[This Transcript is Unedited]
J.D. Morgan Athletics Center
325 Westwood Plaza
Los Angeles, California
Call to Order and Introductions
Diversity of the Asian, Native Hawaiian and other Pacific Islander Populations - Dennis Arguelles
Native Hawaiians and Mainland Hawaiians - Nolan Malone
Data Issues in Asian, Native Hawaiians and other Pacific Islanders - Paul Ong
Use of Census Data and Health Planning and Contextual Community Development in Support of Community Services - Melany Dela Cruz and Bong Vergara
Pharmacologic Differences in Asian Populations - Keh-Ming Lin
Language and Translation Aggregation and Disaggregation - Ninez Ponce
DR. MAYS: Folks, we are going to get started. I'm sure that our colleagues who are joining us are going to do it over time. Our committee is here, so we don't want to miss anything that anyone has to say. We don't want them to have to leave early or anything, some are trying to get started.
I'd like to welcome you to UCLA, first of all, on behalf of UCLA, for those of you who have traveled long and far. I think if we were giving prizes, Dr. Malone probably wins it for coming from Hawaii, in terms of being here. But on behalf of the campus, I welcome you. On behalf of the committee, I also welcome you for being here.
I want to say a few words about the committee, so that we have some -- you have some sense, for those of you that we are new to, you have some sense of what we do and what our agenda is today.
We are the population subcommittee of the National Committee on Vital and Health Statistics. The National Committee on Vital and Health Statistics is an advisory committee, one of the federal FACA committees, so it is an advisory committee, which means that part of our job is actually to look at those issues which are in our charge. For populations, the issue that we have determined that we want to look at for this particular hearing is the collection of data on race and ethnicity, and for this hearing, specific to Asian, Native Hawaiians and other Pacific Islanders.
The reason for our interest in looking at those issues, as many of you know, there has been a long history in HHS of collecting data on racial ethnic groups. HHS has over time developed a number of policies and guidelines for how to collect that data.
In 1997, which was the latest issuance of guidance by the Office of Management and Budget, there were some attempts to change the way data has been collected, so that we have expanded the categories, as well as, as many of you who filled out the census are well aware, allowing people to be able to choose more than one racial category.
Part of what we want to do is get a sense of, with that guidance there, and those recommendations for not only agencies, but as I understand agencies within HHS, but also at other levels, whether or not that seems to be working. In particular we want to get a sense of issues of multiple identification in terms of race and ethnicity by the Asian Pacific Islanders and Native Hawaiians. We wanted to get a sense of whether or not people think that the current guidance may at all lead to misclassification. We are interested in issues of language, the extent to which data is collected and languages that are inclusive of particular populations.
We are also interested in whether or not you feel the data is being collected. As I understand it, today we are going to hear from some individuals that the current system of classification that we are using still does not really seem to result in data that is usable for your particular subpopulation. So we are interested in hearing about those things.
One of the things that I want to make sure I underscore is that we are not HHS. So the ability to automatically change things does not reside with us, but instead, our goal will be after the hearing as a committee to look at what we hear in terms of testimony. We are very interested in receiving specific recommendations from you and pulling those recommendations together to send them forth to HHS, and hope that HHS will move forth on many of them. But we are not HHS ourselves. We are just in an advisory capacity.
The process that we are going to work on throughout the hearing is, we have had a series of hearings, and we try and tailor those hearings based on the particular subpopulation that we are working with. One of the things that we typically do, because we recognize for some populations that there is a wide geographic distribution, and not everybody can travel, we try and make sure that there are ways in which we can get input.
For this particular hearing, we are limited to two ways, as opposed to -- typically we sometimes have three ways. Way number one is for individuals to be here and to be able to make comment. We have an invited set of individuals who are serving as advisors, as well as experts on this topic, but we also have people in the audience who also are experts.
We will hear from the people who are presenting. The committee will open it up to questions, and then we will open it up to the audience to also ask questions, make comments, et cetera. So this is a public hearing, and so we clearly will invite individuals from both the audience as well as my colleagues here at the dias to ask questions, make comments, clarify matters, et cetera. We have tried to build this hearing so that there is sufficient time to be able to do that. We are very open to questions, comments, et cetera, from participants who are presenting, as well as those who are attending.
Is there anything else that I should share? I should ask my colleagues before I do introductions, et cetera, other than to welcome everyone.
What we usually do at our hearings is start off by introducing everyone, so that we have some sense of who is here and that you have some sense of who the experts are in the audience. I'm going to start. I am Vickie Mays. I am a professor here at UCLA in the Department of Psychology. I am also the chair of the Subcommittee on Populations with the National Committee on Vital and Health Statistics.
Dale?
DR. HITCHCOCK: Good morning. My name is Dale Hitchcock, U.S. Department of Health and Human Services. I am staff person to Dr. Mays and the Subcommittee on Populations here.
DR. JACKSON: Debbie Jackson with the National Center for Health Statistics, staff to the National Committee.
DR. BREEN: Nancy Breen. I am an economist specializing in health, especially screening in healthy populations. I work at the National Cancer Institute in the Division of Cancer Control and Population Forensics.
DR. MOOSE: I'm Pat Moose, good morning. I am representing the Pacific Island Council of Leaders. We are a group very interested in the racial category for Pacific Islanders, and we are here to be able to provide input and also to hear the discussion that takes place. Thank you. Data is very important for us.
PARTICIPANT: (Comments off mike.)
DR. ARGUELLES: I am Dennis Arguelles. I am the Assistant Director of the Asian-American Study Center here at UCLA.
DR. MALONE: I am Nolan Malone. I come from the Kamehameha Schools in Honolulu, Hawaii, where I am a research scientist. I am a demographer by training, a sociologist as well, and so I deal with the data issues that we face in our research of Hawaiian well-being.
DR. ONG: Good morning. I am Paul Ong. I am the Director of the Lewis Center for Regional Policy Studies here on this campus.
DR. POESSI: Good morning. My name is June Poessi, and I am the Director of the Office of Samoan Affairs, and also a member of the Pacific Islander Cancer Control Network, which is with UCI and NCI.
MS. WHITE: I'm Gracie White, staff to the NCVHS Subcommittee.
TRANSCRIBER: I am Chanda Chai. I am the transcriber for the committee.
DR. FIELD: My name is Greg Field. I am with Audio Visual Services.
DR. WILHEID: My name is Cheryl Wilheid. I'm with Magna Systems.
DR. MAYS: We just had someone step in. So you might let them introduce themselves, and then we're going to get started.
DR. TAGALOA: Mishi Tagaloa with the Second Samoan Congregational Church.
DR. MAYS: Great. We're going to get started. The presentation we are going to start with first is Dennis Arguelles, who as he introduced himself is the Assistant Director of the UCLA Asian-American Study Center. Dennis has been with the Center and has played some very key roles, both in his work with the community as well as some of the work that the center has been doing with the census.
One of the things that we asked Dennis to do is to help us to understand the diversity. I think it is important both for the committee as well as others to understand the diversity of the populations. I think what that will help us to do is to insure that we have a very good overview of the populations that are involved in -- when we say the terms Asian, Native Hawaiian and other Pacific Islanders, that we understand the task that is before us in terms of the diversity of those groups.
So Dennis, I appreciate your doing this. I know that the Center has been very involved in work in this area, and that the Center actually has -- I think you are presenting from one of the Center's publications, which I think Paul has been involved with. I am appreciative of your spending some time with us today. Thank you on behalf of the committee.
DR. ARGUELLES: Good morning. I am pleased to be able to present this information to the committee this morning. I am also very pleased to be in the presence of some very important community leaders, at least in the Pacific Islander community. I am pleased to see them being out today.
Just a real quick history. The Asian-American Study Center at UCLA is about 30 years old now. Obviously we have a teaching, research and community service mission, but community service is something we really emphasize. As part of that, we about three years ago launched a census information center project, which we are doing in collaboration with some community-based organizations. Basically, it is our attempt to take census data and make it usable for the Asian and Pacific Islander communities for their organizing, advocacy and service delivery efforts.
Much of what I am presenting to you today is -- well, it is all primarily census data, and a lot of is drawn from a new publication that I am going to go ahead and push right now. It is a book that I edited, it just came out a few months ago, it is called The New Face of Asian Pacific America: Diversity and Change in the 21st Century.
It is a comprehensive overview, a demographic profile of the Asian and Pacific Islander communities in the United States, drawing upon the latest census data, through the release of SF-3, which for those of you who are familiar with census data provides a fair amount of detail. But we also drew upon the expertise of numerous scholars and researchers across the country, including the likes of Paul Ong who is here this morning, who I consider probably the national authority on Asian-American demographics.
Finally, I just need to say that the slides that I am presenting this morning were actually prepared by Melanie Dela Cruz, who is the staff person at the center, who is coordinating our census information project, and who I think will be presenting this morning.
When I thought about presenting to professionals who deal with statistics and various sets of data, I wanted to make sure that my presentation this morning wasn't too elementary. So I just tried to think about what were probably the most salient information that I wanted to communicate this morning.
When I think about the Asian and Pacific Islander community in the United States, at least from the work of this book, for me three main things come to mind. The first is the emergence and growth of several newer ethnic populations within the Asian and Pacific Islander racial categories. This doesn't mean that some of these communities don't have a long history here in the United States, and have a history of immigration that goes back several decades, maybe even 100 years, but it does mean that a lot of communities are now reaching critical mass, and obviously that is going to have various policy implications.
The second point is, one thing that is very important to do is to disaggregate data and make sure that individual groups are recognized and treated properly. One of the key findings in our book is that you really need to look at Pacific Islanders separately from Asian-Americans. Their histories, their experiences and their economic and social conditions are much different from that of Asian-Americans. As well as, with the Asian category there is quite a range of diversity.
Finally, one thing that is very important to keep in mind -- and I'll be referring to these points as I go through my presentation this morning, but the significance of the multi-racial Asian and Pacific Islander population is tremendous. As you will see from some of our numbers, it can even alter numbers and populations by significant amounts, which could have some policy implications.
So those are the three points that I will be referring to throughout my presentation this morning.
The Asian and Pacific Islander populations in the United States, what are we talking about? First let's look at some overall numbers right now. As you can see from the census, when you look at race alone, Asian-Americans are a little over ten million, Native Hawaiians and Pacific Islanders, about 400,000, but when you use race in combination, you will see that for Asian-Americans that increases to almost 12 million, and Native Hawaiians and Pacific Islanders more than doubles to over 800,000.
What are we talking about when we say Asian and Pacific Islanders? First, Asian-Americans, we are talking about over ten or 12 distinct groups, at least that data is collected for through the census, but we are also talking about numerous other groups. We are talking about Asian Indians, the huge South Asian population. That includes the countries of Bangladesh, Pakistan. The Chinese population is huge, but that is also divided up between Chinese from Mainland China as well as Chinese from Taiwan, probably for political reasons or lobbying reasons. So as you can see, it is quite a range of groups.
For Pacific Islanders and Native Hawaiians, which I am actually just learning a lot about myself, you see three distinct regional groups: Polynesians, which tend to be associated with Hawaii, Samoa, the Central and South Pacific; Micronesians, who tend to come from farther west in the Pacific, Guam and that area, and Melanesians -- I'm sorry, west and north for Micronesians, and Melanesians, you are talking about the farthest western and southwestern parts of the Pacific, and that would include countries like Fiji as well as places like the Solomon Islands and Papua and New Guinea.
Let's look at some numbers now. The Asian population in the United States since 1980 went from about three and a half million in 1980 to almost seven million in 1990, up to over ten million in 2000. If you include Asians in combination with other races, you are talking almost 12 million.
For Pacific Islanders and Native Hawaiians, you are looking at some large growth between 1980 and 1990, a little bit slower growth since 1990, but again, if you add Native Hawaiian or other Pacific Islander in combination with any other race, you are talking about a big increase in numbers.
As far as specific ethnicities, -- and again, I have all these slides; I have a set I can leave with Dr. Mays before I leave today -- here again, this is where race in combination has a significant impact. If you look at some communities, for instance, Japanese-Americans, Filipino-Americans, you're talking about substantial increases, when you add Asian with race in combination.
These groups I guess can be considered the big five or six. Chinese-Americans still are the largest population and they have been, Filipino-Americans next. Asian-Indians however are probably the fastest-growing category, and I'm sure in the next census will probably surpass Filipinos as the second-biggest Asian group. Vietnamese, Korean-Americans still at about a million, and Japanese-Americans at one time were one of the top two groups in the country, are down to about 800,000.
Other populations with emerging numbers, Cambodian community, Mong, which are an ethnic group from the highlands in Southeast Asia, Laotian, Pakistani. But these groups I believe tend to be most subject to under counts. I know service providers in the Thai community who feel that there are 100,000 Thais just in Southern California. That might be an exaggeration, but these numbers are obviously census numbers and don't account for any undocumented populations. But as you can see, significant numbers in a lot of these ethnicities.
DR. MAYS: May i ask you a question? This is the national data?
DR. ARGUELLES: Right.
DR. MAYS: Can you say something about the distribution of these groups, where these smaller populations tend to cluster?
DR. ARGUELLES: Sure. I have some other slides that will look at states.
DR. MAYS: Okay, that's fine.
DR. ARGUELLES: But actually, even that just gives you more general numbers. I think for these groups, it is real interesting, because a lot of these Southeast Asian populations were subject to refugee resettlement policies in the '60s and '70s and through the '80s, so you will see them in interesting places. But to make a long story short, the Mong community was dispersed throughout the country, as well as the Cambodian, but they ended up congregating in certain areas. The Cambodian community, the largest community outside Cambodia, is here in Long Beach, California, just a short drive down the freeway. The second-largest Cambodian concentration is probably Lowell, Massachusetts.
The Mong community you will find throughout the Central Valley here in California. I think the largest population in the country is in the Fresno area, but you'll also find a big population in Minnesota, some of the Great Lakes areas.
The South Asian population, Indian, Pakistani, you are seeing larger populations developing in the West Coast, where generally, as you will see from some slides later, generally the biggest numbers in terms of Asian-Americans are in the West Coast. But for South Asians, you are seeing a lot of communities developing on the East Coast. New York is a huge -- there is a huge population of Asian Indians there.
So anyway, that hopefully gives you a general idea. Let me get to those slides; those are coming up next. I'm sorry, with Pacific Islanders numbers first.
DR. BREEN: Dennis, will you be talking in a little more detail later about the resettlement policies and the implications of those for social welfare and health care access?
DR. ARGUELLES: I'm not an expert in those areas. But Paul may be able to talk a little bit about that. I'll give you some information about the book. We have some writers that have written extensively on the Southeast Asian communities, and provide a lot of information about these specific populations.
As you can see, Native Hawaiian and Pacific Islanders; Native Hawaiians are a particularly interesting group. I'll get into this a little later when we start talking about quantum, but numbers alone are 140,000. But when you look at numbers in combination with other race, you have 400,000. That is having a lot of implications in Hawaii and other areas, in terms of cervices for Native Hawaiians.
Anyway, you can see the numbers for some of the other communities. In terms of distribution, and you will see this in some slides, the Pacific Islander population tends to be concentrated here on the West Coast. As you can see for Native Hawaiians and Pacific Islanders, California, Hawaii, Washington tend to have huge concentrations, but as you can see, they are moving to other parts of the country as well.
For Asian-Americans, as I mentioned, still primarily concentrating in urban areas on the East and West Coasts, but for specific populations you'll find them in different areas, Minnesota, the Great Lakes, for instance, for Southeast Asians.
DR. BREEN: Excuse me. You mentioned the Asian populations concentrating in the urban areas. Do the Native Hawaiians and other Pacific Islanders also concentrate in urban areas? You mentioned Fresno and other areas that tend to be more rural.
DR. ARGUELLES: Yes, it depends on the population. For a lot of these Southeast Asian communities, again, that is reflected by resettlement policies. A lot of them came from rural backgrounds. For instance, the Mong community, you will find them throughout the Central Valley here in California, which is primarily an agricultural, rural area. But even in L.A. where they live, there is a small town just outside of L.A. called Beaumont, and there is a concentration of Southeast Asians there. It is a little town in the desert, but they moved there because they are comfortable there, as opposed to living in a large urban area.
Generally though, for the big groups, Chinese, Filipinos, Asian Indians, they are moving to Chicago, New York, Los Angeles, San Francisco, Seattle, the big urban areas on each coast.
DR. BREEN: Thank you.
DR. ARGUELLES: So that gives you a general idea of some of the geographic distribution. I mentioned the importance of disaggregating data and understanding the diversity within the Asian and Pacific Islander race categories. Looking at age is something that helps emphasize that.
Generally for the biggest groups in the Asian and Pacific Islander categories, you will see that they don't differ too much from the national median when it comes to age. Maybe just a little bit younger than the national average.
But when you look at some of these newer communities and smaller communities that are now emerging, you see some very significant differences. Cambodian Mong refugee populations, far younger than the national median of 35 or 36. The Mong population, less than half, primarily a young population. And you can see the distribution among some of the other communities.
Pacific Islanders, generally you will see that it is a younger population. It doesn't mean that these communities don't have senior needs, or senior populations that don't have specific health needs, but generally as a trend, probably because of the nature of their immigration, you are seeing some pretty young populations.
I only have one more slide, to emphasize the point that it is important to understand the diversity within the communities, and in particular to understand the needs and conditions of the Native Hawaiian and Asian and Pacific Islander populations, separate from Asian-Americans.
I have this chart on educational attainment. The first bar in each column being for Asian-Americans, the next for Native Hawaiian and Asian and Pacific Islanders, and the third for basically the Hispanic white population in the United States. As you can see, -- well, if you look at this initially, you will see that Asian-Americans in particular are doing quite well in terms of educational attainment. One of the things that it reveals is that the Asian and Pacific Islander community is really struggling. There are some real needs in those communities.
But I think it is also important to understand the diversity within the communities. While as a whole the Asian-American community is achieving educational success, if you look at specific populations, for instance, the refugee population, Samoan and Cambodian, some of the newer immigrant populations like Thai, those populations are really struggling. So it is important to look at more than just the overall numbers.
DR. BREEN: And those are the same groups that are getting the refugee programs, right?
DR. ARGUELLES: The ones that have been subject to refugee resettlement policies, yes. The Vietnamese, Cambodian Mong, Laotian.
DR. BREEN: It sounds like they didn't get everything the Cubans got.
DR. ARGUELLES: Well, I'm not a sociologist, and studying the conditions of these communities is not my specialty, but these are communities that came from rural backgrounds, who didn't immigrate here with any resources, or actually never even intended to come here, and were displaced because of U.S. foreign policy intervention in other parts of the world. So they really came here with a tremendous amount of disadvantage.
My feeling is, the Cuban refugee population was economically a different character or class than the refugees that are coming from Southeast Asia.
DR. BREEN: They also had enormous benefits when they came to this country. I was actually interested in the structure of benefits that was available to the different refugee populations. Maybe Paul can talk about those later, or someone else can.
DR. ARGUELLES: So getting back to my initial points this morning, again, it is really important to look at some of these newer communities, to understand their histories.
For instance, the Asian Indian population, or just the Southeast Asian population in general; it is so tremendously diverse. In New York, you have the phenomenon of Indo-Caribbeans, people of Indian descent who are coming to the U.S. through the Caribbean. They are talking about it having an impact on redistricting there, political reapportionment.
DR. BREEN: These are people who may have lived there for generations, isn't that correct?
DR. ARGUELLES: That's true. There is a long history of immigration of Indians to the Caribbean, as well as to places like Fiji in the Pacific. There is a huge Indian population there. That is also making its way to the United States. So it is important to understand these newer communities and the histories and the current conditions that the majority of their populations face.
Again, it is important to make sure that the Pacific Islanders aren't lumped into the same category with Asian-Americans. Their communities have very different histories. I'm sure some of the leadership here can talk about that history and those conditions in their communities.
Finally, the issue of multi-race Asian and Pacific Islanders. This obviously has a tremendous impact on the Native Hawaiian population. I'm not sure what kind of impact it might have in terms of health care and health delivery systems, but any time a population is increased by almost 60 percent when you take into considerations its numbers, along with people who are of mixed race, that obviously is going to have some type of significant impact.
I should add that as we speak, the Census Bureau, as I'm sure you folks know, is slowly releasing SF-4 data, which will allow it to look at each of these communities in much more detail. We are really excited about that, because we can't wait to find specific ethnic data on Thai-Americans, for instance, or Samoan-Americans. So that we are hoping to be able to compile soon.
Then just finally, this is just a general overview, but our Center's new publication, The New Face, I think will really help provide a comprehensive look at the Asian and Pacific Islander community today, and I think could be a real useful tool.
If you are interested in ordering it, let me put up a slide on how you can do that.
DR. MAYS: Just so you know, I did ask him to do this, so he is not --
DR. ARGUELLES: And all the proceeds of the book go to the Center.
DR. MAYS: I did ask him to do this, because I think that it has a lot of -- as we talked, a lot of useful information, and the committee has tried to be very good about the things that we receive, which we will know about, and everyone else can also benefit from knowing about it. So he is doing it at my request.
DR. ARGUELLES: But if you want to get a copy of the book, you can get it through our Center or you can contact Thao Cha, who is our distribution manager. It is also available in CD-ROM. I forget to mention that our partner in this publication was Asian Week, which is a national Asian-American news magazine. If you want to order the book on CD-ROM, you can order it through them.
Also, I will leave copies of my card with Dr. Mays, if you want to contact me for any reason, as well as a copy of the slides that I just presented today.
That's my presentation this morning. If you have any questions, I'll be more than happy to answer them. If I can't, maybe Paul or someone from the Pacific Islander community can jump in as well.
DR. MAYS: I want to thank Dennis. This is exactly what I had hoped for, in terms of helping the community to have a sense of the diversity. In many of those issues, while the data was very clear for us, what it does is, it raises a lot of issues for us as we think about the health side, in terms of the education level, the age, et cetera. So realize you have given us quite a bit to think about.
Let me see if there are any questions from anyone on the community at this point. Are there questions from anyone in the audience at this point? If so, please, to the mike.
MR. TAGALOA: Mishi Tagaloa, pastor of Second Samoan Congregational Church. You had a slide there that had a distinction, multi-racial. Could you elaborate a little bit on that and say what that means? What does it mean when a census goes multi-racial, and how will that have an influence in how races are counted in the future?
