Medical mistrust: Why representation matters

In this episode of Econ To Go, Stanford physician-economist and MacArthur Fellow Marcella Alsan joins host Neale Mahoney, the Trione Director of SIEPR, to explore how trust and representation shape U.S. health care. Her research shows that historical events like the Tuskegee Syphilis Study continue to affect health care use and health outcomes today, and that trust isn’t abstract, it’s measurable. 

“What we found was that in fact, health care utilization dropped; over time chronic mortality increased; and medical mistrust also increased among Black men,” says Alsan, who is the Thomas J. Davis, Jr. Faculty Scholar at SIEPR. “Trust matters.” 

The conversation also highlights how trust can be built, how underrepresentation in clinical trials can influence both physician behavior and patient trust, and other systemic aspects.

 Watch or listen wherever you get your podcasts. 

     Apple Podcasts     Spotify    Amazon Music      YouTube  

  Follow along as we cover key themes, including:  

• (01:33) The mistrust problem
• (06:50) Representation as remedy
• (12:18) Clinical trials and trust in data
• (24:04) Eroding trust across the system

Alsan is the Annie and Ned Lamont Professor of International Studies and Professor of Economics at the School of Humanities and Sciences, and director of the Health Inequality Lab at Stanford. To learn more about her work on the economics of health disparities and the role of institutions in shaping health outcomes, explore her bio and these links:

• Tuskegee and the Health of Black Men
• Does Diversity Matter for Health? Experimental Evidence from Oakland

• Representation and Extrapolation: Evidence from Clinical Trials

  ---

  Econ To Go brings Stanford economics into your everyday life — served with a side of coffee. Hosted by Neale Mahoney, the Trione Director of SIEPR and TG Wijaya Professor of Economics, the video podcast takes you on a walk across campus with leading thinkers as they unpack the ideas shaping our economy, public policy, and daily lives. Smart, curious, and conversational — it’s Stanford economics, to go. 
  Explore the series

  TRANSCRIPT

  Econ To Go: Episode 04

Neale Mahoney: The United States has never truly had a universal healthcare system, and for some communities, it's more than policy. It's personal and about mistrust.

Marcella Alsan: They were deceived, they were told that they were being treated for bad blood when, in fact, they were just being experimented upon.

Neale Mahoney: I'm Neale Mahoney, Economist and Director of the Stanford Institute for Economic Policy Research. On this episode of Econ To Go, I'm joined by physician economist and MacArthur Genius fellow Marcella Alsan. We grabbed a coffee and talked about how trust and representation shape our healthcare system, not just historically, but today. When people ask you what you do, what do you tell them?

Marcella Alsan: I say that I function as an academic economist right now. And once we solve this health and inequality thing, I'll go back to practicing.

Neale Mahoney: In our conversation, we'll explore Marcella's groundbreaking research on racial bias and medical mistrust. We'll start with the legacy of Tuskegee and how that history still echoes through the healthcare system today. And then we'll look at the role of representation in clinical trials and how it matters more than people realize. Because mistrust isn't just a feeling, it's measurable, and so are its effects. I started by asking Marcella what's the research that's helped her understand trust, not just as a concept, but as a powerful force shaping health itself?

Marcella Alsan: I think what some of the research that I've done is, shown me is that trust is the tagline, is like, "Trust is earned." And I think the difference between basically assuming that something about the utility function and the functional form or whatnot versus really trying to empirically establish situations of distrust or higher trust, whatever it might be, is that there's kind of a this notion that we have that people are rational, that they update when given information. And so, when there are episodes in which a group is exploited or they are not able to access resources that other groups can access, that that can influence their beliefs about whether that institution, whether that system is really for them.

Neale Mahoney: So, let's sort of dig into your research. So, you won a MacArthur Genius Award in 2021. It's huge honor and very well deserved. And that award sort of cited two papers, one on how trust can be lost and one on how trust can be rebuilt. I'd love you to sort of tell us about that research.

Marcella Alsan: The name of the study is The Tuskegee Study of Untreated Syphilis in the Negro Male, so that tells you three things off the bat. The study is going to take place in Tuskegee, Alabama, there'll be individuals who have syphilis who will not be treated for that disease, and that the experiment is specifically targeting Black men. Further history is that it was a study that was carried out by the United States Public Health Service between 1932 and 1972. Halted when a whistleblower, who was a student at the time, was actually working for the Centers for Disease Control, and was appalled based on, you know, his own personal connection with the Holocaust and medical experiments on imprisoned people during World War II. And, you know, first brought it to the attention of the medical authorities, which was dismissed, and then brought it to the media. And so that's why 1972 was actually when the Associated Press ran a story, which was picked up by several news outlets. So, that gave us another, you know, piece of data, which is the time, and so we combined that variation to look before 1972 and after 1972 to look what happened to Black men in the population in terms of their healthcare utilization, in terms of their chronic disease mortality, and in terms of their mistrust. And what we found was that in fact, healthcare utilization dropped, over time, chronic mortality increased, and medical mistrust also increased among Black men.

