Howie and Harlan are joined by Sejal Hathi, a physician, podcaster, and social entrepreneur who recently left a White House role to become New Jersey’s health officer. Harlan asks if ChatGPT’s imagination is a bug or a feature; Howie discusses the toll of motorcycle accidents and makes the case for helmet laws.
Harlan Krumholz: Welcome to Health & Veritas. I’m Harlan Krumholz.
Howard Forman: And I’m Howie Forman. We’re physicians and professors at Yale University. We’re trying to get closer to the truth about health and healthcare. This week we’ll be speaking with Dr. Sajel Hathi, but first we like to check in on current topics in health and healthcare. Harlan, what’s on your mind this week?
Harlan Krumholz: Well, Howie, you’re going to get tired of this. When the pandemic was raging, every week we’d be talking about the pandemic. I’m still stuck on these large language models, the ChatGPTs of the world. I had the privilege of being at a closed meeting in London this week with a lot of people with expertise and was learning so much about this and it’s got me still obsessed with how this is going to impact medicine and what the right places are.
Our colleague Jeffrey Sonnenfeld, who was on with us, you may have seen recently talked about a survey he did at one of his meetings where he had 119 CEOs, 90% of them said they didn’t think that the opportunity of AI is overstated, they actually think it’s immense, but about 40% of them said they thought AI could destroy humanity within the next decade. I mean, that was mind-boggling in itself. It has nothing to do with the meeting I was at, but when Jeff came out with that, 40% of these CEOs—
Howard Forman: I saw that. It’s crazy.
Harlan Krumholz: …said that they were concerned it could destroy humanity within the next decade. That to me seems overstatement, but even if it’s 1%, that’s a concern. 40% are…
Another person we’ve had on the podcast, Eric Topol, was writing about a research letter that was in JAMA that took a bunch of cases. In medicine, what we do sometimes is we’ll assemble the facts of a case and present it to an expert as a way of teaching us how an expert thinks and also as some showmanship to see whether or not someone can, when presented with something that’s quite complex and convoluted and far from obvious, whether they could reason themselves through to the answer.
There was a research letter in JAMA that basically took a bunch of these. The New England Journal of Medicine publishes every week one of these from Massachusetts General Hospital. This is a long, long tradition that they’ve done and something that we all grew up on where a really, really, really hard case, something that seems insolvable is presented to a really, really, really smart person. And then we sort of see whether or not the person figures it out in the end. Most of us mere mortals are fumbling our way through this trying to figure out what this could be. And often the person who’s presented this will come up with the answer. If not the answer, often it’ll at least be on a list of things they say are possibilities. And not uncommonly, even the most expert person in the world fails to really come up with the right answer, in part because they’re not all typical features of a particular condition and so it sort of leads them in on a wrong path.
So naturally, somebody came up with the idea, let’s take these and feed them into ChatGPT-4 and see how they do. And of course, maybe not surprisingly, given that we’ve seen the ChatGPT-4 can pass the medical boards and answer a whole range of questions, even though it wasn’t built specifically on medical knowledge, that it actually did pretty well. Now that doesn’t mean it answered all the things right, but neither do average physicians, put in a position, are able to do that either.
So then Eric goes on to opine about what’s the right place for this? Where does it fit? The authors of the articles were tweeting that this is really ready for testing in real-world situations to sort of figure out what it can do and realize that ChatGPT-4 is still just the first half of the first inning. I mean, we are still in the very beginning of this, and yet these are so powerful. One of the things I saw at this meeting—and we’ve heard about before—are these so-called hallucination and a discussion of what are called hallucinations from these models. And what hallucinations are are sort of answers that are made up. So there was a case of somebody and there was a description of what was in the medical record, and then ChatGPT was asked to summarize that record, and some of the facts of the summary were actually made up.
And part of the discussion that we had was, is this a feature or a bug? Obviously in this particular case, you don’t want it making up facts about a patient, but if you appreciate the creativity of ChatGPT, when you say, write me a poem about something and then do it in the form of a Shakespearean sonnet, and it can actually generate something quite novel that’s never been seen before. In that case, it’s a feature and then when it actually takes a patient case and in a way speculates about a fact—in this case, it speculated on a body mass index that was probably close to what the patient’s real body mass index was, but it wasn’t anywhere in the record. And so we decry that as sort of a hallucination. And so there was a large-scale discussion about where are we going to be able to trust this?
