The Color of Healing: Race and Medicine
By Bob Wells Photography by Will McIntyre
Every day, in hospitals across the United States, patients undergo procedures that were unimaginable a generation or two ago. Thanks to a revolution in biotechnology, coronary artery bypass surgery has become virtually routine. Hearts, kidneys, livers, and other organs are successfully transplanted. Mammography and other new imaging techniques detect cancers at their earliest stages. Artificial hips and knees restore the ability to walk and move free of pain.
Yet, a growing body of evidence indicates that these and other advantages of modern medicine are not reaching all members of society who could benefit from them. In one of the most provocative areas of biomedical research today, investigators are finding that race is still a powerful predictor of who gets treatment and who doesn't, particularly for invasive and aggressive procedures. Simply put, almost a half century after the dawning of the civil rights era, blacks and whites do not get the same health care even when traditional barriers, such as access and insurance status, are not an issue.
Investigators at Duke University Medical Center are among the leading contributors to this new area in minority health research. Drawing upon the resources of the Duke Databank for Cardiovascular Disease, Eric Peterson, MD, and colleagues at the Duke Clinical Research Institute have shown that blacks are significantly less likely to undergo bypass surgery than are whites with comparably severe heart disease. Likewise, in a series of studies conducted throughout the Veterans Affairs hospital system, Eugene Oddone, MD, chief of the Division of General Internal Medicine, and his research team have found similar racial differences in the use of carotid endarterectomy-a surgical procedure to reduce the risk of stroke by removing blockages from the major artery feeding the brain.
No one yet knows precisely what's causing the disparities in these and other treatments. To some extent, clinical differences and socioeconomic factors such as insurance status and the ability to pay clearly play a role. Yet, even when these factors are taken into account, the disparities remain. It's perhaps telling that while Peterson and Oddone are studying two very different procedures, they are heading in the same direction. They believe that much of the explanation probably lies in the physician-patient relationship and the complex process by which patients and their physicians reach treatment decisions.
"That's where my nickel is," says Oddone. "I think that a lot of what's going on is what's happening when physicians are sitting down and talking with patients. What are physicians bringing to the table? What are patients bringing to the table? What information is being given to patients, what do they understand about it, and what context do they put it in? That is the most interesting and most fruitful area of research in this field right now, the physician-patient interaction."
It's More Than Access
Whatever the explanation proves to be, this is very important research, according to Linda George, PhD, a medical sociologist and assistant director of the Duke Center for the Study of Aging and Human Development.
"The good news is that medicine is now bothering to look," says George. "For so long, the only thing that dominated the literature on racial differences in health care was access. That's all you heard. 'If you can get people equal access, then the health differential should be reduced.' But there's a whole lot more to treatment--who gets it, what they get, and how it works-- than simply getting people in the door. And the most important thing about these studies is that they show you that once you're in the door, there's still differences."
It's not only important research, though. It's also extremely sensitive. Even controversial. In the closing years of the 20th century, few subjects can strike such an emotional and visceral response in human beings as does the topic of race.
"This research elicits strong emotions," Oddone says. "People form immediate opinions about what's going on. Doctors get angry and say 'You're saying I'm a racist.' And no, I'm not saying doctors are racists. I'm saying let's look at what it is we bring to this and what patients bring to this. This is all very subtle.
"Whether we like it or not, we and other researchers have observed these differences. They are real, and they are affecting our patients. We can explain some of it, but not all of it. And the overarching issue is that we need to make sure that everyone is at least getting the same opportunity to undergo these procedures."
Typically, physicians cannot believe, at least initially, that race has any bearing whatsoever on the care that patients receive or on how physicians deal with patients, Peterson says.
"I don't think the answer is a simple one," he says. "If you ask me if I believe there's widespread racism in health care as we normally define the term, the answer is absolutely not. But if you ask me if physicians act with patients differently depending on a patient's socioeconomic status, education level, financial background, job status, and on whether the patient appears interested in being involved with treatment decisions, then the answer is yes, absolutely. All these factors affect how we deal with patients."
Opening a New Door
The treatment disparities being documented by researchers are projected against a backdrop of long-known differences in the overall health status of blacks and whites in America. Indeed, as a 1990 article in the British journal Lancet noted: "The poor ranking of America's black population in the indices of good health is a scandal of such long standing that it has lost the power to shock." For virtually every statistical category of health-from infant mortality to overall life expectancy, from the incidence of cancer to hypertension to kidney disease-blacks fare worse than whites. In some categories, Lancet noted, American blacks experience mortality and morbidity rates "that were long ago banished in other industrialised nations." But if these dismaying statistics are well known, so too are the causes.
