Skip to content

Race Based Medicine Essay Topics

Mary-Jo DelVecchio Good, Ph.D., Professor of Social Medicine Cara James,B.A.

Byron J. Good, Ph.D., Professor of Medical Anthropology

Anne E. Becker, M.D, Ph.D., Assistant Professor of Medical Anthropology

Department of Social Medicine, Harvard Medical School


Racial disparities in medical treatments and in health status have been documented in numerous studies over the past two decades. Epstein and Ayanian noted in a recent editorial in the New England Journal of Medicine “there is little evidence that racial disparities in medical care or in measures of health have substantially diminished” (2001:1471). Gary King, in an insightful theoretical analysis of this research, argues that explanations of racial differences in medical care and of participation rates in medical research are grounded in institutional racism and in the professional ideologies of medicine and health care systems that lead to power imbalances between minorities and medicine's elite professionals (1996:35). King identifies three phases of research in this field, which have relevance to our project: (1) initial “exploratory research,” which documented differences between blacks and whites in medical care, utilizing quantitative data; (2) “contemporary” research, which focuses on coronary artery disease (CAD) and other specific diseases, using rigorous methods to investigate causes of disparities in treatment; and (3) most recently, “an incisive period in which researchers attempt to combine theory, methods and policy considerations” (1996:36). King argues that to understand documented differences, one must come to understand covert as well as overt racism and the multifaceted dimensions of institutional racism in medical and health institutions (1996:43). His work and that of numerous other researchers who have documented disparities in health and treatment for ethnic and linguistic minorities, in addition to examining black-white differences, raise questions about how we might best understand the mechanisms – attitudinal, structural, institutional, ideological -- that produce disparities in medical care and in health status.

Our initial study questions for the IOM Group on Racial and Ethnic Disparities focused on the culture of medicine. How does the culture of medicine, the training of medical students and residents, and the organization and delivery of health care affect patient treatment in such a way as to produce obvious and documented disparities in therapeutic action? Whether it is treatment for cardiac disease, asthma, kidney disorders, or mental illnesses, some Americans receive less than optimal or standard best medical therapies. Two decades of research have documented that whether bounded by ethnic or racial identities, immigrant status, English language fluency, educational attainment, poverty, low SES, or urban/ rural residence, minorities and the poor re ceive less care and poorer quality care than their middle class and educated compatriots. The committee invited our group to address the question, “how could well-meaning people (health care providers) provide inequitable care to minority and non-minority patients?”

This question and the Committee's initial larger query about the culture of medicine and its contribution to patterning disparities in medical care and treatment have proved daunting to address. Empirical studies on medical training and racism appear scant in our literature searches. Our own research (B. Good 1994, M. Good 1995, Good and Good 2000) on the socialization of medical students and on the culture of medicine, while suggestive, has not directly addressed disparities in care granted patients. However, in our studies over several decades, we found that “the medical gaze” soon becomes the dominant knowledge frame through medical school, that time and efficiency are highly prized, and that students and their attendings are most caring of patients who are willing to become part of the medical story they wish to tell and the therapeutic activities they hope to pursue. Nevertheless, we identified no clear relationship between medical hierarchy, the culture of training, and professional ideologies, that would readily explain patterned disparities in care by race, SES, and ethnicity (such as the study group's suggestion of a modeled preference for private rather than public patients although clearly such preferences may exist at many institutions). We contend that such relationships are multidimensional and subtle, and that to identify the production of disparities in care requires considerable additional critical observation of our institutions of training and care. For example, differences in value of public and private patients may be conveyed not only in direct but in indirect and subtle process (see below, for a discussion of Whaley [1998] and his concept of aversive racism). In our own academic community, such distinctions may be demonstrated through a common but not hegemonic hierarchy of value. For example, for many physicians, albeit by no means all, practicing at a community clinic may be regarded as less prestigious than practicing at an academic medical center, where private as well as public patients receive care. The diversity of medical professional communities also allows for “the saint” or hero physician who is devoted to the care of the underprivileged or the poor. Professional careers are made in community oriented medical service and in teaching “cultural competence” to one's medical students and residents. Nevertheless, even the charismatic hero physicians, leaders in social medicine internationally and nationally (see Farmer 1999), and the less publicly known academic and community physicians, commendable for their commitment to improving the quality of care provided to the underserved or to ethnic minorities, work within a biomedical knowledge frame. Although tempered with a social medicine perspective and interpretative sensibility, these physicians also employ the medical gaze in their daily clinical work and practice. Thus, we do not wish to underestimate the power of the medical gaze, and the biomedical sciences which inform it to shape patterns of care and thus differences in care as well. In such a hierarchy of valued knowledge, psychological and social data are often regarded as inadmissable evidence by students who are learning to hone their case reports and presentations to focus on the essence of “what medicine cares about” (M. Good 1995). The social data may work in ways to produce patterns of care not otherwise critically examined, even for more practiced and socially concerned clinicians.

This simple picture of “valued knowledge,” while the dominant model and underlying ideology in early medical training, hardly captures the complexities of contemporary medical education. In contrast to a singular biomedical model of training which gives scant attention and value to the social aspects of medicine, many American medical schools over the past quarter century have increasingly incorporated curriculum materials on diverse patient cultures; in addition, students have often been encouraged to engage in activities oriented to caring for under served, poor and minority patients. Our observations of and engagement with medical education since 1976 have exposed us to social medicine sensitive curricula even as we have contributed to it (Good and Good 1980, 1981). Examples include the University of California Davis, which in the late 1970's and early 1980's, developed cultural courses that addressed “the health needs, beliefs, and practices” of California's various ethnic populations and immigrants (Asian, Mexican, Mexican American, Chicano, Vietnamese immigrants and other refugees from Indochina – with a notable lesser emphasis on African Americans). Harvard Medical School's commitment has flourished in the past two decades, increasing curricula in medical anthropology and social medicine, although the school has had a long tradition through the informal curriculum of encouraging students to participate in international as well as local community programs providing health care for the poor or marginally served ethnic groups. A third program, and among the most impressive, was begun in the late 1960s when the University of New Mexico established its medical school in 1968. The highly committed, state sanctioned and legislated programs of the University of New Mexico School of Medicine's teach students about the state's diverse populations and the health care needs of rural as well as urban ethnic communities. The school mandates practice training in underserved areas to provide care to the American Indian (Hopi, Navajo, Zuni and others) and Hispanic rural populations (Good and Good 1998, Report on Innovation in Medical Education). Many schools have produced formal and informal curricula which on paper appear to be promoting cultural competence. Rarely, however, does medical training focus on the culture of medicine itself; rarely do students have the time or the formal sanction to critically analyze the profession and institutions of care to examine how treatment choices, quality of care and research practices are shaped; how medical culture may produce processes that evolve into institutional racism or aversive racism in clinical practice (King 1996, Whaley 1998). Theories about how professional elites and the imbalance of power produce institutional racism, such as King's, call for empirical documentation.

Why do racial and ethnic disparities continue to exist despite the two decades of documented research, of educational efforts to teach medical and health professionals “cultural competence” and social medicine, of programmatic efforts to attend to health needs of underserved communities, and most recently of NIH/NCMHD and previous governmental efforts to redress the inequities in medical care and health status? Are educational and research programs ineffective in changing clinician behavior and institutional and professional culture? Does the culture of medicine as exemplified in the medical gaze and its underlying ideologies and political economy of what constitute legitimate medical knowledge, bioscience, and appropriate medical decision-making, too readily exclude patients whom clinicians assess as likely to pose “problems” and compromise the efficacy and efficiency prized by medical world? Do professionals' concerns about patient compliance and community and patient trust at times serve as justifications for their employing different approaches to treating minority patients, and thereby and perhaps naively, to their providing lesser quality of care? Or are disparities in care ingrained in the social and economic inequalities that are rife in our larger society, feeding inequalities of treatment in America's health care system? Clearly differences in insurance coverage influence the kinds and quality of care patients receive. Do patients contribute to these disparities in care and how? We contend that multidimensional processes are at the root of different types of ethnic and racial disparities in health status and medical treatment. These processes are structural, economic, environmental, political and attitudinal. Individual behavior as well as institutional culture and practices are implicated.

In the remainder of this essay, selected findings from studies documenting disparities are presented. In addition, exploratory qualitative data from recent interviews with physicians in training and with faculty attendings are introduced to suggest possible directions for a critical analysis of the culture of medicine and the political economy upon which it is based and from which it is produced. A cultural analysis of bias in mental health care is presented as one model suggestive for future research in other domains of medicine. In this case, differences in diagnosis and treatment by race and gender in psychiatry are examined, demonstrating how cultural and social analyses contribute to an explanation of racial and ethnic disparities in care and treatment. The conclusion draws on these materials and the wider literature documenting disparities in health care with the aim that future research findings and interpretations may contribute to formulating policies to redress disparities in the quality of care.

