This video is part of “Innovations In: Solutions for Health Equity,” an editorially independent special report that was produced with financial support from Takeda Pharmaceuticals.
Have you ever gone to the doctor’s office and felt like they were judging you, maybe even before you opened your mouth? Unfortunately that’s probably a pretty common feeling and not because doctors are trying to be jerks. Like all humans, doctors have unconscious biases that can lead them to make unfair judgment calls. But those biases can pose a serious risk to people’s health.
Rachel Feltman: For Scientific American’s Science Quickly, I’m Rachel Feltman, and
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
I’m joined today by Dr. Cristina Gonzalez. She’s a professor of medicine and population health and associate director at the Institute for Excellence in Health Equity at NYU Grossman School of Medicine.
Cristina—can I call you Cristina?
Cristina Gonzalez: Yes. Thank you for asking.
Feltman: Thank you so much for joining us today.
Gonzalez: Thank you for having me.
Feltman: Let’s start with a basic question: What is implicit bias?
Gonzalez: So implicit bias refers to the unconscious and unintentional mental associations that we make about others, often along lines of personal identity factors like race or religion or gender. But they’re unconscious and unintentional.
Feltman: How much does that tend to come up in a clinical setting, and why does it matter?
Gonzalez: It’s more likely to come up when we’re pressed for time; when we’re fatigued; probably when we’re hungry, although that doesn’t exist in the literature, to my knowledge, anyway; when we’re not really knowing the person in front of us very well and we may have incomplete data—honestly kind of every day at work, right, at different times of the day, different time pressures, etcetera.
And the reason why it matters is because it can actually influence our communication behaviors with patients. And so I wanna be clear that implicit bias isn’t, like, a moral indictment. It’s a coincidence of our lived experiences, how our unconscious mental associations go, but because it can influence our behaviors, we wanna—we work on that in, in our lab.
Feltman: Tell me more about the work that you’ve been doing specifically on implicit bias. How did you get started being interested in that topic, and what has your lab been doing?
Gonzalez: So a couple years after I finished residency, one of my mentors, Dr. Paul Marantz, showed me an article that had been written while I was still a resident, so I hadn’t had a chance to read it. And it was about implicit bias and some medical decision-making. And I remember reading it and thinking to myself, “Hmm, so good people with good intentions, including myself, could be hurting patients. I could work on that.”
Seriously [laughs]. And the rest is, is kind of history. But I did a medical education research fellowship and did everything I could to do every homework assignment, in essence, as I got my master’s in medical education and learned how to do research in—whenever I could, affiliated with implicit bias, let’s say, as, as the topic, and then I really focused on education as the intervention to help us gain new skills. And eventually, after a few years, my lab made the framework of implicit bias recognition and management, which sets the foundation for our interventions to date.
Feltman: Well, I definitely wanna talk more about those interventions. But to zoom out a little bit ...
Gonzalez: Sure.
Feltman: What do we know about how implicit bias in a clinical setting can impact patients?
Gonzalez: How do you know when implicit bias has impacted the encounter, right? I know it’s because the vibe—I and others—that the vibe has changed, the nonverbals have changed, the patient may get a little more curt in their answers or shorter, you know?
And we do this in real life. And I keep getting people who say, “How do we, how do we teach that?”
Feltman: Mm.
Gonzalez: And, and I keep consulting other people. I’m like, “How do you teach it? Like, I don’t know.” [Laughs] And so I think we could talk about that as a challenge in case anybody writes in and gives us the answer [laughs]. I would love it so much.
Feltman: Yeah, maybe, maybe there’s, like, a, a body language coach out there ...
Gonzalez: Yes!
Feltman: An acting or movement coach who thinks about micro facial expressions. Like, somebody’s, somebody’s gotta be able to help with that.
Gonzalez: Yes—yes, yes, yes and yes.
If we stay focused on communication behaviors and communication skills, there’s the concept that’s called verbal dominance, meaning that if we have a 15-minute encounter, and if we are gonna center the conversation on racial bias, then if you have higher unconscious—implicit racial bias, more pro-white as a coincidence of your lived experience, then you’re likely to talk more in those 15 minutes when you’re seeing a Black patient compared to a white patient.