"There are no mysteries left about the origins of these disparities," the Lancet article said. "In terms of all the interconnected social and economic conditions associated with good health-income, education, nutrition, housing, family and community stability-blacks generally rank poorly."
Researchers, though, opened an entirely new door about ten years ago, when they began documenting racial differences not in health status but in treatment. Since then, disparities have been documented not only in bypass and carotid endarterectomy, but also in a host of other procedures including kidney transplants, (blacks are also much less likely to be organ donors), cesarean sections, cholecystectomies, laminectomies, appendectomies, and in diagnostic procedures such as mammography.
Cardiac procedures, though, have probably been the most studied and best documented area of research. Some of the earliest work was conducted by Arnold Epstein, MD, of Harvard Medical School, who served as one of Peterson's mentors during his residency at Brigham and Women's Hospital. In 1989, in one of the first studies in the field, Arnold reviewed administrative databases in Massachusetts and found that black patients presenting with chest pain were much less likely than whites to receive a cardiac catheterization or to undergo either bypass or angioplasty.
At the time, most researchers thought that race was probably a proxy for insurance status and the ability to pay and that when those factors were taken into account, the differences would disappear. While he was a fellow at Brigham and Women's, Peterson got the chance to test that theory when he served as principal investigator in a study of patients who had been treated for heart attack at Veterans Affairs hospitals across the country.
"Because the VA provides free medical care to military veterans, insurance status and ability to pay was not an issue," Peterson recalls. "As a result, we thought the VA study would clearly show no racial differences in care. As it was, though, we did the study, and the results came back saying the differences were almost as large as in the non-VA population."
The Perfect Place
When he came to Duke as a cardiology fellow in 1992, Peterson found himself in the perfect setting to continue his research.
"Duke is ideally suited to answer these questions," he says. "We have this incredible, large database on heart patients that doesn't exist anywhere else in the world. We have information on thousands of patients, including a good percentage of minority patients, and we have long-term follow-up data that nobody else has."
Most previous studies were based upon medical claims data, which-unlike the Duke database--doesn't contain the clinical information needed to adjust for differences in severity of disease. Theoretically, if blacks had less extensive heart disease, fewer symptoms, or fewer coexisting illnesses, then it would be expected that they receive fewer bypasses and angioplasties. Also, without knowing long-term outcomes, researchers can't really say whether the differences they are observing reflect an underuse of procedures in blacks or an overuse in whites.
In a study that took three years to complete-the most comprehensive study of its kind ever conducted--Peterson reviewed Databank records on 12,402 patients, about 10 percent of whom were black, who had undergone diagnostic catheterization at Duke between 1984 and 1992 and were found to have at least 70 percent blockages in one or more coronary arteries. The researchers then examined treatment patterns after adjusting for disease severity, angina status, and estimated survival benefit that could be expected from the revascularization. Finally, they looked at five-year survival rates.
In a much-noted article published last February in the New England Journal of Medicine, the Duke researchers reported that after adjusting for disease severity and other characteristics, blacks were marginally less likely-about 13 percent--than whites to undergo angioplasty but were 32 percent less likely to undergo bypass surgery. Even more disturbing, the greatest racial disparities in the use of bypass surgery were found among the patients who stood to benefit most from the procedure.
Every Way They Looked
Whether researchers looked at benefits in terms of symptom relief or improved chances for long term survival, they found that bypass was used consistently less often in blacks than in whites. Overall, 48 percent of blacks with severe disease underwent surgery versus 65 percent of whites. Using a complex formula that incorporated disease severity and a variety of other prognostic factors, researchers calculated how much longer patients could be expected to live if they had a bypass.
They found that, although the use of bypass surgery increased in both blacks and whites as the expected survival benefit increased, the surgery was still used less often in blacks than in whites. For example, for patients who would be expected to have little or no survival benefit from surgery, blacks were only slightly less likely than whites to undergo the procedure ( 8 percent versus 10 percent). But for those whom bypass would be expected to add a year or more of life, only 42 percent of blacks underwent bypass as opposed to 61 percent of whites.
Finally, when they examined long-term outcomes, the researchers found that blacks had significantly higher mortality rates than whites. That difference stemmed partly from higher risks at the time of the initial catheterization, primarily due to higher rates of diabetes and hypertension and worse ventricular function. Even when researchers controlled for those risk factors, though, blacks still remained 18 percent more likely to die than whites during the five years follow-up period.