DR. ARGUELLES: Sure. For the first time in 2000 -- and if you remember, if you filled out the census form, you were able to check that you were more than one race for the first time. In the past, you always had to choose between five categories or so, or Other. Now you can say that you are Asian and white, or Asian and African-American.
So because of that, we are now able to generate data, the Census is able to generate data that not only shows a population in terms of people who are just that race and who identify themselves as just that race, but also people who identify themselves as that race as well as another race.
For instance, for Filipino-Americans, my community, our population increases by -- I can't remember, but it is almost ten percent, if you take into consideration the people who are half Filipino or part Filipino. So that is what is meant by mixed race.
In terms of impact for the communities, I'm sure in the Kamehameha schools they can tell you about it. It has a tremendous impact. The issue of service for instance for Native Hawaiians has primarily been determined by what is called blood concept, meaning that you have some Hawaiian blood. I'm not the expert in this; it is probably better for the gentleman behind you to talk about this. But anyway, that raises issues of who is a Native Hawaiian, especially as the population becomes much more mixed racially.
MR. TAGALOA: So is it conceivable that there will be a time when a lot of people are checking these boxes? I notice that the Native Hawaiian box, when it is unique, it is about 100,000, and when it was combined with other races, it shot up by about 400 percent.
DR. ARGUELLES: Yes.
MR. TAGALOA: So is it conceivable that maybe OMB Directive 15 might begin to speak to the uniqueness of these particular populations?
I also noticed that you have a very shorter list of Pacific Islanders. You just have the basic Samoans, but you skipped a whole lot of them, Tokalauans, Marshallese, Paluoans, the Minuatoans. They are equally as diverse as the Asian list.
DR. ARGUELLES: To answer your first question, I'm not an expert on OMB Directive 15, so maybe some other folks can address that after me. But you're right, we left off a lot of Pacific Islander categories. Unfortunately, those groups you mentioned kind of get clumped into this other Pacific Islander category, as well as in the Asian-American community. There is Burmese, there is Sri Lankan, there is Nepalese, people whose communities are emerging, but right now are left in the other category because they don't have a critical mass.
But that is important, to try to examine these distinctions. I know for instance that researchers here at UCLA, Professor Margie Singer in the school of public health, when she disaggregates data on Asian-Americans, Asian-American women in particular, and looks at cancer rates, for instance, she finds amazing disparity between different Asian ethnicities.
I should mention though that a lot of the categories that you see listed today, they fortunately or unfortunately are listed today because of advocacy and because of lobbying. Taiwanese, for instance, separated out from Chinese-Americans, only happened because -- I believe because of the strong Taiwanese lobby here in the United States. So it speaks to the need for us to work and help organize our communities so that we can get the recognition that each of them deserves.
DR. MAYS: Some of the individuals he mentioned we actually will be hearing from. Dr. Tagaua Singer will be with us tomorrow. I think as we get other files, we will actually learn a little bit more about some of the diversity that exists within some of the groups.
Thank you very much. I really appreciate you doing this. I think it is helpful for all of us.
Dr. Malone, can we ask you to come up? You can choose however -- Dennis has a lavaliere mike, so you can choose to either sit or walk.
While he does that, there are a couple of others that have come in, that it might be useful to have them introduce themselves.
(Whereupon, introductions were performed.)
DR. MAYS: Dr. Malone as he introduced himself is a demographer by training, a sociologist in terms of his discipline. He comes to us having been at the Census Bureau. When he was at the Census Bureau, as I understand, he worked in the division where issues of ethnicity -- so he was working with the Hispanic data. Since then he is with the PASE Kamehameha School.
One of the areas that they have focused on quite a bit has been that of the analysis of data in Hawaiian, so we are very pleased to have him with us. Thank you very much for making this long trip and being with us. Thank you.
DR. MALONE: Thank you. Thank everyone for coming today. I'd like to thank UCLA for hosting this nice event. It was a lot easier for me to travel to Los Angeles than it would be to go to D.C. Thank you to the committee for listening to us. It is very, very nice to know that folks are interested in issues that deal with the entire Hawaiian community, including Native Hawaiians, because as you will see from the slides, it is much more than Native Hawaiians.
I am very, very happy to see folks from Samoa here, because they play an important role in Hawaii as well.
I'll begin by simply saying that PASE or policy analysis and system evaluation, is a small research arm of Kamehameha Schools. The Kamehameha Schools is a large not-for-profit charitable trust that was established to further the well-being of Native Hawaiians, especially through the education of their children. We maintain three campuses on three different islands in the State of Hawaii that provide fantastic K-12 education, as well as early childhood education and extension programs for adults as well.
We do research not only on this program, but also on the general well-being of Hawaiians in the state and in the country and in the world, as a matter of fact. You will see from some of the data that I will present today that it is not so easy to assess the well-being of Native Hawaiians outside the state, owing to data deficiencies.
This is a general overview of what I will be discussing today. It can be divided into three main areas, the issues of data definitions, second, data sources or the data sets we have, and finally data quality, that is, the questions we ask.
The list could go on for days. However, given the mandates of this meeting and what we hope to address here, I thought it was important to narrow it down. Because there are other experts on the Pacific Island population, I focus mostly on Native Hawaiians. And because there are issues of vital statistics and health, I focus mostly on those areas, even though our specialty tends to be in general well-being and education.
We will begin with data definitions, which I have subdivided into three areas, the group race categories. I should stop and thank Dennis for really laying a great groundwork or foundation for my presentation, because I will be echoing lots of the recommendations that he made. We will also discuss multiple races, and we will discuss the bridges between old and new definitions. We talk about group data and we talk about definitions.
As everyone knows, some statistical organizations continue to report Asian and Pacific Islanders as one grouped category. I would probably be preaching to the choir to suggest this isn't great. I think it is important -- it is convenient to do that, and institutions like the Bureau of Labor Statistics and Census Bureau who administer the CPS, the Current Population Survey, on an annual basis, would say that their sample size is simply insufficient to report data at a smaller level for individual race groups.
I do know that they are making attempts to do three-year averages for certain race groups, to provide statistically reliable data, because that is what we want. We don't want just numbers; we want statistically reliable numbers, numbers from which we can draw some salient and important inferences.
However, that said, it is important to note that Asian and Pacific Islanders lumped together are not necessarily unique. They do the same for Hispanics. My history is in migration as well, so I have lots of experience in the Hispanic population. So to see statistics for the Hispanic population that lump Cubans with Puerto Ricans and Haitians can also be incredibly misleading. The same could be said for other racial groups as well. So it is important that in general, we consider encouraging all statistical agencies, all organizations that might be releasing data, to consider that detailed race groups are very, very essential for the work that we need to do.
As mentioned earlier, the OMB Directive 15 did a wonderful thing. It disaggregated that API group into broader Asians, which is still fraught with problems, as I'm sure we will learn later today, into Asians and Native Hawaiians and other Asian and Pacific Islanders. It is not ideal, but it is certainly much better than it was before.
I offer an example which we have already seen basically. I focus solely on those who achieved a bachelor's degree or more, and I do this because there was a recent press release about the most recent Current Population Survey data, and it listed those broad race groups. It said Asian and Pacific Islanders together have a very, very high educational attainment post high school.
However, when we look at Native Hawaiian and other Pacific Islanders alone, exclusive of Asians, we see that it much lower. Then using some special tabulations that we conducted at Kamehameha Schools on Census 2000 data, we can focus solely on Native Hawaiians. In this case, it is Native Hawaiians who may be sole Native Hawaiians or in combination with another race. We see that there is only a slight change.
But this is to reinforce how important it is to have the Native Hawaiian and other Pacific Islanders teased out from the broader API group, and also to show that we can even get greater precision through census data. SF-4 which is currently being released, and Hawaii has been released, these data, and hopefully will come from SF-4, those data are very, very important. They show that because Hawaiians dominate that NHOPI group in the state of Hawaii, that is why the statistics are so similar. But it would be a different story if we included the mainland, because there are educational disparities between mainland Hawaiians and those in the state of Hawaii, as well as Pacific Islanders in Hawaii and Pacific Islanders on the mainland.
By the way, a little plug for Aloho Counts. That is the name of our special tabulations that we did on the Census 2000 data. So if you see that in these slides, you will know what I am talking about.
The next broad area dealing with definitions deals with multiple races. As Dennis mentioned, it was hugely important for the Native Hawaiian population to have this benefit of the OMB Directive 15 revisions. Only one-third of all Hawaiians report a single race, so looking at those graphs that Dennis showed, or those numbers, we saw that there was a huge increase when you added alone or in combination to the mix.
Although this is a recent phenomenon, multiple race categories in federal statistics and on the mainland, having worked at Census, I do know that the age distribution of the two or more race population is very young. It is typically children of interracial marriages. However, there is a long history of interracial marriage and multiracial populations in the state of Hawaii. One of the very first books on interracial marriage comes from the 1960s, and it was written in Hawaiian by a Hawaiian researcher, and it talks about interracial marriage in Hawaii and how prevalent it was. So although it is a new phenomenon in general, it is a very, very historically rooted issue in Hawaii.
The example is basically a graphical version of what Dennis already presented you. These are 100 percent graphs, so although the magnitude of these populations differs in size dramatically, this shows you how much of that population was one race and how much of it was two or more races. The bottom part of each bar, the green part, is the percentage with the number written in the bar of that race group that reported one race. The top part or the orange part is the percentage or the number written, those who reported two or more races. You will see in white and black, very, very small percentages. I succumbed to the pressure and did American Indian combined with Alaska Native, which was not necessarily right, but you see a large proportion of them, roughly two out of five, report two or more races.
Then we come all the way down here to Hawaiian, and you see that one one-third report one race, and the vast majority report two or more races. This is just a figure, a graphical illustration, of how significant multiple races in the state of Hawaii are.
Actually, these are national figures in the state of Hawaii. They are comparable. Trying to find detail on mainland Hawaiians is actually much more difficult, and I'll get to that in a bit.
DR. BREEN: Paul, do you have this information for other Pacific Islanders?
DR. MALONE: I included Samoans, actually. They were on the graph as well, and I saw that more people were going to be speaking about Pacific Islanders, and so I felt like I should leave it to them. But Samoans had -- around two out of five, I think -- they had a very large percentage as well, rivaling that of American Indians and Alaska Natives of multi-race versus one race.
I didn't look at the others, but we do have those data in SF-4. Actually, you can get those data off SF-1 and SF-2, because having come -- I don't know if everyone is familiar with all these horrible, horrible census products that mean nothing. It is really, really confusing. I went there; it took me months to figure out what everyone was talking about, but they released the data based on what form the data came from.
As you know, everybody in the country gets a form that asks some basic demographic information. It is called the short form or the 100 percent file. That is what SF-1 and SF-2 come from. Race is on that form. So you can get race totals from summary file one or SF-1, because everybody in the country answered that question and they processed those data quickly. Everything from the long form, those extensive questions about nativity and education and income and housing costs and so forth take longer to process, so those data are reserved for summary file three and four, which only came out recently, and summary file four is currently being released.
What is the difference between summary file one and summary file two? Summary file two iterates everything by race, which is really important. Everything that is on summary file one for the most part gets iterated by individual race groups in summary file two. So if you want to know specifics for a race group, go look at summary file two, if it was something that was found on the short form. If you want to look at the long form, you go to summary file three. If you want to look at information from the long form for individual race groups, go to summary field four. That is kind of a key of looking at those four data products from Census.
DR. HITCHCOCK: Dennis, will you refresh my memory as to how these items were collected? I see the standard racial categories, and Hispanic would be separate, I presume, but then the -- well, no, Chinese, Filipino, Japanese, is that nativity or birthplace?
DR. MALONE: These are simply race. Place of birth has actually nothing to do with the race questions or the ethnicity questions. There was a test done later. Members of the REAC committee can tell you that they did do tests to see, for those who reported some other race, they looked at place of birth to see if they could determine what their race category would have been had they used the non-some-other-race categories. So these all come directly from the race question on the Census 2000.
I tried to list -- as a former migration researcher, I listed the more populous groups in the United States from the Asian community and compared them to Hawaiians, as well as to the primary or predominant racial categories.
The third area of definitions that I wanted to discuss was another recommendation that was made in this OMB Directive 15 revision. I will mention that the OMB Directive 15 -- if you go to the OMB website, www.omb.gov, and simply search for Directive 15, they have a nice, convenient, small, uncomplicated page, very different from the Census Bureau, dedicated to this. There are three files that you can look at. They show the original call for information, asking for researchers and agencies to provide their insight as to whether or not we should revisit how we classify race, a summary of all those comments that they received, and then the final recommendations, three very simple, easily readable files. I encourage everyone to read them. They are very insightful.
As we know, these racial classifications change over time. OMB Directive 15 looked at that, but there is this conflict. We want our data to be consistent. We have data going back to the turn of the century, let's say, from vital statistics or something. We want our data to be comparable so that we can do time series analyses and to look at trends over time, and we can't do that if we start changing the way we define things.
One of the recommendations that OMB Directive 15 made was, not only should the new racial classifications be instituted by 2003 for at least federal agencies, but also that they should provide bridges to the data users so that they can map back to old definitions.
For example, if I want to do comparisons between Native Hawaiians from the decennial censuses, I have a hard time doing that right now, because as we saw, the multiple race option dramatically changed the way we look at Native Hawaiians. There is no bridge that is offered by the Census Bureau at this time for me to go back and determine how many people from the 1990 census for example would have been multiple race and Native Hawaiians; how many people from the 1990 census who didn't identify as Hawaiian at all could in fact be Hawaiian because they were multi race, and were forced to choose a single race.
This is an important issue. When we talk about data needs, we need to face the fact that we need to be able to map back to early data sources.
Back to the overview, we are just dropping into the second section, where I talk about data sets, data sources. There are three issues here, that of small populations, that of the actual respondents, and how they deal with the data sets, how they are asked, and finally, cross-sectional versus longitudinal data, and I'll elaborate on those.
This is just a description of what I mean to get at with this section, what are the current obstacles to collecting data on Native Hawaiians, furthermore, are these barriers unique to Native Hawaiians, or could we say the same thing about other ethnic groups as well, and finally, what types of data do we need.
The first big issue, small populations. As you saw from the charts that Dennis presented and some of the statistics I've shown so far, the Native Hawaiian is a very, very small proportion of the entire U.S. population. Although it is difficult and expensive to capture Native Hawaiians using random sampling techniques, it is really, really important, because we need sufficient sample sizes in order to conduct statistical analyses. I am alluding to the CPS argument earlier, that they can't present Native Hawaiian data because there are insufficient sample sizes to maintain statistical rigor.
So what do we do to alleviate this problem? Do we simply throw more money at it and try to sample more people and hope that we can grab more Native Hawaiians or other groups, other small population groups? Or do we simply re-look at the way that we actually sample people?
There are different techniques. You can develop sampling weights for samples that are not randomly selected. There are cluster techniques, there are stratified techniques that are difficult, yes, they are complicated, yes, but there are other techniques out there that need to be investigated.
Furthermore, over sampling when it is done is simply done at a geographic level. You will see in an example that I will present to you that Native Hawaiians, unlike other ethnic groups, not all other ethnic groups, but some other ethnic groups, are not necessarily residentially segregated. Therefore, doing geographic over sampling, for example, in a certain census tract, will not necessarily be effective in capturing more Native Hawaiians.
This example might seem a little confusing. The REAC Committee, the Race and Ethnicity Advisory Council to the Census Bureau, suggested this very same argument, that we need to over sample Native Hawaiians and other Pacific Islanders in order to do better analyses. So the Census Bureau performed a special tabulation, and out of all the bloc groups, which are very, very, very small geographic levels, and census tracts which are also small but larger than bloc groups, they looked at every single one in the United States on the mainland. They found these 20 percent as a threshold, how many of these had 20 percent or more Native Hawaiian or Pacific Islander. Five, that were Native Hawaiian or other Pacific Islander alone, out of all the bloc groups in the entire country on the mainland.
When they expanded the definition to that maximum, alone or in combination, which is the phrase, which simply means that it is Native Hawaiian or other Pacific Islander with any other race, we get nine of the entire mainland. So this is very, very discouraging for the issues of over sampling based on geography.
DR. HITCHCOCK: I apologize, I think I called you Dennis the first time, I'm sorry.
DR. MALONE: That's okay.
DR. HITCHCOCK: We are running behind time, I guess. Do you remember where those tracts are?
DR. MALONE: They listed them by code, and luckily I know the codes. I seem to recognize California and Washington State. It is either Washington or Oregon. No, I think it was Oregon. For these five, when you lower the threshold down to ten percent, you do get a broader range, but most of them are on the West Coast.
DR. HITCHCOCK: So if you were to do a survey of Native Hawaiians and you wanted to make sure you captured most of the population on the mainland as well as living on the islands, then you would probably want to go to these areas as well as the islands, and you would probably get 90, 95 percent of the population, right?
DR. MALONE: Actually, roughly half of the Native Hawaiian population lives on the mainland, and we know very little about them, because we can't find them. So we can go into these census tracts, we can't go to specific areas and over sample there, because we are finding that they are not geographically segregated. They are not residentially segregated.
DR. HITCHCOCK: Interesting. Thanks.
DR. MALONE: Unlike other groups. There is lots of literature out there about other ethnic groups that are residentially segregated, be it voluntary or involuntary.
DR. HITCHCOCK: Good. Thanks.
DR. MALONE: Does this make sense to everyone? I had a hard time figuring out how to present this.
The second area of data sources deals with respondents, the actual people that we survey, that we ask questions of. This gets to some cultural issues which I will touch on a little later as well. But among Hawaiians, culturally, in-person efforts are much more salient, are much more culturally acceptable than the detached mail-in or telephone survey methods. It is just an issue of being sensitive to cultural differences. I'll touch on this a little bit later.
Our own surveys that we conduct through Kamehameha Schools recognize that. Although we contract out to market research firms to perform much of our surveying, we have very explicit instructions on how that survey methodology is to be conducted. The first contact is always in person. I'll touch on it again later, but I just wanted to mention it when we are talking about respondents.
Second, concepts of O'hana pono malama affect relationships with respondents. Respondents in Hawaii, Native Hawaiian respondents especially, don't see it as a detached person coming in and asking questions; they see it as a relationship. So that is something that we also recognize. O'hana means family, pono means righteousness or doing the right thing, and malama means caring.
DR. BREEN: Means?
DR. MALONE: Caring or nurturing. That leads to the third point, which is basically that attention to these cultural differences can actually result in successful coverage. If you establish a relationship with your respondents, then you have less probability of loss to followup if you need to do a longitudinal study. You have greater probability of response, or higher response rates. So there are ways where culture actually benefits those of us who do survey work or research in this area.
DR. MAYS: Let me ask a question before you move on. Do you find that it depends upon who says they are doing the survey, the extent to which you actually get the participation? If the federal government comes in and they say, as IRBs are now making us do, we have to really identify who is doing the study. So if it says it is the federal government versus for example your school, would you get greater cooperation? And is there a sense of mistrust, distrust, et cetera in terms of participating?
DR. MALONE: It is funny, because it varies by island. People on Oahu are very different from the people on Hawaii, for example, on the big island, so the response rates vary. But I think it does matter.
Informed consent, the human subjects protocols have really come to the forefront, that we always have, but have really come to the forefront, really do benefit us in Hawaii, because informed consent means talking story with your respondents and explaining to them what is going on. That is very, very appealing, and that is very conductive to establishing these relationships.
For that reason, we always say up front that we are Kamehameha School, which is very meaningful among the Native Hawaiian population. It is very, very meaningful. It is a very powerful organization culturally speaking to the people of Hawaii. It was founded by one of their last members of royalty, who left all of her money in a charitable estate to benefit Native Hawaiians. So it is a very, very powerful thing, to say we are calling from Kamehameha Schools. Even when we contract out, our contractors will always identify themselves and who they are, and for whom they are conducting the survey.
So it does matter. I think saying that you are contacting them from the federal government would be important, because many Native Hawaiians would be pleased to hear that the federal government cares. But it is more than saying I am from the federal government, you have to hep me, you have to answer these questions. I was involved in the CPS. I was actually a subject in the CPS, which is very odd; how many of us get to actually be a subject in these national data that we actually analyze? It was a nice relationship that I had with this person, but lots of that was based on the fact that we were both Census employees. So I think establishing that relationship does lend itself to greater participation.
Moving along, I would simply say my big plug for longitudinal data. I do this everywhere I go, it seems. Most current data sources are cross-sectional. That means they just take a snapshot of a population at a time, T. However, when we take a second snapshot ten years later, one year later, however long, it is rarely of the same people. It can be, but often it is not. So we are making general inferences, or we are drawing inferences from aggregate measures of maybe the same population, they are all Native Hawaiian, for example, but they are all different people.
In order for us to see what trends there are, we really need to ask the same people the same questions over time. That is where longitudinal data comes in. It is very expensive, it is very difficult to develop and implement and get funded, but I think it is very, very important, especially in the case of small populations.
Trends are especially crucial to the ongoing analysis of Native Hawaiian well-being. I'll present some data in a moment that shows some disturbing trends. But also, playing off of what I have discussed about cultural differences, I think it is important to realize the benefit of those cultural differences again.
There is a theory that longitudinal analysis might be especially suited to Native Hawaiian populations, owing to their relative geographic isolation. Many of the problems associated with longitudinal analysis is called lost to followup, meaning you can't find that person that you asked a question of a year ago, because they have moved or their phone number changed or something to that effect. When you have an isolated community such as those in the state of Hawaii, then that process tends to be a little easier.
DR. BREEN: That seems inconsistent with what we were just talking about, where there is an enormous amount of out-marriage within Native Hawaiians, and there are half of Native Hawaiians living in a dispersed fashion throughout the mainland of the United States. So my guess is, you might have thought about how to do this a little more, but it seems like there would be enormous loss to followup.
DR. MALONE: There is enormous loss to followup on the mainland. Most Native Hawaiians --
DR. BREEN: But just in Hawaii, let's just focus on that.
DR. MALONE: Yes, we're just talking about Hawaii.
DR. BREEN: To be on the mainland, they have to have left Hawaii, right?
DR. MALONE: Right.
DR. BREEN: So some people are leaving.
DR. MALONE: I think these are second-generation mainland Hawaiians as well, those who have never been to Hawaii. We would like to know more about them. Many people leave for school. Some leave for work, but mostly for school. Many return. At this point I think it is 40 percent or so, 40-60 split.