When you’re talking more, that means they’re talking less.
Feltman: Right.
Gonzalez: That means we’re likely to be asking their opinion less; we’re likely to be doing less shared decision-making, meaning getting their input on what the, what the treatment plan is, is it acceptable to them; asking if they have questions. Patients perceive less patient-centeredness. They perceive, in essence, a colder affect or vibe, if you will, in the encounter. And, and we also end up using more words that relay anxiety.
And I’ve learned from some cognitive psychologists that interracial anxiety is actually contagious if you’re white and you have interracial anxiety and you’re, you’re taking care of a Black patient or interacting with a Black person. But if a Black person has interracial anxiety, it’s actually not contagious, if you will, to the white person.
But it’s not a Black-white thing—I just wanna be clear. It’s the way we’re socialized and our unconscious ...
Feltman: Sure.
Gonzalez: Mental associations. So it’s just humans.
Feltman: Yeah. So let’s talk about what you’re doing at your lab. What interventions have you been working on, and what’s been working?
Gonzalez: So we’re interventionalists, right? But instead of needles or devices or pills, we use education. And so what we work on is, is, as I said, the framework of implicit bias recognition and management. And I wanna bring us back to that because it’s exactly what it sounds like. I want things to be easy to remember and easy to implement.
So we recognize when implicit bias may have impacted the patient encounter, right? And then we teach people skills to be able to manage that negative influence, negative impact—partner with the patient and then restore rapport, you know, discuss ways of moving forward, etcetera—to be able to have the positive outcomes we wanted in the first place. And so a lot of what we talk about is basic “humaning,” to be quite honest. And so—but people get nervous when it’s something about race or religion or gender or sexual orientation, and people worry.
Feltman: Mm.
Gonzalez: So if I can take a step back and explain that the—we were the first to study—we weren’t the first to study patient perceptions of bias and discrimination in their encounters, but to our knowledge in the literature, our lab was actually the first to study it and then stop and say, “Okay, great. Not great that it’s happening, but great that we’re talking about it, that you’re talking about it.” And then, “What should I teach myself and the medical students?” ’Cause at the time it was just—medical students was my audience. And they, they said, “Just apologize.” And I was like, “I’m sorry, what?” [Laughs] They were like, “Just apologize.” And I was like, “Oh—I can teach that! I can do that,” right?
And so it, it, of course, evolved from there, but we did a, a focus group study with Black and Hispanic patients in English and Spanish across New York City, across the socioeconomic spectrum, and over and over again they were so generous with us. They said, “We know you’re human; we don’t need the encounter to be perfect and no bias, etcetera, but once it happens we can’t have it ignored,” ’cause that second insult was a real attack on their, like, their, their core, their dignity.
And sometimes we may also—they may also perceive bias—we, too, when we’re patients—we also may perceive bias when actually it’s a routine question ...
Feltman: Sure.
Gonzalez: But based on our lived experiences, right—being followed around in stores, being accused of wrongdoing, being questioned about our identity, etcetera—we may take it as bias. And that can be hard as a physician or a medical student, being like, “I’m supposed to ask you this.”
And so we teach skills to step back and depersonalize it so that you can, again, partner, apologize if necessary or explain why and be able to move forward together. But it’s the ignoring part that was really, really hard for patients.
Feltman: Sure. Well, and I think a, a lot of folks who are familiar with implicit bias, you know, thanks to a great surge in research over the last few years, they might think that most of the battle is just becoming aware that we all have implicit bias because that’s really revelatory and difficult for a lot of people ...
Gonzalez: Sure.
Feltman: But it sounds like that is very much just the first step in actually mitigating it. Would you say that’s right?
Gonzalez: Absolutely. And so there’s been a lot of talk in the literature, and in the lay press as well, about awareness being enough, right? Because if we become aware, then our good intentions will prevail. If it were that easy, I would happily be studying something else, right?
And so I think that, for us, it’s becoming aware, and then—but it’s not fair to have people be aware and then say, “Go be better,” you know, because, of course, they were trying to be their best at the very beginning. And so that’s where the skill comes in. And I just wanna be clear that it’s unfair to make people aware and then not give them skills because then they worry that they’re hurting people more, and that’s really unfair.
