR. NEWACHECK: Yes, are there particular populations, subpopulations, that come to mind that we should be sure to include in that kind of a plan?
MS. FREEMAN: Marjorie mentioned one, actually the survey center researched unusual success rate with finding out about women who had been subject to abuse, and that's a population that isn't generally too easy to survey, but they had a very interesting experience and learned a lot of useful information.
DR. IEZZONI: Hortensia has a question or a comment.
DR. AMARO: I was going to follow up on your comment, that in our response to the recommended changes on race and ethnic categories, the committee commented also on the issue of needing to clarify the use of race, especially I think from my perspective, not only hypothesis, but also the framework the researcher is using for the use of race and ethnicity, theoretical framework, that a lot of the research is used, and race and ethnicity been not articulated clearly.
DR. IEZZONI: Hortensia is referring to her very nice response that she wrote for our committee to the O&B director, "15 Changes in the Race and Ethnic Classification" last September I think it was, and I think it might be on our web page.
MR. ROBBINS: Are you making progress in that area? Are people responding, Hortensia, positively to that kind of thinking?
DR. AMARO: I think, to tell you the truth, and it's a very small number of people in the field who are thinking about this, and there are some people who I think made important contributions, like David Williams and other people, you know, and in the area of socioeconomic status, Nancy Kruger (phonetic) and other folks, who are keeping this debate alive and helping us think along the ways we thought, that I really think it's a limited number of people who continue to write in public health and medical research using race and ethnicity without a good articulation of what is the framework, what is the understanding, what is the serving as a proxy for. And, also, I think the field is making some progress in articulating the many things in a more explicit way that race and ethnicity have stood in for, so some of Nancy Kruger's work and the experience of racism, for example, in relation to health outcomes or the work of further, you know, infants mortality, for example, and political representation and the relationship of those. So I think we're beginning to push the traditional edges, but I think as a whole, the field has a long way to go.
DR. TURNER: As far as I know, just to respond in part to that and what sounds like a trend, that what I know of the epidemiological data and the sociological data and the psychological data, that in spite of whatever controls one might have, social class, education, occupation, certainly residential value and so forth, repeatedly there is some part of the variance that is not accounted for in those things, and I think that we make a mistake if, in fact, we imagine that there is a one-to-one surrogate relationship between social class and race, and that if we don't get it, that somehow the race issue will go away. It seems to me that we need to know what part of the variance is being accounted for by something that has nothing to do with how much money you have.
DR. IEZZONI: Good. Well, a thought-provoking discussion, and we thank the panelists very much for spending time with us this afternoon and sharing your thoughts. I think we'll adjourn, and we will reconvene tomorrow morning at 9:30.
(Whereupon the hearing was adjourned at 5:00 p.m.)
C E R T I F I C A T E
We, Suzanne M. Bruce and Robin Gilbert, Professional Shorthand Reporters, do hereby certify that the foregoing transcript, Volume I, is a true and accurate transcription of our stenographic notes taken on Tuesday, April 14, 1998.
_________________________________
Robin Gilbert
Suzanne M. Bruce
Professional Shorthand Reporters
COPLEY COURT REPORTING