The Culture of Medicine: Insights from Physicians in Academic Teaching Hospitals

Clearly the financing of health care coverage, type of health insurance or accessibility to government programs, as well as lack of coverage, are relevant to understanding what happens within health care institutions and in the intimate exchanges that we characterize as clinical interactions, and how these interactions result in decisions that produce racial and ethnic disparities in treatment. The Harvard Medical Practice Study analyzed over 32,000 records from New York State hospitals located in metropolitan/ urban town / suburban/ rural areas. The project found that patients who had no health insurance and who lived in poor urban areas were more likely regardless of race or ethnicity to experience an adverse event (Weiler et al. 1993). Where one receives health care, including how one enters a health care institution, via an ER or via a controlled appointment process, influences the type and scope of care provided. It may also influence how health care providers, from nurses and physicians to medical students and technologists, interpret who their patients are, what life experiences they carry with them, and what problems they may cause for those who will be delivering care. These processes, whether stereotyping, prototyping or profiling, clearly have consequences for treatment choices and medical decisions. Disparities in care and differences in therapeutic actions, regardless of provider race in certain cases (Chen et al 2001), rests in part on these larger social processes within our complex medical institutions. The following discussion examines findings from our pilot interviews and seeks to raise what we believe are some reasons disparities persist.

Our past research on medical education and the socialization of medical students, carried out at Harvard Medical School in 1986-1991, addressed how medical students and physicians in training learn to see, present and write up patient cases (B Good 1994, M. Good 1995, Good and Good 2000, and other jointly authored articles -- see bibiliography). Students come to embody the medical gaze as they learn to see what is relevant data and to speak the language of medicine. Early in training they enter the molecular worlds of disease and therapeutic interventions, the world of medical practice and medical culture. They also learn socially acceptable behaviors: when to speak, how to listen, what is relevant to the clinical task. Students struggle to learn "what medicine cares about.” Students are also socialized to attend to social and economic issues in health care; some embrace the 'social medicine' perspective even as they come to speak and be in the molecular medical world. Those who do so incorporate into their studies projects in international health, urban health, and volunteer work providing basic health care for poor or immigrant or minority populations. These social medicine projects become less central in students' education as they move into the clinical clerkships and become responsible for the clinical care of patients and for decisions of therapeutic consequence. Other research on medical education from the classic work of Robert K. Merton's Student Physician (1957) to Howard Becker's et al. Boys in White (1961) to Renee Fox's many essays (1988), to more contemporary projects such as Fred Hafferty's Into the Valley: Death and the Socialization of Medical Students (1999) and Kenneth Ludmerer's Time to Heal (1999) document similar experiences, suggesting historical depth and continuity in the culture of medical education. This continuity of medical educational culture persists despite a sea change in the gender and to a lesser extent racial and ethnic profile of medical students. In addition extraordinary developments in medical technology, biomedical science and the political economy and financing of medicine and delivery of health care appear to be subsumed into this culture and ways of learning medicine.

Scholars interested in narrative analysis such as Katherine Hunter (1991), as well as many physicians interested in medical narratives, explore ways physicians present and speak about patients. Narrative forms of the culture of medicine are ingrained; they have historical depth and substance and make for continuity of professionalism despite changes in practice environments. They are reinforced by the modeling that occurs through the hierarchy of medicine, through interactions between students and interns, interns and residents, and residents and attendings. What is important in medicine is learned through daily interactions with peers and with the hierarchy, and students are rewarded as they come to behave as competent reliable and responsible clinicians that have learned appropriate professional behavior. How does this professional socialization affect physician behavior that it impacts the treatments offered to patients of different social groups?

First, the culture of medicine emphasizes the dismantling of patient life narratives, and the reconstitution of patient concerns and experiences of illness and associated social context, into medically meaningful narratives of relevance that allow physicians to determine a diagnosis and formulate plans for therapeutic actions and procedures. In our research, patient life issues were often regarded by students and attendings, as “inadmissible evidence;” senior clinicians modeled for juniors how to streamline medical narratives and to edit out data irrelevant to the clinical decision or task currently at hand. Patients are not ignored however. As students and residents mature as clinicians they learn to create clinical narratives for patients; these clinical narratives are biomedical stories through which physicians explain to patients disease diagnoses and processes, therapeutic options, and treatment courses and goals. Good doctors engage patients in these clinical stories, teaching and guiding and helping patients own what is happening. Some clinicians describe this very simply as empowering patients, others reflect little but work hard at fitting patients into the medical world of treatment, helping patients understand and accept what may be difficult treatments for frightening and life threatening diseases. These narratives smooth the working of what one of our junior colleagues calls “the medical machine” (personal communication, Herskowits 2001).

Disruptions in the Medical Machine

Recent interviews with attending physicians, residents, and medical students in the Boston area suggest several ways the medical gaze may lead to disparities in health care. Interviews were carried out with physicians from Euro-American, Asian and African American background and with men and women. A second year medicine resident at a major teaching hospital who had recently completed a Ph. D. in medical anthropology made the following observations as we explored what might lay behind the reasons for disparities in health care treatments for minorities, immigrants, and lower class patients. She had been thinking about these issues, in large measure due to publication of recent research on health disparities and to the publicity of NIH minority research activities:

"People who don't fit into the medical machine" are ones who may not get offered the latest therapeutic interventions. Here at [x hospital] is a medical machine -- we are all cogs in it, not just the docs but the patients too. and the more we fit into our role, the smoother the machine runs. In internship -- it is relearning to listen to patients without listening to them. It is painful, because it is the opposite of anthropology. You learn to do a better job by not listening to your patients. … We become different types of subjects, disciplined (after Foucault), to fit into the machine. Or Fordist. The machine model breaks down when patients don't get referred into high tech medical therapies, the most cutting edge of medicine. When physicians experience difficulties in interacting with patients, it befuddles the doctor, and derails them. In ER shifts, there is the discipline of time, and when a patient derails you, it is glaringly obvious. In the ER, you have an immediate problem --what do I do with this person -- the faster you make a decision the better you are as an ED doctor; it is a different yardstick to figure out what is troubling them."

Patients derail physicians when they present with what an attending described as "socially complex problems." The notions of the medical machine, of derailment and befuddlement, are relevant to explaining why disparities exist. The comments of a cardiac specialist, who has practiced in community clinics as well as in a major teaching hospital highlight similar issues. He remarked: “One needs to attend to the more mundane aspects of doctoring. In today's practice environment, "we need cooperative patients because of the tightness of time."

Another attending noted how language problems and family issues can disrupt practice flow and influence choice of therapeutic options. Her work with patients at a community mental health clinic became most difficult when she found that many of her patients of color had social situations that were "so dismal that it far outweighed the clinical problems" she was expected to address. She remarked that her work over the four years at this clinic led her to reevaluate

“why I became a clinician… when patients would ask for disability papers when I thought they should have a job and structure; when I had to turn in many patients for abusing their children. I felt I was in an adversarial role that I did not want to be -- court papers, reports, and people needed social interventions and no one was offering it. In the past, clinicians had the luxury of time to teach and do paper work, now they are scheduling brief therapeutic sessions instead of 50 minute sessions to allow time for insurance paperwork.”

A fourth year medical student had experienced similar pressures of time that the attending and resident discussed. He explained:

“a measure of success (as a student) is extracting a history of illness and developing a treatment regimen that allows you to discharge the patient with some improvement. Things that interfere with the above make physicians uncomfortable. If you have a minority patient who allows you to do the history and treatment plan (in a timely fashion) then the interaction is not problematic for the most part. I am of course speaking for myself.”

When we discussed if he detected any bias in the curriculum or in modeled behavior by attendings or residents that might have an impact on how minority patients were treated he reflected:

“One modeled message currently about hospital care is rapid assessment, efficiency, and treatment. One thing about minority, indigent, low income patients, is that their social situations are complex in negative terms and not fixable by any intervention that can happen in the hospital. As a result, one becomes almost discouraged to begin to explore these things. With some people you work with a translator and even when not, there is a reluctance rather than an inclination to get a good social history and explore the social roots of the illness.”

When I asked whether he observed any differences by racial or ethnic groups in terms of their trust in the medical system he remarked that he had not interacted with that many African Americans but his impression was that

“Absolutely. African American patients do not come into the hospital expecting to be treated well. White lower class patients, for example at the MGH, expect to be and they have been treated well. Asians -- so many different kinds. New immigrants versus academics.”