Feltman: Yeah, that’s a really good point.
So like you said, I think most people really want to address implicit bias once they learn about it. And as you said, it’s not a moral failing to have implicit bias; we all have it. But what happens when someone is resistant to the idea that this is something they need training for?
Gonzalez: There are a few things you can do to address very resistant people. People who are explicitly biased, meaning on-purpose racist or on-purpose don’t like a religious group, for example, they’re beyond the scope of our interventions ...
Feltman: Sure.
Gonzalez: But for implicit bias, I think that you have to have safe places to fail. And in the educational literature that’s called a critical incident, and in a critical incident you may practice something and then have that internal aha! moment, and internal aha! moments happen in private. The fancy way of saying it is you capitalize on the aversive arousal of private hypocrisy, which basically means that you get this reaction that you don’t like when your actions don’t match your values. If you could have that privately, right—maybe with a interactive computer case, for example—then you’re more likely to change your actions to match your values. If you have it publicly, you’re actually at risk of changing your values to match your actions ’cause you feel the need to save face.
Feltman: Right, that shame response, it’s really hard to avoid.
Gonzalez: And so I share my failings. Like, I do this for a living, and I still catch myself and have to go back and have conversations with patients [or] students, etcetera.
The other thing is that for the people that are doing this work—I learned this, and so I hope, I hope it’s helpful to other people—I kept trying to get that very resistant group. And there’s this probably 25 percent that no matter what I do, they’re gonna come back ’cause they’re into it; they’re social justice people. It’s that middle 50 percent that is sort of like, “Hmm, I’m not sure. All right, I’ll listen to this lady,” and then they keep going, and then you can really bring them in. But if you spend a lot of time trying to get that one or two, which still happens—we just did a seminar, like, seven or eight days ago where it, it definitely has a very resistant person—if I kept trying to teach to that person, then you, you lose everybody else ...
Feltman: Mm.
Gonzalez: And so we just do our best but really focus on the people who are open to change and just want some skills.
Feltman: Yeah, absolutely. It sounds like a lot of it is just about recognizing that we’re all human ...
Gonzalez: Mm-hmm!
Feltman: And trying to do our best and giving people the tools to actually do so. So speaking of those tools, how do you go about measuring implicit bias in, in a clinical setting? Is there a way for you to see how much work there is to do?
Gonzalez: There are tests, like the Implicit Association Test; it’s a free and publicly available test. There are other tests, and they’re what’s—what are called latency response tests, meaning they see how quickly you react to two opposite concepts—joy, evil; male, female; just thinking about possibilities—down to the millisecond. So see how quickly you react to different concepts. And that is supposed to measure your implicit bias.
They’re interesting, and I don’t use ’em diagnostically. They tend to mirror what we would expect in terms of the way we’re socialized.
Feltman: Sure.
Gonzalez: But if you change someone’s implicit bias, you have to do all these experiments, they’re very time-consuming, and then it doesn’t actually change behavior. What we strive for is, is to see the communication skills and the—so we can lead to those aha! moments.
So in the clinical setting, as you were saying, how would you measure it? We made high-fidelity, meaning very realistic, simulations that have the types of stressors you would have in the clinic, so interruptions, a very pleasant but somewhat meandering patient—because they’re not a professional historian, okay ...
Feltman: Yeah.
Gonzalez: They’re telling the story the way it happened to them—and various things in, in the clinic that you would see, in a simulated encounter.
Physicians didn’t know why they were going into it. Standardized patients, I hope, are not watching because they were blinded to [laughs] the purpose of our study and were phenomenal. And so nobody knew what was going on. And so we actually were able to measure the physicians’ racial implicit bias on that Race Implicit Association Test at the end so that they—all the behaviors had already happened and correlate it with communication behaviors.
And so you’ll see it, again, in patient education, establishing rapport, eliciting all a patient’s concerns, active listening—like, “How’re you doing right now?”—it actually makes a difference in those types of behaviors. And also things like interpersonal distance.
Feltman: Mm-hmm.
Gonzalez: So when you have higher bias, I would say, against the person or pro the opposite of that personal identity factor, you’ll end up being further away.