Can you actually somehow tune it so that it’s not going to be creating wrong information? Where is the best place to be putting this? But I remain remarkably enthusiastic about the possibilities and believe that it can be transformative. And not ChatGPT alone but ChatGPT in the proper workflow as a co-pilot to healthcare professionals and as an assist to patients in ways that can help everyone achieve better performance. It’s just going to be a matter of finding the right place for this and being sure that we’re guarding against unintended adverse consequences. But it remains a very fascinating story to me.
Howard Forman: We’re going to learn a lot over the next year. It’s going to be a very steep learning curve for everybody. And I think the best advice that you can have about it now is to be a little bit cautious and not to overextend what we know into the future until we can start to see where it goes.
Harlan Krumholz: Howie, let’s get to the main course here with our guest, and so excited to have her here today. So why don’t you go ahead.
Howard Forman: Dr. Sejal Hathi is the former senior policy advisor for public health at the White House. She’ll soon start as New Jersey’s Health Officer and Deputy Commissioner of Health. An internal medicine physician, she’s an assistant professor and joint faculty member at both Johns Hopkins School of Medicine and the Bloomberg School of Public Health.
Among her many achievements, Dr. Hathi has founded and led two nonprofits that have collectively empowered over 30,000 young women globally, has served on numerous boards and advisory groups, including the UN Secretary General Ban Ki-moon’s expert advisory group on women’s and children’s health and produced and hosted the health equity–centered podcast Civic Rx during the height of the COVID-19 pandemic. Dr. Hathi holds a BS from Yale University, where I first had the privilege of meeting her 13 years ago, and an MD/MBA from Stanford University, where she studied as a Harry S Truman Scholar and Paul and Daisy Soros fellow.
She completed her residency at Massachusetts General Hospital and served as a clinical fellow at Harvard Medical School. So first, I want to welcome you, but in preparation for this, I did look back at the letter of recommendation that I was fortunate enough to write for you 11 years ago. And oh my God, you have fulfilled everything that I could have ever hoped for. You are an enormous success. You are a clinician. You are a leader in public health. You remain connected to the same spaces that you were passionate about back then. And so I want to first start off and ask you about this journey, about your passion for wellbeing and for mental health that started, I think, maybe even in high school, that you’ve developed this passion for it and have continued to work toward improving the wellbeing of society. So first of all, welcome to the podcast, Sejal.
Sejal Hathi: Thanks so much, Howie. And to answer your first question, Howie, I’ll say that you’re absolutely right. This journey did commence many years ago, more than half a lifetime ago, at the age of 14, 15, when I was a high school student in the Bay Area of California. And my journey flowed from my own personal experience, my own personal struggle with the healthcare issue. And so when I was 15, I was diagnosed with an eating disorder, anorexia, and I should share that I come from a very traditional Indian-American family. My parents were both born and raised in East Africa, came to this country as refugees, knowing very little English, having very little sense of the healthcare system, let alone what eating disorders such as anorexia was. And so for the longest period of time for them as well as for me, myself, this was a very difficult diagnosis to accept and grapple with.
It was really my years-long journey toward recovery, that one, inspired me to want to pay it forward in the same way that my physicians held my hand and pulled me and my family through those hard years. I wanted to be able to do the same for young people like myself. It also taught me the indispensable, really foundational role of health to achieving self-actualization. And I knew that one could not possibly aspire to the types of impact that I wanted to achieve without first being healthy herself. And I never wanted that absence of health to be a barrier for others like me.
And then thirdly, and I think related to public health and my interest in population health, I recognize that our healthcare disorders or healthcare conditions like anorexia like others are really patterned by the social, cultural, economic, political choices and forces that we’re surrounded by. I was socialized to believe I was not pretty enough, good enough, perfect enough, as are thousands of young people, and we’re struggling with the mental health crisis among young people today like me. And I wanted to be able to step outside of the four walls of a hospital and really decry, identify, address the structural barriers to achieving health and achieving self-actualization. And that’s why I decided to pursue this dual career in healthcare but also public health and public policy.