This very thoughtful student and I discussed how readily one distinguished different groups of Asians, some difficult to care for because of language barriers and social situations, others more similar in educational status and class to the medical community and easy to communicate with but that it was easy to slip into grouping African American patients into one category.

A resident also noted the stereotype of African American patients are that "they are dreadfully sick and their social life is so disorganized, they are "non-compliant" and living in a state of chaos, with a disorganized household or they are socially isolated. And incredibly sick and incredibly difficult to manage." In these cases "it is really hard to make progress and take care of patients," if for example diabetes is out of control and vague symptoms are difficult to manage." The resident concluded however that she did not have experiences that fit the stereotype she had just elaborated; her experiences were more varied. When asked if she had stereotypes for Asian patients, she noted that she did not have any, although she thought of East African patients (Somalians) as different from African Americans and Asians. This resident reported that she and her colleagues have a profile for young Hispanic women with total body pain --"they hurt every where" and "you simply cannot interview them. You take it on their terms or abdicate helping them. Some people work with them, I use the notion of nervious with patients, and it speaks to the political and domestic violence in their lives, if I think it makes sense; a sense of culture bound category. Everybody calls them crazy including Hispanic doctors." In this case, the resident's advanced degrees in medical anthropology have added to her ability to bridge cultural gaps between this population of patients and the medical world.

Race and ethnicity continue to define responses of clinicians. Assumptions are often made that may have inadvertent influence on how treatment interventions are offered to patients and how disease etiology is explored. One mode of profiling racial groups is through case formulations -- the way certain diseases are discussed in relationship to certain minorities. Examples include African American patients associated with hypertension and diabetes; Asian patients with hepatitis; poor Hispanics, Puerto Ricans and Mexicans or Central Americans, with diabetes and obesity; working class Irish, with alcoholism. Case examples may follow an epidemiological pattern -- diseases more prevalent in certain groups will be described in terms of patient demographics associated with those groups. Thus, though in a subtle way, race, ethnicity and class become part of the world of the medical gaze and the standard patient formulation. Does this lead to explicit racism or bias in what is offered to patients in terms of medical interventions? An African American attending psychiatrist noted that in his consultations on psychotropic medications he often finds primary care physicians prescribe less current drugs for their black patients. When he asks them why, they are baffled and unaware. Current antipsychotics and psychoactive drugs are not only an improvement over older medications in treating mental disorders but also are better tolerated than older drugs.

The Bias of Efficacy

All physicians we spoke with expressed concern about how effective their interventions may be when patients may be "train wrecks" -- the student's term, or victims of complex social problems -- the term used by residents and attendings. The following example recounted by a resident suggests how complex responses by clinicians may be and how clinicians may take actions to offset what is the perceived bias:

“Last night I had an elderly African American woman [in the ER], 'the classic invisible cardiac patient who does not get referred to cath'. I was trying to make a strong case that she needed to get cathed, to see if blockages [were the cause of ] her intermittent episodes of shortness of breath. Her story was fuzzy in the way it came out. We push people to answer questions in a format we phrase for them. We guide and teach them how to answer the questions. If we try to coach them and they don't respond to it or submit, it throws us off. Her job was to put her words into mine.”

The attempt was successful. The resident commented on a second patient, with whom she was having less success. She had grown familiar with her through the patient's many visits to the ER.

“I was thinking about how to help these patients and figure out what was wrong with them. I had another African American lady, a huge part of her problem is chaos in her life, with teenagers living in her house who use drugs. She does not have keys to her own apartment, the kids stole the keys, she has to be let in. I got the social workers to impose their external presence to intimidate the kids, and get them out of the house. She [the patient] comes less often into the ER. I also gave her a tranquilizer. I feel bad about dosing a social problem. If chaos is in their life, patients such as this woman don't get referred to high tech care, to cath, because they have a "difficult social situation." There is a danger of quickly moving to that interpretation and physicians are biased in looking at the patient and saying 'oh difficult social situation' and sure black physicians do this too. Some people do this and some don't -- it needs to be documented.”

Political Correctness, the Medical Machine and the Meaning of Bias

In the clinical contexts we discussed, political correctness appears to be the normative order in public discussion. Medical students with whom we spoke note they never hear overtly negative racist comments in the hospital nor among classmates. This sensitivity is new to the late twentieth century generation of medical students and faculty in our study area. However, when race is not a category of response, certain groups of patients are fair game for jokes and occasional expressions of dismay and amazement about discrepancies in expectations, behavior, and their treatment of physicians. Patients from the former Soviet Bloc, Russians and others, are in particular maddening for physicians and because "race" is not part of the package, physicians feel freer to comment on what they regard as strange behavior inappropriate in our society's wider medical culture. Examples which are perceived as disrespectful behavior toward physicians and the health care system include not showing up for appointments and not notifying the clinics, calling in the middle of the night for minor problems, choosing to go on vacation instead of keeping surgical schedules, demanding particular treatments when paying without regard for physician recommendations and expertise. No shows are calculated into the time schedules for many community clinics and ambulatory care units; surgical units and surgeons are aware they may have problems with such patients as well. These problems are also encountered with other new immigrants many of whom do not speak English. The community clinics with large immigrant non-English speaking populations appear to be scheduled differently than are units at the teaching hospitals.

The gap between the culture of medicine and the social and cultural resources, contexts and frames of reference of certain social groups clearly is related to how health care is delivered and how therapeutic options are offered and chosen by clinicians. The boundaries are fluid; they are sometimes associated with race; sometimes with class; sometimes with immigrant status; sometimes with disease state and age (the “train wrecks” -- the old heroin addict for whom little can be done that is effective and efficient.)

Anne Fadiman's popular A Spirit Catches You and You Fall Down (1998) is an account of misunderstandings and mistakes by the medical system, by well meaning physicians, and by the state in the care of a young Hmong girl suffering from epilepsy. The story Fadiman tells is not only dense but it has some facile answers to "the problem": it is cognitive, get translators, increase understanding, listen to patients and their families. However, as we review the larger picture of disparities in health care, the issues are more resistant to analysis, and rest not only in relationships between physicians and patients but also in larger organizational practices, cultures, and the financing of health care systems.

The labor force in American medical institutions today, at least in the Northeast and in California, is remarkable for its ethnic diversity (see Appendix 2, US maps of physicians and medical staff by race which we prepared for this discussion). In the metropolitan area where the authors work, the hospitals' staffs include many new immigrants some whom speak English as a second language and with limited fluency. The impression of our interviewees as well as of the authors is that the majority of the medical teams in the local teaching hospitals, the nurses and doctors, are of European background and considered "white." However, medical staff are a minority of the hospital labor force. The majority of the hospital labor force are mixed race and ethnicity and reflective of community surrounds. Interestingly, a major shift has also occurred in the race and ethnic composition of attendings and residents in these hospitals, a sea change that has occurred in the past decade. When interviewees identified the residents with whom they worked by race and ethnicity, the picture was of great diversity. And even greater diversity is evident in the local medical schools. For example, one fourth year medical student in recent rotations worked with several Jamaican residents, with African Americans from New York and Alabama, with Asian Indians, Chinese Americans, Chinese, Koreans Americans, Hispanics, Mexican - Americans. He noted reflecting on his response to this question, that he may have had more "minority" residents than not in his rotations. As a white male, he found himself in the minority. A woman resident in medicine noted that her colleagues were African American and Hispanic as well as Asian American and South Asian Indians. Whites included many Jewish physicians and nearly half of the residents with whom she worked were women. This sea change in resident color, gender and ethnicity is the future face of medicine in the United States. However, color, race and gender do not make medical culture. The profession of medicine is powerful in the reproduction of culture and practice, expectations and "gazes" and in defining what is important and significant in medicine. The financial and organizational shifts also are of import and recent changes in the financing and organization of medical care have had great impact. Universalism in patient care becomes more difficult to achieve when pressures of time and money shape clinical interactions and treatment choices. Thus, attitudes of clinicians are but a limited part of the culture of medicine and reasons for discrepancies in care. Institutional practices that favored the privately insured patient over the publicly insured patient, that favor patients with greater social and personal resources -- education, money, and social position and respond more readily to patient demands – are likely to neglect or give less attention to patients who are poorer and who may be less socially and psychologically integrated.