Feltman: Mm.
Gonzalez: And so that’s something that we can actually monitor. I always tell people, like, you know, “Take—put a, put a coffee mug, and make sure you don’t go behind it.” The patient’s not gonna know why there’s a coffee mug. They were always drinking coffee, right [laughs]?
Feltman: Yeah.
Gonzalez: But, but that’s a nice little cue for yourself of, like, “Oh, let me lean in a little bit and make sure that I’m looking at the person, connecting, making that eye contact, etcetera.” And so it’s in—honestly general communication skills, but it’s different with different patients.
And so we video-record them and then are able to analyze and find the inflection points of where to teach to be our best selves.
Feltman: That’s so interesting because I, I think there’s a real stigma in general around having to learn and practice things that maybe we’re told are just: “That’s just being a person, acting like a person. That’s having a normal conversation.” What does that teaching and learning process look like for medical professionals?
Gonzalez: We call them critical humanism skills. Some people like to call communication skills “soft skills,” but they’re not—I mean, I love warm and fuzzy, and so—but they’re not, right, ’cause that implies that other skills are harder, whereas it’s, it’s very difficult to communicate effectively. And so we call them critical humanism skills.
Last week, actually, with some first-year medical students—they were so good—we practiced role-plays ...
Feltman: Mm-hmm.
Gonzalez: Where a student had the role of the physician, the—another student had the role of the patient, and when you had the role of the patient, of course, you’re not gonna be a 19-year-old Black man or a 22-year-old, you know, young woman who identifies as a lesbian if you’re not, but it—we practice perspective-believing, so they don’t see what each other has and the instructions each other has. And I ask them to act very well, [as] if you’re trying to win a Tony Award.
And so they act out the role-plays. And in the—in one of the scenarios, the, the physician asks a routine question, but the—based on the patient’s lived experience, they perceive it as bias. It’s, you know, “Do you smoke cigarettes?” kind of a question to the 19-year-old-young man who hurt his arm playing tennis, right? And they practice what we call verbal procedures, which are verbatim statements of actually apologizing, or acknowledging if you’re not comfortable apologizing, and restoring that rapport.
Feltman: Mm.
Gonzalez: They say it to the other student word for word, the other student gives feedback, and then they practice it again. And then they, they articulate it verbatim again, and then the student—and then they decide on a final. But that do-over has been called a “gift.”
The other one, if I may, is a case where it’s a young woman—this actually happened; it’s de-identified, but it actually happened to a young woman—she is on oral contraceptives for endometriosis, which is a painful condition in your uterus. Long story short: low-dose, in essence, birth control pills will help the symptoms.
Feltman: Mm-hmm.
Gonzalez: And we’re helping her, and she went to a new doctor and asked for this, and they said, “Well, this is better at preventing pregnancy.” And she’s like, “No, I don’t really need it for preventing pregnancy.” And [they’re] like, “Well, no, this is really better.” And she’s like, “No, I only have sex with women.” And the physician’s like, “Well, it’s New York; you never know.”
Feltman: [Inhales deeply]
Gonzalez: And, and I’m sure the physician wasn’t a terrible person. I’m sure she was flustered or who knows, right ...
Feltman: Right.
Gonzalez: But, but that—it’s not to vilify the physician at all ...
Feltman: No, no ...
Gonzalez: But, but it gives—and I understand your reaction ’cause that’s how the patient felt.
Feltman: Yeah.
Gonzalez: Right?She’s like, “Well, no, I actually, I actually do know.” [Laughs] And so, so I—so we don’t make, make the students make the actual biased statement, right, the assumption, but when they role-play it, you can always tell when they get to that part ’cause someone’s always like, “[Gasps] I am so sorry I made the assumption.” And, and they are able to practice in a safe space.
But I was talking to a surgeon, and she was making the analogy of building muscle memory. So if you’re like, “What would you say?” You’d be like, “Oh, I’d, I’d apologize.” “No, no—what would you actually say?” And people are always like, “Hmm.”
And so you build the muscle memory so that you’ve got those in your clinical skills toolbox, and you’ve already practiced restoring rapport or apologizing when you actually make an assumption, right—either way. And then you’re ready for when—you’re more ready than you would have been when it happens in real life.
