Study Points to Emergency Room Bias Against Blacks A new study from Harvard suggests that racial bias affects the treatment that many African-American patients receive in hospital emergency rooms. Two authors of the study discuss the research and its implications.

Study Points to Emergency Room Bias Against Blacks

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript


I'm Cheryl Corley. This is TELL ME MORE from NPR News. Michel Martin is away.

Coming up: a police scandal rocks Puerto Rico, a son keeps his promise, and tracking used clothing. But first, it's safe to say that most of us expect fair treatment and an accurate diagnosis when we have to pay a visit to our doctors or, God forbid, find ourselves in an emergency room.

But what happens when our ethnicity plays a role on the type of treatment we receive? A new study published this month in the Journal of General Internal Medicine says some medical professionals actually do have subconscious biases about race or gender that influence how they diagnose and treat patients.

Well, joining us now are two of the study's authors. From Boston, Dr. Alexander Green of Massachusetts General Hospital, and with us from Seattle is Dr. Mahzarin Banaji. She helped design the study measuring bias. Thanks for joining us.

Dr. ALEXANDER GREEN (Senior Scientist, Institute for Health Policy, Massachusetts General Hospital): Thank you.

Dr. MAHZARIN BANAJI (Social Psychologist, Harvard Medical School): Thank you.

CORLEY: Dr. Green, let's start with you, with an overview, perhaps, of this subject. Can you explain how biases about race or gender might influence the ways that doctors or nurses diagnose or even treat certain patients?

Dr. GREEN: Well, we have a very serious problem in health care today. The problem of racial and ethnic disparities in care - that is that the quality of care that minority patients get in this country is not as good as the quality of care that white patients receive. And it's not just an issue of socio-economic status or insurance. We know that from a number of studies, this has been extensively documented in hundreds of studies - probably close to a thousand now.

Now our study focused particularly on heart disease. And so we were interested in the number of studies that had documented that, in fact, African-American patients who present with a heart attack to the emergency department are less likely to receive a potentially life-saving, clot-busting medication, or a treatment called thrombolysis.

CORLEY: Dr. Banaji, you designed the survey that measured these biases. Tell me why these findings are significant.

Dr. BANAJI: First of all, let me just say that this is actually a very different test than what we might call a typical survey.

CORLEY: Mm-hmm.

Dr. BANAJI: In a survey, you would pose questions to the respondent - what did you eat for breakfast this morning? Are you racially biased? Things like that. And what this test tries to do is explicitly get away from posing those kinds of questions. We do that in part because we know that people may be not very willing to reveal the truth about what goes on in their minds and hearts. But especially because we don't think that most human beings or any - really have direct access to the contents of what goes on in their minds.

So it's not that they're trying to lie to us. It's not that they're being insincere. It is that we simply don't understand what goes on in our minds. So these tests are a part of a movement in the mind sciences to try to get access to what goes on in our heads that we may not be aware of. And so it is actually a test that is quite different from a survey. What it tries to do is ask people to very quickly associate two things. And we measure how quickly they can put them together. And from that speed and from the accuracy with which they can or can't put two things together, we can deduce whether those concepts are associated in their mind or not.

So for example, in your mind, Dr. Green and I have come to be associated because we're colleagues, and you're interviewing us together.


Dr. BANAJI: Those two entities for you can, over time, become one. When you think of Dr. Green, you'll think of me, and vice versa. So what we do is try to test how that might actually work in the case of something like a race bias. When you are asked to rapidly associate the category or the group African-American with cooperative, will take their medication, good people, things like that, how quickly do you respond to that? Versus how quickly do you do that when you have to do that with white? How quickly do you associate the category of white people with cooperative, nice, good, et cetera?

And what we've discovered - this is well before the study was conducted, based on research from millions of test takers. We know that most white Americans and Asian-Americans have a very difficult time associating black with good in general. So what we did is pose the question, well, would doctors show this kind of bias? It was our sense that doctors should be like most other people and show it. But we had to do the test. And we discover, first of all, yes indeed, they're like everybody else. They have the bias.

But the second part of it is the more interesting part. What we did is look at the level of bias in a given doctor, and we then tested whether the level of the bias that they have - let's say a doctor has a low level of anti-black bias. Does the doctor with low bias actually treat people better - treat African-Americans better than the doctor with medium bias? And does the medium bias doctor do a better job than the doctor with a high level of bias.

CORLEY: And what did you find out?

Dr. BANAJI: What our study has shown is, indeed, that that's the case. That level of bias in the mind corresponds to the likelihood that you will prescribe this particular blood-thinning procedure called thrombolysis.

CORLEY: That's very interesting. I must say that I took the test - or portion of the test, since it is online, to find out what my bias was.