This diversion on ethnicity and race in the health care labor force and in the medical profession suggests that whatever the causes underlying disparities in the past, that in the present, ongoing disparities in care cannot simply be explained by racial differences between providers and patients. A recent NEJM article by Chen et al (2001) documented that racial difference in cardiac catherization after an MI was not related to physician race; the study was carried out with population of white and black physicians and black and white patients. The importance of a sea change in the ethnic and racial diversity of American physicians may be measured in part by closer examination of disparities in care and by a concern to address these issues in government sponsored NIH research. The training of health care professionals is also of high priority in reducing disparities in care. The following section addresses ways that health professionals have been trained thus far and the limits of current approaches.

Addressing Health Care Disparities Through the Training of Health Care Professionals

The literature on health disparities among ethnic minority populations includes discussion of sources of inequality based on institutional, clinician-centered, and patient-centered factors. A substantial literature on cultural sensitivity and cultural competence—particularly in relation to clinician training—has thus evolved. In the recent past, the vast majority of the scientific publications on cultural competence have been in the nursing literature. Most of these publications focus on the compelling reasons to train “culturally competent” clinicians (e.g., the pronounced ethnic and racial disparities in health care access and outcomes, the burgeoning ethnic minority population within the United States, and the well-documented underrepresentation of ethnic mi nority practitioners in health care professions (Nickens, 1992; Stoddard et al., 2000; Gonzalez et al., 2000). Curricula or principles supporting the development of cultural sensitivity or cultural competence in health care professionals are outlined in these articles. Notwithstanding the indisputable face validity to developing such curricula and principles, there has been an unfortunate relative dearth of studies that systematically investigate either (1) effective strategies for training clinicians and (2) how such training improves patient and clinician satisfaction and health care access and outcomes (Brach and Fraser, 2000). This literature on cultural competence is no doubt invaluable in mobilizing interest in promoting culturally sensitivity through moral rhetoric (e.g., Richardson, 1999), but also underscores the shortage of evidence-based data in addressing and resolving health care disparities.

The developing interest in cultural competence in clinician training is both a pragmatic response to the increasing proportion of ethnic minority individuals in the United States population and the failure of a strictly biomedical model in achieving uniform outcomes in this diverse population and a moral response to the inequities of health care among ethnically diverse populations. Although legitimated by epidemiologic studies of disparities in health care access and outcomes, the specific features of cultural competence have drawn substantially from the tradition of cultural relativism (initially promoted by social anthropologists, e.g., Franz Boas, Margaret Mead, and Ruth Benedict) and developed and applied within the more recent anthropologic sub-discipline of medical anthropology. Specifically, the seminal work, Patients and Healers in the Context of Culture (Kleinman 1980), introduced and popularized the concept of “explanatory models” of illness. The frequent divergence of explanatory models between clinician and patient and the clinician's failure to appreciate and negotiate this, were arguably a primary source of non-adherence to treatment recommendations as well as clinician and patient dissatisfaction with the clinical encounter (Good and Good 1980, 1981). Subsequent concepts developed within the field of medical anthropology, such as ‘semantic networks of illness meanings” (B. Good 1977) “idioms of distress” (Nichter, 1981), the distinctions between disease and illness (Kleinman, 1988), and the social course of illness (Ware and Kleinman, 1992) provided the theoretical underpinnings of an informed approach to cultural diversity in the clinical encounter, allowing both culturally sensitive and strategic health care delivery to ethnically diverse populations. Eventually, the impact of this theoretical body of knowledge has been manifest in clinical materials; for example, in an appendix outlining a means of formulating relevant cultural details to enhance understanding of psychiatric illness presentation among diverse populations in the DSM-IV (American Psychiatric Association, 1994) used by most practicing psychiatrists.

Several tensions are identified by educators in addressing cultural sensitivity in clinician training. For instance, Good (1994) has described one way in which medical students are socialized into the culture of biomedicine by being encouraged to edit out so-called extraneous details (often the very details that illuminate relevant sociocultural background that may impact upon communication and adherence in the clinical encounter) in the preparation of oral presentations. In this case, the growing pressures to be efficient in the evaluation, triage, and disposition of patients, reduce a patient to his or her physiologic condition, and encourage exclusion of social context in negotiating clinical care. Despite this prevailing tendency in traditional medical education, data from a recent study on physician communication patterns support that patient satisfaction is significantly higher in clinical encounters during which the physician practices a “psychosocial” communication pattern (i.e., in which physician talk is almost evenly divided between psychosocial and biomedical issues). This study also found that the frequency of this communication style is relatively low (< 10%), possibly because physicians felt that such an approach is more time-consuming; however, the study documented that psychosocially oriented clinical interactions did not, in fact, significantly increase the length of the patient visit (Roter et al. 1997). Failure to attend to social context may not only have adverse consequences in case formulation and treatment decision making, but may contribute to the disenfranchisement of ethnic minority populations vis a vis their health care. That is, exclusion of psychosocial context may contribute to diminished opportunities for collaboration in the clinical encounter, especially between ethnically dissimilar clinicians and patients. For example, a recent study demonstrated that ethnic minority individuals report less positive perceptions of their physicians than whites (Doescher et al 2000) and another found that race-concordance in the physician-patient encounter was associated with higher participatory decision-making (Cooper et al. 1999).

Another dilemma in training clinicians identified by two psychiatrists teaching psychiatry residents about cultural competence in a Boston teaching hospital is in negotiating the tension around exposing clinicians to the diversity of cultural traditions while resisting the tendency to stereotype or racially profile patients. Whereas it is useful to teach clinicians about the diversity of and patterns in culturally-based help-seeking practices and traditional remedies—particularly with immigrant populations with which they will have contact—it is important to maintain a perspective that fits this information into heterogeneous personal and social contexts. A failure to do so risks clinical reductionism that resembles prejudice, and more importantly, misses the opportunity to grasp the complexity of cultural, social, and personal variables that come to bear on the health problem at issue. These two psychiatrists agreed that case studies provide an excellent avenue for residents to explore “the multiple layers” and “complexity of clinical decision-making.”

Another paramount concern identified in the education of clinicians on cultural sensitivity and competence is in helping clinicians in training move beyond a mastery of the catalogue of diverse health care related practices to an examination of their own preconceived notions and feelings in clinical encounters with patients from ethnically diverse backgrounds. A recent paper on nursing education illustrated the distinction between a certain competence with respect to ethnically diverse patient practices and the unfortunate and persistent racism still manifest in various clinical interactions. The paper's author concludes that a frank examination of racism needs to accompany exposure to cultural diversity (Tullmann 1992). Similarly, the two above-mentioned psychiatrists note that they have observed that, because of a concern among residents “about being politically correct” in their seminar and specifically, because “no one wants to be on record with peers or with you about certain [racially sensitive] things,” creating a safe environment in which to explore and discuss countertransferential feelings generated in racially and ethnically diverse clinical encounters remains “a training challenge.” They report that “although we are not living in an era in which there is complete closure to this subject [of cultural sensitivity]” that they have noted an increasing sophistication among residents with respect to awareness of cultural diversity, in part due to greater exposure through, for example, experiences working in other countries. However, increasing opportunities for exposure to other cultures can also contribute to foreclosure to self-examination of one's deeply held attitudes about the Other. These psychiatrists thus frame their seminar as “not about knowing what to say or even what to think” but rather, as a process of deepening the process of self-examination vis a vis one's attitudes toward diverse peoples. Without such self-examination, two psychotherapists point out that, within the context of therapy, “it is possible for acts of racism to occur ‘innocently,' routinely, and with little detection or accountability” (Laszloffy and Hardy 2000). Finally, based on a case study of a medical school course addressing race issues in medicine, Fischbach and Hunt (1999) suggest the need for “proactive” and ongoing programs to address racial and cultural sensitivity in medical education. The concern about creating space to tackle the problem of racism in clinical encounters is consistent with a recent policy statement issued by The American Public Health Association (Anonymous, 2001) calling for government funding to research the impact of racism on racial and ethnically-based health care disparities in the United States.

Increasing time constraints on clinicians arguably place pressure on clinicians to seek “shortcuts” in managing complex clinical problems. In a recent review of biases in clinical judgment, Lopez (1989) writes that clinician error may occur in the absence of clinician prejudice as a result of selective information processing. That is, the time-pressed clinician uses available information and past experience about patient characteristics such as race and social class to arrive at a clinical hypothesis. Unfortunately, this practice may lead to systematic over- or under-diagnosis of certain illnesses among certain populations. Lopez argues that this conceptual framework for understanding clinician bias suggests specific strategies for training clinicians that contrast with more traditional emphasis on examining and changing prejudicial attitudes. Specifically, he suggests that education will need to focus on how clinicians process information to avoid such errors.