Feltman: In what ways does our medical system or, you know, the current establishment fuel implicit bias? What needs to change there?
Gonzalez: Huh.
Feltman: I know that’s kind of a big question [laughs].
Gonzalez: I, I was gonna say, “How much time do we have?” No, but, but—well, so, again, it’s not intentional, right? No one’s in a back room being like, “Heh, heh, heh, we hurt 10 patients today,” right? So I get that. But it—there’s a couple of things. There’s education.
Feltman: Mm-hmm.
Gonzalez: So we end up pathologizing race or pathologizing sexual orientation, meaning, meaning that, like, we’ll say, you know, “Race is—Black race is a risk factor for high blood pressure.” Actually, no, it’s not. Racism is, but race is not. Or we’ll do things like whenever there’s a sexually transmitted infection type of case, it’ll be, you know, a, a gay man. And, and we make these stereotyping cases, which I don’t think anyone’s doing intentionally, but you have to intentionally not do it and make sure that, you know, we put different humans in, in all kinds of different cases when people are learning.
Feltman: Yeah.
Gonzalez: So that we don’t reinforce stereotypes. And—because that’ll actually limit our diagnostic abilities later.
The time pressures, because we’re all socialized in, in the system—I’ve only ever lived in the United States, so if we’re thinking about how we’re socialized here, you think about movies, commercials, newscasts, all the things that we see, [that] could become ingrained in our minds, and then we end up having those unconscious mental associations, implicit bias.
And so what ends up happening is appointments are, like, 15 minutes long, sometimes 10 minutes long. The way that people are forced to churn, for lack of a better—terrible word, sorry—through patients because of systemic pressures, right—and it’s not even like your hospital that’s saying that; it’s really just the way the whole system is—makes you have to have these snap judgments.
You know, we were just at a National Academy of Medicine diagnostic equity meeting, and we were talking about, “How can we make it—a change so that we have more time?” And that way—I think that’d be more cost-effective, right, because you’d be able to actually have time with your patient.
And the last thing I’ll say is that the burnout is, is real. And so many physicians and all clinicians—I happen to be having a physician kind of centric conversation right now, but we know nurses are working really hard, the respiratory therapists; I mean, everybody’s working really hard. It sets us up to make snap judgments because efficiency and what’s called throughput, right, getting patients through, is the pressures that we’re—under which we’re working.
Feltman: Yeah. You know, beyond the obvious individual risks to a patient, how is this impacting health care?
Gonzalez: Think about how implicit bias influences communication skills. And then we think about how—the patient experience, right, patient satisfaction, the patient experience scores, for example, have an influence on reimbursement rates. And really what we want is to help patients have a better experience, right? But if we can get that system-wide driver to say, “Ooh, and it’ll help your reimbursement,” maybe we could get some buy-in.
On the missed—misdiagnosis and thinking about making assumptions about patients, not believing patients, there’s a concept of testimonial injustice, meaning that you’re less likely to be believed for what you’re actually saying if you’re, you know, a woman, if you’re of a minoritized race or, or ethnic background. So if you don’t believe the patient or you kind of dismiss them—there’s lots of literature in them for women, right—and then you end up having delayed diagnosis, all those costs to the person and to the system could be avoided.
Addressing implicit bias isn’t—I’m not trying to make it a—paint it as a cure-all, right ...
Feltman: Sure.
Gonzalez: It’s a—health equity and health disparities is a very what’s called a wicked problem, right? It’s very complex. So I think it is a important tool to help us take excellent care of all of our patients and achieve health equity.
Feltman: Has the conversation around implicit bias changed in, in the time you’ve been doing this work, and how do you want the conversation to change in the future?
Gonzalez: It has changed—for an unexpectedly not so great way, in one sense. Before, when I started—I did my fellowship from 2010 to 2013. So it’s been—so I’ve had my—working in my lab since fellowship, for about 10, 11 years now—11. And so in the beginning people were very upset about it, about the thought of having bias. They assumed other people did, “But not me; I’m a good person,” right? So we worked a lot to get that moral judgment out of, out of the conversation. There were some really difficult kind of yelling [laughs uncomfortably] conversations back then.