Harlan Krumholz: It’s such a pleasure to see you. I wanted to just ask you a question about your process because somebody could be looking at what you’ve done and just wonder, how in the heck does she do all those things? And it’s such a broad range of areas and areas where you’ve gotten traction. So let me just, for example, what I was just wondering is, of course you’ve got these jobs. I mean this major job at the White House and this job that you’re starting or already started in New Jersey, but how do you decide what problems that you want to solve?
I mean, there’s so many needs in the world. And then one of the things that’s so impressive about your accomplishments is the way you bring people together. Howie said, “empowering 30,000 people.” It’s one thing to say, this thing has gotten traction and really engaged a lot of people. It’s hard to engage people, especially around something new. So I’m just wondering, can you tell a little bit about your process where you decide to direct your energy, and then how do you go about in enlisting people, motivating them, and getting people on board, and then ultimately the people you’re trying to help to get them engaged in ways so you can fulfill the promise that you’re trying to achieve?
Sejal Hathi: Yeah, those are excellent questions and I will caveat or, rather, preface the response I offer with the very real concession that I’m still figuring out the answers. And a lot of the progress, again, that I’ve made has been serendipitous. And some days I don’t get it right. I’m not sure that I have mastered that balance that you speak so generously of, but I’ll try and answer your question by saying that in terms of selecting what problems to solve, I have tried to look to the Japanese concept of ikigai, and I don’t know if you all are familiar with that, but it’s like a four-part Venn diagram and it asks you to consider, One, what does the world need? Two, what are you good at? Three, what will you be paid for? And four, what do you love? What are you passionate about? And the concept calls upon you to identify and commit to something that can fulfill as many of those questions as possible.
And that’s what I have tried to look to in deciding what roles I assume in terms of the problems that I addressed. I think to be honest, when you serve in the president’s administration, you address the problems that the president wants you to address. You were there to… I served on the Domestic Policy Council to execute the president’s agenda, and I tried to do that faithfully at the White House, I will try to do that faithfully in New Jersey as a member of the administration of Governor Phil Murphy. And beyond that, it’s whatever makes me angry. When I was a 15-year-old, 16-year-old high school student, I wanted to create a sisterhood of change makers, in my teenager’s parlance, so that no other girl would have to suffer the way I did. So that other young women would have a community of young women they could look to for support.
The first organization I started was called Girls Helping Girls. And the absence of that for me, the struggle that my own family faced in both accessing quality, frankly eating disorder care, and then overcoming the emotional, physical, psychological trauma that eating disorder recovery necessarily entails, was infuriating for me. And I didn’t want anyone else to go through that. When I decided to continue to do the work I was doing for global women’s rights, I recognized that the first and most frequently violated basic human right for young women, for women generally from birth to motherhood to beyond, is quality healthcare, and that’s extremely upsetting.
And so in my first organization, we used to offer a social change curriculum to young women, and the question I asked them was, what makes your blood boil? What kindles this insatiable hunger to address the problem, make it better so that no one else has to go through it? Any problem that you are so passionate about, that you can work the whole day and it just keeps you up at night? That’s the problem that you have to solve because frankly, these problems are all hard. They’re intractable, even the ones I’m working on now, they’re not going to be solved overnight. So you have to be committed to them. And so if you found something that just really keeps you entranced in that way, then you know that’s what you should focus on.
Howard Forman: Can you give our listeners a little understanding of what you’ve learned by working inside the White House, inside the executive branch of our government? And also let our listeners know, I don’t think we talked about it in the intro, but you’ve had a chance to work in state health, I think, in North Carolina under our incoming CDC director, Mandy Cohen. So you have experience at the state level; you’ve had this experience at the federal level. What are your thoughts about how the levers of government can be used to advance the objectives of better health for our population?
Sejal Hathi: That’s a great question, Howie. And I’ll say that working in the federal government, what have I learned? I’ve learned that there are really great committed, dedicated public servants in this administration that are trying to do the right thing, and sometimes certain problems, as alluded to earlier, take longer to address than others. And to address those problems often also requires building consensus among a wide variety of people and special interests and different offices and divisions and agencies of government, so things can take a little bit longer. But at the end of the day, people are there in the federal government because they believe in making life fairer and better for all Americans, and their heart is in the right place. And I think it can be rare to belong to and work among kind of a community of individuals that are singularly united in mission space, as I enjoyed working in the Biden-Harris administration.