Notwithstanding strong impressions based on anecdotal and epidemiologic data, limited research has investigated whether there are systematic deficits in the education of clinicians with respect to cultural competence. Two studies have assessed the prevalence on formal instruction on cultural sensitivity and cultural competence in medical schools and found that such courses to be present in only a minority of institutions. Lum and Korenman (1994) surveyed American medical schools in 1991-92 and only identified 13 schools offering cultural sensitivity training; similarly, Loudan and colleagues (1999) identified 13 programs with such training in North America, less than half of that were compulsory. A study assessing the prevalence of cross-cultural content within psychiatry residency training programs showed a much higher prevalence of cultural content integrated into training materials, with 92% of programs surveyed reporting inclusion of such content (Baker et al., 1997). The health professional education literature contains relatively few studies on whether inclusion of cultural sensitivity or cultural competence material enhances skills of trainees. Robins and colleagues (2001) developed two standardized patient cases and assessment instruments to explore how medical students responded to cultural data in a clinical encounter. Their study did demonstrate differences in cultural sensitivity based on the ethnic background of the student; based on their results, the authors advocate curricular intervention, but did not use their intervention to test outcome in acquisition of cultural competence skills. One study was able to document an increase in language skills and cultural knowledge among medical student participants in a didactic and experientially-based program (including an educational trip to learn firsthand about health practices in Mexico) as compared with controls. Another study found higher levels of cultural competence (as measured by knowledge of, tolerance of, and comfort with diverse populations) among preclinical medical student participants in a “global multiculturalism track” as compared with non-participants (Baker et al. 1997). Two outcome studies on cultural competency training in nursing school showed somewhat contradictory results. One study examining the impact of specific instruction on culturally appropriate care to nursing students found that students who received the instruction actually felt less prepared to provide culturally sensitive care (Alpers and Zoucha 1996) and another showed that both recipients and non-recipients of a cultural sensitivity intervention improved self-reported cultural competency skills (Napholz 1999). Further research will be necessary to deter mine whether increased cultural sensitivity can be achieved among health care professionals with various curricular programs and if so, which types of programs are most effective.

In the following case analysis, the use of cultural analysis to explain clinician bias is discussed. Examples of research such as the work on mental health services may suggest ways to pursue analyses of reasons for disparities in medical treatment in the non-psychiatric domains of medicine.

A Case Analysis of Disparities in Mental Health Services: Evidence for the Role of Clinician

“Bias” and the Culture of Mental Health Institutions

Just as we were completing this review of the role of health professionals and the culture of institutions in reproducing health disparities, Surgeon General David Satcher announced the release of his office's report on disparities in mental health care in America. Entitled Mental Health: Culture, Race, and Ethnicity. A Supplement to Mental Health: A Report of the Surgeon General (Surgeon General 2001), the report sets out to assess evidence for disparities in mental health care and the burden of mental illness for racial and ethnic minorities in America. After providing a wide-ranging review of current data, the report provides the stark conclusion that “ethnic minorities collectively experience a greater disability burden from mental illness than do whites. The higher level of burden stems from minorities receiving less care and poorer quality of care, rather than from their illnesses being more severe or more prevalent in the community.” The report argues that minorities have less access to mental health services and are less likely to receive needed services, that minorities in treatment often receive a poorer quality of mental health care, and that minorities are under-represented in mental health research. It argues that while all Americans face fragmented and inadequate levels of mental health services, as well as societal stigma toward mental health care, racial and ethnic minorities experience a particular constellation of barriers that deters them from reaching services: mistrust and fear of treatment, racism and discrimination, and cultural and language barriers. The report argues that mental health disparities are grounded in historical and present day struggles with racism and discrimination, and that minorities are over-represented among vulnerable, high-need subgroups such as persons who are homeless, incarcerated, or institutionalized. But its strongest conclusions focus on disparities in access to and use of services and the quality of care members of minority groups may expect to receive.

In a section of the Executive Summary entitled “Main Message,” the Report sets out one of its primary claims: “The main message of this supplement is that ‘culture counts.'” Culture, it argues, is the “common heritage or set of beliefs, norms, and values” of members of a group. Language and culture are particularly significant for mental health care. Since mental disorders affect thoughts, moods, and integrative aspects of behavior, the report notes, the diagnosis and treatment of such disorders greatly depend on verbal communication and trust between patient and clinician. At the same time, the report notes explicitly that mental health professionals also constitute a culture, and makes the provocative claim that “the culture of the clinician and the larger health care system govern the societal response to a patient with mental illness.” Given the stark indictment of the mental health system, this claim would suggest that racism and discrimination may well be a part of ‘the culture of the clinician and the larger health care system,' responsible for a significant proportion of the disparities identified, even though broader social experiences of violence, poverty, and discrimination increase minorities' vulnerability to mental illness.

What exactly is meant by suggesting that racism and discrimination are a part of the culture of the mental health profession and the health care system, given that community mental health professionals are often among the leading advocates for the poor and for minority persons suffering mental illnesses? Rather than attempting to address this broad question comprehensively, we focus here on two very specific debates – concerning the over-diagnosis of schizophrenia among African Americans, and the elevated rates of involuntary commitment of African Americans – that may provide a model for thinking more generally about how the ‘culture' of clinicians and health institutions may contribute to reproducing disparities in health services and outcomes. In particular, review of these issues will make clear the importance of close attention to the particularities of the medical conditions and health services being examined.

African Americans and the ‘Over-diagnosis' of Schizophrenia

In 1981, Adebimpe, an African-American psychiatrist, reviewed studies of psychiatric treatment of black patients and reported that "a modest body of circumstantial evidence" suggested that members of minority subcultures are at particularly high risk for error in psychiatric diagnosis and assessment. Although it is not easy to carry out empirical studies, evidence continues to mount that specific patterns of misdiagnosis may put members of minority communities at special risk for poor medical care. In particular, research dating back to the 1960's provides strong evidence that African Americans in mental health settings are diagnosed with schizophrenia at much greater rates that white Americans and that whites are diagnosed with affective disorders at much higher rates than African Americans, with data suggesting a similar though less severe pattern among Hispanics. For example, a review of health services literature by Snowden and Cheung (1991) found that black Americans in in-patient psychiatric services are diagnosed with schizophrenia at almost twice the rate of white Americans, and that Hispanics are diagnosed with schizophrenia at one and a half times the rate of non-Hispanic whites. Most researchers believe these statistics reflect over-diagnosis of schizophrenia and under-diagnosis of affective disorders, rather than differences in prevalence or care-seeking (for reviews of this literature, see Neighbors et al. 1989, Good 1992, Good 1997). Since a diagnosis of schizophrenia carries powerful implications for treatment with anti-psychotic medications, assignment to mental health services designed for the chronically mental ill, and social stigmatization, and since failure to diagnose bipolar disorder may lead to the failure to use effective medications, these findings are particularly troubling.

Researchers over the past decade have continued to investigate this phenomenon. The most recent data continue to show great discrepancies between black and white patients in diagnoses. For example, Strakowski, Shelton and Kolbrener (1993) found that 79% of African Americans compared with 43% of whites in a public sector hospital were diagnosed with schizophrenia, and Strakowski et al. (1995) found that 28% of African Americans compared with 20% of whites in a university hospital emergency room were given such a diagnosis. Furthermore, treatment patterns seems to reflect these judgments concerning the severity of illness. Strakowski et al. (1995) found that black patients seen in an emergency room were significantly more likely to be hospitalized, and Segal, Bola and Watson (1996) found that African American patients in an emergency room received 50% higher doses of antipsychotic medications, while their doctors devoted less time to assessing them and scored significantly lower on an Art of Care Scale. Although these studies do not prove that black patients are misdiagnosed, they are consistent with a conviction among many researchers that African Americans are at serious risk for receiving a misdiagnosis of schizophrenia, for being provided inappropriate and inappropriately high levels of antipsychotic medications, and for receiving poorer care than white patients in the same settings. Research has pursued several lines of explanation for this apparent phenomenon.

