jewish white guy finally feels marginalized. [laughter] >> well, i hope it was by design but i suspect it was just an accident. i'm a physician in new york city where i do health services research. my research discipline uses a hybrid of techniques from economics, epidemiology and statistics in order to gain an understanding of the healthcare delivery system. .. relevant.

decade ago, my colleagues and i wondered if a higher mortality rate from cancer seen among blacks could be due to the receiving less a effective treatment, specifically in the setting of cancer. we chose to study a single cancer procedure to address our questions. surgery for early stage one cancer. we focused on this procedure because it treats the no. one cancer killer, you lung cancer, and is enormously effective. in an analysis we published in the new england journal in 1999, a decade ago, prior to the report that has been referenced many times, we showed that in medicare, blacks received this surgery 15% less often than bites with the same diagnosis. we showed that this was not due to greater morbidity amongst blacks or differences in socioeconomic status, we showed that we believed the achievement gap was the explanation for their survival lot comes.

the study is memorable for me, one of the first major analyses published in the medicare database which is a cornerstone of cancer care and one of the few studies that demonstrates that treatment gaps were important in terms of disease outcomes. prior to that, treatment gaps were illustrated without a link to outcomes. we were unable to determine why treatment rates were lower for blacks. our study was not designed with that question in mind. and data was insufficient to address the granular patient little question. we had used national data covering many areas and tens of thousands of patients, we had little individual level information. other disparities, notably the opposite. sometimes covering a few patients and doctors in a single practice setting from which a lot can be learned about that setting but less about the universe of care settings. the fallout from the publication was a educational for me.

the number of pundits stepped on top of our findings to use as a platform to decry the house care -- health care system is racist and doctors as racist. the new york times week in review carried on their front page article about their study which was titled not just another case of health racism. it is framed in my bed room. too many people concluded too quickly that the explanation for our findings was that doctors discriminated against minority patients. the new england journal of medicine reviewing the report on equal treatment, the invocation moves the problem. our research group studies the potential for another explanation that dr. chandra has referred to which blamed the system rather than the doctor. we hypothesize the reason blacks

receive lower quality care than whites could be because they went to doctors who for a variety of reasons for less able to provide the care they routinely received by whites. this could be the case because doctors were less well-trained, less will resources or less knowledgeable. a few years later in 2004, we published another study in the new england journal providing evidence supporting our explanation, we documented 2 conditions, the key precondition for our hypophysis existed. we were able to show the care black patients was clustered among a small group of doctors, it took 20% of primary care doctors in the u.s. to account for 80% of all care received by blacks. whites were different, their air was mostly with other doctors. the doctors at the level of patient visits were different. we asked the question, violet typical medicare patient who is black, what are the features of my doctor compared to a typical

medicare patient who is white. week found doctors looked different. a black patient was less likely to have a doctor who is board certified in their primary specialty. that was important because having board certification has been shedding decades of research to be a key predictor at delivering high-quality care. we found a primary-care doctors who treated blacks and fewer resources to direct care of their patients. they had our referrals for all their patients, they had to -- to the local hospital and getting imaging tests. more interesting, the financing of the practices was different. blacks went to doctors who more often provide free care to medicare patients and the effect was they had lower revenues per patient and therefore less resources to support their practices and the other caregivers in them. they were more likely hurried, something more recent research has shown. around the time of this paper and over the years since, these findings have been reproduced numerous times including by dr.

chandra and his terrific colleagues. doctors, hospitals, surgeons, managed-care insurance companies, lower quality overall disassociated with more black patients and your route -- whites in the care setting. will get more detailed aspects of practices to treat large numbers of minority patients and estimated the impact of the payment rates for medicaid itself were sizable contributors to access problems and lead to shorter patient visits too. my colleagues have some unpublished findings i can give you a top line review of. we are finding a medicare patients, the important predictors of getting lower quality care are, first and foremost, socioeconomic status or the regional or local area economic climate in which you live. how could the quality of care that your doctor gives his or her other patients, meaning his or her white patients, we are unable to detect any consistent evidence that doctors are treating black or white patients differently.

findings are consistent with the plausible hypothesis that care for blacks is of lower quality primarily because blacks are accessing part of the system that is poorly functioning. little if any of the treatment appears to be doctors singling out minorities for lower quality care. neither my colleagues nor i take the challenges posed by this alternative explanation for health disparities lightly. in ways, they will be harder to ameliorate. but the payoff will be more durable and robust. you have asked me here to talk about health disparities and what our research found, research in a social context in which many people arrived early at a conclusion that discrimination, conscious or unconscious, lay at the heart of treatment disparities, our work has provided a different explanation in which we have a poorly distributed health-care system, in which the lowest quality resources are in the neighborhoods with the most needy individuals. if correct, the mechanism suggests an approach that targets these high risk areas will be the best way to improve care and outcomes for patients

in them. thank you for inviting me. i look forward to your questions. >> i would like to thank all of the participants. this has been a fantastic presentation. i would like to open up the floor for questions from commissioners. don't jump in all at one time, folks. commissioner arlan melendez. >> thank you for being here today. i want to ask a question of dr. graham. as far as data collection, you mentioned for the native american population, one of the issues we always had was the census and there were tremendous problems with the statistics, especially in 1990.