The New England Journal article was entitled "Racial Variation in the Use of Coronary Revascularization Procedures: Are the Differences Real? Do They Matter?" The answer to both questions, the researchers found, was 'yes."
The Carotid Studies
While Peterson and others looked at cardiac procedures, elsewhere at Duke, Oddone and his team have been examining the use of carotid endarterectomy. A relatively new procedure, carotid endarterectomy has emerged in recent years as the treatment of choice for patients at risk for stroke from extensive blockages-70 percent or more-of the carotid artery, the major vessel in the neck that carries blood to the brain. In appropriate patients who are at low operative risk, the procedure can cut the risk of stroke in half.
Oddone began his research on the procedure about eight years ago, when a colleague and research collaborator, David Matchar, MD, was giving a presentation on the procedure's effectiveness in patients in general at 12 hospitals. It was one of those classic "forest and tree" moments, when something that should have been obvious but wasn't, suddenly came into focus.
"I was giving this presentation," Matchar recalls, "and someone in the audience stood up and said 'You can't present that. That's inflammatory. Your patients are 96 percent white.' I had been looking at the overall effectiveness and had never even thought about that. I went back to Gene Oddone and Ronnie Horner, our research partner, and said, 'We have to figure this out.'"
What made the number so stunning was not only that blacks accounted for much more than 4 percent of patients at the 12 hospitals but also that blacks generally are at higher risk of stroke than whites and suffer more disability and death as a result of stroke.
To Oddone and his research team, three possible explanations existed for the small percentage of black patients receiving the procedure. It could be caused by access-to-care issues such as insurance coverage and ability to pay. It could reflect clinical differences in the way the disease manifests in blacks and whites. Or it could somehow involve the patient-physician relationship and the decision-making process.
One Step at a Time
In a series of studies, the researchers began methodically testing each of those possibilities.
Like Peterson, in their first study, published in 1993, the Oddone group turned to the VA to test the access issues of insurance coverage and ability to pay. Reviewing all hospitalizations throughout the entire VA system for Fiscal Year 1989, they identified a group of almost 36,000 veterans diagnosed with either stroke or transient ischemic attack (so-called "mini strokes") who would be likely candidates for angiography to screen for carotid blockages and for endarterectomy. Of those, 3,535 underwent angiography during the year-long study period and 1,249 went on to have an endarterectomy.
Although blacks accounted for 18.2 percent of the veterans who had a history of stroke or TIA, they were only 9.8 percent of the patients who received an angiography and only 4.2 percent of those who had the endarterectomy procedure. Whites, on the other hand, constituted 77.1 percent of patients with stroke or TIA; 86.1 percent of patients who had an angiography; and 93 percent of those who underwent an endarterectomy. In all, blacks were half as likely to get the screening angiography and a third as likely to undergo a carotid endarterectomy. In a health care system where care was provided to veterans at no charge, ability to pay and insurance status were clearly not a factor.
As to the possibility that clinical differences accounted for at least part of the racial disparity, a number of studies over the years had suggested that atherosclerotic disease in general--and in the head and neck in particular--is different in blacks and whites. Very simply, scientists have long believed that the types of lesions that form in the smaller arteries inside the skull are associated with elevated blood pressure and blood sugar levels, while lesions in the larger arteries outside the skull are caused primarily by the accumulation of fats.
Because blacks are more likely than whites to have hypertension and diabetes, they tend to be more likely to form blockages in the smaller, inter-cranial arteries that cannot be reached by surgeons. To some extent, then, blacks overall should not be getting as many carotid endarterectomies as whites.
And Then There Was One
In a recently completed study now under review, the Duke researchers put the clinical difference hypothesis to the test. Working with the RAND Corporation, Oddone and Matchar developed a computer program that analyzed the entire complex array of clinical factors and then classified patients as either appropriate, uncertain (neither appropriate nor inappropriate), or clearly inappropriate candidates for carotid endarterectomy.
In the study, they found that a significant percentage of the black patients at risk for stroke were indeed inappropriate candidates for carotid endarterectomy based on clinical differences. But that explained only about half of the difference in use of the procedure by blacks and whites.
"Even when you account for appropriateness and the entire constellation of all clinical symptoms, there's still a significant racial difference," Oddone says. "And that's where we are now. By the process of exclusion, we're now zeroing in on the patient-physician interaction and how they reach treatment decisions."