However, when talking just about the state of Hawaii, I think it is just theory, but loss to followup would be minimized owing to the geographic isolation and the cultural bonds that keep people there as well, and the social ties.
Now we will move on to the final meaty section, which is basically the questions we ask, data quality. I break it up into two specific groups, health issues and language and culture. I am not an epidemiologist or a health researcher, so I can't speak to some of the statistics I will be presenting you, but health is an important element of the discussions of Native Hawaiian well-being, especially in the state.
As far as health issues go, current research shows that Native Hawaiians are less likely to be insured, less likely to seek treatment. They are more likely to experience domestic violence and substance abuse.
This is just a snapshot of some of the important issues going on right now in Native Hawaiian health. I'll be presenting some data, and I would encourage everyone to contact the publishers of where those data come from, because it is a really great resource. I have one copy here, I'll gladly leave my copy behind for folks to look through, but it is a great source of current information on Native Hawaiian health.
Some of the important measures that it discusses are help seeking and risk behaviors among Native Hawaiians, chronic illnesses, cancer and obesity, and detection and treatment and how they differ among Native Hawaiians relative to other ethnic groups in the state of Hawaii.
One thing that is unique is that Hawaii is one of the most well-insured states in the Union, as far as health care goes. But when you look at the actual ethnic breakdown within the state, you find that Native Hawaiians are among the lowest, but nicely, still above the national average.
Here is the example I was talking about. It comes from a journal called Pacific Health Dialogue, the most recent copy, that is produced by (foreign phrase), an organization in Hawaii that deals with Native Hawaiian health. This simply shows breast cancer mortality rates from two different data-gathering periods.
The big red line represents Native Hawaiians. You see that the trend among Native Hawaiians is very, very disturbing relative to the other ethnic groups represented here.
This is simply to underscore the importance of health statistics among the Native Hawaiian community and other ethnic groups within the state.
DR. BREEN: Can I just ask you a question? I don't know if you know about the denominators of these underlying rates, but they make a huge difference in small populations. It is something that -- one of the main data sources is the surveillance, epidemiology and results survey, which is a registry, actually, of all cancer patients. Hawaii is one of the sites for SEER data, as we abbreviate it.
With changes in the 2000 census, and particularly not knowing how to bridge the '90 and 2000 census, enormous changes have been occurring, and there is a lot of uncertainty in terms of what the actual rates are now. They look pretty stable in the other populations. Are you comfortable with the consistency between those two periods in terms of the vital statistics, the underlying mortality rates?
DR. MALONE: Well, my feeling is, as I mentioned, not being a health researcher, when I look at data such as these and I worry about methodology, the one consolation that I would take from it is that the same methodology was applied to each group.
One thing that is important to note is that these were well before OMB Directive 15 instituted the multi race thing. However, DOH uses a different algorithm for determining race than the U.S. Census Bureau would have done in 1970. So for that reason, you might have somebody classified as a race, and your denominator is actually dealing with people classified in a different manner. The only consolation I would take from that is that that same method is being applied to each of these groups. I would imagine that that enormous gap would still be significant if we looked at it in other ways as well, applying the same methodology to each group.
DR. BREEN: The other thing I see, those are just straight mortality rates, the denominator-numerator. I was thinking survival when I made those comments, but instructions to people who are filling out the death certificates can make a difference.
I'm not saying those are right or those are wrong. I'm just bringing up a lot of data issues that I know NCI has struggled with to try to understand them. When you see something like that, immediately it is like, oh my God, what happened, why did that rise so quickly. Sometimes it didn't. Sometimes there was something with the data that caused the rise, rather than an actual rise in mortality in this case.
DR. MALONE: In the article, Joanne Sark does mention some probable explanations for this trend, as well as some of the others.
DR. BREEN: Do you recall what she said?
DR. MALONE: Some of them dealt with issues in how things were reported, and greater advertising. It could just simply be greater accuracy in reporting those vital statistics, those causes of mortality and so forth.
DR. BREEN: At NCI we do think that Native Hawaiians have high rates of breast cancer mortality, and we do think that there is a lot of late stage diagnosis of breast cancer.
DR. MALONE: It's a great article. The data seem a bit old in my opinion, but I don't know, vital statistics are odd, and trying to amass them and do an analysis on them, sometimes you get them much later than when they are actually compiled. So for that reason, I think it was enlightening to me to read that article as well as many of the others.
Let me leave that journal with you, because I have more in my office.
DR. BREEN: Oh, okay, great, thank you.
DR. MALONE: But I encourage everyone to write down the site and contact (foreign phrase) for it. They are happy to send them out.
Are there any more questions on this? I'm taking a little too long, but I'll quickly go over language and culture, the other issue of the questions we asked or data quality that I wanted to discuss.
I just wanted to mention that there is a language revival in Hawaii. There were proscriptions against speaking Hawaiian in Hawaii, so there is an entire generation of Native Hawaiians who actually are unable to speak Hawaiian. However, that is changing. We now have emergent schools, we have adult education classes to teach Hawaiian, so there is a resurgence. That is important in the questions we ask, using colloquial expressions, using Native Hawaiian terms are important in some of the research we do.
Also, Native Hawaiian health practices. There is a separate area of Native Hawaiian health practices, health practices that are traditional methods that Native Hawaiians use that are overlooked in national surveys and things that just do broad-scope needs assessments.
There is also differing family structure. A lot of the research we do in social sciences and hard sciences deal with family structures and how family ties affect treatment and prevention and so forth. For that reason, it is really important to recognize that a family is not necessarily Ward and June and Beav and Wally; it can go beyond that for all different groups in the United States, and even regionally it is very different.
One thing that has always bothered me about national statistics in the United States is that we refuse to accept that child fosterage occurs. The child has to be yours or there is no emotional bond if it is not biologically yours. We know that that is not true, we see it around the globe. Child fosterage occurs, and it occurs in the United States. We can't ignore that. It occurs a great deal, and it is very, very meaningful in Hawaii as well.
Finally, a summary of what --
DR. BREEN: Excuse me, but to summarize on the different family structures, I don't want to reduce this, but would you say the family is not necessarily a nuclear family if there is a much broader range of possibilities, extended families, fostering?
DR. MALONE: Yes. Aunties, equivalent of godparents, things of that nature. If you ask many people who is in your family, they would list all the names.
Finally, this all gets to the conflict between Western and traditional values between certain groups, and Hawaiian values and priorities may be very, very different from that Western model.
Finally, we have some recommendations, and we say what we do in the state of Hawaii. Recommendations would be briefly, to improve coverage of Native Hawaiians, do that by oversampling or doing targeted surveys, for example. Especially on the mainland we know very, very little about Native Hawaiians on the mainland, except from the decennial censuses, which is only a snapshot, and it is only limited information.
Please collect and report in detail race groups. Disaggregate the traditional API group. Detailed race groups are very, very important. You just get a wash effect when we combine very, very broad and different groups into a single lump category.
Third, develop longitudinal data sets so that we can analyze trends and we can track individual development. Finally, it is really important that we consider cultural factors. Language, community and aloha, caring, love, are very, very salient to the Native Hawaiian community.
What do we do in the state of Hawaii? In Hawaii we have some really great data sources. The Department of Health is amazing in the state of Hawaii. They have the Hawaii Health Survey, which is a huge survey that asks a myriad of questions on health. They also ask detailed ethnicity, not only of you, but of your parents. Furthermore, they allow folks like Kamehameha Schools to insert questions on the survey as well. They are really wonderful.
There is a behavioral and risk factor surveillance system which looks at at-risk behaviors, treatments and so forth. There is still annual vital statistics as well. There are other data sources as well. Kamehameha Schools conducts the Hawaiian community survey, which is a random sample of the state of Hawaii. We are also trying to do that on the mainland, and we are having very, very little success.
The Hawaiian community survey is longitudinal. We have around a 60 percent followup rate with our respondents. The state of Hawaii data book is secondary data sources compiled. This is the national statistical abstract. The Hawaii tumor registry, which is an invaluable source of information on cancer among Native Hawaiians. The Native Hawaiian data book, another source of valuable information.
DR. HITCHCOCK: Is that CDC?
DR. MALONE: No, the Native Hawaiian data book comes from OHA, Office of Hawaiian Affairs.
DR. HITCHCOCK: There are a series of CDC state data books. They call it just that.
DR. MALONE: Really? I do know that CDC, their vital statistics do take out Native Hawaiians. They send data CD's free of charge.
DR. HITCHCOCK: What about the risk factor surveillance system? What about the one for kids?
DR. MALONE: Yes, there is a team.
DR. HITCHCOCK: Thanks.
DR. MALONE: Sure. In the state of Hawaii we also talk story. We have face to face interviews. We have language flexibility. Our interviewers speak several different languages. They speak not only Hawaiian, they speak pidgin as well, because many Hawaiians speak pidgin, which is kind of a slang. Togala, Chinese, Japanese, Mandarin, Cantonese. It is important, we talk story. We don't say, all right, sit down, answer my questions, let's go. We say hi, how are you, we take time.
Then we communicate our results. We don't leave and never talk to them again. We contact them and let them know that we have finished the survey, would they like to see the results, do they want it on a CD, do they want it in a report, do they want to come to a public meeting where we discuss it.
Also, we have a sense of community in Hawaii. The Hawaiian DOH as I mentioned is fantastic. They share that data, they are very, very helpful. (Foreign phrase) is a nonprofit organization that deals with Native Hawaiian health issues. They are a great source of getting people together and sharing information. The University of Hawaii's John Abern School of Medicine has a great team of researchers that look into Native Hawaiian issues regarding health. The Kamehameha Schools, we do lots of research on issues of Native Hawaiian well-being and education.
The Hawaii DOE, even though they are financially strapped and face a very, very difficult time, they have never denied our requests for data, and they work collaboratively with us on lots of our reports, even though they are not necessarily favorable for the DOE. And the Office of Hawaiian Affairs, the umbrella group that does an amazing job at getting people together and sharing resources.
This is all to say that in Hawaii we have a really good system. We communicate with each other. It is a very tight-knit research community. However, what is lacking there is what we know about the mainland, those on the mainland, the Native Hawaiians on the mainland. It is a big source of -- it is just a black hole. We don't know what is going on with the Native Hawaiian on the mainland. Although we have a sense of community within the state, on each island, mostly on Oahu, it is a real big problem, getting to the mainland.
Any questions?
DR. MAYS: I think because the committee has asked several and we are limited for time, why don't we take them from the audience? If you could step up to the mike, please, and identify yourself? If there is anybody else that wants to ask a question, if they will just step up to the mike.
MR. POUESI: Thank you very much. My name is Jim Pouesi. I am a Native American Samoan. I want to thank this committee for this opportunity and this hearing. It is very much needed in our communities. I would also like to thank Dennis and also Nolan for the great presentation.
My question is, you were saying that over sampling and a targeted type of survey is what you are recommending. For the Native Americans, they are a very small group. What is now presently being used by them in the Census or throughout the federal government that is allowing them to capture their information? Can we not use the same?
Thank you.
DR. MALONE: Actually, yes, I think these methods can be tested for different groups. As I mentioned, some groups are geographically isolated. For example, if somebody wanted to know about Native Hawaiians in the state of Hawaii, they won't have a very difficult time finding them. if they wanted to know about Puerto Ricans in New York City, they might not have a very difficult time finding them, either, because there is lots of geographic segregation among them. If they wanted to know about Native Americans on reservations, they might not have a difficult time. But looking for those that fall outside those boundaries, it is difficult. Looking for those ethnic groups that don't geographically concentrate, it is difficult. Over sampling isn't necessarily going to do that. Targeted surveys would.
I don't necessarily have the answers, but I think that it is a question that needs to be asked, and it is a methodology that needs to be researched. I do know that Census 2000 had overwhelming response, much better than they anticipated. I was there when they released those results. That was owing to a multi-billion dollar ad campaign, where census forms were left in 7-11s for folks who may not have gotten them at their home address or didn't have a home address. There were TV commercials, there were billboards and so forth. These types of things would be effective in targeting certain populations, especially language minorities, those who are linguistically isolated. Having public displays, an ad campaign in a specific language, saying, be counted, we want to know what you feel, what you are experiencing, in that language. There is a connection there. If they could call up a number and speak to somebody who actually speaks their language and who can hear their story, it would probably be very effective.
But it is just one option. I don't have the answers, but I do know that relying on archaic methods isn't working.
DR. BREEN: Could I follow up to that?
DR. MAYS: If it is quick.
DR. BREEN: The answer isn't quick, but the question is really quick. Do you have any thoughts on the American Community Survey and how that might contribute to improving data on small populations?
DR. MALONE: On small populations, the jury is still out. Unfortunately, it is just not going to do it. I was there when C2SS, the Census 2000 Supplementary Survey, which was the 2000 version of the American Community Survey. They had greatly expanded the number of households surveyed, and they duplicated the long form. I was there when those data came in, and I did lots of analyses on them. We are going to have these data every year, so cool, yay. But then I had to do some work on Native Hawaiians, and I couldn't find them. They were teased apart. Because of confidentiality issues, they didn't offer an easy way of finding Native Hawaiians alone. You could find Native Hawaiians and other Pacific Islanders as a group, but you couldn't find Native Hawaiians alone. The way you had to do it was go through every single code and look for Native Hawaiian. It would say, white, black, yadda yadda, white and black, white, black and Asian, white, black and specific groups and so forth, and you had to go through every single code and look for Native Hawaiian. Then there was this catchall group, and you didn't know how many Native Hawaiians were in there. So I'm concerned about that.
They are supposed to bump up to three million households, ideally, if they are funded. It looks like that funding is going to come through. When I left Census, it didn't look like that funding was coming through. That will certainly offer a benefit, but for some of the smaller populations at the national level maybe, looking for specific geographies, like the Southeast, for example, or Central South area or something like that might be a statistic. At the state level it will probably be a huge problem.
Even doing statistically reliable analyses in the state of Hawaii for Native Hawaiians would probably pose a problem using the ACS at this point. We would have to do multi-year averages.
DR. MAYS: Thank you very much. Each of these presentations have been excellent, in the sense of how much I think we are all learning. What I really appreciate is, even though you don't have answers, you have suggestions and recommendations. I think it is helpful to us as a committee to hear them.
Dr. Ong, first let me apologize, because I know you are here, so we might be keeping you from something. I have to be very careful, because my colleagues are darting in and out of here, so I apologize for the fact that we are running a bit behind.
Dr. Ong as he introduced himself is here at UCLA with us. He is a professor in urban planning. In addition, he has served on the race and ethnicity census committee, and I think just recently came off of that committee, so he brings to us a very thorough knowledge about some of the census issues. But also what he brings to us is a lot of knowledge about the contextual issues. So for those of us in health and how to understand a lot of these issues, Dr. Ong has been an excellent source of providing things like asset mapping, when we are trying to look at things like our colleague, Dr. Singer, is trying to look at issues like cancer. He is knowledgeable -- one of his areas is transportation, so he brings a very interesting set of skills that we normally don't have at these meetings. So I am very happy that he agreed to do this.
So Paul, thank you very much. Again, my apology if someone is waiting on you somewhere.
DR. ONG: Thank you. Thank you for the invitation to be here. I have to commend Dennis and Nolan for great presentations. Some of what I am going to say will repeat, so I apologize for that.
I am also present without much numbers, and with the right side of my presentation cut off. But that's okay, I can make it through.
Good morning. What I want to do is make some comments. I don't want to present particular statistics or particular numbers. If you have questions, if I can recall the numbers, then I'll be happy to share them with you, but it is not really about the numbers.
Before I start, I ant to make sure we acknowledge that there has been remarkable progress. It is easy for us to talk about all the problems and all the shortcomings, and certainly there are numerous ones, and we need to deal with those. But at the same time, there has been remarkable progress, particularly with census data and with other data as well. There are better accounts.
Nolan mentioned the outreach. Matter of fact, Nolan's comments created flashbacks for me. Sara who is back there and I had the privilege of being on REAC for about six or eight years. That overlapped the 2000 census and the many friendly battles with the staff of getting things done.
But certainly there are better counts. Overall, the under count rate is very small by historical measures. By some measures it is even close to zero, if you believe some of the estimates.
There seems to be a great deal more sensitivity about differences in population. Certainly we have seen that in terms of the response to people who are of multi-racial heritage, and the change in the way we collect questions on race. There is certainly sensitivity about different types of household structures, and particularly for structures of where there are members of the same sex. We have seen acknowledge for example of the increasing role of grandparents as head of households.
So in some ways, there is the sensitivity that I saw and that we worked with in the Census Bureau, in terms of understanding the diverse populations that are out there. I'm not just talking about ethnic and racial diversity.
There s also better reporting that is taking advantage of new technology to making reports available. It still may be a little bit cumbersome, so when you go to the Census Bureau's website, you may not be able to navigate it very well until you have done it a few times, but nonetheless, there is better reporting.
There is also better access to dissemination, particularly of public data sets. Those in the universities always had one way or another access to that data. It may have been somewhat delayed, but now particularly if you have a fast line, a DSL, if you're lucky you have a TS-1 or some of the faster ones now, you can download these very large data sets. I do that at home. When I forget to bring my data set from campus, I download them directly at home and do my work. That is remarkable.
There has also been an enhancement in terms of community capacity, particularly the monies invested in the census information centers. The numbers that they are now supporting are very large. If you are a community organization that is willing to dedicate some resources to this, you could become a CIC. Maybe I am making a little bit too much of a blanket statement, but there is this creation and acknowledgement that there ought to be other avenues and capacities in communities, not just the state data centers, for example, but well beyond that.
And there are more research tools. They moved into GIS and have GIS interface. They are packaging more materials that way. There are also regional centers. These are the centers where researchers can access the confidential data. So for example, rather than relying either on the aggregate data that is published in the SF files, are using the public use micro sample. If you can get approval from the Census Bureau, you can actually for example identify all of the Native Hawaiians that filled out the long form, whether it is in Hawaii or on the mainland. You can do a number of different things that you cannot do with public data, and that is remarkable.
In the past, what you had to do was go through a very long, cumbersome process and go off to Suitland, where you were a nonpaid employee. But now you can do that. We have one of those centers here on this campus. There is a center up in Berkeley, Michigan is applying for one, they have always had one in Boston, there is one at Carnegie Mellon. I think those centers will be increasing.
It is still cumbersome to apply for access, but certainly there is greater access than there was a decade ago. So I just want to make sure that we acknowledge, and we ought to applaud, the progress that has been made. Part of the reason I do that is that that is how you encourage people to continue making progress. Lesson 101, raising my son.
I want to frame my discussion as a numerator and denominator problem. In a sense, the numerator is -- at the risk of simplifying this, and I apologize to those in the health field, I will simplify this. The numerator in many instances is the data we collect on health, the occurrences that happen. The denominator really is a problem about the base population, how do we understand that base population. I realize that there are other ways of thinking about this, and this only captures part of the issues, but for me it captures the issue in an easy conceptual way that is meaningful for many of the problems we face in terms of understanding health data.
In part because it is important to understand prevalence, prevalence is the current census divided by the base population, a very simple concept. It is amazing, how simple these things ought to be. The reason we do that is, clearly we want to identify group disparities, where there are high prevalences in particular. From that, we also want to be able to understand and determine causal factors, what are the factors that relate to these outcomes that we are concerned about.
Why do we do that? Clearly it is to design policies, strategies and programs that addresses the inequality and the disparities and problems in the health field. Clearly, there is a challenge of doing all of this with ethnic specific data. Dennis and Nolan talked in depth and wonderfully about the need for ethnic specific data.
Both the numerator and denominator, there are common problems. Some were already mentioned this morning. One is the consistency in grouping, that is, how do you group populations into meaningful subpopulations. In some ways that is simple and in some ways it is very difficult. It is difficult if you are a demographer or a sociologist,with all sorts of issues about what are the relevant dimensions that form the grouping, but the grouping is also, I have to acknowledge, a critical issue.
So for example, where do you place Hispanics who are African-Americans? The process that we do it here on the West Coast by and large is very different than what happens in New York. Part of it is a politics of numbers. So let's also acknowledge that, put that on the table.
There is also issues about stability of definition of categories over time. Again, that has been mentioned. I want to talk a little bit more about it later. Another issue is timeliness, how do we get the data in a timely fashion and updated regularly. We talked about accuracy and coverage, and that is a big issue certainly in this decennial census, but it is a big issue with any data collection effort.
This is on top of the standards things that we worry about when we do surveys on the census about validity of responses and all those, non-response rates, item non-response or biases and so forth. But again, I want to talk about the four common criteria, consistency of groupings, stability of definition, timeliness and accuracy of coverage.
I want to focus a little bit first on the denominator, because that is where my expertise is, much more than the numerator. There is no question that the decennial census is in one form or another related to the denominator. It is either related because that is what they use as the denominator.
If I read the cancer incidence graph right, they used the decennial population as the base. You could modify that. Between census you could try to do population estimates, but those population estimates always start from the benchmark, which is the decennial census. So you can't get away from it, even if you are doing population estimates and population projections.
So let's look at the four criteria about the census based denominator, what we know. There is an inconsistency in grouping. Part of the complexity of grouping is that there are many different ways of doing it. Actually, the census itself provides a number of different ways of doing it. How you do it makes a big difference.
We collect what we nominally call race information on the short form and the long form. But it always confused me, why Chinese is a racial category, or why Japanese is. Formally it is a racial category, I know, because we fought for it, so I know why it is there. But conceptually it is not a racial category, but it is there.
Then there is the issue of ethnicity. In some sense, if you want to use the Chinese and Japanese, the Filipino and so forth on the race question, they are really ethnic identities, but not the ethnic identity that we may think about, because it is confounded with ancestry. What do we mean by ancestry? For Asians and for Pacific Islanders, that may be very difficult. So what happens if you are of Chinese descent, many generations from Vietnam? What does that mean about ancestry? How do you get classified, or how do you classify yourself?
Then more recently it has come to the fore about mixed race. Like I said, I personally think it is an advancement to allow people to check off one or more category if it is appropriate, given their background. I know I was somewhat in the minority in my subcommittee on that, but I think it is important. It is important to acknowledge that people have different heritage, and they have different background. But that also poses a complexity.