And so that element of public service, despite its other perhaps bureaucratic woes, was incredibly special. And to the extent that your students are listening to this, I would urge them, if possible, to take advantage of such an opportunity through an internship or fellowship to join the federal government if only to taste of that sense of camaraderie and shared mission. In terms of your second question, which was how can I think the policymaking apparatus be leveraged to advance the public health agenda?
I would say that I think health and public health is inherently political. And democratizing health and improving public health requires remaking decades of public policy choices that have perpetuated racial inequity and disproportionately punished the poor. And so it is precisely by leveraging public policy, working in government, whether at the local, the state, or the federal level, that we can address these larger social forces that bring our patients back into the clinic, into the hospital over and over again.
What we see in the hospital are often the physiological manifestations of social woes. I’m just kind of preaching what one of the fathers of modern medicine, Rudolph Virchow, first kind of called I think social medicine in the 19th century. But I think one of the longest-lasting interventions that we can look to to address these medical challenges we see in the healthcare setting is through public policy. And one of the silver linings, I think, of the COVID-19 pandemic is that it occasioned kind of an awakening in physicians who might otherwise have embraced the comfort afforded by the perhaps apolitical symbol of the white coat. It induced… it awakened them to realize this fact and to speak up and speak out and advocate for the political and policy changes that would improve their patient’s lives.
Howard Forman: So you finished your residency, but you did your residency very much during the pandemic. I mean, the peak of the pandemic coincided with much of your residency, and that’s not an easy time, even if there’s no pandemic. And yet you started doing a podcast during that time as well. And Harlan and I have now come to realize that a podcast is not merely sitting down for 45 minutes a week. Why did you make that a priority of your work at that time? And what did you learn from that experience?
Sejal Hathi: Hosting a podcast, producing a podcast—we were a team of two—is hard work. And kudos to both of you for maintaining this as long as you have. I realized that my co-residents and I were on the front lines of a once-in-a-century pandemic, and yet our voices were not really being heard or integrated in the types of conversations that were happening at both the hospital level and the national level in terms of how we should address this pandemic socially, medically, politically.
And so I started the podcast actually in the form of a series of fireside chats wherein I brought some of these experts—I cold-emailed them—that I was seeing on nightly CNN from Dr. Fauci to our mutual friend, Dr. Murthy, to Dr. Lena Wen. And I said, “You’re having these conversations on live TV. You’re having these conversations with one another. Why don’t you come and have them with some of our frontline providers?” And they gladly agreed. And I hosted a number of these, including with my now former boss Ambassador Susan Rice, between these luminaries of public health and of politics and residents, trainees, even faculty across Mass General Brigham. And I realized these were such rich conversations, why keep them contained to just the community of Mass General Brigham? Why not spread them more widely? And it was actually one of my first podcast guests, I won’t say who, recommended, in fact encouraged me to translate this into a podcast.
So with that individual’s help, I decided, you know what? Let me just give this a go, and it’s not going to be perfect. It’s not going to necessarily going to be professional, but let me try and share these conversations more broadly. And frankly, they served as a forcing function—I don’t know how you feel—but to stay abreast of the developing issues and also cultivate kind of a perspective of my own on potential solutions and paths forward, which inevitably helped me in my work subsequently with the Biden-Harris transition and the administration.
So I’m very grateful for the experience and for the distraction too, that afforded me from often arduous COVID-19 clinical care. But it started really in this desire to bring some of these elusive primetime TV conversations down to the ground level with me and my trainee colleagues.
Harlan Krumholz: Well, let me say that I’m grateful for our podcast because we get a chance to have someone like you on. And by the way, the reason our podcast is any good is because of Howie and Miranda and Inez and Sophia, and actually they put in so much in into this. Let me ask you kind of a parting question. If you were to give your younger self advice, what might that be, based on where you are now?
Sejal Hathi: I’ll just say two things. One, don’t listen to the haters, and don’t be afraid to be different. And at every stage of my journey from high school, when my parents refused to engage initially in this work or to back me when I wanted to start my nonprofit. The first adult co-signer on my NGO bank account was my high school principal. To medical school when I was told by a senior administration official at Stanford that I was throwing my potential away and I shouldn’t be doing what I was doing with the UN or the WHO or maintaining my entrepreneurial proclivities because no residency would want me. To residency, when I was doing these podcasts, and I used to get hate messages once I decided to translate my fireside chat into a podcast from other faculty as well as co-residents who said, “Who do you think you are? You’re just a resident starting this podcast. Why do you think anyone would want to listen to you?” Or “Clearly you got too big for your britches.” I would say there have been people who have doubted and denigrated at every step of the way, and there will always be those people.