And then, and then as we continued to adjust our messaging as well, right, to make—to help engage people, then the pandemic happened. And during the COVID pandemic, there was the twin pandemic of, of racism being more obvious to the rest of the—to the general population ...
Feltman: Sure.
Gonzalez: I don’t wanna say that racism was a new pandemic ’cause it was there the whole time; it just became very obvious. And so then we had this thing of people really wanting to help. And all these implicit bias programs started springing up. And at times the good intentions were more than the actual rigor or scientific methodology, if you will.
And that actually has made it really frustrating from, like, I would say, ’21—2021 to 2023 to have these conversations ’cause now people will roll their eyes. They’re like, “Ugh, I’ve been to one of those,” you know? And I often tell people—and I appreciate that people care about racism, and it’s okay that—if they didn’t—it didn’t affect them before, personally, right, that they’re aware of it now and they wanna do something; I am all for that. I think it’s just: partner with experts and people who’ve been doing the work forever. Just as if your family member got leukemia—God forbid—your family member got leukemia, you wouldn’t grab, like, one of those pipettes, run into a lab and just start discovering, right ...
Feltman: Right.
Gonzalez: [Laughs] You would know what you could and couldn’t do. And so I think that we should respect the work—I don’t even mean my lab; I just mean in general—respect the work of the scholars that have been doing this forever, brilliant, like, Black feminist theory and all the things that have been happening before, because then we can have really good programs that are effective and capitalize on, on the people’s interest right now. And hopefully we can make this moment a movement.
Feltman: Yeah.
Gonzalez: And actually make change, right?
Feltman: Yeah, absolutely ...
Gonzalez: Yeah.
Feltman: And in an ideal world, you know, five, 10 years from now ...
Gonzalez: I’m gonna put on my ideal word—world hat. I’m ready.
Feltman: How do you want medical schools and hospitals to be thinking about implicit bias?
Gonzalez: So I want this to be just like anything else. When students and, and residents get the wrong diagnosis, we help them get the right diagnosis. Even when we get the wrong diagnosis, we work on getting the right diagnosis. I want this to be just like any other clinical skills, any other communication skill, any other diagnostic and treatment skill. I want implicit bias, whether it’s race or, or any of the constructs it can be, right, ’cause it’s not—we don’t have to have a special class for race if we were having that communication change.
And so I would say that I would love it just to be another clinical tool and that we can have, if I—if you may indulge me an example, that we could have what we have now, which—analogous to what I call, I should make it more gender-inclusive, but it’s—we call it the “I got a guy” consult. And you’re sitting there, and everybody’s really busy, and you’re all on the computer after you’ve seen your, your patients and writing your notes, and you’re like, “Hey, I got a guy.” And the least busy people roll their chairs over, and you run the case by them. And I love when they roll their chairs to me, and I love rolling my chair to them.
I really wish that one day we could have, you know, bias buddies and these “I got a guy” consults—and, again, forgive the, the term—but the “I got a guy” consults could also be about: “I was talking with someone today, and I think maybe I said a biased thing, or they maybe perceived a biased thing.” And I won’t worry that you’re—you think I’m racist, just like I never worry you think I’m a bad physician when I do my “I got a guy” consult.
Feltman: Yeah, absolutely.
Gonzalez: That would be great [laughs].
Feltman: Cristina, thank you so much for coming in to chat today ...
Gonzalez: Thank you as well.
Feltman: And for the extremely important work that you do.
Gonzalez: Thank you so much for having me. It’s been an honor. Thank you.
Feltman: That’s all for today’s episode. But if you want to hear more of my conversation with Cristina, you’re in luck: a longer version of our chat is available on Scientific American’s YouTube channel. You’ll find the link to that video in our show notes.
Science Quickly is produced by me, Rachel Feltman, along with Fonda Mwangi, Kelso Harper, Madison Goldberg and Jeff DelViscio. Shayna Posses and Aaron Shattuck fact-check our show. Our theme music was composed by Dominic Smith. Subscribe to Scientific American for more up-to-date and in-depth science news.
For Scientific American, this is Rachel Feltman. See you next time!