Culture and the Expression of Mental Illness

One leading hypothesis about the source of high levels of apparent misdiagnosis in minority and immigrant populations is that it may result from cultural differences in the experience and expression of symptoms between such patients and 'typical' majority patients. Since the current diagnostic manual is based explicitly on symptom criteria, the experience of culturally distinctive symptoms – of nervios and ataques among Mexican American patients, of hearing voices of the dead among bereaved Indians, of hallucinations not associated with psychosis among Puerto Ricans or African Americans – may lead clinicians to misunderstand patients whose culture is different from majority culture norms, resulting in mistaken diagnoses. Current diagnostic categories are derived largely from research among majority populations, particularly those found in hospitals or specialty psychiatric clinics, and thus tend to lend support for the impression that such expressions of illness are universal (Mezzich et al. 1996). Cross-cultural research seriously challenges the validity and universality of some diagnostic criteria and diagnostic categories derived from such research, indicating a series of hypotheses that have important implications for medical care for immigrant populations or minority communities (Kleinman 1988). Thus one hypothesis, consistent with the message that “culture counts” in the Surgeon General's report, is that cultural differences between majority norms and normative ways of experiencing and communication symptoms among minority persons, including African Americans, may lead to mistaken diagnoses. Although there is now a large literature on cultural shaping of psychiatric symptoms, few studies have systematically examined explicitly whether these differences lead to increased levels of misdiagnosis for cultural or ethnic minority patients, such as difficulties in assessing and diagnosing African American patients who suffer psychoses, depression or anxiety disorders.

“Clinician Bias," “Aversive Racism,” and Misdiagnosis

A second set of hypotheses suggests that systematic patterns of misdiagnosis may result from clinician bias, and that the social and cultural context of diagnosis and diagnostic judgments should be submitted to sustained research. Psychological investigations, begun in the 1960's, focused on patterns of "overpathologizing bias" when clinicians treat women, the elderly, members of racial and ethnic minority groups, the poor, and the mentally retarded, using experimental techniques for investigating how both clinician and patient variables influence assessment (see Lopez 1989 for an excellent review and meta-analysis of this literature; cf. Good 1992). Lopez argues that research on social class shows the most consistent findings of bias, and that "bias is also consistently revealed in diagnostic judgments of Black and White patients" (1989:191). He finds little evidence to date for bias toward Hispanic patients (see Malgady, Ro gler and Costantino 1987 for a discussion of linguistic bias in evaluation of Hispanics; cf. Lopez 1988 for a critique of this paper). Lopez concludes from his review that "systematic errors in judgment based on patient variables may pertain to all clinicians and not just to those clinicians with prejudicial attitudes," and that "investigators should give careful consideration to the symptoms or disorders used as their clinical stimuli" (1989:194).

Support for these conclusions comes from an interesting study, conducted by Loring and Powell (1988) (but not reviewed by Lopez). Loring and Powell mailed questionnaires to psychiatrists, stratified by gender and race, asking them to participate in a study of the reliability of DSM-III by making diagnoses of two case vignettes. Two hundred ninety psychiatrists responded. The case narratives, written in the style of the DSM-III Casebook, included clear cut diagnostic criteria to support diagnoses of undifferentiated schizophrenic disorder (an Axis I diagnosis) and dependent personality disorder (an Axis II diagnosis). Descriptors identifying the gender and race of the case, or excluding such characteristics, were randomized, allowing the researchers to analyze how diagnosis is influenced by race and gender of both psychiatrist and patient. Findings are suggestive. First, "correct" diagnoses were given for only 38% and 45% of the two cases, respectively. Correct diagnoses were most often given when no identifying characteristics of the client were provided. Second, a highly complex pattern of interaction between race and gender of psychiatrist and race and gender of the case emerged. For example, white female psychiatrists diagnosed the first case as brief reactive psychosis 50% of the time when the client was identified as a white female, and paranoid schizophrenic disorder 53% of the time when the client was identified as a black male. No single interpretation of the overall response pattern can be given. However, black patients were given significantly more severe diagnoses, and black psychiatrists shared in this pattern of rating. Third, one clear pattern emerged. Black male clients were diagnosed as paranoid schizophrenic disorder in 43% of the first cases (compared with 6%, 10% and 12% for white males, white females and black females), and as paranoid personality disorder in 50% of the second cases (compared with 26% for the total sample). As the authors conclude, "clinicians appear to ascribe violence, suspiciousness, and dangerousness to black clients even though the case studies are the same as the case studies for the white clients" (p.18).

Whaley (1998) provides an insightful, complementary perspective on sources of racism in mental health services. Providing a comprehensive view of social-cognitive models of racism, Whaley elaborates a theory of “aversive racism” which may be relevant. Drawing on the work of Gaertner and Davidio (1986), he ascribes aversive racism to “low prejudice” white subjects who hold negative stereotypes of black people. He argues that “whites who identify with a liberal political agenda (e.g., endorse public policies that promote racial equality and combat racism, view themselves as nonprejudiced and nondiscriminatory, and sympathize with blacks' history of victimization) but who harbor negative perceptions of black people” will experience “prejudice-related conflict” rather than hostility or hate, expressing “discomfort, disgust, uneasiness, or fear” indirectly or in situations that do not threaten their liberal self-image (Whaley 1998:49). He cites evidence that liberal whites are as likely as conservative whites to discriminate against blacks in situations that do not implicate racial prejudice as the basis for their actions, then goes on to examine how such “aversive racism” might function in mental health settings.

Whaley focuses particular attention on research that identifies “bias in mental health professionals' judgments associated with the racial stereotype of blacks as violent” (1998:51). He reviews evidence that African Americans are likely to be sent to local correctional facilities, while white patients with similar levels of psychopathology and violent behavior are more likely to be referred to a mental health hospital. He interprets the Loring and Powell (1988) study cited above as evidence that black patients are more likely to be given a more severe diagnosis “because they are stereotyped as more dangerous.” And he links this specifically to the literature on over-diagnosis of schizophrenia for black patients. However, since violence is not among the diagnostic criteria for schizophrenia, he is forced to make an indirect argument. “The racial stereotype of violence may set off a chain reaction in the mental health evaluation and treatment process for black persons seeking help,” he writes (1998:52). He cites evidence that a diagnosis of psychotic disorder “is associated with fewer sessions with a primary therapist, greater likelihood of being treated with medication, less likelihood of being in outpatient treatment, and a lower chance of being in treatment with a professional therapist (Flaskerud & Hu 1992),” arguing that “severe diagnoses and restrictive interventions,” both forms of “therapeutic social control,” are linked to the view that mentally ill persons are dangerous. “Thus, the stereotype of violence is the common denominator in perceptions of black individuals and the diagnosis of schizophrenic disorders,” accounting for both the over-diagnosis of schizophrenia for African-Americans and a “chain reaction in the mental health evaluation and treatment process for black persons seeking care” (Whaley 1998:52).

Race, Perceptions of Violence, Involuntary Commitment, and Diagnosis of Schizophrenia

Lower socioeconomic classes and minority persons, in particular African Americans, are over-represented in public mental health institutions. However, perhaps even more troubling is a consistent finding that African Americans are over-represented among those who are committed involuntarily. After an extensive review, Lindsey and Paul (1989:172) conclude that after individuals "have reached the public system for admission, Blacks have been and continue to be even more over-represented in comparison with Whites among those who are involuntarily committed. Such over-representation appears across all periods of data collection and all areas of the country, without regional differences (e.g., North vs. South)." Recent accounts by patients and patient advocates provide reminders that involuntary commitment is still often associated with abusive treatment and abrogation of human rights (Nordhoff and Bates 1989), and those experiences of coercion influence inclination to participate in treatment (Lidz et al. 1995). (There is a large literature on elevated rates of schizophrenia among Afro-Caribbeans in Britain, as well as higher rates of involuntary confinement. See Coid et al. 2000 for a recent example. We do not review that literature here.)

Involuntarily commitment is linked in part to assessment of severity of a person's illness and to diagnosis, suggesting an added reason for concern about misdiagnosis and its over-representation in minority communities. However, involuntary commitment is also linked to assessment of "dangerousness," formally defined as the likelihood that an individual will use violence against self or others. The study of actual rates of violence among persons identified as suffering mental illness has recently been the focus of research and sharp debate. Analysis of data from the NIMH Epidemiological Catchment Area studies found that the claim that the mentally ill are no more likely to be violent than those who are not ill is untrue (Swanson et al. 1990; Link, Andrews and Cullen 1992; Monahan 1992). However, research also indicates that "excess risk for violence among mental patients is modest compared to the effects of other factors," in particular when compared with the effects of alcohol and drug abuse, and "only patients with current psychotic symptoms have elevated rates of violent behavior and it may be that inappropriate reactions by others to psychotic symptoms are involved in producing the violent/illegal behavior" (Link, Andrews and Cullen 1992:290). The sources and types of risk of violence among persons suffering mental illness and those who engage in substance abuse, as well as the implications for mental health services and involuntary commitment, are significant issues in ongoing research (Link and Stueve 1995; Junginger 1996).