I also want to say that the mixed race is not new in the census, it really isn't. If you go back before 1950 and prior, and look at the state of Hawaii data, there is always off and on two categories, Native Hawaiian and part Native Hawaiian. So it is not the first time we collected information on people who are part Native Hawaiian. If you go back to different time periods, you could see some of the same sort of things for American Indian, so it is not new, in the sense that we collect through the Census Bureau people with multiracial background.
It is not also new in terms of how people responded to the census. One of the remarkable things and in some ways very funny episodes that we have learned about the census is that prior to the year 2000, despite explicit instructions to individuals of checking only one race, many individuals checked two or more races. In some ways, the incidence of that is not that much lower than the multiracial checkoff that we see in 2000. It is lower, but it is not magnitudes lower.
What was also very funny was how they went about to force a single answer out of it. So there are strange rules they told us about, and other things that they do to enforce a single race, despite the fact that people claim to be multiracial. So I just want to make sure that we get out there that the whole idea of multiracial checkoff identity is certainly not new for individuals, nor is it new in the census, in terms of past historical practice, nor is it new in terms of the data they gathered prior to the year 2000.
But certainly there is an inconsistency of how we group people together. It makes a difference. We talked about for example the difference if we use single race and multi race for Native Hawaiians and American Indians. The discrepancies are even bigger if you include ancestry. So on the long form there are ancestry questions, so if you look at American Indians, for example, if you do a count by single race or you do a count by multi race or you do a count by ancestry, the factors go up to not twice as large from single to multi, but when you include ancestry you bump it up another two or three times.
I'm sure these must have meaning. They must have meaning to peoples' lives and how they live it. We don't quite understand what it means analytically, but we ought to. But again, from a policy point of view, how you count these things makes a huge difference in terms of the population we are talking about.
There is certainly instability in the definition of the categories. I talked a little bit ago about the change, the new directive from OMB, the collecting multiracial categories. I also talked about the changes that -- individuals did the multiracial prior to the 2000 change. But there is a different type of instability. That is, racial identities seem to be very fluid. It is remarkably fluid.
One of the things that we did, and I was part of the subcommittee, was to recommend to the Census Bureau that they do a post 2000 survey of a sample of people who check off multiracial, and over sample the multiracial, ask them the question again, ask half of them the question based on the 1999 question, ask half of them based on the year 2000 question, and then go back a third time, and you count the census itself as the first time, and rotate the questions.
I'm no longer in the REAC, but in the last meeting we had, they started presenting some of the preliminary findings. One of the things that we are finding is that for example, those who checked multiracial, even if they were given the option to check multiracial again on the second round, about half of them checked single race. Then people who were single race in 2000, if they were given the option to check multiracial again, many of them did.
So it is not just -- we focused a lot on the instability of the formal institution, how we go about defining or allowing people to identify themselves, but there seems to be also an inherent instability in individuals and how they perceive themselves.
We also know that these things change over time systematically, as things evolve. A good example is what happened with the Vietnamese population, whatever that means. When we look at the data for 1980 and look at the data for 1990, and we haven't had a chance to look at the data for 2000, one of the things you find the data indicates is that Vietnamese when they first came here, many more of them identified with being Vietnamese. But by the year 1980, many more of them identified as being Chinese.
Matter of fact, that surprised the Filipinos. Remember, the Filipinos were betting that they would be number one. Suddenly there is this influx of Chinese; where did they come from? They just showed up. Part of it is that a huge segment of the Southeast Asian population were of Chinese descent. They were the merchants who got ran out, a big segment of the boat people. But their identity was fluid. It was influenced by their experience in the U.S.
So if you were Chinese Southeast Asian -- if you go to Chinatown now in Los Angeles, it has been transformed. It is no longer the Cantonese Chinese there. It is mainly the Southeast Asian Vietnamese Chinese, but they now begin to identify themselves not as Vietnamese, but as Chinese. So the instability in personal identity is not a random phenomenon, just taking the post 2000 survey as a random phenomenon. It is not, it is a systematic one.
The same thing is happening with Native Hawaiians as far as we can tell from the data. There is an historical period where identifying or not identifying Native Hawaiian fluctuates with social movements and what it means to be Native Hawaiian and the kind of Native Hawaiian.
So for example, when I looked at the Hawaiian health interview survey, it is a remarkable survey, because it does ask about your parents and so forth, grandparents. People who were less than one quarter by parentage Native Hawaiians, a large number of them in recent years begin identifying themselves as either full or part Hawaiian, or one-eighth, many of them identify. There is something happening there that we don't quite understand. So this is instability.
Let's also talk about the under count. I gave at the beginning the Census a great deal of credit for attacking the under count and mounting an outreach program which was remarkably successful. What I tell people, it is not about the over or under count rates that makes a difference for policy and allocation; it is what we call a differential under count. Even if the net under count is zero for the country, but the variance around that by groups is huge, that is, if you happen to be minority or you happen to be poor or you happen to be an immigrant without English language ability, if your risk of being under counted is greater, but if you happen to be white, you happen to be suburban with kids in college who tend to report their own residence as well as parents claiming them living at home, so they over count and net out to zero, there is still this difference. It is this difference that gets played out in terms of population and geography. It is also this difference that gets played out in differences in allocations of resources.
So it is not just the overall count rate that we ought to be concerned about, it is also the differential under count. As far as I can tell, there are still large differential under counts in the year 2000 that we ought to be concerned about.
Then on timeliness in the census-based data, once every ten years. Much of the data is coming out faster, but still it is coming out slowly. We are what at 2003 now? We are getting the SF-4, which is the more detailed, broken down by very small groups in terms of the socioeconomics. The public use micro sample, five percent, won't come out for probably another year. So there is still the timeliness issue about the census data.
There are also issues that are -- I talked a little bit about the denominator, which is census based one way or another, whether we update or not. There are also issues about discrepancy between the numerator and the denominator in this data that we collect on health occurrences versus our estimates or our counts of the population. Some of it are very clear. That is, there are differences in grouping. That is the way we group data quite often for collecting information doesn't coincide with the way we group data for census based. There is also differences in timing, that is, we almost on a continuous basis collect some forms of the health data, whether for vital statistics or other monitoring effort. We have ongoing efforts to have that data. We don't have a similar process for the population, although we do publish estimations that fills in between the census.
How good those two are is a big question, because if one drift round the other, you may get things that are artificial in terms of prevalence. Is there reason to suspect that there is a drift? The answer is yes, for some subgroups. For some groups we are better for doing population estimates than we are for other groups. If we can't improve that, then we are going to have problems for some groups in terms of coming up with understanding of trends and patterns.
There are differences in coverage. One is population-based, one is event driven. There are also differences -- and this is also partly definition and one that I am somewhat concerned about in my current research -- there is a difference between legal administrative eligibility versus self reported information. We see that for example, in eligibility for American Indian health services, that definition about who participates is really driven by legal definition of who is eligible. It may or may not coincide with self reported American Indians in the census. In some ways, we really don't know how much data are a part. We don't know for example how many American Indians who qualify for health services do not count themselves or identify themselves as American Indians in the census. We don't know how many who self identified themselves as American Indian in the census really are not eligible. So in both directions you have problems.
We did a small research project with the United American Indian involvement here in Los Angeles, where we tried to get a handle on that. It clearly shows that there seems to be some sort of discrepancy when you try to compare those numbers, because they have a different definition and a different basis for identifying who belongs and who doesn't belong.
I also want to say a little bit about our geographic biases, some of which is picked up, this morning, about difficulty of collecting information in certain areas, Native Hawaiians on the mainland and so forth. But on the flip side of that there is also a danger, because it is not uniform in terms of being able to get data.
The Hawaiian health interview survey is remarkable. I just started using it, and I am greatly appreciative of it. In California we had the California health interview survey which over samples Asians, so we get data in certain states and certain locations, that we are able to look at small populations. But the problem is that when you do that, people ask the question, or they project that it has meaning for the nation.
We see what is in front of us. We don't see what is not in front of us. Quite often, there is a danger of making a mistake on what is in front of us for everything. That is all I mean by it. I don't mean to put down these efforts. Matter of fact, I am greatly appreciative of these efforts. But then the flip side of not being able to detect other data, that is, our vision is driven by what is there, and we ought to be cognizant of that.
So how severe are these problems? The answer is that in some ways, the issues that are brought up are conceptual issues. That is, in practice they may or may not make any difference. So what if you have incomplete coverage, so long as the missing people are uniformly distributed. In a practical sense, it doesn't matter. We wish we had better counts, but it doesn't matter.
But in some situations, the outcomes are much more extreme; then we ought to be concerned about this. So the consistency in grouping, I think that matters for certain populations and not for others in matters for multiracial groups, groups such as Native Hawaiians.
For Native Hawaiians, an example. I don't want to go over what was presented this morning. It is a wonderful example where it does matter you decide to put in, who you decide not to put in. But it is not just that. We also have in the Asian population the question of Hispanic origin, and Filipinos. So it is not just the racial issues.
Stability definition of categories, single versus racial; we need to do the bridging. One of the things that we ought to do is put much more into the bridging by getting the Bureau of the Census and other agencies to come up at least in the short term, with ways of figuring out how to do the bridging.
I'm not sure what that does about self reporting. That is, there is this instability, we know that there is this instability in self identification. At the minimum, what we ought to do is be cautious in interpreting some of the population data from decade to decade or from period to period, because there is this in some populations systematic drift. I made fun of it this morning about the appearance of so many hundreds of thousands of Chinese out of nowhere, but it makes a difference. I think the Filipinos were a little bit surprised, but it didn't happen.
It makes a difference. For example, if we know the drift is happening but don't build in our population estimates that denominator, then you are not going to have the right denominator. So people who identify themselves when they go for health services are a part of that drift. They are reforming their identity when they go in. So rather than saying they are Vietnamese, they are now saying they are Chinese. If you don't build that into the estimate between the decades, then you are going to have unusual numbers. or some populations, this has more meaning than others. Timeliness. I think the population estimates are very problematic for immigrants and for populations that are multiracial in percentages. I got a rude introduction into population estimate politics here in Southern California. When I first came here, we first started publishing some estimates of Asians. Our estimate was that the Asian population in L.A. County -- I forgot what year it was, surpassed the African-American population. So being a good colleague with my fellow demographers in county government and state government, we said that, and they refused to acknowledge it. Some of the better ones told me, they can't acknowledge that, because politically it won't go.
I understand that. I understand politics. I get involved in politics and so forth. I understand that. But from a demographic point of view, it is not good. So there were official numbers that were being put out by the county and so forth about the population estimates by race. There was stuff we were doing which indicated maybe those numbers are off a little bit, particularly with the Asian population growing very quickly because of immigration and so forth. But they continued with the official numbers. Lo and behold, by the time we got to the next census, Asians were a much larger -- not larger, but they surpassed the African-American population.
My position is that we ought to try to do the best job possible. But it is difficult. There are certain populations we have a great deal of difficulty with number of immigrants. In fact, your assumption about number of immigrants, particularly undocumented immigrants for example, may make or break the size of the under count rate.
There are several ways you can estimate under count. One is based on comparison to demographic models with the count and the difference. But what assumptions you make about immigration makes a big difference in that comparative number.
So there are certain populations in terms of timeliness which came between the census, there could be huge problems.
Accuracy in coverage. Again, it may be small, but what we ought to be concerned is the differential under count rates across groups. There is also the issue about program versus self reported. I mentioned that a little bit earlier. I mentioned it in the context of American Indians, but we have the same issue for Native Hawaiian, that is, how are you eligible.
For example, Kamehameha now have an open policy. That is, you don't have to be Native Hawaiian. But it is interesting, this is a comment, not a criticism. If you go and look at the population that attends this large number of students who, at first blush they look Chinese or they look Japanese compared to Native Hawaiians. There is no question there is a certain amount of gaming that goes on when you have eligibility requirements.
That is the reality, but I wonder if we get that reality because it does have an impact on how we count people and how we compare people and how we do prevalence incidence. That is, if we do gaming in one arena but not necessarily in another arena, then you are going to have a difference in how you count these numbers.
It becomes even more important, because many of these ethnic groups, particularly for indigenous populations, these programs are coming under attack. It is part of the neoconservative civil rights movement. That is, civil rights protect -- my apologies to anybody who is offended -- protect white males. Therefore, any program that is targeted towards minorities and so forth is illegal. It is a reality. In some ways we want to laugh at it, but it is real. There are challenges for example in the state of Hawaii right now on Native Hawaiian programs and so forth.
So again, those things are real and important, but part of what we ought to do is try to understand what is happening, because I think there are big differences in programmatic identity versus self identity that we can get from the census.
Some recommendations. Certainly we should encourage more research on the race ethnic identification. I think we should take more advantage of the survey data that the Bureau of the Census collected following that cohort of people who responded in the 2000 census, and see how their identity varies, and what factors influence variations in identity.
We mentioned earlier, we ought to encourage the full implementation of the American Community Survey. It is not going to be fully implemented this round in the fiscal budget, hopefully a little bit further down, but do it in such a way that it over samples Asian-Americans and Pacific Islanders.
We need somehow to assess and improve our administrative system. That is, I talked a great deal about the census information. At the same time, we need to look at how administrative data is collected, do an assessment about particularly the race ethnic identities, how they are collected, how consistent they are. We also need to improve these systems, because there is a great deal of variation from for example one tribe to another tribe, if you talk about American Indians. So there is huge inconsistencies, and how do we get some sort of standardization there.
I am also a big component at least during the bridging period that these administrative data sets are -- surveys are done in conjunction with administrative programs, that they ought to collect questions that help us to bridge not only the multiracial question, but also help us bridge the way data is collected for administrative purpose, with the way the data is collect for census purposes. That means for example included a few questions that are modeled after the census, and see what results you get, and use that information again to improve the way we impose greater consistency between the numerator and the denominator.
We also need to develop better sampling tools for health interviews for race as well as the American Community Survey. So for example, I'm not saying this is a panacea, but one of the things we could do is improve -- put more resources and time in getting a good Asian surname list that could be used for survey. But we also need to do the fundamental assessments of those lists, that is, how good are they, how good is each name in terms of its consistency with identity with being Asian or with being Pacific Islander, where are the problems and so forth.
So we need to develop those tools. I also want to say that for populations that are spatially concentrated, maybe if we are interested in immigrant groups and so forth, we ought to do a better job of mastering and using administrative data that tells us where that population is amassing and concentrating between the census. That means better use of school records as well as other records.
There are other things that we ought to think about. I actually wrote one based on the comments this morning. One of the things that we lost in the last couple of census is the ability to look at generational changes. All we have left is nativity, that is, whether you are born in the U.S. or whether you were born abroad.
But we do know that change is much more gradual from the first generation to the one and a half generation to the second generation to two and a half generation and third generation. I think we need to know how those generational differences get played out in terms of health status, in terms of socioeconomic status.
The reason we need to know that from a practical point of view, at least in Los Angeles, what we are seeing is that although the huge demographic changes, the huge impacts of the '80s and maybe into the early '90s, was driven by immigration and the increase in the foreign-born population. We have seen, or at least we suspect we are seeing in the demographic data a demographic transition, where the driving force for a number of outcomes are now going to be driven by the second generation, that is, the children of the immigrants who are going through the schools, entering the labor market, increasingly becoming a larger part of the population.
A second generation, as far as we could tell from analysis that is being done here in Los Angeles, are different from third generation. They are somewhere in between. But we need to know where are they in between, in terms of socioeconomic status.
So I would also encourage that these efforts, including the American Community Survey, the health interview surveys, include questions that would allow us to look at generational differences.
DR. MAYS: Could you tell me how?
DR. ONG: I wasn't on the committee at the time we lost it. Parents answered questions that we lost.
DR. MALONE: Owing to lots of lobbying from European Americans, people of European descent, the parental nativity questions were dropped from the census in 1980 and replaced with the ancestry question, because individuals of European heritage felt that their heritage was not being represented, in a nutshell. That is probably not the politically correct way of stating that.
DR. MAYS: That's okay. We're not on the Internet, so it's okay.
DR. ONG: Thank you.
DR. MAYS: I'm sorry, did I cut you off?
DR. ONG: No, no.
DR. MAYS: Shall we open it up then for questions from my colleagues up here first?
DR. BREEN: I had one question. Paul, just now as you were winding up, you talked about the ancestry versus nativity question. You said that demographic changes in the '80s were due to immigration, and then in the current census, what you are finding in the '90s, they were due to the second generation. You said the second generation is different from the third.
Could you talk about that a little bit more? I wasn't really following what your meaning was, or maybe you could help us understand the importance of that a little bit more.
DR. ONG: There are three strands of research that inform me about this. I can't speak definitively, but there are three strands. One is that in our population projection model that we do for Asians a few years back, we had projected that in the early part of this century, the relative components of growth would shift for Asians.
That is, in the earlier time periods, it is immigration driven. But immigrants come and they do wonderful things, including having families. So what happened is that that component of the children of immigrants and subsequent generations, they become an increasing larger part of the growth components. But in another sense, what is more important is that that cohort, that generation, is also maturing. My particular concern as a labor economist is that as they move into the labor market, they become an increasingly large component of the labor market. So in the '80s, maybe in the '70s but certainly in the '80s, a great deal of the growth in the labor force, particularly in this state and this region, is driven by the increasing number of immigrants.
They will still increase. I'm not saying that stops, but what we see is a shift in the increasing importance of the children of immigrants moving into the labor market. So that is one stream. It is just the population models that we looked at.
The second one is that we are just completing a project on the trajectory of poor neighborhoods in Southern California, which we are doing with several regional agencies. We looked at what has happened with poor neighborhoods as they evolved, and looked at the demographics.
So if you look at the very poor neighborhoods or the poor neighborhoods, in the '70s but particularly in the '80s, one of the things we noticed was the rapid increase in the number of the foreign born in neighborhoods that were poor and stayed poor. So that speaks to us the role of immigration in driving that process. But when we compare the '90 data to the 2000 data, foreign born population were still increasing in poor neighborhoods, but not necessarily at a much more rapid rate than other types of non-poor neighborhoods, and the increase is actually much smaller than the previous decade.
So we interpret that as, we are seeing the beginning of a demographic transition, where immigration is no longer the major driving force in terms of defining these neighborhoods.
The third piece of research that bears on this is that we are doing -- it is headed by some colleagues at U.C.-Irvine, and I am part of the team. It is called the second generation project for Los Angeles, and it builds off the second generation project that was done for New York City.
One of the things that we are analyzing is the degree of economic mobility across generations, and essentially comparing what we call the 1.5 generation, that is, children who have immigrant parents but they came here as kids, so essentially they were educated and socialized here, with the second generation, which means U.S.-born kids who have two immigrant parents, the 2.5, U.S.-born with one immigrant parent and one U.S.-born parent, and the third-plus generation.
First of all, the numbers indicate that the second generation, the 1.5 and two and 2.5 generations were growing in size. So we know they are growing in size as part of the labor force. We also know that the 1.5 in the second generation, they are making progress, but they are not comparable to the third-plus generation in terms of earnings, in terms of education and so forth. So we are concerned about the degree of mobility that we see across generations.
I also have to say that generational mobility in Los Angeles appears -- and this is preliminary work, but appears to be much smaller than in New York. We don't know why, we are speculating. So if you look at the gap between the third generation and the second generation, it is much larger here in Los Angeles than in New York. This is after you account for differences in age, gender and education.
These streams of research have convinced me that we ought to be more sensitive to generations, and that from a public policy point of view, because this population will grow over the next decade or two, that we need to understand this population, understand their needs, understand the barriers they are facing, the outcomes they are experiencing, so we could have good public policy and good programs.
DR. BREEN: Thank you.
DR. MAYS: Any other questions?
DR. JACKSON: A followup to the administrative data sets. Were you referring to the particular sets that could be compared in certain systems? Often these sets are beleaguered because they don't have enough information, often they are the set of last resort. But you have that as part of your recommendations. How do you see using that? In your recommendation, it wasn't just using it, but actually bridging it to other sets. I'd like to explore your use of that.
DR. ONG: I think we need to improve the use of administrative data for several things. One of the things that the Bureau of the Census started exploring is to supplement their efforts to administrative data, because the costs of collecting data through Census surveys is getting so expensive. So they want to figure ways to augment that in a cost effective way. One is using administrative data.
We also do policy evaluation, for example, welfare to work, as heavily anchored in using administrative data. So for example, we could take information on welfare recipients before 1996 or in California, 1998 when the welfare reform was implemented here in California, to look at who left welfare, how many didn't file for work and so forth.
To the degree that race makes a difference in outcomes, race and ethnicity makes a difference in outcome, one ought to have good data about that. One ought to know about the relative incidence or prevalence of welfare usage as well as decline in welfare usage by different groups.
Then also, administrative data quite often is used for population projections. So in the California Department of Finance, they use a number of data to try to fill in between the decades of the use of the Department of Motor Vehicle license information, but they also use school enrollment. To the degree that these things all fit in terms of evaluation, population estimates, monitoring of outcomes, if you want to break it down by different groups, you want to have some sort of consistency, so you don't have any false outcomes or false trends. That is what I mean.
Now, the quality of these things varies tremendously, because they are collected differently. In the past, when you go in and collect unemployment insurance, we actually collect information on your race and ethnicity.
Now if you collect unemployment insurance, you may not recall filling that in. At one time, the way it was done is that I am the intake person, you come in.
There are also issues about school race information and how it is collected and the quality of that, and we ought to know that. I know analyzing some of the data that Nolan has access to, I was actually surprised that when we tracked Native Hawaiians from one grade to another grade, that there is some shifting. I was surprised by that. I don't know enough detail about how it is collected, but if you looked at how race and ethnicity was reported in the third grade and you compare it to the eighth grade, there are some changes going on.
I could say it occurs. I don't know why it occurs, and that is part of the assessment that we ought to do to figure it out. But if we are thinking seriously -- and we ought to think seriously about using administrative for a number of public policy purposes, we need to know much more about these administrative data sets.
DR. MAYS: Yes?
MR. POUIESI: I am very grateful to have heard your presentation and very delighted about this opportunity. I can't state it more than that.
But I do have a question, Paul. I heard from Dennis the idea of disaggregating the Pacific Islander information. The charge of this committee was to look at the collection of data to see what we have in reference to disparities in our communities.
If we are talking about disaggregating the Pacific Islander group from there, and you are talking about bridging where there is a continuation or stability of groupings, if I heard it correctly, and then I am seeing API, I now think of Nolan's presentation, of the OMB Directive 15.