Neale Mahoney: So, trust matters.

Marcella Alsan: Trust matters.

Neale Mahoney: It's not just a feeling. It has real consequences on people's health-seeking behavior and on their health outcomes.

Marcella Alsan: And may I add, like, what end policies matter. How we treat or fail to treat a group matters for not just... You know, in the Tuskegee study itself, there were 600 men. And syphilis is a sexually transmitted disease, easily curable, even today with penicillin. Those men were denied that care. They were deceived, they were told that they were being treated for bad blood when, in fact, they were just being experimented upon. And then when they died, their burial costs were taken care of, so that was the financial incentives. But this paper is not a paper about those 600 men, the wives, or partners they infected, and the children that were stillborn because it's also maternal to child transmission. This study is that when you harm a group in that way, rationally, they update. They update on physicians and what they are doing to them or what they are not doing for them. So, I think that's a very microeconomic-founded thing. And so, it's empirically based mistrust.

Neale Mahoney: And at least it makes me think, you know, if this event had all these consequences, then all of the other betrayals of trust we see in the society probably also have consequences. And if you start accumulating across all of those, you can start to explain the broad-based lack of trust and disparities we see, you know, in our healthcare system. So, that research, I think, is quite deflating, but you've also done work on how we can repair trust, I know, in Oakland. Do you wanna tell us a little bit about that research?

Marcella Alsan: We randomized men from Oakland, Black men, to receive physicians of different racial backgrounds, so they were all male physicians, but some of them were Black men, some of them were not Black men. And to be honest, our initial idea was, let's just partner with clinics in the Bay area, in the Oakland area. But the scarcity was so severe. We couldn't find one existing clinic that had one Black male physician, much less more than one, which you would need to randomize. So, we ended up standing up our own clinic. In the jargon, we did a double-blind randomized trial, which means that neither the physicians nor the patients knew that we were actually randomizing them on this dimension of whether they saw a racially concordant or discordant physician. And what we were interested in was the take-up of preventative care, because preventative care is one of those situations where the patient, him or herself, feels quite fine, but they have to trust that the physician actually has their best interest in mind or whomever the expert is. So, the types of care were, you know, basic screening for cardiovascular disease, so blood pressure, body mass index, which is essentially standing on a scale and taking your height. So, those were what we considered the sort of non-invasive tests. Slightly more invasive but still pretty benign are things like diabetes screens and cholesterol screening. That actually just required a finger prick of blood. And then maybe the most kind of invasive was actually an injection, so a vaccine against flu. So, those were the five preventatives that we offered.

Neale Mahoney: And what was the impact of seeing sort of an own-race physician on uptake of these preventative measures?

Marcella Alsan: Yeah, so, the study was in two stages. So, in the first stage, where they just were seated in their physician's consultation room and had a picture of the doctor they were going to see and had an opportunity to select what services they would like on a tablet, there was actually no effect. So, there's no effect of being randomly assigned to, you know, a racially concordant physician for Black men if it was just kind of a hypothetical thing where you would look at a tablet. But then when the doctor, you know, on the screen actually came in and talked to the patient and they were able to re-optimize or re-choose what sort of services they demanded, then we saw pretty substantial effects, and they were much larger on the order of 20 to 30% came-

Neale Mahoney: Those are huge effects.

Marcella Alsan: Those are very large effects, you know, off of a smaller base, but substantial, I think, from a medical standpoint. But basically, as the invasiveness of the test increased, so, too, did the effectiveness of the racially concordant physician in terms of, you know, convincing or improving demand for those items.

Neale Mahoney: Makes complete sense to me, that if you're doing something which is a bigger lift that you need more trust to be willing to go through the hassle or go through the pain of, you know, seeking that care. Were there impacts that went beyond what you could measure in that physician-patient relationship?

Marcella Alsan: The whole point of doing the work in the community is that these are people that are often outside the system, you know, and we don't have a universal system. So, it would've been lovely to connect it to all sorts of administrative databases, but I think the linkage rates would've been shockingly low.

Neale Mahoney: Right, it's sort of a fundamental problem, the people where there are the largest gaps in trust. Those are gonna be the situations where we have least coverage in our traditional data sources, and that's why you have to go to those heroic efforts of forming your own sort of healthcare company, recruiting physicians, so that we could learn about that understudied population. Looking at healthcare through both a medical and economic lens, Marcella uncovers some of the system's deepest truths. During our campus coffee break, I asked her how our clinical work compares to academic research.

Marcella Alsan: I mean, I think I really enjoy clinical work. I love it. There are some similarities with teaching actually in the sense that when you see that sort of moment of epiphany in a student's eyes, it's not dissimilar to cinching a diagnosis.