Although Lindsey and Paul (1989:179) conclude that "empirical data to date provide no direct assistance in narrowing the range of explanations proposed for the over-representation of Blacks" among those who are involuntarily committed, the data are consistent with Whaley's theory of aversive racism discussed above. Both epidemiological and ethnographic research are needed to understand how this explanation may fit with other social and institutional factors, and to extend the analysis to other minority groups. Detailed ethnographic and clinical research will be required to identify how assessment and commitment processes function in particular settings and with particular populations – for rural Indians jailed for drinking, African Americans apprehended by the police, or persons assessed in psychiatric emergency rooms of general hospitals or state institutions. Findings from such research would have important implications for improving mental health services in community settings, correctional institutions, and psychiatric institutions, and for eliminating racial and ethnic bias in assessment and treatment.

For the purposes of this review, the linking of perceptions of violence, psychiatric diagnosis, and involuntary confinement may provide a model for understanding how clinical judgments and treatment decisions which have discriminatory impact, resulting in disparities in health services, may be made by persons who are not overtly racist, indeed by persons who are strong advocates for minority communities and patients. The linkages in this case are highly specific, rather than a result of generalized attitudes or tendencies to discriminate. Indeed, because good care requires assessment of levels of dangerousness to self and others, the role of stereotyping of those who are to be considered “violent” may well ‘slip under the radar' of efforts to provide training to support ‘cultural sensitivity' or culturally competent care. To be useful, this model needs to be verified in research and extended to other types of ethnic stereotyping. For example, research suggests that Asians may receive inadequate levels of psychotropic medications precisely because they are viewed as relatively free of emotional and mental health problems (see Hull et al. 2001 for a recent study in British primary medical care settings). And rather than hypothesizing attitudes such as “aversive racism,” research should be directed at measuring them.

Cautions: Health Systems Issues, and the Complexities of Mental Health Phenomena

It is important to temper the hopes that a single line of reasoning, such as that outlined above, is adequate to fully explain disparities in mental health services. No mention has yet been made in this section of disparities in access to health insurance, or of the tendency for persons who belong to ethnic and racial minorities to be over-represented among those who are poor and homeless and grow up in settings where violence is a part of the routine social environment, and no mention has been made about how managed care, or “managed behavioral health services,” is affecting the relationships formed between patients and clinicians. Yet we know that health insurance is closely linked to access to mental health care. McAlpine and Mechanic (2000) recently demonstrated that for a sample of persons with severe mental illness, who are dispropor tionately African American, unmarried, male, less educated, and low income, almost three fifths received no specialty mental health care in a 12 month period, one in five were uninsured, and only 37% were insured by Medicare or Medicaid. Because “person covered by these public programs are over six times more likely to have access to specialty care than the uninsured,” any explanations that ignore formal mechanisms for access to care are likely to extremely partial. Furthermore, privatization of mental health services as managed mental health care has often led to disbanding of clinics devoted to the care of high risk children, youth and families, and psychiatrists have far fewer visits to develop relationships with patients that allow them to move beyond initial stereotypes – both on the part of the clinician and the patient – and develop truly therapeutic and understanding relationships than is true under other forms of mental health services. Thus, great care should be taken in developing explanations that focus on decision-making processes of individual clinicians.

Finally, in the specific case we have been discussing here, that of the apparent ‘over-diagnosis of schizophrenia' among African Americans, the robustness of the phenomenon remains in question. Because there is no gold standard for diagnosing schizophrenia, no biological marker, one line of research attempting to determine reasons for higher rates of schizophrenia among hospitalized African Americans has compared diagnoses using standardized diagnostic interviews with recorded hospital or emergency room diagnoses. The best such study, carried out by Neighbors et al. (1999), found that use of a standardized diagnostic interview reduced the percentage of black patients diagnosed with schizophrenia in a Michigan hospital from 58% to 39%. However, the study found that white patients diagnosed with schizophrenia were reduced from 49% to 31%, and that levels of ‘misdiagnosis' (assuming the research diagnoses were correct) for white patients was somewhat higher for whites than for blacks. Misdiagnosis in this study was shown to be as high for whites as blacks, and even the most carefully designed study continued to find higher rates of schizophrenia among hospitalized African Americans than among hospitalized white Americans. Thus, although mental health care provides an important model for how to approach the issues to be addressed in this review, it remains a particularly difficult domain to make claims with great certainty.

Concluding Questions

The multidimensionality of reasons for disparities in health care and medical treatment for racial and ethnic minorities and the lack of data focused explicitly on the role of the culture of professionals and health institutions in producing these disparities makes it difficult to fully respond to the charge given to these authors by the IOM Committee. "Culture counts," as Satcher's report notes. Until recently, when cultural analyses were proposed, the focus was largely on patient culture. Burdens of difference were on patient communities, and medicine and health professionals were expected to learn to be culturally competent in attending to diverse populations which make up American society. When we are challenged to examine the culture of medicine and of our health care institutions, we are also challenged to bring a critical perspective which has largely been ignored by most research to date or which has circumscribed cultural inquiry to the differences between patient and physicians “beliefs.” Disparities in medical treatment are not simply matters of differences in “beliefs.” Clearly, political and economic factors that shape our medical commons and our large society are implicated in the production of these disparities. Physicians interviewed as well as research reviewed indicate that societal racism and persistent inequalities may be responsible for many of the differences, now so widely documented. In a recent study sponsored by the Robert Wood Johnson Foundation, Hargraves et al. (August 2001) found that minority physicians who were more likely to be in solo practice were also less likely to be able to obtain referrals to specialists and had greater difficulty admitting patients to hospital. Given that minority physicians care for a greater proportion of minority patients, differences in high technology care may be related to environmental practice factors. As Hargrave et al. (2001) note, “Minority physicians' inability to arrange important medical services for their patients may be positively associated with problems minority patients have with access to care.” Minority physicians in this study included Hispanic and African American physicians. (See also Gray and Stoddard 1997 and Komaromy et al 1996.) Minority physicians also feel there is discrimination in peer review, hospital promotions, Medicaid and Medicare reimbursement, malpractice suits and private insurance oversight (Byrd et al 1994). Bias appears in the awarding of managed care contracts as well, with disparities between white and minority, in particular Asian, physicians documented in a national survey (Mackenize et al 1999). These complexities of bias and practice environment clearly indicate that differences are found not simply in cultural diversity or in practice “beliefs.”

In sum, we recommend the following approaches to further understanding of reasons for disparities.


Attend to a critical analysis of the culture of medicine in its broadest meaning and in different practice and training environments and geographical regions. The dimensions of time, efficiency and efficacy and the medical gaze may be useful starting points, but analyses should examine behavioral modeling and hierarchical relationships that may influence patterns of care and choice of treatments as well.


Examine the political economy of cultural practices in medicine, from the arrangement of health care delivery systems to the financing of biomedical innovations and practices, to the justification for choice of treatment and care.


Examine the practice arrangements of minority physicians by ethnicity, age and region (urban/rural and state).


Explore how the sea change in the ethnicity and race of medical students, physicians, nurses, and health care staff impacts provision of care to ethnic and racial minorities, new immigrants, and the poor.


Identify interventions and programs that have been successful in medical and nursing education that have influenced the way care is provided to ethnic and racial minority patients.


Assess the success of programs directed to redress imbalances in care such as minority outreach programs and clinics; what are positive lessons, negative if unintended consequences and avoidable difficulties?

These are elementary suggestions, and the questions with which we began this chapter are but partially addressed. Unfortunately, we have not been able to conclude with a clear set of findings that answer our initial questions. Clearly, more work focused explicitly on these questions is required, and such research will have to include depthful, qualitative work -- observations, indepth confidential interviews with health practitioners, including those in training, and similar indepth interviews with patients from diverse minority groups about their experiences in health care. In terms of policy responses to eliminate racial and ethnic health disparities, these will not only have to come from innovative research and programs in the medical commons and the health care arena, but as David Williams suggests, from larger societal changes (Williams and Rucker 2000). And such responses will have to use new and innovative understandings of culture, ethnicity and racism to develop multidimensional results.


Appreciation to Giuseppe Raviola HMS IV and Elizabeth Herskovits for interviews and analyses of contemporary academic hospital culture; to Jennifer Clark for data analysis, charts and maps; to Naomi Ko for research assistance, and to Martha MacLeish Fuller for bibliographic assistance.


  • Adebimpe VR. 1981. Overview: White Norms and Psychiatric Diagnosis of Black Patients. American Journal of Psychiatry 138:279-285. [PubMed: 7008631]

  • Alpers RR, Zoucha R. 1996. Comparison of cultural competence and cultural confidence of senior nursing students in a private southern university. Journal of Cultural Diversity 3(1):915. [PubMed: 8788835]

  • American Psychiatric Association. 1994. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC: American Psychiatric Association. pp.843-49.