As with any good research, as investigators move closer to finding an answer, their task doesn't necessarily become simpler, but more complex than ever. To say that the answer may lie within the physician-patient relationship only raises a multitude of other questions. As Oddone puts it: What do doctors bring to the table? What do patients bring to the table? What information is conveyed and how? What context do patients put it in?
The process by which two human beings attempt to talk about and decide upon medical procedures that carry a certain amount of risk in what are often life-and-death circumstances is an extraordinarily complicated endeavor-even more so when the two individuals are from very different racial and cultural backgrounds. And given the fact that blacks account for only about 2.6 percent of all physicians in the United States, that cross-cultural scenario is not a mere possibility but a virtual certainty for most black patients.
"No matter how culturally attuned and sensitive a physician is, the truth is most physicians don't look like these patients and don't understand things in the same context they do," Oddone says.
James Carter, MD, professor of psychiatry and one of the first blacks on the Duke medical faculty, has studied and written about psychosocial and cultural issues in medicine and psychiatry and how they affect the health care of black patients. He believes the Duke researchers and their counterparts at other institutions are on the right track and that the difficulty of cross-cultural communications-and the woeful lack of physician training in that area-plays a large role.
"I don't think physicians are capricious," he says. "I don't think this is being capriciously done. It's a matter of lack of awareness. There is very clearly a problem with communications between white physicians and black patients. It is human nature that people feel closer to people who are more like themselves in terms of education, socioeconomic status, and regrettably, race.
"But when African-Americans have it explained to them and understand that a procedure is necessary, I submit to you that you will not have African-Americans dropping out of treatment any more frequently than members of any other group."
It Takes Trust
In the interplay between doctor and patient, Carter says, particularly in a cross-cultural context, a whole host of factors may come together to make it more likely that white patients undergo certain medical procedures and blacks do not. Social support, overall familiarity with technology, and trust all come into play. While it is perhaps less so than in the past, Carter-along with several other researchers--theorizes that many blacks are distrustful of modern medicine, in sort of a lingering effect from past medical abuses such as the notorious "Tuskegee" studies. Indeed, some studies have indicated that blacks are more likely than whites to decline physician recommendations for treatment, particularly for aggressive, invasive procedures.
"When you stop and think about it, it's remarkable that so many patients, either black or white, are willing to undergo these procedures as readily as they do," Peterson says. "It takes a great deal of trust to allow someone whom you barely know to walk into your hospital room and tell you that you need an operation, and you're supposed to say, 'Okay, I'm going to let you operate on me and open up my chest and take my heart in your hands and make me better just because you say that's what I need to do.'"
At the same time, from the physician side of the encounter, when discussing complex procedures that carry a risk of death, doctors generally want and need to believe that patients want to have the procedure done, Peterson says. If they sense reluctance on the patient's part, they may be less willing to push the procedure.
"So, patients you don't relate to well, or patients you don't have a bond with, you're less likely to engage in the conversation that's necessary in order to convince them to have the procedure done," Peterson says. "To me, that's some of the most important part of this research. Physicians have to be aware that spending time with patients is valuable and getting patients involved in decision-making is all-important."
George, the medical sociologist, points out that the whole phenomenon becomes very self-reinforcing. If patients are part of a culture or a community with few physicians and whose members generally don't get many bypasses, they're not as likely to know many people who have had successful experiences with bypass and, as a result, become less likely to undergo a bypass themselves.
Ideally, trust is developed over time, by getting to know someone through repeated encounters. But when time is not available to get to know another person at a leisurely pace-as in the specialty care relationships where most such treatment decisions are made-people tend to fall back on the old standby, social similarity: Is this person of my religion, my race, my age? Do they look like me?
"It's just so much easier to assume trust with someone who is socially similar to us," she says. "And when you have poor black patients seeing socially dissimilar health care providers, who talk in words they don't understand, and who are pressed for time, and who probably want to get this meeting concluded and get on to the next patient, the conditions for creating trust are extraordinarily poor."
Eugene Wright, MD, a 1978 Duke medical graduate who joined the faculty recently after 15 years as a primary care practitioner in Fayetteville, N.C., says he was frequently called by black patients he did not know and asked to explain what some other doctor had told them.