There is no such critter as API. We now have NAOPI, which is Native American and other Pacific Islander. I am stating to this committee here that you already have a system that is supposed to be implemented as of 2003, January, and that is the new race classification called Native American and other Pacific Islander. It has moved the Pacific Islanders out of the Asian component.
What we are looking at is direct, across the board federal departments to be able to track our disaggregated -- what we have been stating; your presenters have stated it. All I'm stating here at this point in time, please, make that a reality. It should have been here in January. If you institute that particular race category under Native American-Pacific Islander, you will be able to track Pacific Islanders. I'm only talking about Pacific Islanders per se.
You may wonder why I call myself a Native American Samoan. The Department of Health and Human Services made the definition that Native Americans are those who are indigenous to the continent and have treaties with the United States. American Samoa has two treaties, one in 1900 with the paramount chiefs at Tutuela, and one in 1904 with the Tuee or kingly individual in Manua Islands in 1904.
So in closing, I appreciate all of the information, and I would like to reiterate again, OMB Directive 15 is there. If you are looking at categories, cataloguing what is happening in data reference our population, institute OMB Directive 15.
Thank you.
DR. MAYS: Thank you. Do we have any other comments from the audience? Anyone else want to comment?
MR. NISHIMOTO: Sero Nishimoto, a former REAC committee member. I just want to let Paul know that because of his and our efforts at trying to get the Census Bureau to do Asian surname lists, it is finally really going forward. So if your recommendation is for them to have better surname lists, I think the Census Bureau is going to be an excellent source for that, especially because in this last census, they were able to link surnames with the race data.
So I just wanted to add that.
DR. ONG: It is good to hear that. Actually, you should know that when the Census first say they will do something, it is well worth monitoring to make sure there is followup.
DR. MAYS: Nancy, you wanted to comment?
DR. BREEN: Yes. This question comes out of an empirical problem that I have had, and it is a followup to your question. Now that we no longer include Native Hawaiians and other Pacific Islanders in with Asians, what has happened with the data is, now we can look at Asians separately, and often those estimates aren't very robust, but Native Hawaiians and other Pacific Islanders get mixed in with the others.
So the question was raised -- I had been thinking about this independently, and you brought it up, would it make sense to put for analysis purposes, and this isn't data collection purposes, but for analysis purposes, to put Native Hawaiian and other Pacific Islanders with indigenous people and have a larger category which would be American Indian, Alaska Native, Native Hawaiians and other Pacific Islanders?
Does that make more sense as a grouping, so that at least we have a category that is more meaningful than Other?
DR. ONG: Let me just preface that by saying one thing. I fully acknowledge that for OMB and for analytical purposes, we separate Asians from Native Hawaiians and other Pacific Islanders. But whether you group people together or not is really in context. So if one does a study on minorities, there is no racial group called minorities, but we use the term minorities. Rooted in the history of Asian-American studies, we have included Pacific Islanders. So in a sense, there is still a field of study where API has meaning, but in that field of study we need to be sensitive and understand that we need to break it apart.
So the use of the API is not meant to say that they exist as a group; it means to encompass the group of populations that we are concerned about analytically. That being said, are there better ways of reconfiguring it? Because actually when we look at more details about OMB 15, there is always that caveat, that is, they would not release data that is not meaningful. So it is not an absolute mandate, even though they create these categories, that everything will necessarily be reported in these categories.
That being said, what do you do with groups that fall below the threshold for different sorts of statistical reporting? Do you group them into other populations? The answer there is, there is a tradeoff. I'm sure the American Indian population would see this as a way of diluting -- if you were reporting singly for them, it may dilute reporting for them. I do know that that was one of the sensitive questions that occurred in the REAC as we debated about separate Native Hawaiian and other Pacific Islander grouping. So there is the issue about that.
When you are part of the other group, regardless of whether it is within the Asian population or within grouping Native Hawaiian and other Pacific Islanders with American Indians and Alaska Natives, you are still part of the Other, and it is not comfortable.
The state of California for example, even after Asian-Americans became much larger, their largest group, the Department of Finance continued putting them in with others, not because Asians were small groups, but they couldn't figure out what to do with the residual others, so it had to be added to some population. So it always got attached on to Asian-Americans.
The truth of the matter is that in some ways there are benefits. You at least get something, particularly if populations are similar and face similar problems. So you group them if they have similar problems, similar socioeconomic status and so forth. You get something under the current threshold, but there is always this bad taste. I understand that it is always residual, that you are buried somewhere. So it is a compromise.
If there is a compromise, it offers one particular compromise. But I would argue, and I suspect other people would argue, that you need to do a better job for these populations, and not bury them. That is the position that I come from, because Asians were buried. There are similarities in the history. Even in California Asians were buried in the statistics of others for a long time. That is recent history, not decades ago.
DR. MAYS: I want to thank you for your presentations. It is incredible. You all have worried a little bit about overlap, and it is the like, the overlap has helped for us to get better and better understanding. So I really appreciate your also talking about some of the conceptual issues that I think will help us tremendously as we try and think through these problems.
As a committee who is supposed to be concerned about health, I should be concerned about your bladders. So I am going to take a short break. Let's try and keep it, because we are over time about eight minutes, so let people get out of here. The staff in the back will tell you where the restrooms are around here, and help yourself to water, stretch your legs, and then we'll be back in a very short time.
(Brief recess.)
DR. MAYS: We'll start, in case some of you came in late, by welcoming you here, welcoming you on behalf of UCLA, since this is my home, so we're going to do that. Also, welcoming you on behalf of the committee.
We are very happy that you are here, that you have taken time to be with us. But more importantly, we are happy that you are sharing with us. That is part of why the committee will go to different places, to not just fire a bunch of questions, but to hear from everyone, both at the table as well as in the audience. So I appreciate the way everybody is participating. I'm very happy and I really appreciate it, because it really helps us to do our work better. We really are trying as hard as we can to get this right, to make a set of recommendations that will make a difference, so it is in that spirit that we welcome all of your comments.
We are going to turn to our last set of presenters for this morning, who also are going to continue in much the same vein that we have already been talking, as they too are both aware of the census data, as well as what happens when we start translating that into the community.
I almost feel like we are saying we are turning the mantle over here in terms of the younger generation kind of thing, because Bong was actually a student on a committee I was at, so he reminded me about that. I'm thrilled to see how well he is doing.
What we are going to do is, Melany is going to start. To my surprise, Melany was here, and I have met her via this process, which shows you how big this campus is. One of the things that we talked about, and I think it was Paul that actually commented on it, is that we have here at UCLA one of the community development data centers. You can see that one of the benefits is that that center has the mission of actually working with community groups. So we are very happy for Melany to be here, to talk with us about the kind of work they do, and for those of you who are actually part of the community here to also get a sense of what this resource is like, and the kind of interaction that you might have with this group.
Bong is here for Special Services for Groups. Special Services for Groups is also a community group that has been around for many years and has been providing information, technical service, looking at economic needs, more on the development side. What you should realize is that we have people who talk health and people also who talk context, so we think it is important to have a group such as Special Services for Groups here.
So without further ado, I welcome you both, and it is a real pleasure to have you here. So thank you very much.
MS. DELA CRUZ: Good morning, everybody. We want to get started by just giving you a little bit of background on the census information centers. We are two of three in Los Angeles. The Asian Pacific American Community Data Center that I run, which I want to bring your attention to with this yellow pamphlet right here, that gives more of an overview of the services and the programs that we provide.
We have partnered up -- we were designated by the Census Bureau in 2000 the UCLA Asian-American Study Center, and we have recently partnered up with the National Coalition of Asian Pacific American Community Developments. This blue sheet provides some fact sheets about our community and our community development issues.
Currently there are 52 programs nationwide. We serve as repositories of census data reports and products, free of charge. They provide us with this data, and we try to disseminate it to the communities that we serve.
Do you want to say something about the Special Services for Groups?
MR. VERGARA: Special Services for Groups is a private nonprofit. We provide the full continuum of social services to the diverse groups of greatest need in L.A. County. As you can read on this page that I prepared for you, SSG is a network of about 25 different social service providers and about five consortia that advocate for health access and health policy for several disenfranchised communities. By disenfranchised, what I mean by that term is, we are not only referring to racial and ethnic minorities, we also provide services to persons living with HIV and AIDS, persons with disabilities, mental health and other groups like that.
Our CIC, the census data and geographic information services, was designated by the Bureau of the Census as a CIC in 1991, back when there were still ten CICs nationwide. There are now -- as Melany will tell you, there are now 52 CICs nationwide.
MS. DELA CRUZ: What we want to do is provide you with some of the examples that we have -- the trainings that we have produced for our member community organizations and also show you how the census data is important to the work that they do.
DR. MAYS: I'm going to ask you to pull the mike a little closer, or ask him to turn it up a little bit.
MS. DELA CRUZ: How's that?
DR. MAYS: That's better, thank you.
MS. DELA CRUZ: We are going to first start with Bong.
MR. VERGARA: What you see on the screen is one of two presentations given at two community trainings that we have had in our partnership in the last year.
At the risk of repeating what she just said, the three CICs in L.A. County, Green American Coalition, SSG and Asian-American Study Center, we partnered up with them, and the legal center; they are not a CIC, but they are part of the partnership. What we did is, we put together two community trainings to help local community-based organizations and service providers and community workers understand how they could use Census 2000 data in light of all the changes that happened when it was collected in their community work.
So this is one of the -- we only copied the first three slides to show you, to give you a flavor of what the presentations were. But this particular one had to do with how to use census data and designing effectiveness based and evidence based social service programs, and how to use census data to demonstrate program effectiveness.
Much of the discussion centered around how they can use census data and client data that they themselves collect to pursue those two goals, which is to demonstrate program effectiveness.
The second presentation is helping local CBOs and other nonprofits use census data to be able to discuss health disparities in whatever target community they are designing their programs and conducting their services. That is actually -- this one is particularly targeted on the Asian alone population in L.A., Riverside and Orange County. That is actually what I want to talk about.
If you want to follow me, this is the one thing I am going to be reading from. Special Services for Groups, SSG, is a private nonprofit agency established in 1952, aimed at providing culturally relevant community-based solutions to the social and economic issues of groups in greatest need. SSG is a network of 25 human development programs and five social service consortia aimed at meeting the diverse needs of multiple disenfranchised groups in L.A. County.
In its 50 years of service, SSG has garnered a distinguished status as a significant community stakeholder and health advocate on the relevant health and community issues faced by the African-American, Latino and APIA communities in L.A. County.
The key to SSG's success is closely linked with its bias towards a participatory approach to service provision, that locates the participation of indigenous stakeholders and community leaders to the center, and effectiveness based and evidence based program planning and evaluation.
SSG puts a high premium on data, especially data on minority populations, that enables it to meet the relevant health and community needs of its various target populations. It is for this reason that SSG is working in close partnership with the U.S. Bureau of the Census as a census information center, through its census data and geographic information services, or CDGIS.
Since 1991, CDGIS has provided data access to local minority community-based organizations and service providers, local government units, and the general public. CDGIS routinely conducts data analysis and data display in service of program planning and design for various social service providers and health policy advocates in L.A. County, and increasingly across the state and the country. So we have been working with Asian Pacific American Health Forum and the Arab American Institute in analyzing their target populations, and helping them discuss the issues they want to discuss using data.
CDGIS provides population tables and maps on racial and ethnic communities and other technical assistance services. See flyer, which I am referring to our yellow flyer that has our local on there. In partnership with other local CICs, CDGIS has been providing census data trainings in the past year to help the general public, particularly the APIA community, make use of Census 2000 data. The ability to use race and ethnic data from the decennial census is critical in social work, especially minority social work, for which we continually find a dearth in research with regard to the health disparities being faced by minority populations, especially a lot of the newer Asian and Pacific Islander communities, like Tongan Samoans and Southeast Asians.
The ability to use race and ethnic data from the decennial census is critical in the design of social service programs that are responsive to the demographic changes in all communities, particular minority communities such as the APIA community. Census data that form the basis for discussing poverty, language isolation, overcrowding, unemployment, receipt of supplementary income and race and ethnicity, to name a few, allows social service providers and health advocates to assess areas of health disparity.
The role and impact of race and ethnic data in the provision of health services for minority populations, particularly the APIA community, cannot be overemphasized. SSG and its partners rely on census data for racial and ethnic populations in providing quality service and community advocacy for the communities they serve.
DR. MAYS: Do you just use the census data, or could you also -- for example if there were other data sets, like we were just talking over here, like some of the NCHS data sets or something? Are there others that you could with resources do the same thing?
MR. VERGARA: Right. What we do is, we use census data as the basis first for the way we talk about population, any target population that we are interested in. We relate census data, either at the block level, tract level or place level, with relevant databases like GIS, if the geographic levels of the data are available in that database, as relatable to census data. We use the data that are programs can collect themselves. We also use the local DHS data, United Way data, and some data put together by research units like the Asian-American Study Center.
MS. DELA CRUZ: One of the most predominant requests that I get from a lot of the member organizations is how I use the census data in program planning and especially grant writing. So one of the trainings I developed was looking at how we could use the census website for gathering the data, the race ethnic, economic, housing data that would be useful for their grants.
I want to just highlight some of the tools. The American Fact Finders site has thematic and reference maps. Back in the day you were using these maps, that you pulled from Thomas Guide or something, or the AAA book. Now, if you don't have the sophistication of creating geographic information systems maps, which our data provides, you can look up these maps on the American Fact Finder site.
In this example, you are able to find the percent of persons below the poverty level. We try to locate them in Alhambra, Rosemeade and El Monte. You can see on the left, there is a legend, and you can click on the legend, and you can also graph schools in the area. I guess I didn't capture that part, but you can basically do some mapping that can be significant and act as a visual representation for your grants. In addition, you can just pull data together at the statewide level, the national level, county and city, and just have it as a comparison.
So that was just some very quick examples, but we have been able to use this data to support work in creating home ownership opportunities, to create better literacy programs for a lot of our immigrant populations, also provide small businesses with data that will help them locate where would be a good neighborhood for them to provide services and so forth. Also for community centers. Going back to this map, this tells us where are the concentrations of Filipinos in L.A. County; maybe we should develop some community centers in those areas.
So having this race and ethnic data has been useful in helping our community development organizations serve their populations better.
So basically in conclusion, this plays a significant role in both the social work and community development communities, and it allows for this evidence based analysis necessary in providing effective health and community development programs and public policy or advocacy.
MR. VERGARA: If I could add something else, the other thing that we have been finding out, and what we have been encouraging at SSG is the use of data to raise the capacity of small to medium sized minority CBOs and service providers, to raise their capacity to do effectiveness based program design and evaluation.
Oftentimes, those tasks that they need to do are viewed as something very distinct to them. Frankly, no one has shown them that it is so easy to access census data. You go to American Fact Finder, and you can even map the population you are interested in down to the census tract level.
Having census data that collects a diverse list of different items that CBOs and nonprofits can use in being able to talk about their target populations and their issues intelligently is such a great thing to have. A lot of the small nonprofits that we have been working with are now much more aggressive in going after funding that before the trainings, they stayed away from, because they thought they can't possibly put together an analysis that is compelling enough, using data.
That is probably one of the most fulfilling outcomes that we could immediately see from helping out these local CBOs look at and use census data, the level of confidence that they have in analyzing health issues, discussing health disparities in their specific target population.
So without raising ethnic data, I think these local nonprofits would have even more difficulty talking about their own communities, because it would be so hard to talk about their issues.
DR. MAYS: I have questions, I don't know about my colleagues. One of the things that I would be interested in, because this is very exciting, to think about how do we do this, not just with the census, but how do we do this with some of these other data sets. Some have specialty information. There are community service providers who specialize in diabetes or specialize in particular things. So I'm sitting here trying to figure out as a committee what kind of recommendations can we make to get more data out to whoever wants it.
So I want to understand what it is you all do in the capacity that you have. My questions are the following. Could you ramp up to do more of what you do? Could you do it at a level where you actually are the group that works directly with the data? I haven't used American Fact Finders, so that makes it very simple. But a lot of these other data sets, I don't know if we could get it to be that way. Ask CHIS, for example, has developed a software that makes it very easy for you to get some cross tabs. I guess we could explore whether they could do that, but if they couldn't, is there capacity in groups such as yours to be able to do that, or is that something where we think more about, we do that here, and then we send it to another level and you do the training or something like that.
Help me to think through how we could have more data available, and to have you all do what you do.
MR. VERGARA: Sure. Let me try to answer that in two ways. What has been happening in the Asian-American community in L.A. County is that there is actually a pretty established history of API CBOs and service providers coming together yearly through the Asian Pacific American Research Roundtable, where they bring researchers, faculty, students and community workers and service providers together to discuss the prominent health disparity issue or the prominent community issue for a given time span.
So that is happening, and that happens yearly. The other way that I want to answer the question of capacity is, the partnership that we have started is really aimed at helping local nonprofits use census data, not only out of our own realization that that is a need, but also because it is an obligation that we have when we agreed to be a CIC with the Census Bureau.
DR. MAYS: So you are funded by the Census Bureau?
MR. VERGARA: No, ma'am, we're not funded. What we do get from the Census Bureau is, we get first access to the data as they release them. So as the different files are released, we get access to them and with that, we are able to look at some data if we so desire.
The other thing we get from being a CIC is, we get training, we get annual training from the Census Bureau. In fact, the most recent one was last week in Vegas, where they show us all the different interfaces that are working, that provide access to data, so that we can use them in not only working with census data, but relating census data with other databases, like Data Ferret, POMs, what is the other one?
MS. DELA CRUZ: Events Query.
MR. VERGARA: Events Query, right.
DR. MAYS: I'm sorry, what was the other one?
MS. DELA CRUZ: The Census Bureau is developing several ways of displaying data for its census information center users. So there is an advanced query system that we currently have access to that allows us to cross tabulate data, up to three data variables, that you can't do with American Fact Finder. So it allows us to do much more with finding out information on our racial and ethnic groups.
It is a very sophisticated system, but the problem is, as Paul has mentioned, it has to pass a population threshold, and because of confidentiality filters, you usually cannot find data on Pacific Islander groups or smaller populations.
MR. VERGARA: The other issue with census data is the extent to which we can use census data to talk about health issues and health disparities is limited to the extent that data items can be related together to allow us to talk about health data.
For example, to be able to talk about health disparity, what we have been doing is looking at the data on whether a person is able to speak English well or not at all, relating that with poverty data, relating that with unemployment data, to be able to talk about the likelihood that immigrant populations are facing health disparities, that they are facing gaps when it comes to health access. If they can't speak English well or not at all, they cannot access the health services that exist in their communities. If they cannot provide to pay for services, they cannot access services, similarly.
So in an ideal world, I guess what would be ideal is if we could think about or maybe even revise some data items to more directly collect health data from the Census. That may be a pipe dream, but since we are thinking about ways to use census data in talking about health data, that would be an ideal situation, where you don't have to use proxies and relate them together to talk about health access or health disparity, because the data item would directly be referring to that issue.
DR. BREEN: Are you suggesting that some health questions might be added to the census long form? I'm not sure what that suggestion was, something like self reporter health status, or what were you thinking?
MR. VERGARA: Yes, I am, basically.
DR. MAYS: His wish list.
MS. DELA CRUZ: As far as the capacity for this L.A. Asian-American census information center, we only have one fulltime staff, which is myself. So I think a lot of assembling the data would work well if that was then on your end, and then more of the dissemination portion would fall on our part, as you were mentioning earlier, us partnering up by doing the trainings and making the trainings available to the community and to our networks, so that that way, the data can reach beyond the university or the people who know about it at the moment.
MR. VERGARA: Another thing, short of including a data item that asks health status on the long form or the short form, perhaps what might be more feasible is if there is a way that DHS could partner up with the Census Bureau and its annual CIC training, to look at ways in which those two departments can help train CICs in being able to talk about health disparities and health issues using data and census data. I think that would be a very productive partnership.
DR. BREEN: Can I ask another question?
DR. MAYS: Sure.
DR. BREEN: I see on this that you also use the current population survey, and the current population survey of course, besides being our main source of data on unemployment in this country, because it is collected frequently and we don't rely on the decennial census for that, of course, also has information on health insurance in the March supplement.
I was wondering, is that adequate for small populations? Are you able to get health insurance estimates from the CPS?
MS. DELA CRUZ: Not really, because it is limited. As far as finding that information at the local level, usually you can find statistics at the state and national.
Also, I think a lot of the supplemental information from the CPS reports just reports it by -- doesn't disaggregate the data, so you don't get the breakouts of Asians and Pacific Islanders. So I haven't really used it much to offer that data to our community-based organizations. Mostly what we have been offering is trainings on how to use Census 2000.
DR. BREEN: Thank you.
DR. MAYS: Any other questions up here? Let me turn tot he audience, in terms of questions. Any questions?
Could I ask you to do one thing, which is to share with us these? We can see them, but I want to make sure that they are also shared with them, so that they understand how this came to be.
MS. DELA CRUZ: (Comments off mike.)
DR. MAYS: If you grab a mike, because he is trying to record this.
MS. DELA CRUZ: There are actually maps and charts that were produced in the New Face book that has just been published by the UCLA Asian-American Study Center, and also Asian Week magazine in partnership with National Capacity and the Organization of Chinese-Americans.
Basically, it just shows the importance of understanding our communities. We have been able to map over time the growth of the populations. This map right here shows the immigration patterns of the Asian community. We have another one on income, that deals with the poverty rates, median household income, and then also by individual income. The last map right here looks at employment and national attainment of Native Hawaiians and Pacific Islanders.
So definitely, feel free to come up -- or I don't know where they will be later, but please take a look at them.
DR. BREEN: I have one more question.
DR. MAYS: Go ahead.
DR. BREEN: We have been talking about how to group data. How do you group your data, or does it depend on the organization and what they want? Have you figured out any groupings that seem to work better or maybe provide better information, better guides for scientific research?
MS. DELA CRUZ: For our member organizations, we have separated Asians from Native Hawaiians and Pacific Islanders. We have also offered a minimum and a maximum. So they are showing the race alone data as well as the race alone or in combination, so that they can decide whether they want to show the minimum or the maximum of that population.
But we caution them and tell them that the maximum does include only responses and not persons per se, respondents.
MR. VERGARA: Are you referring to how to group Asians and Native Hawaiians, if we group them together or if we separate them?