  • Anderson LM, May DS. 1995. Has the use of cervical, breast, and colorectal cancer screening increased in the United States?American Journal of Public Health 85(6): 840-42. [PMC free article: PMC1615482] [PubMed: 7762721]

  • Atkins K. 2001, August 10. A racial gap in the health-care system. Boston Globe.

  • Ayanian JZ, Cleary PD, Weissman JS, Epstein AM. 1999. The effect of patients' preferences on racial differences in access to renal transplantation. New England Journal of Medicine 341(22): 1661-1669. [PubMed: 10572155]

  • Bach PB, Cramer LD, Warren JL, Begg CB. 1999. Racial differences in the treatment of early-stage lung cancer. New England Journal of Medicine 341: 1198-205. [PubMed: 10519898]

  • Baker FM, Bondurant A, Pinderhughes C, Fuller R, Kelley S, Kim SP, Triffleman E, Spurlock J. 1997. Survey of the cross-cultural content of U.S. psychiatry residency training programs. Cultural Diversity & Mental Health 3(3):215-18. [PubMed: 9277022]

  • Balsa A and McGuire T. “Statistical Discrimination in Health Care,” January, 2001.

  • Barrett RJ. 1996. The Psychiatric Team and the Social Definition of Schizophrenia: An Anthropological Study of Person and Illness. Cambridge: Cambridge University Press.

  • Barrett B, Shadick K, Schilling R, Spencer L, del Rosario S, Moua K, Vang M. 1998. Hmong/medicine interactions: Improving cross-cultural health care. Family Medicine 30(3): 179-84. [PubMed: 9532439]

  • Becker H, Geer B, Hughes EC, Strauss A. 1961. Boys in White: Student Culture in MedicalSchool. Chicago: University of Chicago Press.

  • Bhui K. 2001. Over-representation of Black people in secure psychiatric facilities. British Jour nal of Psychiatry. 178:575. [PubMed: 11388982]

  • Bird JA, McPhee SJ, Ha N-T, Le B, Davis T, Jenkins CNH. 1998. Opening pathways to cancer screening for Vietnamese-American women: Lay health workers hold a key. Preventive Medicine 27:821-29. [PubMed: 9922064]

  • Bolen JC, Rhodes L, Powell-Griner EE, Bland SD, Holzman D. 2000. State-specific prevalence of selected health behaviors, by race and ethnicity—Behavioral Risk Factor Surveillance System, 1997. Morbidity & Mortality Weekly Report. CDC Surveillance Summaries 49(2):1. [PubMed: 10965781]

  • Borowsky SJ, Rubenstein LV, Meredith LS, Camp P, Jackson-Triche M, Wells KB. 2000. Who is at risk of nondetection of mental health problems in primary care?Journal of General In ternal Medicine 15(6):381-8. [PMC free article: PMC1495467] [PubMed: 10886472]

  • Bottero A. 1991. Consumption by Semen Loss in India, and Elsewhere. Culture, Medicine and Psychiatry 15:303-320. [PubMed: 1935181]

  • Brach C, Fraser I. 2000. Can cultural competency reduce racial and ethnic health disparities? A review and conceptual model. Medical Care Research & Review 57 (Suppl 1): 181-217. [PMC free article: PMC5091811] [PubMed: 11092163]

  • Braverman P, Egerter S, Edmonston F, Verdon M. 1995.

Genetics researchers often discover certain snips and pieces of the human genome that are important for health and development, such as the genetic mutations that cause cystic fibrosis or sickle cell anemia. And scientists noticed that genetic variants are more common in some races, which makes it seem like race is important in genetics research.

But some researchers say that we've taken the concept too far. To find out what that means, we've talked to two of the authors of an article published Thursday in the journal Science. Sarah Tishkoff is a human population geneticist and a professor at the University of Pennsylvania. Dorothy Roberts is a legal scholar, sociologist and a professor at the University of Pennsylvania's Africana Studies department. This interview has been edited for length and clarity.

How do geneticists use race now, and how does that cause problems for science?

Sarah Tishkoff: We know people don't group according to so called races based purely on genetic data. Whenever the topic comes up, we have to address, how are we going to define race? I have never ever seen anybody come to a consensus at any of these human genetics meetings.

Dorothy Roberts: That's because race is based on cultural, legal, social and political determinations, and those groupings have changed over time. As a social scientist, looking at biologists treating these groupings as if they were determined by innate genetic distinctions, I'm dumbfounded. There's so much evidence that they're invented social categories. How you can say this is a biological race is just absurd. It's absurd. It violates the scientific evidence about human beings.

Tishkoff: But I as a human geneticist wouldn't want to imply that there are no differences —but among different ethnic groups, not racial classifications. For example, I'm Ashkenazi Jewish. I have a much higher risk of getting certain genetic diseases that are common in certain Ashkenazi Jewish populations. That was an important question when I was having children.

There was a drug, called BiDil, that somebody claimed is more effective with African-Americans than other races — which was not true. But there are genes that play a role in drug metabolism. So if a doctor was prescribing drug treatment based on her identification of race she'd say, "You should use drug A because that's better for people of European descent." But the patient might not carry the right gene. That might have negative consequences. That might be the wrong treatment for her.

Roberts: Race isn't a good category to use to understand those differences or the commonalities. It in many cases leads researchers down the wrong path and leads to harmful results for patients. For example, black patients who have the symptoms of cystic fibrosis aren't diagnosed because doctors see it as a white disease.

So part of the problem is that when we see a high frequency of a medically relevant gene in one racial population, we start to assume that all members of that race have that gene?

Tishkoff: Yeah, I think that's right.

Roberts: People take what's a difference in [gene] frequency and turn it into a categorical difference that interprets it as if one race has one gene and another race doesn't have the gene. You can't reach the conclusion that because you know someone's race you know what their genes are. It's not the case that there are populations where 100 percent, everyone, has those genes and nobody in other populations have those genes. It's a crude way and unhelpful way of figuring out what the disease risk is.

Tishkoff: That's not to say that genetic risk in disease isn't important. I do think geography is important, and I think that people historically during evolutionary history have adapted to different environments.

Is it that the science of genetics and the science of human populations are racist? Or is it that the numbers are there and, as a society, we're interpreting these things in a racist way?

Roberts: There is a long history of justifying the subordination of different groups and social groupings based on myths about their biologic or genetic predispositions. It's not only that there's scientific evidence that humans aren't divided into discrete biological categories we'd call races. But there's also evidence of the harm these biological meanings of race have caused for centuries. It's one of the reasons why it's difficult for human geneticists today to grapple with the meaning of race. You can't talk about race without also considering the history of racism.

Tishkoff: But modern human geneticists, we're not trying to say they have a racist agenda. It's a positive thing to try and increase studies of genetic diversity that may differ across different ethnicities or ancestries.

Roberts: Yes. I'm not trying to say anything about the motivations or what scientists are trying now to do. Our paper is a call for scientists to come up with better ways of understanding human genetic diversity without relying on this antiquated concept of race There is a failure of imagination for people to think, what is there something better that we can use? Let's develop that.

Is it that difficult, though? What are the things holding scientists back from developing something better?

Tishkoff: If I want a grant from the National Institutes of Health, I am required to check off the racial classification according to the U.S. government's census categories. I study very diverse people from all over Africa, but I believe the classification is African American or Black. I always feel awkward.

Roberts: The NIH guidelines require the use of race in recruiting research subjects. There's a history of advocating for that in order to increase the participation of minorities in clinical research. Then it gets confusing, because the researchers continue to use these categories in conducting the research. Scientists must conform their research to these admittedly social categories of race.

Tishkoff: One also has to take into account that you need a way to identify your study population. Ideally you want ethnically diverse populations, so obviously you have to have some way of identifying research subjects. And that's fine. But they don't need to say based on race. The language and terminology does matter.

Roberts: Except if the research question has to do with investigating the effects of racism — race as a social category that does affect people's lives and health and future because of the impact of social inequality. I often get the justification from doctors that "I know it's crude but it's the best we have given the limited resources."

Tishkoff: To some extent I think that's true. If a doctor doesn't have a readily available genetic test to look at ancestry or to look at individual genotypes of that person, race will be their best proxy. But the language matters. We need to move away from racial terminology, particularly in the field of medical genetics. That should just be eliminated. It reinforces the notion that there's a genetic basis to this classification system. We as scientists have to set an example.