Neale Mahoney: It's a telling insight, because when trust is real, when patients feel seen, engagement shifts. That type of trust doesn't come from theory. It comes from proximity, representation, and human connection. Next, Marcella and I turn to clinical trials and why who gets included in the data can reshape outcomes across the board.

So, you've, you know, done foundational work on representation in clinical trials. Tell us a little bit about the gaps in representation we see now, what they may arise from, and what the consequences of those gaps are for sort of people's healthcare.

Marcella Alsan: So, if that Oakland study is about sort of representation on the physician side, let's now think about representation on the patient side. For medical trials, for medical products, the key input into the test data are actually people. Things may change, but I think the standard still is, for the foreseeable future, how does this actually work? Not just in mice or not just in, you know, in avatars. How does it work in actually humans?

Neale Mahoney: And correct me if I'm wrong, but Black people are 13% of the population and something like 5% of the population in clinical trials. I get that-

Marcella Alsan: Yes.

Neale Mahoney: Roughly correct then?

Marcella Alsan: That's roughly correct, and that's actually very similar to Black representation on the physician side, too. You could have said the same statistic.

Neale Mahoney: And if we, like, weighted by disease prevalence, right? That gap would be even larger.

Marcella Alsan: Even larger, yes, for many diseases. So, the title of the article is "Representation and Extrapolation," and I think that framing is very useful because it asks a fundamental question, which is, does who is in the data affect how I interpret it?

Neale Mahoney: Yeah, and so what's the potential mechanism? Why would low representation matter?

Marcella Alsan: We weren't sure it did. You can just ask the question. You don't know if it does or not. The reason why it might is because, humans being humans, or people being people, the way we tend to react to information is we tend to extract more, update more if you will, or extract more from a signal if we have something in common with that data set or with the people-

Neale Mahoney: And people in marketing have known this for generations, right?

Marcella Alsan: Right.

Neale Mahoney: People like you like Coke. People like you-

Marcella Alsan: Yes.

Neale Mahoney: Like to drive an F-150.

Marcella Alsan: Right.

Neale Mahoney: That makes you more likely to want to buy a Coke or an F-150 or whatever it is.

Marcella Alsan: Yes, yes.

Neale Mahoney: But you must know, you know, from your clinical experience, that patients will ask, "But does it work for people like me?"

Marcella Alsan: That is exactly the question that often comes up. And physicians, when they are actually being, you know, trained, they're asked. It's not just about whether something is, quote, "efficacious," so does it work on paper? But is it gonna be effective in the real world? And one of the questions you ask when making that translation is, "Well, you know, is it gonna work on my patients? How would I know? Was it tested on patients like mine?" So, either patients asking that directly of physicians or physicians just having to do that extrapolation exercise every day all the time patient after patient, are kind of naturally attuned to think about those types of things.

Neale Mahoney: Right. So, you knew as a clinician that patients come in and ask, "Does it work for people like me?" You knew in your clinical training that doctors-

Marcella Alsan: Evidence-based medicine. Yes.

Neale Mahoney: Are asked to think about whether it works for their patient. And then you wanted to see how important is this in the data. So, how important is this extrapolation mechanism in the data you looked at?

Marcella Alsan: So, we wanted to look at it and we wanted to assess whether or not it was important, and to what degree if so. So, the way that we took that approach was, well, first, let's go to the clinical trials database.

Neale Mahoney: All the data. Basically all of it-

Marcella Alsan: Yeah, we had all the data,

Neale Mahoney: Which is crazy.

Marcella Alsan: All of it since about 2000, I wanna say, when this registry was started. And we thought, again, you're gonna notice a pattern here, "Let's just go to the observed data."

Neale Mahoney: Yep.

Marcella Alsan: "Let's find trials that had more versus less Black patient representation, and feed those results in an experimental fashion to a set of physicians, and see how much they update on the drug." Can you guess what problem we found?

Neale Mahoney: That there weren't enough examples-

Marcella Alsan: Yeah.

Neale Mahoney: Where there was actually equal representation.

Marcella Alsan: Yes, exactly right.

Neale Mahoney: Right, it's the same problem that you found in Oakland-

Marcella Alsan: Correct.

Neale Mahoney: That there weren't enough clinics that had, you know, sufficient Black physicians.

Marcella Alsan: Yes.

Neale Mahoney: Oh, that's amazing.

Marcella Alsan: So, can you imagine the solution?

Neale Mahoney: No, actually. You're more creative than me, so you came up with a solution.

Marcella Alsan: So, we came up with our solution, was to fabricate the data. So, we took a large set of physicians, and we sent them surveys, and we did disclose that the data were, in fact, fabricated. We used a novel sort of diabetic, anti-diabetic drug, so we had told them that these were not real drug trials that they were looking at, but we didn't tell them that the goal was anything to do with representation.

