Are physicians prejudiced? Do their attitudes contribute to the poorer health of black Americans – their higher rates of infant mortality, kidney and heart disease and more? According to the media’s coverage of a new report, the answer is yes. “Color-Blind Care is Not What Minorities are Getting” (U.S. Newsday); “Race Bias in Health Care” (USA Today); “Separate and Unequal” (St. Louis Times Dispatch), “Subtle Racism in Medicine” (The New York Times). These recent headlines -and the uncritical coverage that followed– were prompted by a report from the Institute of Medicine (IOM) called Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Virtually every story picked up the triumphant remark of Dr. Lucille Perez, president of the National Medical Association, which represents minority physicians, “It validates what many of us have been saying for so long — that racism is a major culprit in the mix of health disparities and has had a devastating impact on African-Americans.” Yet I could find only one news event, CNN Live Today, that elicited comments from physicians who questioned the bias charge. The truth is less inflammatory than Dr Perez and the press would have us believe. The Institute’s case for prejudice in the report, released March 20, is weak. Its very own flaccid conclusion tells the story: “Some evidence suggests that bias, prejudice and stereotyping on the part of health care providers may contribute to differences in care.” Some. Suggests. May. Such limp language is indeed in order when data are flimsy and reputations are at stake. A look at the 191 page report shows why IOM authors hedged their bets so carefully. The report was written at the request of Congress which asked the IOM, a government-sponsored think tank, to assess the extent of racial differences in receipt of treatments with a special emphasis on “evaluating ?the role of bias, discrimination and stereotyping” in explaining those differences. To meet its charge, the IOM examined data on the frequency with which white and minority patients underwent certain procedures and received specific treatments. The report also reviewed evidence specifically aimed at measuring doctor bias, discrimination and stereotyping. Evaluating the evidence had to be extremely difficult. In reviewing the substantial literature of academic studies comparing treatment of white and black patients, the IOM selected studies that took into account obvious reasons for differential care, such as insurance status, income or age. But even the best of the studies were limited by fact that they are retrospective, meaning that the journal articles the IOM reviewed contained data already collected for other purposes (such as patient billing). Thus, depending on the nature of data amassed, investigators could not obtain certain details vital to understanding why patients might be treated differently. For example, a data set showing that African American patients underwent a certain procedure less often than whites might be missing crucial information on refusal of treatment (this is relevant because in some studies, though not all, refusal rates among blacks are often higher than among whites). Subtle medical details that determine the appropriateness of care might not have made it into data sets. Most likely, a constellation of these and other factors, not a single one, explains the treatment gap between white and black. A study of lung cancer treatment by Peter Bach and colleagues at Memorial Sloan Kettering Cancer Center was cited by virtually every newspaper report as an example of unfairness. Bach looked at records of over 10,000 Medicare patients who received diagnoses of operable lung cancer. Seventy seven percent of the white patients underwent surgery compared to 64 percent of the black patients. Five years later, one third of white patients but only one quarter of the black patients were still alive according to the study which was published in the New England Journal of Medicine in 1999. Those numbers understandably arouse concern but many unanswered questions remain. Among them, did the black patients refuse surgery more often than whites? (Bach did not take account of this) Did black patients in the study have higher rates of other conditions, such as poor lung function, that would have prohibited surgery in the first place or contributed to an earlier demise? (e.g., some important measures of lung function were not noted in the data base upon which Bach relied) Were there fewer married patients among the black group? (this is relevant because married men tend to fare better after surgery than equally ill unmarried ones; black men over 65 are more likely to be single than white men over 65) It is possible that prejudice — that is, doctors’ giving patients less desirable treatment on the sole basis of their race or ethnicity — played a role in the lower rate of surgery among the black cohort but the available evidence does not make the case. Of note are studies, some of which were mentioned in the IOM report, showing that white and black patients served by military health plans are treated comparably. This should not surprise. After all, in the military, access to care is not an issue. Also, military personnel and their families are a relatively homogenous group in terms of economic status and education – two powerful correlates of health – irrespective of race. The part of the IOM report that focused on doctors’ attitudes toward patients relied upon studies that are even more difficult to interpret. The authors’ conclusions of stereotyping were based on studies that asked doctors their impressions of a hypothetical black or white patient. In these surveys of doctors the black patient were more likely to be judged as less financially well off, less likely to comply with treatment, less attractive as a potential friend and so on. Yet what doctors might pencil in on surveys bears questionably on their actual practice because doctors don’t deal with made-up clinical cases; they respond to concrete situations with flesh and blood patients with detailed medical histories. Moreover, these studies contain negligible evidence that doctors’ impressions led them to deny or offer second rate care. Again, these inferences are too shaky to support the allegation that doctors fail to give appropriate care on the basis of the patient’s race. In view of this, the IOM’s calling for expanding the Office of Civil Rights at the Department of Health and Human Services, only fuels the perception that doctors are prejudiced. Am I saying that individual physicians never base clinical actions on impressions? No. In fact, they, like other professionals, are guided by experience and statistical norms. When it comes to an actual clinical encounter, surely individual practitioners derive impressions partly on the basis the patients race (or sex or class, for that matter): impressions that have relevance to how they treat patients. For example, if a physician thinks that a patient will not comply with triple therapy for HIV he might do one of two things: (a) forgo the medication, assuming that the patient wouldn’t follow through anyway, (b) give the medication while putting more effort into monitoring the patient and having his nurse explain in great detail the importance of regular dosing. To the extent that a physician does the former, the IOM report serves a consciousness-raising function, prompting doctors to ask themselves whether they are giving every patient the opportunity to benefit from treatment. But to elevate the phenomenon of making clinical generalizations – based on experience and statistical norms – to the level of a civil rights violation is a large and questionable leap. Largely absent from the literature is a body of evidence describing the care that black patients receive when they are treated by white doctors compared to black doctors. After all, we wouldn’t expect black doctors to be prejudiced against black patients. A group of Yale cardiologists, in one of the rare studies that looked at this matter, found that that all black patients underwent catheterization at the same rate no matter what the race of their doctor. Also, the black patients had slightly better outcomes – more likely to be alive at one and three years — than the white patients, even though, as a group, they received the procedure less often. (Incidentally, this brings up another important point: disparities in treatment do not necessarily lead to disparities in outcome.) Despite the anemic evidence the IOM presented, the press and some advocacy groups ran with the notion that “bias, stereotyping and prejudice” make a large contribution to racial and ethnic disparities in health care – and, by extension, the poorer heath status of black Americans. This is disturbing because it diverts us from the far more important causes of the health gap (1) socioeconomic status (this relates to health insurance and access to care); (2) health literacy (knowledge about preventing or controlling chronic disease, asking questions about their treatment, negotiating HMO rules) and (3) attitudes toward health (the extent to which patients seek timely care and comply with treatment). To better the health of black Americans we should put our efforts here. The IOM has contributed to the press hype. By calling the report Unequal Treatment, rather than, say, the more neutral Treatment Differences, it helped create the impression that care is being denied patients on the basis of their race. Not only does the allegation go far beyond the evidence, it prompts alarmists like Dr. Perez to proclaim “racism as a major culprit” in the poorer health of African Americans. There is much we can do to improve the health of African Americans. Inciting their distrust of the medical profession with misleading claims about physician prejudice can only hurt.