"I thought that was odd at first," he says. "But then I realized this or that person felt comfortable calling me as an African-American physician and asking me about what the doctor they were seeing was telling them. Inevitably, what the doctor told them was correct, but their effectiveness in communicating might not have been up to par. Whether it was a timing issue, or a cultural issue, or whether they used the right words, or whether it was body language, or eye contact, I don't know, but I suspect all those things came to bear."
Turning the Light On
Under managed care, Wright points out, all the pressures that make for poor cross-cultural communications will only get worse. It's an approach to health care that assumes all patients are basically the same and that physicians treat disease states and populations of people, not individuals.
"If 98 percent of your patients are white and middle- or upper-class and have the appropriate support systems and someone who lives next door they can talk with about whether or not to get a bypass, it doesn't take long to establish communication and get that light to come on. But if you have a percentage of patients without that background and historical perspective, managed care has the potential not to be responsive."
For physicians today, it's more important than ever to know everything possible about a patient and his or her background, Wright says. Like many Duke-trained internists, Wright points to Eugene Stead, MD, emeritus chairman of medicine, as the prototype.
"Dr. Stead wanted to know all about the patient. 'Where does he live? Does he have running water? Electricity? Does he own his own home? Where does he work? Who are his family members?' He wanted to know all that, and with good reason. We need to understand patients' fears and hopes and get to know them and their illness in the context of their lives."
Of course, all of this-trust, cross-cultural communication, social support, and so on-is speculation, just some of the most likely scenarios that knowledgeable people cite to explain how disparities in treatment may be rooted to some extent in the physician-patient relationship and the decision-making process. Until now, few studies have gone beyond simply identifying variations in treatment within general categories of race and culture. But that is about to change, and Duke will be among those doing the changing.
Getting at the Differences
Under a new VA research program that seeks to explain ethnic disparities in treatment, Oddone and his team of investigators have been awarded a $1 million, three-year study of patient preferences and racial differences in the use of carotid endarterectomy.
"Basically, we're going to try to get at cultural differences in a scientific way," Oddone says of the new study. "We're going to interview likely candidates for the procedure and their physicians and follow them over time to see what factors predict who gets a carotid endarterectomy."
Oddone's hypothesis is that blacks have a greater aversion to the procedure than do whites and that racial differences in aversion are a function of trust in physicians, perceptions regarding previous personal or family members' experience with surgery, and "locus of control," an overall measure of one's belief in the role of chance, fate, or 'powerful others' in determining health. Some studies have suggested that people who have a more deterministic or fatalistic outlook on life and who believe life is essentially out of their control are less likely to undergo surgery.
In a pilot study, Oddone has already documented that blacks are more averse to undergoing the procedure. Using a theory called "standard gamble measurement," researchers presented 100 patients with a hypothetical carotid situation and asked whether they would rather undergo the surgery to remove the blockages or take a magic pill that could also clear up the blockages, but which had a significant risk of immediate death.
The researchers found that blacks had much higher aversion scores. About 90 percent of white patients were willing to accept the risk of carotid endarterectomy and were not willing to accept much risk of immediate death from the pill. While blacks, overall, also preferred undergoing the procedure, a much higher percentage--almost 50 percent-preferred taking the pill, with its immediate risk of death, rather than undergo the surgery.
"What we're going to try and do now is tease out what's behind this aversion factor," Oddone says. "We're going to sit down and interview physicians and patients and then follow them forward in time to see who gets the procedure and what predicts it. We're going to measure trust in physicians, social context, this idea of 'locus of control.' And at the end of this, we want to be able to say what's driving these decisions."
The First Priority
But for now, for most physicians, the first and perhaps most important step, Wright says, is simply to recognize that the problem exists.
"We have to recognize that we still live in a country that has subtle racism, some unintended and some not. If we approach it in less offensive, less polarizing terms, people come to grips with their own differences and insecurities. But even so, when we locked up Japanese Americans in detention camps during World War II, was that 'racism' or just 'cultural differences?'"
In any event, as Surgeon General-nominee David Satcher, MD, PhD, director of the federal Centers for Disease Control, has written, it's past time to start talking about the problem and get on with finding the answers:
"Much has been written about the doctor-patient relationship and its many challenges and ramifications. However, almost nothing is written about the effects of race on this relationship. This is unfortunate, since we live in an era of increasing concern for community medicine and for improving the delivery of medical care to all people. In discussing the effects of race on the doctor-patient relationship, our aim must not be merely to accuse or to place blame but to analyze critically the problem, with the goal of improving medical care delivery. Our silence on this issue tends to deny that we see racism as a major barrier to good medical care for many people in our society."