Neale Mahoney: Yep.

Marcella Alsan: But we said that we were really interested in understanding how they interpreted trial information, and we also did this sort of revealed preference approach, real stakes that economists are doing these days with survey experiments.

Neale Mahoney: You asked these doctors how their decisions would change-

Marcella Alsan: Yeah.

Neale Mahoney: If the studies had different representation. And what did you find?

Marcella Alsan: So, what we found is what I think you would hope to see, the more efficacious the drug is, the more likely the doctor was willing to prescribe the drug, but we saw a level shift down. As the data became more represented, there was actually no effect on the prescribing behavior for physicians that saw white patients.

Neale Mahoney: Which sort of makes sense-

Marcella Alsan: Which sort makes sense.

Neale Mahoney: If the study is 99 versus 85% white patients. You can extrapolate from that information probably with equal validity.

Marcella Alsan: Right. If given a trial that had very few Black patients, there was a gap in the prescribing behavior between physicians that treated Black patients and physicians that didn't. So, even though they both kind of prescribed more, if it was more efficacious, there was still this pretty pronounced gap in prescribing recommendations between physicians that treated Black versus white patients. And that gap was able to be closed by just providing more representative data to the physicians. Next question we asked was, does it matter to patients themselves? So we also recruited patients and we were able to find one hypertensive trial that occurred in two different places. One place where there was, you know, a more diverse clinical patient population to draw from, one less. And the drug, this particular anti-hypertensive was similarly effective in both instances, and so we were able to basically randomly assign Black and white patients to either data, again, holding efficacy this time constant to different samples. And we saw exactly the same thing. For the white patients, they updated positively on taking the drug, no matter what's the situation. For black patients, there was a gap that one could close with more representative data. So, physicians are being good agents for their patients. They kind of internalize that patients won't trust those data to the same degree if they can't see themselves in it. And these gaps are, in fact, they exist and they can be closed.

Neale Mahoney: Marcella's research shows how deeply trust and representation shape patient behavior. Over coffee, we keep coming back to this idea of trust, what it takes to earn it, and why it matters so much in healthcare.

Marcella Alsan: Kind of have to think about trust as being fundamental because that's deciding that, I feel fine in this moment, but I trust Neale. I'm gonna ask Neale for his advice. And he says that my blood pressure is high and taking this pill would help me, then I'll do it. That's secondary prevention. Or if he tells me that I'm of the age that I should get vaccinated for, you know, zoster, I will do it.

Neale Mahoney: If you're feeling good, why would you spend time and money, unless you really trusted the advice of-

Marcella Alsan: Absolutely

Neale Mahoney: It's a small moment, but it captures something big. Trust isn't just a clinical input, it's a precondition for action. And it's not limited to healthcare. Marcella and I got to talking about where else it shows up and what happens when it disappears.

So, healthcare is not the only institution where we're seeing a breakdown in trust. We're seeing eroding trust for many of the institutions in society. Ask you two questions, which I realize are really hard, one, to what extent can we think about what's happening in healthcare by itself? And two, what are the lessons that you and others have learned from healthcare, which, you know, we think are revealing for people who are thinking about these in different domains?

Marcella Alsan: If people feel like the system, and here, I'll say the health system is there for them, has their back. Fentanyl is not ravaging their communities. Gun violence is not preventing them from taking their children to the street, to school. That if there's a new cancer drug or a new anti-obesity medication, they'll be in the data and it'll be available to them, and that their doctor or their healthcare provider is invested in them and not have some other agenda about keeping costs down or pushing them into things they don't need. I think if people feel secure in that, that can have massive benefits for the productivity of the economy and the social capital in the economy. That could really be a solve. Where is my data? I don't have an experiment on that, but I have a strong intuition based on some of the work that's coming out of the gov department at Harvard that shows that when people have, you know, higher satisfaction in their health system, they also have higher satisfaction in their government. It's correlational. I have lived experience from going to very remote parts of the world and working with a non-profit, and I would say like, "Let's randomize where we roll out." And they said, "Absolutely not. We first have to establish that we are there for them. We need to go to places where their longest queues are because, you know, there hasn't been healthcare provision, and we need to treat them for what they have problems with. And then they'll come to us and we can spread out from there." It's sort of that sort of incremental evidence that I think is pretty powerful that it establishes how important health is to everyday life.

Neale Mahoney: When people trust the system, the benefits ripple far beyond healthcare. It strengthens communities, it builds social capital, and it boosts the economy. A big thank you to Marcella for such an important conversation, and to you for joining us. I'm Neale Mahoney, and Econ to Go is where we bring Stanford economics into your everyday life. If you liked this episode, subscribe or follow wherever you get your podcast. We've got more smart, curious conversations on the way. From the Stanford University campus to wherever you're listening from, until next time.