[A talk given by H. Jack Geiger, MD at a forum April 27, 2004, sponsored by the NY Metro Chapter of Physicians for a National Health Program, in cooperation with the Joanne Lukomnik Fund for Health Care Reform. A shorter, edited version can be found in Forum Report #5.]
H. Jack Geiger, MD
“Equity in access does not guarantee equity in health status. Health is the complex result of multiple interacting factors, of which access to appropriate and timely care is a necessary but not sufficient component. Addressing the many non-financial barriers to care, as well as expanding financial coverage for all Americans, is essential if equity of access is to be achieved”
— Joanne Lukomnik, “Non-Financial Barriers to Health Care,” 1992
The phrase “health disparities” is thrown around in multiple ways without appropriate clarification. As it is usually used, it refers to two quite different things, and it is important to know which is being discussed. In one instance, it refers to disparities in health status — the burden of illness, living sick or dying younger — in African-Americans, Hispanic Americans, Native Americans, and a variety of the Asian subgroups, groups which have never at any point in our history even approximated the health status of the white majority. And we all know that the major causes of those differentials are the social determinants of health, the way we have organized our society on social class and racial grounds — nutrition, housing, segregation, income, wealth — all of those variables that are the profound determinants of who gets sick and who dies young.
The other meaning of the phrase “health disparities” is racial and ethnic disparity in the quality, comprehensiveness, adequacy, and intensity of medical care itself. Again, if you are African-American, Hispanic, Native American, or a member of a number of the Asian subgroups, you have a strong probability of being treated differently, less intensively, with less quality than a white counterpart who is identical in all of the other variables of insurance status, income, education, severity of disease, co-morbidity, hospital resources, and the like.
I will not spend hardly any time on the evidence of racial and ethnic disparities in medical care because, as far as I’m concerned, the case has been proven over and over and over again. There are at least eight major reviews of the peer-reviewed literature that I can count and probably at least several more since I stopped. Over the last four years, my colleagues and I have abstracted and annotated close to a thousand peer-reviewed articles on racial and ethnic disparities in care that met adequate criteria for sample size, control of confounding variables, and all the other things that health care research studies and epidemiologic studies properly worry about. This report, “The Right to Equal Treatment”, and a complete annotated bibliography, organized by disease, are available at the web site of Physicians for Human Rights (http://www.phrusa.org/research/domestic/race/race_report/index.html).
The case is proved. There is no longer any reasonable question that these disparities in treatment exist. One of the interesting things is how far back the peer-reviewed literature goes, and how long it took for this issue to get to the front burner. Gunnar Myrdal and his colleagues, in An American Dilemma published in 1944, observed that “area for area, class for class, the Negro cannot get the same level or access to curative and preventive care as can the white person. Discrimination, both conscious and unconscious, contributes to Negro morbidity and mortality.” In terms of what we know now, that is an astonishingly prescient statement. We know that this exists, and exists in a major way, and has existed for a very long time in every major category of disease and every major category of medical and surgical procedure.
What else do we know? We know that the causes are multi-factorial and complex, which is what we might expect. There is no single smoking gun. A great deal has been written about negative racial and ethnic stereotyping on the part of physicians and other health care providers, and we’ll come back to that — it’s clear that it does happen, but it is, in fact, just one of multiple contributing causes, some of which we know about, most of which we need to know more about (though I hasten to add that doesn’t mean there aren’t things we should be doing now).
Everything I will say in these next few minutes is based on good evidence in the literature. I’m not going to cite chapter and verse but take my word for it if you will. Broadly, we can group the causative elements in this discriminatory differential treatment into three categories, in no rank order.
First of all: patient-level factors. These include patients’ knowledge and beliefs about health and illness, as well as the issue of patients’ trust in the physician and in the health care system. I started out in looking at this literature on the issue of trust thinking, “Well, here is another victim-blaming attempt to say that all minorities are paranoid and it’s their own damn fault.” In fact, the evidence is compelling. A recent survey of African-American respondents — we need similar studies of Hispanics and other groups — found a significant plurality of African-American respondents said that their own physician was capable of experimenting on them without their consent, their own physician was capable of giving them potentially dangerous experimental drugs without telling them, even in the face of evidence that continuity of relationship does mitigate this kind of mistrust. The message of the data on mistrust is that this is yet another element that physicians, other health providers, and institutions and organizations of health care need to be aware of and need to address.