DR. BREEN: I was just wondering how you did tend to group the data, because that is one of the issues that we have been talking about here, because we have been looking at Asians and Pacific Islanders together, when in fact we have a new way that federally we are mandated to look at them. I think we are also trying to explore what may be ideally the best way and how we would want to improve data collection, but also what is feasible, given the current data.
MR. VERGARA: When we talk about those two communities without any reference to census data, we talk about them as one community. This, in light of the fact that the term came out of some need to coalesce these two communities together back in the '60s. We want to stay true to that history.
We separate the communities when we refer to Census 2000 data, because that is in fact the only accurate way to refer to the data, since they are separated from each other.
DR. MAYS: Do we have any other questions? This is very exciting. We are going to try to figure out something here. I can tell this is going to be food for our discussion. I think it is a good model for getting data out. Again, as Dr. Ong said, one of the things is congratulating the Census where they have done well, and this is like a good example of ways in which they have seeded lots of places to get their information out. I think we just need to think about this relative to some other things. So this has been very helpful to us.
I thank both of you for taking your time and presenting to us. One of the things that I will ask is, we are about to break, but for those of you who are community groups in the audience, if you have not talked with them, please do, because this is your source for actually getting some of the information that they might have. To some extent, the more that it is to their ears as they go back to the Census Bureau and try and see if they can get it even down into finer groups, if there is any way to get around some of the confidentiality concerns.
So we thank you both very much. We really appreciate your spending your time with us today, because it has been very helpful in terms of giving us food for thought about what we might make recommendations on. So thank you for a great job. Thank you.
One of the things that we are going to do, even though we are behind time, we are going to take a lunch break. We have to think a little bit about health. So we are going to reconvene in exactly an hour, so why don't we say that we will be back here at 1:15? I'll give you a little less than an hour, but we will be back at 1:15. So we hope all of you will join us for the afternoon.
Thank you very much. This has been just a phenomenal set of presentations. I think all of our presenters are here, so I'd like to actually give them a hand and thank them for all of their work.
(The meeting recessed for lunch at 12:20 p.m., to reconvene at 1:30 p.m.)
A F T E R N O O N S E S S I O N (1:30 p.m.)
DR. MAYS: Good afternoon. Thank you, everyone, for being with us this afternoon. One of the things that I'd like to do is ask someone to help me. We have a couple of new people here. If you would introduce the people who have joined us, we appreciate having you here and we just want to make sure you let us know who you are, and then we will begin.
DR. CHANG: John Chang, UCLA Department of Medicine.
DR. MAYS: Welcome.
DR. HIJASHI: Taka Hijashi, the same, UCLA Department of Medicine.
DR. MAYS: Welcome.
DR. YOUNG: Ryan Young, Asian Pacific American Families Against Substance Abuse.
DR. MAYS: Welcome. We are back for our afternoon session. I appreciate all of those of you who have stayed with us. Our apologies for getting back a little late. We were running a little bit over, so my apologies particularly to Dr. Keh-Ming Lin forgetting back here late.
I am particularly pleased that he has agreed to be with us. One of the things that we don't do sometimes is deal with the mental health side as well as deal very specifically with what happens when we see racial ethnic differences in the actual physiologic processes. Here he is going to talk about the psychobiology of ethnicity by focusing on pharmacologic differences that we see in Asian populations.
Dr. Lin is a treasured colleague here. He is at UC Harper General. He directs one of the NIMH centers on minority mental health, and his center focuses on the psychobiology of ethnicity. So welcome. We got to him late, so I really appreciate you doing this. He has also lectured in my class, so I already know what we are in for. Thank you very much.
DR. LIN: Thanks for the invitation. I got here a little bit earlier and have been staring at these beautiful flowers and enjoying that.
DR. MAYS: Peaceful.
DR. LIN: Yes. One thought came to mind. I thought maybe we could change the title to the uniqueness and variation of yellow tulips. Perhaps that could be for a different presentation.
Let me move ahead with my slides. As you can see, this is slightly modified from an earlier presentation. I didn't get around to changing this first slide. You probably covered this already, but I thought I would throw in these slides about demographic changes.
By 2050 there will be no majority in this country. It is projected that more than ten percent of the nation's population will be Asians, broadly defined. In Los Angeles County as well as in California, this is already the case. For both the state and the county, 12, 13 or 15 percent of the population are of Asian origin. If you look at the world, Asians represent more than half of the world's population, so it is not actually a minority group; it is a majority group.
The next major point I want to make is, it may sound ridiculous, but it dawned on me relatively late that in anything we do, one size does not fit all. There is great variation, both individual and cross-ethnic variation in terms of everything we do. What we are doing here is focusing on treatment response.
This has a personal origin. In 1974, I finished part of my residency in Taiwan and started my residency in Seattle. In a span of two or three weeks, in terms of one of the drugs that was just recently available for psychiatry in Taiwan, the dosage of that medicine that seemed to be effective jumped in Taiwan from less than one milligram to something like 100 milligrams. That really surprised me and got me started thinking about what may be the reason, whether the American physicians were being too aggressive, or the Taiwanese physicians were too timid, or whether there was biological differences. After 30 years, it is pretty clear that all three were true, all three were there.
We have since come down in the American side in terms of the dosage of medication. There often is still a lot of over medication, but the average dose has come down to a mild degree. In Taiwan unfortunately they learned from the American textbooks, and they became much more aggressive in treatment as well, but because of that experience, when we had a chance later to look at dosage differences, we indeed documented that the differences were there, even for Asian and Caucasian patients treated in the same hospital setting. So this provided some further confirmation that there may be objective differences in how people respond to treatment.
So we then recruited three groups of normal volunteers, foreign born Asians, American born Asians and Caucasians, and give them a small dose or test dose of Haloperidol and measured the blood level. The summary of the results, part of it, is here. The two Asian groups had a much higher blood level adjusted for body weight compared to the Caucasian group. This is both true when you give the medicine intramuscularly as well as orally. The difference especially on the lower panels is quite dramatic.
But this is also seen when you graph these blood levels individually. Each bar represents an individual subject, and you see that the majority of Asians tend to have higher blood level. The majority of Caucasians have much lower level, and the difference without statistical analysis, you can be pretty sure is highly significant.
But I think this graph also serves to show two other very important points. Irrespective of ethnicity, each group has very substantial individual variations.
The third point is of course, you should not interpret any discussion or any ideas about ethnic differences stereotypically, because there are always exceptions and there are always overlaps, and you really have to treat everybody on an individual basis, while recognizing that in aggregate there are substantial ethnic cultural differences in almost everything that we do.
This is just another example of a subsequent study. In the last 30 years, there have been hundreds of studies that document often very substantial ethnic variations.
What we did earlier was the Haloperidol level, which has to do with how drugs are metabolized. This is a graph which shows neuroendocrine response to Haloperidol, which also shows that Asians had much higher (word lost) compared to Caucasians. This remains statistically significantly different after you control for Haloperidol level.
So it means that after you take out the difference of the drug metabolism, the brain's response to the medication is still there. There are two levels of differences that we could look at, and these are the most important factors that we need to examine when you look at drug responses. One is how the drug is metabolized, the other one is how the drug has the effects on the target organ.
Another example which has to do with mostly the brain's response to medication is lithium. The therapeutic level for lithium is about two thirds in Asians compare to Caucasians. The exact reason for this we still don't know, but that seems to be a pretty robust finding. It was reported independently in at least four different Asian countries.
How about other kinds of medicine? Here, this is even more clear. This is propranolol, which is sometimes used in the psychiatrist's world, but the primary use of propranolol, Inderol, of course is for the treatment of hypertension. The graph is a summary of a very elegant study which shows that the blood level of propranolol that is required for reducing heart rate and reducing blood pressure to a similar degree.
In this case, for example, on the right panel, to achieve a reduction of ten points in blood pressure, you only need about ten percent of the propranolol concentration for Asians as compared to Caucasians. So this difference is quite substantial, and this actually has been a reason for Asian visitors coming to this country running out of medication and going to the doctor's office, and saying that I only take the lowest dosage, and the lowest unit dose in this country is three or four times higher than the unit dose in their country of origin, and they could go into hypotensive shock, because of that.
I mentioned earlier already that there are two processes that determine how you respond to medication, the pharmacokinetics and pharmacodynamics. To put it in a simple way, pharmacokinetics means what does the body do to the medicine, and pharmacodynamics means what does the medicine do to the target organ, to the body. In this case, in our field, it is the brain, but this applies to other fields of medicine as well.
What I wanted to show here is that this is also absolutely clear, that both processes are controlled by genetics. Also, the way your genetic predisposition, your genes, are expressed. It is very much influenced by environment as well. So this is actually the major challenge in the whole field, which is figuring out how the environment and the genetics interact, and how do they determine how likely you are to suffer from one condition versus another, as well s how you are going to respond to medications or treatment.
These factors are very much determined by who your ancestors are, where they came from, which is in a broad sense ethnicity. The environmental factors are very much tied to your culture of origin, cultural background, and your practices and your beliefs. So in doing this, I think it is important that we do not neglect the influence of ethnicity and culture, because they are tied together with your environment and your genetic background, which then determine everything.
To focus more on the ethnic variations in terms of genetic variance or genetic endowment that vary across ethnic groups, perhaps the best example for this is the way people respond to alcohol. Most of you if you are living in Los Angeles or California, you probably are familiar with the phenomenon called flushing response, which happens in 40 to 50 percent of Asians. This has als o been something that has been figured out in the last 20, 30 years. The reason that Asians often respond to a very small amount of alcohol with facial flushing, palpitation, dizziness and sedation and vomiting. It could be very unpleasant. It is because they do not have an enzyme that metabolizes acid aldehyde. Very often, about 90 percent of these Asians furthermore has another mutation that is especially efficient in turning alcohol into acid aldehyde, so they have this double jeopardy in producing acid aldehyde very fast and not being able to get rid of acid aldehyde. This exists in about 40 percent of Asians, corresponding to the prevalence of flushing response. The prevalence of this is very low in other groups, with the exception of some American Indian groups.
So the genetics of that is well characterized. We can use a simple test and determine whether you are homozygotic or heterozygotic for this mutation.
But how about drug metabolism? There are of course many factors that determine how efficient you are in terms of handling medications or getting rid of medications which are regarded by our bodies as foreign substances, potentially harmful, potentially dangerous, potentially toxic. So we are very good at getting rid of any drugs. For most drugs, 99.9 percent of the drugs do not reach the bloodstream; they disappear between the intestine and the liver. Only very small amount reach different parts of the body.
In psychiatry, what we are interested in is what reaches the brain. The major step appears to be a group of enzymes called cytochrome p450 enzymes. Again, this is achievement of the field in the last 20 years. We have pretty much characterized all these important -- the most important of these enzymes. Of these, the most important ones are listed here, the SIP-2D6, SIP-3A4 especially, as well as 1A2, 2C19 and 2E1.
We are talking about genetics and environment. Here, 2D6 is the best example of genetic variation that determines how much enzyme you have. 3A4 is especially responsive to environmental influences, so we will focus on these two enzymes.
These are just summaries, slides, that show that practically all the drugs that we use in psychiatry and also in many fields of medicine, they are either metabolized by one or the other of these enzymes or combination of some of these enzymes. This is true with antidepressants as well, so there is a reason to pay attention to the cytochrome p450 enzymes which determines how fast you are able to get rid of these drugs, and whether you need a higher dose or lower dose or whether you have side effects.
How about 2D6? The reason this has been the focus of a lot of attention has to do with not only that it metabolizes about 40 percent of the drugs that we prescribe, but because it is highly variable, and it is bimodally distributed, especially in the Caucasian group. About seven percent of Caucasians actually do not have the enzyme, but the rest are efficient in terms of these enzyme activities.
But when you look at this in different groups, in this case I think we will focus on the Asian group, the Asian group, the distribution of enzyme activities is on the upper panel. You see that Asians have very few who do not have this enzyme at all, but the majority of them are slow in terms of the enzyme activity compared to the Caucasian group.
Now we know exactly why Asians tend to be slow; 17 percent of Asians have a single base pair mutation out of these 10,000 or so base pairs. It just takes one single mutation which slows down the enzyme activity. It is called star ten. For the seven percent of Caucasians who don't have the enzyme, the reason was because of the star four mutation. These are all single tests that we can just run within one day turnaround. We know what kind of genotype you have and what kind of enzyme activity you have. This in turn could be used to predict what kind of dosage would be good for you in terms of drugs that are metabolized by 2D6.
The interesting thing about 2D6 is that in other ethnic groups, there are mutations that result in duplication or multiplication of the enzymes. So when that happens, you have many times more enzyme activities compared to those who have two copies. So they often will be not responding to the regular dose of medication that physicians typically prescribe.
Incidentally, you look at the graph, you see that from one to another, the difference is about 10,000 fold. So this is one reason why people's response to medication, even in the same ethnic group, could be quite substantial.
This may be a bit out of order, but I am just using this to show that there are many factors that determine treatment response. This is called genetics, and many factors modify the expression of the genes. On top of that, culture influences many factors which then determines whether you have greater response or not.
The most basic of it is, if you don't take the medicine, you are not going to respond to treatment at all. This may be a best-kept secret, but if you just look at anybody who needs to continue to take medication for more than ten days, it is guaranteed that 50 percent of them are noncompliant. How do you maximize compliance or adherence is a major issue which may be equally important or even more important than all the genetic or biological factors that we are talking about.
Now I remember why I am showing this. This is to show transition to saying that even though genetic variations are important, environmental factors are also very important. That also could determine ethnic variations in treatment response.
In this example, the blood level of a drug was measured in British whites and Asian Indians. Again, this is similar to the Haloperidol slide I showed earlier, which shows that there are both individual variation within group as well as differences between the two groups. The British whites in general are much faster metabolizing Nifedipine, so that they have no level, as compared to Asian Indians, but there are also overlaps.
Since 3A4 activity is not determined by genetic factors, the assumption of these ethnic variations was that it has something to do with dietary differences between the two groups. This is also true with another drug. I will not belabor the point. There is another drug called Comipramine, also studied between South Asian Indians and British whites.
The interesting point about Comipramine is that they did a subsequent study looking at the Asian Indians who have lived in London for awhile, and started changing their doctor practices. They abandoned their vegetarian diet and started eating beef and high protein diet. When they do that, the metabolic profile changed. They became identical to the British whites. So it shows that the difference was due to dietary practices.
Here you usually ask a question, what is this? Does anybody know what this is? This is St. John's wort. We don't usually see it in this forum, but St. John's wort is interesting. It may yet be proven to be an effective antidepressant, even though there were several failed trials; there were also trials that showed efficacy. But even more important than the question of whether this is an antidepressant or not is that it is clear that it has a major impact on enzyme activity.
If you take St. John's wort along with one of the HIV medications, the St. John's wort inhibits the enzyme that is responsible for metabolizing the HIV medication, the blood level drops by more than 50 percent. So there is a potential major implication here, being that so many HIV patients who are taking HIV medication may suffer from depression. When they start to take St. John's wort, they may suffer from a relapse of the HIV infection also.
This is the same point, just a different study, which also shows that when you take St. John's wort, blood level drops significantly. Whereas if you take other herbs it may have the opposite effect of decreasing the metabolism so that the blood level increases.
How about the brain's response? Here I focus just on antidepressant. There may be many mechanisms that mediate antidepressant's effects, but the major focus of research of course are the transporters, especially the serotonin transporters. Meaning that most of the SSRI's exert their effects primarily by inhibiting serotonin transporter.
In the past, before this became more clear, most people predicted that there should be less ethnic variation or individual variations in terms of the transporters or receptors in the brain, but that tends not to be the case. It turns out that the variance of genes controlling these target organs, in this example, the serotonin transporter, will determine -- they are associated with how fast you are likely to respond to antidepressants. This was whether the medication will be more effective for you or not. For the serotonin transporter, it is pretty clear now that a long variant confers a much better response than the short variant.
This is represented here from one of the studies. If you have two long alleles, you have a much better reduction of depression scores, as compared to if you have your homozygotic for the short alleles.
So this is very promising. Possibly in the future we can look at this and decide what kind of medication would be good for a particular patient. A challenge that remains to be resolved is that the prevalence of the long alleles is quite different across ethnic groups. Does it mean that some groups may be more responsive to antidepressants or SSIs as others? We don't know yet. That is a focus of some of the research that we are conducting currently.
The point here is that all these variants that are important in terms of determining how likely you are going to develop practically any kind of diseases, as well as whether you are likely to respond to treatment, varies across ethnic groups. This is just another example, an allele in D2 receptor, which of course is very important in the etiology of schizophrenia, addition, and maybe responses to treatment for these conditions. It has various dramatic variation across ethnic groups that have been studied.
Another example is an enzyme called COMT, which is crucial in the metabolism or regulating the neurotransmitters. That also has substantial ethnic variations.
What is the issue here? I think the important point is that variations are very common, and there are individual variations as well as cross-ethnic variations in these genetic traits which then translate to biological differences, and then it translates to behavior, phenotypic differences. These are actually the characteristic of any populations. You need variations. Variations are important for the survival of the species. They are responsible, but at the same time, they are responsible for different traits which then lead to different kinds of risks for different kinds of health risks.
The interaction of all these genes remain to be defined, remain to be characterized. How they interact with environment furthermore is even more fascinating. But the important thing right now, or one of the important issues is that we cannot afford to ignore these variations, both between individuals and cross ethnically, for many reasons, including predicting treatment responses.
I don't think we are going to talk too much about the other side, which may be even more important, which is the more direct causal influences on patients' behavior that then determine treatment response. Again, I just want to emphasize that on top of everything, if the patient doesn't take the medication, there will be no effect, and that is probably the more important factor. Culture influences adherence in ways that still remain to be examined in a systematic manner, and that is an even larger challenge for the field.
On top of that, I want to end with this slide by saying that culture does not only influence patients, culture influences clinicians and the health care system as well. It is the interaction among these that would really benefit from the efforts of everybody, including the efforts of this committee.
So I end here. Thank you very much.
DR. MAYS: Thank you. Questions? As usual I have some. Can you talk a little bit in terms of the cultural beliefs, a little bit more about the cultural contexts? Are there cultural beliefs about medications also that come into play? Would this be a group that would have more or less of an influence in terms of like a placebo effect?
DR. LIN: The placebo effect is actually very often neglected and unappreciated. Anytime when I say that 40 percent of treatment response is due to placebo effect, it raises a lot of skepticism, and people argue about it. I think it is an issue that may be arguable, but the fact is that in any clinical trials when you look at the placebo response rate, it is typically at 40 percent. This is not just in psychiatry, but in other fields of medicine as well.
Amazingly, what we do most of the time is just treat it as nuisance. The effort, the major effort in the field, is to control it, to wash it out. The effort for the major focus of pharmaceutical companies in developing drugs is to manage to show on top of this 40 percent placebo effect, another 20 percent. If you can manage to show 20 percent additional effects, then the drugs are marketable.
So I'll say that there is not enough study on the field in general, and even less information about how culture influences placebo response. Yet theoretically or according to common sense, there is very little doubt, and I think everybody would be convinced that whether you are more likely to benefit from placebo response or not has to do with your belief system and how it coincide or be congruent with the major model of the field coincide with the clinicians' explanations, or whether they clash. When they clash, you have diminished response, even if it is a pharmacological response.
I don't know to what extent that has been objectively documented. Just downloaded a whole book that was published recently, which was available on the website for ten dollars . I haven't read it yet, so I don't know how much information is there.
So that is the placebo response. I think it is our responsibility to move into it, but I guess part of the reason we don't know much also has to do with the difficulty in getting research funding for study in that area.
There may be a little bit more information about the cultural influences on adherence, but still, it is very, very little. So what all we have is more or less personal opinion, and anecdotal information. From my own experience, I can tell you that unless the treatment makes sense to the patient, they are not going to take it. Unless the complication is efficient and unless there is trust, it is not going to work. Even when there is trust, it may still not work, because there may be hidden explanatory model that we don't know about, is this in the patient's mind. Unless there is negotiation and compromise, the compromise will be more of a problem with patients from ethnic minority origins.
DR. MAYS: I want to ask another question in terms of the collection of data in race and ethnicity in the development of drugs, to whatever extent you can answer that.
It is very clear from what you are saying that it behooves us to understand more about how drugs work in various racial ethnic groups, and even then that we need to understand that there is variation. In the process of the FDA process, if I understand correctly, we are now at a point that you are supposed to give information on race and ethnicity, but at what point in the process?
I have heard various points in the process, and it seems like some you don't, and then it is not until it gets to phase three. Maybe you could explain that a little bit more.
DR. LIN: Again, I think it is an evolving process, and there is tension everywhere. The cynical side of me would say that it won't happen until the drug companies see that this could potentially be used for marketing purposes, which might yet happen soon. Either that, or that they need to -- the other cynical point is that if there is enough lawsuits maybe.
But I think the larger issue may be that if you really do it for sure, it will increase the cost, and whether drug companies can afford to do it, whether society would be willing to make that investment or not, in theory sooner or later it should happen, because what we have right now is that we are using the experiences pretty much based on one group and hoping that the information gathered from one group will be applicable to other groups. Then after the drug is marketed it is up to the clinicians to do the adjustment. Fortunately, because of the universality of a lot of things that we all share, to a great extent most of the things that are studied or discovered based on one group to some extent is applicable to other groups. But we don't have objective information to make adjustment to optimize the use of medication for other groups that have not been carefully tested.
The other tension comes from the issue of the drug company and to a large extent also to professionals' need to focus on what is common or what is similar across different people. In that process they ignore -- in addition to ethnic differences or group differences, we also tend to ignore individual differences. So how this tension plays out is something that will be quite fascinating for the future.
Perhaps one day when we have more information about all the factors that determine treatment response, we can truly have an individualized medicine, or so-called personalized medicine, and that would address this question to a greater extent.
DR. MAYS: Do we have questions in the audience?
DR. HIYASHI: Taki Hiyashi from the Department of Medicine. I am very interested in your study. I am a practicing physician in Japan, and whenever I look at the textbook from the American organizations, they tend to specify an optimal dose much higher than we Japanese do. You clearly give me very good research results that explain why.
I was wondering how much this knowledge is reflected in the real clinical practice in America. Does the clinician tend to adjust the dosage of drugs depending on patient's race and ethnicity?