as far as the population of native americans, in that sense this, what do you think, how does that skew the statistics we have, not just native americans, but i am really interested in statistics. >> the other issues, i am here today, instrumental in key issues on data collection. aside from the census bureau, one of the major challenges was the tools that we used to collect data on health statistics. the department of health and human services, state organizations and other agencies, the health status of our country. many times we miss important sub sectors such as the native american population as well as the native hawaiian and other

populations. what that spells out, the data points, instead of actual statistics on those specific populations. we are taking that issue very seriously from the leadership standpoint. we have the data council and all of these infrastructures, we're taking a serious look at how can we capture those demographics? what you don't measure, you don't see. this idea of capturing data in terms of, what is the true incidence of heart disease or cancer, these diseases, the native american population. to reach folks and have them answered the data questions and

be able to put that within some of the data collection systems we have. one of the strategies we are implying is the targeted study where we specifically have specific efforts to reach folks and capture the population. >> if i can just follow up, up one of the things that was of interest to me is as we had this discussion, some of the different conclusion that we reached in terms of health disparities, what is a highlight for me is the need to do more research. quite frankly one of the concerns i had is there is a group, there is a push to eliminate collections of racial and ethnic datas in looking at health outcomes. i believe that will be extremely dangerous. in order to try to find out the

specific reasons why we have these decisions, these wide variations. i am concerned about that. >> first, commissioner todd gaziano. >> my immediate family and their spouses have 8 physicians, i am the only non physician. i am concerned about the existence of the disparity, but i will be glad to yield to h h f its particular share in addressing some aspect of that. our special jurisdiction here is somewhat limited, and that is to focus in or try to focus in on

the causes for which there is implicit or explicit racial discrimination. i will try to focus a few questions, by basic question, the last 3, who tried to get at that, and as i understand this is a terrible overgeneralization. you have testified, there are genetic differences depending on the disease, there are behavioral reasons, socio-economic reasons that have a big role to play before there's any interaction with the health care system. some of these are epigenetic, related with those other factors. and then the impact of the health-care system, we have

heard that some of it, there is at least a concern, some research suggests it has to do with the clinical experience and bias. or the quality of care. and i wondered, particularly the last draft, dr. chandra, you put some of this in words. what percentage, pick a disease if you know it, or some diseases are caused by factors -- that don't really have much to do with the health care delivery system. as for what percentage is related within the health care delivery system to possible bias, what percentage, even if it is a wild guess, just to give

us a sense to the research. >> let me answer that question in the context of the paper i rode with my collaborators that we published in the journal circulation. here's what we found in the circulation paper. black/white differences, in 90 days survival after a heart attack, what we found is if you look at the role, from lighter quality in racial disparity in 90 days survival, provider quality explained 60% of the variation, they are still 40% that you could attribute to the clinical experience or the accused face of the tribute, some of that 40% was the fact that within a particular hospital, treated by different

groups, we weren't, in that study, able to drill down completely, but when you look at something like 3 months survival there is an enormous rule of the care system that you would treat it act, which becomes more important when you narrow the window of for which you are looking at out the differences. if you change the outcome measure to what is the role of geography, provider differences in a 30 day mortality differences, a role of the hospital explains 100% of the treatment disparity which is what you would expect because the hospital is very important for the first week. once you are discharged from hospital, a bunch of other factors start to affect racial differences in mortality. >> it is hard to separate by

disease. we don't think that difference in cancer incidences which are 10% adjusted between blacks and whites, 10% or higher, have anything to do with health care system, we have a preventive cancer measures we distribute, smoking, sitting outside. the paper alluded to, i regret that it is not yet published because i could talk in more detail when you could review it. we look at the delivery of 6 preventive services. we were able to assess that on average, half of the gap, socioeconomic status, different patients, another 30 plus% or 40% was within provider differences so that blacks were going to doctors treated more of them and provided lower quality care and there was a residual that had to do with morbidity and things like that. a few% in one direction and a few in the other, were due to

differences between black and white patients by this and not to which we would argue are statistical anomalies. diabetic eye exams, monitoring lipid levels and things like that. we are not seeing any evidence in the doctor's office. that is where patients are going. >> thank you, mr. chair, i commend the panel for a fine presentation. i have 2 small questions. the first one comes from research that has been done in the asian american community. there are a lot of perceptions as far as health education benefits, most of them tend to be on the bell curve in terms of distribution of income, access, what have you.