So, patient-level factors include knowledge and beliefs about health and illness, trust and mistrust, as well as affordability (the cost out of pocket, independent of the adequacy of health insurance), fluency in English, health care literacy (knowing something about your disease, disease in general, and your recommended regimen), and what are called “preferences,” which is kind of a code word for “I refuse to do what the doctor is recommending.” There is good evidence for the existence of these preferences. However, in the studies that have looked at it prospectively and elicited the evidence of patient preference as a contributor to differential treatment — in no case anywhere in the literature has that factor been anywhere near sufficient to account for the magnitude of the existing disparities. So this is a contributing factor, and again it’s something that health care people need to know about, but it is not the main element of what’s going on.
Now if we go to provider factors. Let me start out with negative racial and ethnic stereotyping. And the classic study — there aren’t a whole lot that have looked in detail at this, this is one of the areas that needs more work — the classic study of negative racial and ethnic stereotyping by physicians is by Michelle Van Ryn and Jane Burke (Social Science & Medicine, v. 50, March 2000, pp. 813-828) in which prospectively — which means they could talk to both the patients and the decision-making cardio-thoracic surgeons after angiograms had been performed — they found a profound differential by race in the decision about whether to perform angioplasty or bypass grafting, all other things being the same (how many lesions, where the lesions were, income, health insurance status, co-morbidity, all the rest). They were able to determine, first of all, whether patient preference was a factor. They were also able, subtly and cleverly, without the decision-making physicians knowing that race was the area of inquiry, to determine some of the attitudes that dictated those decisions. And it turned out that the decision-making clinicians, usually on the basis of an average 12-minute interview, had decided that “oh, this patient” — the African-American — “this patient doesn’t have the energy to do cardiac rehab, isn’t smart enough to stay with rehab and really understand all the things that he or she needs to do to manage the disease, doesn’t have a good social support system, doesn’t have a stressful job.” You will recognize the relationship of these decision-driving stereotypic views to classic stereotypes provided by our culture about African-Americans — and they have their analogues, of course, for other minority groups — that African-Americans are lazy, less energetic, less intelligent, etc., etc. This is a double violation. It’s a violation of what the society professes to be about, but it’s also a violation of medical ethics and the commitment of medicine to treat everybody equally.[1]
I also refer those interested in the literature to one other study, from the Cleveland Veterans Administration Hospital, again in the area of cardiovascular disease. In this study the cardiology fellow worked up the patient and presented all of the clinical and cath lab and history and echocardiograms and other findings to a panel of cardiologists and surgeons. They then discussed each case — some 900 in all -- and reached a consensus as to what should be done — medical treatment, angioplasty, bypass grafting. And in those 900 cases, there was complete presentation of all of the data, but no mention of race. And when that happened, there was no difference between blacks and whites in who got angioplasty or bypass grafting. Removing the one variable and the disparity disappears is pretty powerful evidence for what is going on.
How does this play out in less dramatic circumstances than bypass grafting? Let me mention another article. When you have the opportunity, call up the Journal of Health Affairs and look in the March/April issue for an article called “Learning About Medicine and Race,” a five-page article by a young African-American junior faculty member at Emory University (http://content.healthaffairs.org/cgi/content/full/23/2/220). He wrote about an incident he remembers as a medical student, in which a working-class African-American patient who — quite sensibly given his physical state at the moment — refused to go along with a recommended diagnostic procedure, and was then classified as crazy by his own physicians and referred for psychiatric treatment until the student himself was able to intervene. It is in manifold subtler ways that the stereotyping can have an effect.
Other provider factors include clinical uncertainty, stemming from lack of cultural competence or linguistic competence, competing demands and cognitive overload (leading to shorthand thinking and stereotyping), and whether the physician has a patient-friendly style or an authoritarian approach when interacting with patients. In all of these, there is evidence of differentials in care.