DR. LIN: I think that may be the trend. It is hard for me to judge, because the people that I talk to probably are already familiar with the literature. But 20 years ago whenever we present the information, the audience tended to be very skeptical. They said everything is due to weight and weight differences, or body composition differences. But people no longer believe that. People are in general more convinced that there are differences. Whether this gets translated into practice or not is a different matter.
When this becomes widespread, I worry about the opposite side of the problem, which is stereotyping, because then I worry about the clinicians being reluctant to give some of the Asian patients who may for the Asian standard be more or less outliers, and they may need the higher dose of medication as well.
So as in any field of culture or ethnic studies, the major issue is on the one hand not appreciating enough ethnic differences, and on the other hand stereotyping. This is not only biology. Behavioral science, the same thing. If you assume all Asians will have family support, that may not be the case for some of the people who may turn out to be different. Or assume all Asians are shy; that is not true all of the time, either.
DR. BREEN: Nancy Breen, the National Cancer Institute. Thank you for a very erudite presentation, Dr. Lin; it is very impressive.
I wanted to follow up on this gentleman's question to you. I am a social scientist. I am an economist working in health services research. One thing we don't have much information on is physician behavior towards patients. I was wondering, in response to your question, do we have documentation on the prescribing practices of psychiatrists towards their patients? Or are you just guessing based on -- I'm sure you have talked to a lot of psychiatrists, given that you are a professor of psychiatry, but is there a literature that would help us understand what the practices are of physicians towards their patients?
DR. LIN: That is another whole field, area of research. The thing in the review papers that I distributed, some of them may have information. Actually, there are a lot of earlier reports that show regional differences, ethnic differences in terms of medication prescription patterns.
DR. BREEN: Would those be international or would those be within the United States towards different --
DR. LIN: More international, but also in the United States. I don't think there is that much information in terms of looking at -- most of these studies as I remember -- because we have moved beyond that and started looking at prescription pattern and review of evidence for that as the basis for our moving towards trying to find reasons for differences. But the earlier literature mostly showed relationship between patients' ethnicity and the dosage or side effect profiles. There might be a few, but not much, in terms of looking at the physicians' ethnic and cultural background, and whether that influences prescription profiles or not. I assume that it may have some effect, but it may not be in the literature. So that would be another important research area also.
DR. MAYS: Let me ask you a question. Again, this is pure speculation, but your sense on collecting data on race and ethnicity. One of the problem that we have is, quite often it is talked about when and where do you collect the data. So in terms of for instance within a health center or even a private practice, do you think it is possible to either fill out the insurance papers and say something about -- have the psychiatrists ask the person about their race and ethnicity? Do you think that that would at all be feasible? Do you think that psychiatric patients would balk at letting us know what their race and ethnicity is?
DR. LIN: That is beyond my expertise, but I assume that many people may be working on that as well, along with this genetic initiative. Any large group that looks into the area, they typically have a subgroup that looks at the ethical social issues. I don't know how that is going to play out.
There are major things the field is worried about, and we need to worry about, but within that framework, I would think that as much as possible we should have more information, but not infringing on patients' freedom. There are also sticky issues that people bring up all the time about, what happens when somebody is of mixed parentage, how do you define that and what does that mean. Then there are problems of -- I'm sure you have presentations on misclassification. That is an issue probably. I don't know what is the extent of misclassification. In the emergency room I see often Asians being misclassified as Caucasians. I ran across literature saying that the reason some reports show a lower cancer rate in Filipinos was that often they were misclassified as Caucasians, and inconsistency between admission and discharge, and things like that. I think that is a major issue that needs to be looked at.
But my worry is that people say, this is so sticky so we don't look at it. But if we don't look at it, then the field will not move ahead, or have a major problem.
The example is, the genetic research until ten years ago, this was not an issue for geneticists. But then because in many of the studies that they looked at ethnicity turned out to be a confounding factor. So now in every study, if you don't look at ethnicity, it is the study's fault and the results are not reliable. Very often these differences in alleles will more likely be because of ethnicity, not because of disease or other issues that they are looking at.
DR. MAYS: Do we have any more questions in the audience? What about my colleagues here? Thank you very much. Thank you for taking time to come and spend this with us. Our goal is to utilize this information to try and make a set of recommendations, so your work makes that job a little bit easier in the sense of actually having some data about what happens in the processes. So thank you.
DR. LIN: Thank you.
DR. MAYS: We are going to take about a ten-minute break, and then we will reconvene again, and we will have Dr. Ponce with us.
(Brief recess.)
DR. MAYS: We are going to get started for the afternoon. I appreciate those of you who are here. We have had a very full day, and I would say very good day. It is going to continue to be good. We have Dr. Ponce with us. Many of you know her. As I now discover, she brings her own following with her. Thank you very much.
One of the things that this committee had wanted is to actually hear a bit about CHIS, the California Health Interview Survey. As many of you know, its model is the national health interview survey, but the thing that is different for the California Health Interview Survey is that they have very successfully in 55,000 household and beyond been able to have a number of racial and ethnic minorities participating.
One of the questions that often arises in doing surveys is, what about all these other groups, and decisions get made about language. As far as I know, CHIS is probably one of the few that has translated its survey into several languages that are consistent with the Asian, Native Hawaiian and other Pacific Islander populations. So we are delighted that you were able to take your time to be here with us. I also appreciate the fact that you took some time to prepare some data for us. i am thrilled that you were able to do that, because I think this is actually important for us to be able to say it can be done. So the committee is grateful to both the time you have put in and taking that time to present to us.
So without further ado, Dr. Ponce is going to talk about language, and she will talk to us about the issue of aggregation and disaggregation. Thank you, Ninez.
DR. PONCE: Thank you, Madam Chair, Professor Mays, and also your distinguished -- are you all from NCHS or NCI, the committee?
DR. MAYS: They are all part of the subcommittee. They are from different places.
DR. PONCE: Okay, distinguished federal visitors to California. Welcome to California.
DR. MAYS: See, you can tell she is a good researcher. She gets those categories quite quick.
DR. PONCE: I'm going to talk about the California Health Interview Survey. The presentation is broken up into three parts. First, for those of you who have not heard about it, or who have heard about it but want to know more of the juicy details, I'm going to give the background on CHIS. Most importantly, as is relevant for this particular subcommittee, on sample sizes for Asian subgroups.
I am also going to talk about the cultural linguistic adaptation of the California Health Interview Survey, which I led along with other staff folks, and with the advice of a very valuable multicultural technical advisory committee. Then lastly, so this is the treat if you do end up staying here -- and if you have heard this before, I am going to present some preliminary estimates of CHIS using the Asian oversample data juxtaposed with the random sample. I'll put that more in detail, but there will be some treats at the end.
What is CHIS? It is California's new assessment tool to meet state and local needs for population-based health data. It is principally to improve the health of Californians. So it informs policy analysis, development and advocacy. It helps service providers in terms of program planning, and it helps researchers who want to try to understand why there are some disparities, some assets and where catchup has to be done across the diverse population of California.
It is a collaborative project of UCLA Center for Health Policy Research, California Department of Health Services, Public Health Institute. We have over 150 people in planning, content and methods. As I mentioned, the multicultural technical advisory group is one that I have particularly benefitted from in terms of advice, in terms of the cultural and language translation. We also have other technical advisory committees, and several working groups.
What I don't have, which I now regret, since Nancy Breen from NCI is here, I don't have Rick Brown's funding slide. The funding sources for CHIS come from several sources, the California Department of Health and Human Services, National Cancer Institute is one of the biggest funding sources, federal funding sources for CHIS, which funded the cancer control supplement in 2001. We also got funding from the California Endowment, a private foundation in California. We had some funding from Indian Health Services, another federal source, in terms of specifically for the oversample of the American Indian-Alaska Native population. There were some other funding sources directly related to questions in CHIS, the Prop 10, I don' t know if any of you have heard of the Prop 10 Commission, which is interested in looking at work that advances understanding of early child development, and there were also some researchers who contributed additional questions on adolescence and gun violence.
CHIS collects information on health status, chronic conditions, asthma, diabetes, arthritis, high blood pressure, cholesterol and heart disease. I actually have some slides in the end that look at these by Asian subgroup. Cancer history, risk prevention and screening, health behaviors, firearms and violence, particularly in the adolescent questionnaire.
It also collects information on access to care and utilization. If the respondent has a usual source of care, health care utilization in the past 12 months, both medical and mental health, receipt of screening tests and other preventive care, access barriers. These barriers also include whether it was delayed treatment related to a specific chronic condition that that person may have reported.
We have health insurance coverage, the current, that means at the time of interview, and the past 12 months, so we can get at stability of coverage of that Californian, program participation and eligibility. One of the biggest concerns in California is that there could be -- many of the uninsured could actually be eligible but are not participating, and so we model eligibility in CHIS based on some income, asset and family composition variables.
Demographic characteristics. It is hefty in CHIS. Race, ethnicity, citizenship, immigration status, nativity, English proficiency. We also have birthplace of mother, birthplace of father. We also include a question on sexual orientation.
There are also child and adolescent specific issues that are -- for example, for the child issues, whether you read to your child at night, so child well-being. So it encompasses not just directly chronic conditions, but on child well-being.
CHIS had two objectives that in some ways were a little bit competing, but we tried to address both objectives as best as we could. That was that California wanted local level estimates for counties. So a lot of the national surveys such as the national health interview survey or the current population survey, would be fine for a California sample, because California is large enough. But when you drill down to the county level, to the local level, we wouldn't have enough samples.
So one of the prime objectives is local level estimates. But another objective -- and this I have to say was championed and propelled by National Cancer Institute -- was to get information on California's major ethnic groups and some specific ethnic subgroups. National Cancer Institute was very interested in getting information on cancer control, information on Asian subgroups and American Indian-Alaska Native groups, because those are the groups that are not really well represented in the national health interview survey.
So we attempted to fulfill both goals through sample design in a large sample. The sample is a telephone survey of 55,000 households, drawn from every county in the state to insure that we actually had estimates for each county. We had 41 geographically defined sampling strata. There are 59 counties in California, so some of the counties we have to group.
Then here, where it is different is that we actually did some ethnic oversampling. The oversamples were groups that in my knowledge haven't been oversampled before. Asian-American subgroups, including Japanese, Vietnamese, South Asians, Koreans and Cambodians, American Indians, that included an urban and rural oversample, as well as some counties, for example, San Francisco and Santa Barbara.
DR. MAYS: Ninez, can I ask you a question?
DR. PONCE: Sure.
DR. MAYS: In the previous slide you talked about oversampling in a rural area. Can you talk a little bit about that, how that occurs? It is apparently for the American Indians, so you oversample also a rural population?
DR. PONCE: Well, the intent actually was to oversample urban American Indians, because a lot of the reservation based Indians -- there is not enough information on the American Indians in Los Angeles County. So the intent was to get at urban=dwelling American Indians. But the nature of the oversample was based through IHS data and IHS clinical data, which then is tied to facilities that are more likely to be linked closer to rural reservation areas.
So again, the intent was mostly urban, but by the nature of the administrative list used for the oversample, we ended up also getting an oversample of rural American Indians.
Relatedly, the good news for American Indian-Alaska Natives is that because we made sure we had enough observations for smaller counties, which tended to be rural, so it is mostly the northern California counties, and there was a higher proportion of American Indians that would live there. So we actually then ended up increasing our American Indian-Alaska Native sample, because we oversampled the smaller rural counties.
I don't know if Paul Ong has gone over the demographics. He usually does.
DR. BREEN: Dennis kind of did.
DR. MAYS: I think Dennis did.
DR. PONCE: Oh, Dennis did? So this is just to show how different California is from the United States. We have a tremendously high Latino population, 32.4 percent compared to 12.5 percent. The other thing of note is that the Asian population in terms of racial categories comes in second to whites, and African-American are third. So the Asian population, which is about 11 percent in California, nationally the distribution is less than four percent. African-American it is 6.7 versus 12.3 nationally. American Indian-Alaska Native, about the same. Native Hawaiian, other Pacific Islander, much higher in California. In fact, there were more Native Hawaiian and other Pacific Islanders that lived in California than in the state of Hawaii.
This is a schematic of how we sought to collect information on race and ethnicity. Of course we wanted to comply with the new federal standards for race and ethnicity reporting. I am calling it for short OMG 15. I know it is supposed to be called the new federal standards for race and ethnicity reporting. We separated Native Hawaiian and other Pacific Islander from Asian.
We also -- again, adherent tot hose new standards, we allowed the respondent to reply as more than one race, similar to the census. Similar to the national health interview survey, we also allowed the multiracial respondent to identify with a primary race. Then we pushed the frontier a little bit, in that since we do have a large Latino population, and typically Latinos would put other race after the Latino Hispanic category, and then they would put other, we also asked if they were Latino and white, a single race, white, we also asked if they identified more with being Latino or if they identified more with being white. So this is a little bit frontier, some controversy, but trying to unpeel this complex construct of race and ethnicity.
Data collection results. We finished in about a nine-month period, 55,428 adults, almost 6,000 adolescents and about 13,000 children. We also have an ethnic oversample as I had said on American Indian and Alaska Natives. The ethnic oversample included 540 Vietnamese, over 400 South Asians, over 300 Japanese, over 300 Koreans, 126 Cambodians, and the based on an administrative list from the IHS, 321 American Indian-Alaska Natives.
This is the random digit dial, just to sort things out here. There is a random digit dial component of CHIS and there is an oversample component. So this is just a random digit dial component. For adults 18 plus, single race categories, we completed almost 4,000 interviews amongst respondents who said they were single race Asians, and then 234 single race Native Hawaiians and other Pacific Islanders.
I want to note that this is possibly to my knowledge the largest sample of Asians and Native Hawaiians and other Pacific Islanders in a population-based survey.
Adolescents ages 12 to 17, 353 single race Asians, 35 single race Native Hawaiian-other Pacific Islander. Please note, this isn't all mentions, which I will talk about in another slide.
For children, 842 Asians and 47 Pacific Islanders.
Madam Chairman, I want to point you to this slide, because this is where the complexity of the different samples come in. This slide shows -- the Asian ethnicity is on the very left-hand column, with Chinese, Filipino, Japanese, Vietnamese, South Asian, Korean, Cambodian, Southeast Asian and other Asian. So if the single or most identify -- if the respondent said they were for example Chinese, single race, or they were multiracial Chinese but they mostly identified with being Chinese, the sample is 1229. In 191, moving over to the right on the third column, on Chinese row into the third column, 191 corresponds to that they are multiracial Chinese and they most identify with an ethnic group that is not Chinese.
The RDD total of any ethnicity is 1418. There are no supplemental samples for Chinese, as there is no supplemental samples for Filipino, because we met the minimum threshold of 800 for these two groups. So in the very last column, the total Asian ethnic group sample, any mention is 1418 for Chinese, 978 for Filipinos. So those are just the top two groups to illustrate.
We estimated that we would not reach the 800 threshold just by doing a random population-based telephone survey for Japanese, Vietnamese, South Asian, Korean and Cambodian. So in addition to the random digit dial which yielded for Japanese 572, for Vietnamese 339, South Asians 405, Korean 483, Cambodian 80, we oversampled these groups using list-assisted surnames, and we got the additional 330 for Japanese, 540 for Vietnamese, 449 for South Asians, 326 for Koreans, 126 for Cambodians. So the final column, the furthermost right column, are the sample yields for these different ethnic groups that reflect two strategies of random digit dial and list-assisted oversample, and these are any mentions.
DR. MAYS: I wanted to ask a question. I don't understand why you thought that you were going to -- this means that at the outset, you had some sense of projections of who you thought you would get your 800 from, right?
DR. PONCE: Right, that's correct.
DR. MAYS: I'm surprised, and I want to ask a couple of questions. You thought you would not get Japanese, but you thought you would get Filipinos. Is that based on population? Is that based on something about who participate or who does not participate?
DR. PONCE: It was strictly based on population. It was 1990. The planning stage predated the Census 2000, so this was based on 1990 public use micro sample, the PUMS. At that time, Filipinos actually were the number-one group, and so they have declined over the last decade. So that is one reason.
DR. MAYS: Just one other question. Are there any behavioral differences in any of these subpopulations about willingness to participate in studies? Are there any other groups that have more skepticism than others?
DR. HITCHCOCK: Or is there any other phone coverage among those groups?
DR. PONCE: Phone coverage is actually quite high among Asians and across subgroups. In fact, it is also the most connected through the Internet. The group just based on our experience that was most resistance to telephone interview were the Cambodians. One reason was -- well, there were several reasons, because we convened a group of community-based organizations -- is that there is still a fear of reprisal, that some information would go to the homeland. They also get telemarketed a lot to vote, because they are a very concentrated group. For example, ten percent of Long Beach are Cambodians and I have heard that you are going to find 60 percent of them in one apartment building. So it is a very concentrated group. They are telemarketed, very political, but also fearing some political consequence of reprisal, or thinking the home country government is trying to identify them and find them. So there was -- we had to regroup and think about a strategy of contacting them, and what we ended up doing is just resending a personal letter to them.
The Vietnamese are also another group that has been surveyed a lot through telephones and through household surveys. That is primarily because there is an excellent research team in UC-San Francisco called Health is Gold, that have been cleaning up the list, working on the Vietnamese list, working on telephone surveys. So a combination; the survey lists were a little bit better, and they were used to being surveyed. It was actually quite a success for Vietnamese. I'll have response rates later on.
DR. HITCHCOCK: What popped up on their Caller IDs when you all called these homes?
DR. PONCE: It depended where they called. We had -- the vendor was based in Virginia, but had Colorado, mostly for the Spanish calls, and Sacramento as well. So I don't know.
DR. HITCHCOCK: (Comments off mike.)
DR. PONCE: That is a good question. I can find out. For Caller ID, I don't have Caller ID, so it would put the vendor's name on it?
DR. HITCHCOCK: It is various combinations. I'm just curious about if you had any information that showed up on the screens in the Caller IDs.
DR. PONCE: I can get that information for you. I don't have it here.
I think one thing to note is, Filipinos weren't oversampled, and we also did not conduct a survey in the Filipino language. Chinese, we did conduct it in Cantonese and Mandarin. The South Asian group, we didn't have much experience, or at least the team members who have experience in survey research weren't sure about how that would play out, but it actually was not -- there was not as much friction as we thought there would be.
In the Koreans, there wasn't a problem with the Koreans. So we can look at the response rate later on about it, but I think what comes to mind the most are the Cambodian.
So that is an overview of CHIS. Do you have any other questions before I go on explaining about the cultural and language adaptation?
DR. MAYS: Is it okay to ask about cost, or is that bad?
DR. PONCE: It is on our website.
DR. MAYS: No, I just wanted to know, because I was going to suggest that when we get to the language part, just some sense of the cost for translating. I have translated stuff, and it is enormous, so if you did all these languages, we just want to get a sense of how you did it and maybe what the cost was, if that is okay. If it's not, it's okay.
DR. PONCE: Can I give you the cost after I talk about what we did?
DR. MAYS: Oh, sure.
DR. PONCE: How about that? I thought you meant the entire survey.
DR. MAYS: No, I just wanted to know about the language.
DR. PONCE: About the language, okay. This is the first slide on why we did a cultural and language adaptation process in CHIS.
Again, California versus the U.S., foreign-born much higher -- that is the red bar on the left -- much higher than the U.S., more than a quarter of the population, compared to less than ten percent of the U.S. population.
Census collects -- and I apologize, this is 1990 census, because the PUMS has just come out, but there is a measure of, does anybody in the household over the age of five, and another measure is over the age of 14, can anybody communicate in English. If no one can, then that household is deemed linguistically isolated.
So using this measure of linguistic isolation, again, California has a much higher rate, close to ten percent, compared to the U.S., which is much less than five percent.
DR. BREEN: Is that using the measure of linguistically isolated because you have a five-year-old or a 14-year-old in the house? Or don't you know?
DR. PONCE: That one was using a five-year-old.
DR. BREEN: Boy, I'd hate to rely on that.
DR. PONCE: The California demographics, using Census 1990 and 2000 data, these are just the proportions of what I showed you. In terms of translating language, in terms of translating surveys into different languages, the first question is which ones. Definitely Spanish, but if you want to consider translating other languages, not necessarily Asian, but other languages, you have to have some criteria.
So what we did was, we wanted to set the criteria pretty solidly so that we wouldn't have to much later on defend why we didn't do -- as I am defending right now -- why we didn't do Tagalog, for example. So the criteria was the prevalence of linguistic isolation, again using that measure of a household member over the age of five to speak English well or very well. So if no one in the household can speak English well or very well, no one in the household age five and over, then that house was considered linguistically isolated. So that was actually one of the main criteria.
The second criteria was, if there was a chance that we would reach at least 100 completed interviews. So for example, even though Yu Mien or Mong would be quite high in linguistic isolation, based on the numbers that we had -- again we were dealing with 1990 census -- if they were too few and based on a probability sample that we wouldn't even be able to complete 100 interviews, we wouldn't choose them.
So first it was linguistic isolation, and then it was the numbers. In terms of the numbers, then Tagalog -- I use it to illustrate because it is one of the sensitive languages -- Tagalog would definitely be considered, because it is one of the top five languages spoken in California. But in terms of linguistic isolation, because immigrant Filipinos have been educated in a American colonial system where English is used, the linguistic isolation rate in 1990 was seven percent for Filipinos. So that compared to for example nine percent for those who speak Italian, so it is a very low rate, and that was one of the reasons why Tagalog did not make it.
The selection for oversampling, that actually was specifically for including Khmei translation, the Cambodian language; because we were oversampling Cambodians, it behooved us then to translate the survey in Khmei. And of course, the costs of translation, as Chairman Mays says.
So in the end, we translated the survey into Spanish, Chinese, both Cantonese, Mandarin dialects. It was translated into Chinese, so written it was Chinese, but we had reviewers who were -- we had Cantonese reviewers and Mandarin reviewers review the written translated Chinese version, and Korean, Vietnamese, and Khmei.
Before we actually went into the translation though, we had the English survey go through a process of review by several experts, to make sure that any stereotypes or biases or offensive language were removed from the English survey. So that was the first step, was a review to make sure that it wasn't offending any particular group.
Then the second was, after they removed the offensive language, then we started accommodation. So for example, in questions of what fruit juices do you drink, that we don't just have orange and apple juice, but other tropical juices are included, such as guava and mango.
So it was both to -- the first step was to rid of stereotypical language or offensive language, then the second was to start a process of accommodation for different groups.