this goes to immigrant populations as well. to what extent is linguistic and cultural competence, has an impact on equality and treatment of health care for those who need it. >> i will take the first one. as i mentioned earlier, that has definite if not significant tremendous impact in terms of the factors we mentioned here today. in terms of presenting some of the data, we alluded to this issue. implicit with this ability to provide cultural linguistic care, a patient understands what you are talking about, not only that the patient understands, you understand the patient is important because it is something that is harder to capture significantly in the

hard-core outcomes that peter and others have spoken about. other parts of the data, we see these intrinsic, implicit factors. i want to allude to the other thing you pointed out, which is this myth that asian-americans are uniformly healthy. you see specific health disparities in populations of asian-americans that are troubling. it is up to us to make sure we educate the broader u.s. population. >> i would say the data that the system has his overwhelming, we define the issue. in research it is easy to define a problem by how you set the

premise forward. i served on boat board of grady hospital in atlanta. we have a tremendous problem because we have a growing hispanic population. and that issue is one that we are dealing with. we don't have enough interpreters to speak to the spanish population that we are serving. we have controlled by leaders in the hispanic community that many don't come to the hospital because they don't feel it is a welcome environment. that is a factor that way is on the care the individuals received. in the new england journal of medicine, studies from the university of california showed that black or hispanic positions were 5 times more likely to establish their practices in

black or hispanic areas. the help data in this areas. a one to make sure this is bias in the system. when someone looks at a lot of caveat, you may not find differences if you look. there are gross discrepancies in the health-care system that have an impact on people's lives. as we are working to improve the health of americans, that we do not dispose of those factors that have a system that really makes it difficult for a growing segment of the population to receive the care that they need. this is important for everyone because one of the challenges we face as a society is a growing

health burden as well as problems in our educational system, that is going to erode the strength of our country. so we have to look as broadly as we need to, but not dismissed those systemic problems that interfere with the ability of people to receive care, on the basis of income, education and bias. i happen to be old enough to have grown up in a south where my family drove 41 miles to see a black physician rather than go in a room that said black patients versus white patients. i define that as bias. my parents defined that as bias that they would not accept. i think we have to be careful

not to try to dismiss the factors in our system, that do impair the ability of our patients to receive the care that they need. >> mr. chair, my second question was more along those lines. i don't tend to think as much, certainly i think a lot of you a firm that to the degree that at the micro level, there is individual decisions going on in some physician's mind, i will not give this person x or y but the fact that these disparities exist and they are documented and they seem to be consistent over time, whether it is with native american healthcare, whether it is with the monk populations, asian-american with

latino or african-american, as secretary sullivan, i call you by your highest title, secretary sullivan, maybe he thinks -- >> one of his other achievements is higher. >> i go with straight protocol. secretary sullivan seems to indicate that there's still something of this. the question is not whether someone is saying i am going to stick it to this group or that group but nevertheless, it comes into resource allocation, where people decide to put their resources, in what neighborhoods, in what clinics, in what areas of further study for research. from your point of view, where is it in that decision track that we still keep coming up

with. there's the 60% we can't explain or what happened, that to me is the core of what this hearing is about, because hopefully we are mainly pass the times when secretary sullivan doesn't want to say whites only, blacks only, there's still something wrong. the question is what is it? in this health care debate going on, tremendous change is going to occur. how that change will impact, what we are struggling with is going to be very important. >> you made a good point. one of the things i would hate to walk away from this meeting, the health-care industry is somehow biased towards taking care of patients which it is not. the overwhelming health care providers in this country try to

provide the best possible care and the most equitable care to all of their patients. i have had the opportunity of working in 6 different geographic regions and have met nothing but the most ethically and morally highest individuals in this country. having said that, we have to understand we all come into this profession, health care profession, with our own life experiences. decisionmaking is based a little bit on that life experience and we have to make judgments. some of those judgments on our interpretation, our ability to figure out whether our patients will be able to carry out certain treatment plans, or be able to understand, some of the decisions that are made are based on some of those types of mindsets that go into this interaction. one of the things you have to be

concerned about is the interaction between a provider adaptation, according to various studies, between 6, and 9 minutes. when you add the issue of language, it creates more complexities. having an interpreter doesn't solve the issue. one of the things that was find out initially was family members were serving as interpreters and many time, patients don't want to tell their family members what is going on so they have phone calls. those were not found to be adequate. that speaks to a couple issues. one has to do with workforce diversity and this is an import aspect that you need to look at. the second deals with poverty and educational level. just by being poor, you are likely to live 6 to 9 years less, those determiners are very

critical. as the uniformed answer, we need to look at literacy as the bigger picture rather than language because many patients -- come to our offices and facilities with not the best literacy understanding especially when you talk about medical. so we have to look at the complexity of the health-care industry and put the resources and all of this -- other areas that impact the care of individuals rather than focus on the patient or provider information we have to utilize. >> commissioner harry it? >> i think dr. satel -- >> the answer to your question is we going to -- we don't know what makes up a lot of that noise. at that fine grain level, you need prospective studies and