And finally, there are institutional or organizational factors that contribute to racial and ethnic disparities in the quality of care. We all know about the location of physicians’ offices, the neighborhoods where resources are and are not. These factors also include interactions with office staff, the experiences that minority patients report. It’s not just the physician. It’s the triage nurse, the clerk, all kinds of other people in the health care apparatus who may be contributing in significant ways. So what might be called the organizational culture of the health care-providing institution is important, for instance, the reading level and cultural appropriateness of written or other materials that are provided. I taught for a long time on the faculty of Tufts’ Medical School, which sits right on the edge of Boston’s Chinatown, and I was there six years before it occurred to Tufts that it might be a sensible thing to put up some signs in Chinese. There’s the matter of how co-pays and deductibles are handled. And there’s the question of whether a quality oversight and improvement program is in effect at the health care institution.
And I would add community-level factors. Are there safety net providers in the community? What are the levels of insurance coverage in that community? What is the level of what might be called “community trust” in the health care system?
We need to know much more about the variations in disparities. If we want to find effective interventions, we need to know the relative contribution of each of the multiple factors I’ve mentioned. If we don’t know what’s really going on in a given set of circumstances, our intervention — however well-meaning — may totally miss the mark. We don’t know what works for cultural competence training, which is all the rage at this point. Many of the early cultural competence curricula are simply lists of the strange behaviors of other people, stereotype-reinforcing items one after the other, never addressing the idea that medical and other health professional students bring two cultures to their work: their own culture, including biases conscious and unconscious, and the culture of medicine. Cultural competence begins with self-examination, rather than the idea that culture is something that’s just an attribute of other people. We have no knowledge as to whether training in cultural competence makes any difference in outcomes, in the quality of the care that is provided to different groups.
We don’t know very much about how to intervene in physician/patient communication styles. We do know that it is better when there is physician/patient racial or ethnic concordance. We do know, probably sensibly, that African-Americans and Hispanics, if they have the chance to vote with their feet, will almost always choose preferentially for physicians and providers of their own racial or ethnic or language group. These are just some of the things that need intensively to be worked on.
And we don’t know very much about community variations. Why is it, for example, that African-Americans do as well or better than whites at getting pap smears but much worse than the white elderly at getting flu vaccines or pneumococcal immunizations?
But we don’t have to wait to do some things right now. First of all, we have to see to it that all clinical data in every organized setting of care is recorded by race and ethnicity. Minority group respondents are understandably skeptical of doing that, anticipating that the data will be used for some nefarious purpose. However, there is no way we can monitor what is happening by race and ethnicity, in terms of the quality of care, unless we are collecting data in this way, and it is legal and even federally-mandated to do so. A lot of physicians and a lot of hospitals still somehow feel (inaccurately) that it is not legal or that they will get in trouble, but it is essential and critical.
Secondly, we need to push every organized provider of care to have a quality improvement program and to use practice guidelines for all their patients. There are several reasons for this: (1) There is evidence that it works. There are now about a hundred community health centers, to give one example, that are part of an effort called the Health Disparities Collaborative, in which quality improvement guidelines and a set of related mechanisms have reduced disparities in treatment, as well as having a real effect on outcomes. One of the reasons quality improvement is useful, aside from the fact that it does indeed seem to work, is that it is something that every physician almost has to subscribe to, and it is less challenging than going head-on with many physician groups about the issues of negative racial and ethnic stereotypes. So quality improvement and the use of practice guidelines is something that we can do right away.
We should also be working on patient activation and assertiveness. Training patients, in effect, to be more assertive, to be more demanding, particularly patients from the minority groups. And there is evidence that it makes a difference. It changes the nature of physician-patient communication and has an effect on what happens, what is prescribed, and on the quality of the follow-up.
And that leads me to my final point: I don’t think change in all of this is going to happen if it’s all top-down. Everything that I’ve mentioned has been pretty top-down. This effort has to also involve community-based advocacy groups and community organizations.
We’ve been at this for a decade — we, collectively: people researching in this
area — we have obviously a lot of work yet to do, but this is a problem that
we can and will solve.
[1] I should note that physicians are worse than the general public. There are Kaiser polls, and in response to the question “Do African-Americans get about the same access to and quality of medical care as the majority population?,” a majority of physicians say “Sure they do,” unless they are African-American or Hispanic physicians, in which case they respond with some polite version of “bullshit.” But the physicians’ responses to this question are worse than the general public’s. The general public is more aware of, or willing to concede, that there are these differentials than are white physicians in the United States.