This is TELL ME MORE. I'm Cheryl Corley. We are speaking with two doctors who wrote a study about subconscious racial bias among medical professionals - Dr. Alexander Green and Dr. Mahzarin Banaji.

And let me ask you this: Why does bias have this potential to change the way doctors might provide treatment? You say that you're actually seeing this and have been able to measure this. Why does it have that potential?

Dr. BANAJI: It has the potential to affect our behavior because, really, you know, what goes on in our minds is quite invisible to us. We can't see it. So it's very hard for us to imagine the things that we don't have access to in our minds could be affecting our behavior.

CORLEY: But, is it, for example, doctor, that some professionals might have an idea that African-Americans might not follow up on medication schedules or aftercare or things like that?

Dr. BANAJI: I'm going to argue that all of us carry those associations around in our minds whether we want to or not. Our culture creates those meanings for us, and we have them.

CORLEY: Dr. Green, some patients are reluctant to speak honestly with doctors because of the big power differential in the relationship, you know, doctor-is-God syndrome. That can put a lot of people off. Does that play any role in this dynamic?

Dr. GREEN: You know, it may. It certainly may. I think that patients certainly have a responsibility to themselves to try to speak up for themselves, and we do a lot of education - both for physicians and more broadly - to try to get people of all cultural backgrounds to know how to work with their physician better and to get physicians of all different backgrounds to know how to work with patients from different perspectives better. And so I'm co-chair of a committee at Harvard Med School focusing specifically on educating Harvard medical students how to do better in their interactions with patients.

CORLEY: Is it ever okay - and this is for both of you - is it ever okay for doctors to do this kind of profiling of patients in any sort of way?

Dr. GREEN: Well, in general, I think that profiling is a bad thing. I think that profiling takes away the use of objective information and allows people to use their subjective experiences to influence the way they make decisions. And I think it's incumbent upon all of us as physicians to try to use as much objective information as we can and not let that objective information be swayed too much by the subjective.

Dr. BANAJI: Although I would agree with what Alex said, but let me give a little bit of a twist to this answer about is profiling always bad. Part of the problem with any kind of profiling is that everybody in a given culture uses exactly the same sorts of rules to do the profiling. That's, in part, what makes it difficult.

So, in a sense, if doctors in multicultural society want to actually think about doing a better job, it wouldn't hurt, in a sense, to multiply the kinds of profilings that they do. A Vietnamese woman comes into their office, maybe if they knew something about the culture of how women in Vietnam think and do their stuff, maybe they can ask the right kinds of questions. In that sense, we would say relying partly on knowledge about the group could actually help the doctor draw something out.

I think what's difficult about profiling is - whether it's police officers or doctors or teachers - is that all of us use the same two variables to respond to people in the same simplistic way, and that's where some of the problems lie.

CORLEY: Were the doctors surprised, that took these tests - were they surprised by the results?

Dr. GREEN: You know, it's interesting that you asked that, because I expected that they would be very surprised. But when we started the test, we actually asked them a question around the likelihood that they thought that unconscious biases might impact their decisions. And 60 percent of them either agreed or strongly agreed that that was true, that unconscious biases might affect their decisions.

After the study, it was more like 72 percent who said that they agreed with that statement. And so, we saw that there was a certain amount of learning of the effect of subconscious biases on decision making by being a part of the test.

Dr. BANAJI: What Alex is bringing up is, is really the crux of the issue. We know what goes on in our minds consciously, so it's very easy to say, well, after this phone call, I'm going to go downstairs and eat some breakfast. That's in my mind. I can feel it tangibly. I then see it driving my behavior, and there is a nice one-to-one correspondence between what my brain is saying and what my behavior is doing.

The reason the unconscious stuff throws us all off so much is that stuff that we can't palpably feel or know is, in a sense, affecting our behavior. So it's really - it takes elite to be able to recognize that that's the case, and so I gave doctors and anybody who can come to terms with this a great deal of credit when they say, yes, unconscious biases do indeed affect my judgment and my behavior because I regard that to be the very first step. If we have to do anything about this problem that we're discussing, we're going to need some very unusual kind of awareness about what it is that we need to do. And the first step is going to be to say, yes, it's the first step of any program, I guess, and this is no different.

CORLEY: That was Dr. Mahzarin Banaji, who helped develop the groundbreaking survey that allowed her and the other researchers to measure subconscious biases in medical doctors. Also joining us was Dr. Alexander Green of Massachusetts General Hospital, where he sees patients and helps direct the Disparities Solutions Center. Thanks to you both.

Dr. GREEN: Thank you very much. It was a pleasure.

Dr. BANAJI: Thank you for having me.

Copyright © 2007 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.