Again, you have to set parameters. We can't do it for every single group in California, so we had to set criteria. Broadly, the groups that we were targeting were groups that typically are not represented in population-based surveys. These are communities of color. So our target groups were American Indian-Alaska Natives, African-Americans, broadly Asians, and Native Hawaiians and other Pacific Islanders.
Specifically in instructions to our reviewers, we were very interested in Mexicans. This was particularly for the language translation. So translating the Spanish for example to that that is used mostly by Mexicans versus Central Americans or South Americans, Chinese, Vietnamese, Koreans and Cambodians.
The overview is that we had language simplification by ten national experts, Leickert scale, two-day discussion meeting. This occurred before the cultural review, so this was mostly to also make sure that the language was understandable, at least at the eighth grade comprehensive level.
Then the cultural review which I described in the previous slide, ten bilingual, bicultural experts representing target groups, another Leickert scale evaluation, one week review, two-day discussion meeting. Then we also had mini-groups with American Indian and Alaska Natives and African-Americans, where we tried to get mostly the adolescent groups and trying out some selected items in the English survey.
These are some language simplification examples. Was your child ever diagnosed with asthma, was the original version, and the simplified version is, did your doctor ever tell you that your child has asthma.
Now in this tedious process, the English surveys are ready for translation. In this case we had to decide what the translation method should be. We departed from the translation-back translation method. Mainly we wanted to save time. As soon as we secured funding, we had to be out in the field, and we wanted to have the translated versions out at the same time or very soon after the English surveys. So with forward and backward translation, you have the source language, English say translated to Spanish, and then the Spanish version translated back to English. So there would be a waiting time and a log that we couldn't afford.
We did a lit review, and we decided on something called refereed multiple forward translations. I changed this, because it used to be called translation by committee, and every time I said we tried to save time by doing translation by committee, I always got giggles, like you really thought a committee would make it go faster.
It actually did save us a few weeks. It didn't save us a tremendous amount of time. The main thing behind this was that we wanted several people to look at the translations and to give us feedback. But we didn't want to give it to everybody. We didn't want to ask ad hoc, your cousin, your aunt or another researcher, but we identified that we wanted two forward translations conducted by two independent firms which didn't know each other, and then we hired a third firm. The third firm was connected with a federally qualified community health center.
The third firm then looked at the two forward translations and evaluated item by item the quality of the translations. We then all got on the phone, and the ones that were problematic we resolved. So with the two translators that were competing, and then a third referee, and then some UCLA staff, we were on the phone going item by item.
I have to say that this was -- it took nine weeks per language. We thought it would take a lot shorter, but it took nine weeks per language.
Finally, our vendor, Westat, they reviewed the instrument to make sure it was a more conversational tone.
DR. HITCHCOCK: Years ago I was involved in a translation process with the Hispanic health and nutrition examination survey. That is the process we ended up eventually following. We did it first through the forward-back translation, and people threw up their hands; it was kind of difficult. We wanted to the best extent possible a universal Spanish that could be understood by Puerto Ricans, Cubans and Mexican-Americans.
I think -- and see if you agree with this -- that that last item, the Westat interviewers for conversational tone, turned out to be very important. Not only was it more conversational tone, but they were the ones that turned it around into what you referred to as simpler language earlier on. They changed the question, has your child ever been diagnosed with asthma, to, has a doctor ever told you, and put it more into a survey research framework that we thought was a big improvement.
DR. PONCE: I think that is true. I think the criticism though with Westat being viewed as a non-California firm with recruiting interviewers that may be Latino Hispanic, for example, but may not necessarily be Mexican-American. There was a little bit of some regional concern that the non-Californian interviewers who were looking at the survey were having the last say. So I think some of that can be resolved, because they are setting up big time now in California so the interviewers will be here.
DR. HITCHCOCK: Do you have small -- is this where you had to put alternate words into the questions, so that everybody could choose? This is something that maybe arose between the three Hispanic dialects that we were looking at. But there were some things that just didn't make any sense.
DR. PONCE: Yes, actually. One word was county. County was one. I had some examples. There were some words that -- the one I always say is in the Pap test. There was one translator that -- when we ask a question about the Pap test, we explain what it is; it is when a doctor takes a swab -- swab was hard to translate. Some of the translators wanted it to be Q-tip, because Q-tip is pretty common in Latin America, and somebody translated it into a little brush, but it ended up -- but another person said, little brush, in Peru it is a broom. That is my favorite.
DR. HITCHCOCK: We had babies being translated into Spanish as creaturas, and then translated back to English as creatures.
DR. PONCE: Again, the main thing is, we wanted to have a plurality of comments on this, but to still put a parameter so that we didn't leave it open, and showed it to everyone we knew who we thought was fabulously bilingual in these different languages, and to review it. So I call it a full blind process.
The other thing too is, with the translation-back translation, for Spanish where you have auto-translation. A word of caution. When the first Spanish forward translations came back, they were nearly identical, identical to the point where you either say we don't have to have a reconciliation. But it raised the suspicion that they used auto-translators.
DR. HITCHCOCK: The same software.
DR. PONCE: Yes, the same software to do it. So that is one thing why it is good to have this refereed process.
These are screener interview times by language. The average screener time is in lilac and average interview time is in yellow. I point more to looking at -- in terms of comparability to look at average screener time, because it is basically when you screen through the household, it is the same short screener. When you start looking at the average interview time, what is hidden behind that is that, for example, Khmei looks like it was the longest. It reflects not only that Khmei might be more chatty or more difficult to administer over the telephone or the Cambodian respondent less cooperative. But it could be that that respondent didn't get skipped out of a lot of questions. So they ended up saying yes to all the disease questions, yes to all the food and security questions.
So that is what is behind that. I actually like to point more towards the average screener time to see. So average screener time, with the identical screening script, it still took longer. It is an average six minutes, compared to about three minutes for the English. The Mandarin, Cantonese, Korean, Spanish, generally the non-English languages took a little longer.
The response rates. This is after they have been contacted, 59 percent after they have been contacted, so it is more like a cooperation rate here. We have American Indian-Native American, very high cooperation rate. I want to note that the other aspect of that was not in terms of just using the IHS list, but there were actually presentations done by our staff to several American Indian-Alaska Native groups. So it was very high participation rate there. As I said, Koreans was the least friction, amongst Korean and Asian groups.
This is the oversamples, the listed oversamples. We don't have Chinese and Filipino, because they weren't oversampled here, but you do have -- once the Cambodians didn't hang up, once you got them on the phone, they actually were very cooperative, once they understood what this was about. It was 76 percent.
This slide shows that of the RDD, the random digit dial for CHIS, over ten percent, about 11 percent were conducted in a non-English language, and that is substantial. What we are trying to show here is that had we not done it in Spanish and all the Asian languages, that this 10.7 percent would not be represented in this group. And of course, in the list assisted sample it is a much higher group, because we were targeting these Asian groups.
This is the percent of interviews in language, so it is looking at it another way. If you look at the Latino group in the RDD, 42.7 percent of our Latino sample completed the interviews in Spanish. That is again a huge proportion of the Latino sample completed in Spanish. Among Chinese, 39.1 percent, among Koreans 49.7 percent, Vietnamese half, Cambodians 42.4 percent. This is just in the RDD, this is not in the special oversample. We are seeing that a sizeable proportion of the Asian and Latino population preferred the completion of this survey in another language.
This is the combined, the RDD and the listed sample. In combining the two samples, I looked at the percent of the language groups who were limited English proficient. This measure is based out of a question on the person's perception of their ability to speak English, very well, well, not well, not at all; it was a four-item scale.
Among the Cambodian groups, 64.3 percent very high LEP rate. Even among Cantonese, 62.4 percent, Vietnamese, 53.0 percent, Spanish very high rates of LEP. Total all languages including English -- if you include English to everybody, it is 5.5 percent.
I guess what this shows is that we were adherent to our initial criteria of the linguistic isolation. This is what is playing out empirically here.
Really quick lessons and conclusions for this section. I have to say that this linguistic isolation rate really created the ire of Filipinos and South Asians. In their group, you have a really bimodal population, where yes, it is seven percent linguistic isolation, very low for Filipinos, but if you are only going to do English-speaking Filipinos, then you are leaving out a group that is really, really vulnerable and has the greatest health needs. Same thing with the Southeast Asian, that was a similar concern. So that was one of the things to consider in using that criteria.
We do note that because of this LI rate criteria, over 45 percent of our target language groups completed a translated survey version of the CHIS survey. We believe that this process of language simplification, cultural adaptation and translation in a refereed manner increased inclusion of these persons and made the sample more representative.
I'm going to go towards the end. Sorry I'm taking so much time. These are just some preliminary Asian subgroup estimates. Madam Chairman, it would be instructive to see how when you disaggregate, how groups are different. I want to estimate that these are preliminary, they are for illustrative purposes, they are not age adjusted, and they are not tested for statistical significance. This is just really to show that the bars are -- you're not going to get a horizontal straight line for every group, and that what happens if you disaggregate.
Over the very right-most bars are the aggregate group together; this is Asian aggregate group together. This is what I did. It is still up for question if I can do this, but I combined the RDD sample with the Asian oversample of CHIS, because that was what we wanted to do. So I combined the two, and I used -- the way to do it intuitively is to combine those that most identified with that group, so that is why they are mutually exclusive. These groups are those who are either single race Japanese, single race Korean or they are multiracial but identify with being with that group.
The Chinese and Filipinos are from the RDD sample, and the South Asians, Japanese, Koreans, Vietnamese and other single multiracial are from the combined RDD and list-assisted sample. This group, since my whole discussion was on language, I also stratified these charts by limited English proficiency, where zero is not limited English proficient and one is limited English proficient.
This chart shows that percent reporting fair or poor health, and I only looked at non-elderly adults, in the Asian aggregate, the charts to the most right, it does show that those who are LEP report much higher rates or fair or poor health than the non-limited English proficient population. But taken alone, it masks that for some groups like the Vietnamese, a tremendous reporting of percent of fair to poor health among the LEP, and that is the highest tower over there. Also, even among the non-LEP, it is quite higher compared to the Asian aggregate category.
DR. HITCHCOCK: You looked at the cultural implications of this, I suppose, and how this might be the case, and is reporting of say excellent health something that certain culture groups might be afraid or reluctant to have reported.
DR. PONCE: I didn't do any psychometric -- this is really for illustrative purposes of the aggregate. But you're right. On that note, in translation there was this difference in terms of fair, of translating fair in Spanish. Is it passable, or average? Medium. So there is a difference. There is a lot of research on that, on what should be used. Passable was actually the preferred.
This is just for illustrative purposes on the consequences of, if you present or you are only able to present data in the Asian aggregate category. In this case, it is also fair to poor health, but I have the over-65, and I did this because the age distribution is going to be different for these groups. For example, Japanese may be an older group and Vietnamese might be a much younger group. So this is for the elderly, percent reporting fair or poor health, Asian aggregate, LEP much higher than non-LEP, but even among the elderly you have the Vietnamese population reporting much higher percent of fair or poor health both for the LEP and the non-LEP, similar trends for the Chinese that we saw for the non-elderly Chinese.
This is the percent reporting -- these are the percent uninsured. They have no health insurance coverage. Asian aggregate, which is what we see a lot in national statistics and California statistics of uninsurance, the LEP confirms those who are limited English proficient have higher uninsured rates, possibly because of less access to more core quality jobs, than the non-LEP.
Here we see -- and this is what we have suspected for many years, but didn't really have the data, that the Korean rate is much higher. The Korean rate, because most of them are self employed, is a higher rate, and the LEP Korean rate is the highest there. Sorry that is so small, but that was intentional, since this is preliminary data.
This is percent reporting arthritis diagnosis. I was choosing selected indicators that had high prevalence. The Asian aggregate group, if a policy maker were to see that, they would say it is around 25, 27 percent, no difference whether you are LEP or not. Indeed, you do see some differences. It is much less for Korean, much less for Vietnamese. I would discount other single multiracial right now, because it is a small group, so I wouldn't look at that. But it is much lower for Koreans and Vietnamese.
The blood pressure diagnosis, Asian adults, I'm only looking at those who are over 50. In this case, the Asian aggregate category to the far right has about a 40, 42 percent prevalence rate, and you have higher rates among Vietnamese LAP and also higher rates than the aggregate category among Filipinos who are not LEP.
Similarly in diabetes. The aggregate category masks some of the variations within subgroups. The aggregate category reports about a ten to 15 percent rate. In this case, diabetes diagnosis is lower among the LEP, which has not been what we have been seeing in the last couple of slides. We also see a high diabetes diagnosis rate among Filipinos and among Koreans.
I also have others which I will share in terms of cancer screening, because I know that that is of interest.
DR. HITCHCOCK: What about visits to the physician or something like that? Do you have that?
DR. PONCE: I have that too, but I didn't put it in the Power Point. I have last visit, I have delays in care, cancer screening for breast, cervical and colorectal.
That concludes my presentation. I think the main point is that if you are going to have a population-based survey and you want it to represent your population, and it is a population such as California's, that cultural and linguistic adaptation is an essential quality component.
I also want to say that the cost was not insurmountable. Not counting staff time, the cost for our -- we paid a vendor to manage all the different subcontracts, because we were dealing with three translation companies, 20 individual cultural reviewers, the CBOs that dealt with the mini-groups for the African-American, the American Indian. So we subcontracted with a group, and that amount was under $200,000. So I think it is not -- the entire survey is $11 million.
DR. MAYS: That puts it into perspective.
DR. PONCE: Sorry, I wasn't supposed to say it, but just to give a perspective. But again, that less than $200,000 figure does not include staff time, the time of the multicultural technical advisory group and working with us. It doesn't include student time that worked with us. So it is not fully included; it is just a straight subcontract cost.
DR. MAYS: Thank you. Let me start with my colleagues first in terms of questions.
DR. HITCHCOCK: I have some. I think what you have done is really great. I really appreciate the effort that was put into the different languages and the results that you achieved because of that. So I think for the purpose of your presentation here today, you have done an excellent job and convinced us of the importance of this.
Maybe you can't address these questions, but I am just wondering about response rates, in not so much the magnitude or whatever you would call it, but is there any way that you can evaluate the non-response here? I am assuming that when you talk about the response rate to the survey, you are talking about a completed screener and a completed interview for each person or each household, however it went.
For instance, are there screeners that were done that contained information that might let you evaluate people that hung up during the interview or something like that? Or can you use any kind of geo-coded information to get where the areas were where your response was low, and see some of the census characteristics of the people who lived in those areas where the response was poor?
DR. PONCE: It is not my field or responsibility of this. I know that that is a concern, and that has been raised in terms of looking at that non-response data. I know that there are some staff that are working with Westat on reviewing that.
DR. MAYS: Other questions? Let me ask a little bit about the challenge of doing what you did, just so that we have a sense of -- before recommending that others do it, whether or not for example there is something unique about sitting in a university and being able to do a survey like this, because of all the -- you talked about the students, all the labor that is there.
I'm just trying to get a sense, if we were to ask NCHS to pursue attempting to translate several of the surveys, do you think it is feasible? If so, can you say a little bit about how you think they could do it? Costwise it seems feasible. I just want to get a sense of all the work that got put into it. Was a lot of it also just free labor, colleague labor, student labor?
DR. PONCE: I think the translation cost is not the issue, it is the implementation. It was easier to queue up through the telephone if we knew that the respondent -- because it was a telephone survey, versus NHIS, which is an in-person survey. So I think that is a dimension that needs to be considered.
DR. MAYS: That's true.
DR. PONCE: So you bank what the interviewer may guess is a Spanish speaker or suspect Asian, it goes to the Chinese in this queue, and then you say it is not Chinese, I think it is Cambodian. So it is easier logistically in a telephone survey to do it.
I think it would be more expensive when you send interviewers out, unless you go to that one apartment building in Long Beach, and you know for sure. I don't think it is impossible; I think there has got to be tighter analysis beforehand though, on the sample and the household. So I think they could do it if it is close to the census, like right now. But as the census ages and neighborhoods change, it might be more difficult.
The translation as I said, there is a market out there. You could get the highest quality, most competitive bid for a translator, so I don't think that is really an issue, in terms of the money, of getting something translated. The issue is to have an instrument that you feel is high quality, and that a certain number of people agree to it, or think it is possible.
DR. HITCHCOCK: Did you use the Spanish translation of the NCHS does?
DR. PONCE: That was a starting point, but we didn't explicitly convey that to the translators. We use it to look, but we didn't want them to basically copy.
DR. BREEN: One thing I noticed that I thought was very important about the process that you developed was that you had some specific criteria that populations needed to meet in order to have a translation.
I am asking Vickie's question again. It strikes me that at the national level, oftentimes you don't have populations that meet those criteria to make it feasible to do it with a national survey, which actually is smaller than the survey administered by CHIS.
So it strikes me that these kinds of surveys that are more detailed with specific and smaller populations are probably better done at the state level than at the federal level. Do you want to comment on that?
DR. PONCE: -- at one point being an advocate, trying to advocate NCHS to oversample Asians about ten years ago. They said it was impossible. It would be over a million dollars, we wouldn't be able to do it.
Some groups, and I know this has been done for analysis, don't mind -- even though you can't get 800 Native Hawaiians in one year, there is the strategy of pooling data over several years. So I think for groups that usually aren't in these population-based surveys, just knowing that they have a foot in the door, knowing that there is some effort in collecting information. The NHIS for example is a survey of all Americans in the U.S. For the groups that don't show up in these big national surveys, I think there is the feeling they were left out.
So some effort, not a huge effort, but some effort so that we could pool the data at least for several years, could be done. That is one comment from a previous advocate.
Your comment, Nancy, on local is true. I think it is the local areas where that diversity plays out the most. California, we did the Asian oversample because we had a ten percent population here versus a 3.7 percent population nationally. So there is a difference. I don't know if such a plan would be as cost effective with a 3.7 percent population versus a 10.7 percent population. So for smaller groups, I think there is an argument for targeting harder to reach groups in local surveys.
So I guess it is both. It is both wanting to have these small groups be represented nationally, but also saying that maybe there should be some local surveys as well.
DR. HITCHCOCK: It really strikes me that when you start talking about pooling data that it really makes it important that you get during a single year of data collection every single possible person you can from that subgroup into the survey, because if you throw out a Korean family that your interviewer can't translate it, or no one is available to translate it, then every one case that you throw out in the course of a year means that you may have to pool two or three or four years.
We are talking about such small groups. You've got 36 people in some of the groups of people, or some very small numbers. I think the national surveys, if they encounter a situation where they can't do a translation, they just throw up their hands and move on to the next case.
DR. PONCE: No, if they can't, what they do is, --
DR. HITCHCOCK: I have worked with them. I have worked with HANES and NHIS.
DR. PONCE: And I have been surveyed by them. They try their best. They go to the neighbors. There is ad hoc, where they go to the neighbors, they try to find another member who can -- hence, why using linguistic isolation is a criteria.
DR. HITCHCOCK: My point is, every single person counts. I don't know that they realized that in the operation. I'd like to find out whether they do. It's probably a secret. Probably we will never find out.
DR. MAYS: Let me ask you a couple of questions. One is, to what extent will you or have you in this process actually done research on some of the methodological issues? For example, I'm going to go to Dale's question. It is like, we ask this question about health status. We know from the Commonwealth survey, particularly for the Asian population, there was a different response.
Will CHIS for example spend time actually trying to do some methodological -- because you are about to do CHIS 2003.
DR. PONCE: 2003.
DR. MAYS: So is there a plan for any pretesting? Is there a plan to try and work out any of these issues? Is there a methodological -- is there someone who is planning to look at the methodological issues by race and ethnicity? And if not, with resources do you all have the capacity to do it?
DR. PONCE: I can't speak for the principal investigator, Rick Brown, because there may be some plans that are brewing that I am not aware of. I know that it is not -- because we are so busy with collecting the survey, raising the money and getting the survey out for public use, getting the web query system out. Our promise and part of what CHIS is about is that it is data for the people, so we wanted data out as soon as possible.
Then to plan for 2003, I think the research team is really more in trying to operationalize day to day things and making sure that the data files are ready and that the survey is fine, survey will get translated.
So I guess a really quick answer, but again, the caveat is that I don't know if there may be some other activities, is that there doesn't seem to be time to do any methodological review. But Nancy may know of some.
DR. BREEN: I was just going to mention, Gordon Willis at NCI and Elaine Zahn, who is at the Public Health Institute, are working together, and they are planning to do some observational studies. What they are going to do is to basically watch interviewers administer the survey, and note where the relationship is strong, where the respondents balk at the questions or seem not to understand well what is being asked in the question.
I don't know how many interviews they are planning on looking at. It is my understanding that this study is too small to be able to look at a lot of different racial ethnic groups. It may be possible to either look at some racial ethnic groups, or it might be possible to expand that slightly, because the administration and the infrastructure of that study are being set up now.
So it might be possible to expand it. I think you would want to couple racial ethnic groups. Again, some criteria would need to be set up to figure out what would be most useful to look at Koreans, since this was a group with a lot of linguistic isolation. So maybe it could be fit into that.
I know I spoke to Gordon the other day, and he said they hadn't decided which questions they were going to focus on. I know they are planning to do it. They have some money for it, and they are going to move forward, and CHIS is going to go in the field in about a month. So it would need to be thought of quickly, but I think there is some opening for that.
DR. MAYS: My second question is for CHIS 2001. Is the interview up online, in English and in any of the other languages?
DR. PONCE: Not in the other languages, but in English. The other languages are by request.
DR. MAYS: Questions from the audience? Does anyone have any questions? My colleagues?
DR. HITCHCOCK: Thanks again, very nice.
DR. MAYS: I want to thank you again for your time. What we look for in some instances is a business case, to be able to push forward with our recommendations. I think CHIS is one of those models that allows us to push forward, in terms of saying we know it can be done, so let's see if we can't make that -- what you did ten years ago, maybe another couple of years we can reach that point. So we appreciate your hanging in in terms of your advocacy, and now adding the science to the advocacy. So thank you very much for being here.
DR. PONCE: Thank you very much for the invitation.
DR. MAYS: We thank our audience for being here and staying. We will be reconvening again tomorrow morning at 8:30. We hope that you will be able to join us at that point in time. If not, we appreciate the time that you have spent with us today, and if you are here tomorrow, we appreciate you twice. Thank you.
(Whereupon, the meeting was adjourned at 4:02 p.m.)