And I think it’s important to just, as trite as this is, listen to your gut. Find mentors and people who believe in you and are willing to put their faith in you and hold onto them and hold on to that inner voice of your own that can give you the courage and the conviction to commit to things that may seem unorthodox, maybe even ill-advised at first glance but can propel you in the direction longer term that you want to go. Because I wouldn’t be here today if I hadn’t listened to that voice and I hadn’t found a small village of people who told me that was okay.
Howard Forman: Well, it is such a great privilege to have you on our podcast and to have you back so amazed with what you’ve already accomplished and what you will accomplish. And the people of New Jersey are very fortunate to have you fighting for them.
Harlan Krumholz: And we’re big fans. We believe in you. Of course, it’s easy to say now because there’s a lot to, but there was always a lot to believe in.
Howard Forman: That’s for sure.
Sejal Hathi: Thank you, both of you.
Harlan Krumholz: Wow, that was a terrific interview and that last part. Wow, that’s all I can say.
Howard Forman: She’s wonderful.
Harlan Krumholz: Amazing. Now let’s get to another favorite part of the podcast for me, which is to hear what’s on your mind this week.
Howard Forman: So as we’ve entered summer or about to enter summer, we’re seeing the usual dramatic increase in motorcycle trauma. I see this every year as a radiologist in the emergency department. I see the imaging manifestations of rather disturbing trauma. Nothing is more humbling than to see an otherwise young and healthy person physically mangled and often, worse, by a single motor vehicular accident.
Since the introduction of airbags, seat belts, anti-locking brakes, and other accommodations to cars, we have actually seen car accidents contribute significantly less to our devastated trauma patient population. Motor vehicle accidents, deaths per capita are down almost 50% since 1969, and that figure looks even better if you look at it in terms of miles driven or cars in the road. But at absolute numbers, it’s down 50%, mostly due to the innovations I mentioned as well as better medical and surgical care. From 2007 until 2021 alone, motor vehicle deaths have been flat, but motorcycle deaths have increased by 25% on a per mile driven basis, now over 5,000 per year.
I can’t readily explain why this is increased, but I can tell our listeners that there is compelling evidence that helmets reduce the risk of death. The absolute level of protection varies by study, but 37 to 56% of deaths from motorcycle accidents could be prevented through wearing of adequate helmets, and deaths are not the only thing avoided. Helmet use is associated with decreased admissions to ICUs, decreased long-term term morbidity, and decreased costs. Eighteen states have broad helmet mandates, and they are a motley crew of states. So New York, California, West Virginia, Alabama, Louisiana.
Among them three states—Illinois, New Hampshire, and Iowa—have no mandates. And the rest of the states, including our home state of Connecticut, have mandates for select groups generally under the age of 25 or under 18. So my libertarian sensibility would say that people should be allowed to wear a helmet or not and take risks or not, but that sensibility ceases to hold when your lack of a helmet becomes my problem. And the reality is whether people realize it or not, much of healthcare spending comes from taxpayers. And when your excess risk-taking impacts me, I’m allowed to at least be concerned—or maybe even more—about just your own wellbeing. Most of the time, the ability to change such behavior is limited because it intrudes on privacy or it’s very costly. But in this case, a simple solution that has been tried and works in so many states seems a reasonable one to ask for. And I hope more states adopt helmet laws, including our own.
Harlan Krumholz: What do you see about the parallels between this and the vaccine mandates? I mean, there are some parallels here, and of course a lot of pushback on the mandates, but I don’t know, do you see this as a similar issue?
Howard Forman: It’s really not similar now, but one could argue that at the very early stage of our vaccine rollout, it was absolutely an issue. When we first rolled out vaccines and the vaccines at that time during that variant was actually reducing transmission to some degree, you could argue that it was had real externalities, positive externalities that were keeping our hospitals from being overwhelmed and also reducing spread to people who might not have been able to get a vaccine yet or might not be allowed to get a vaccine or have a vaccine.
We’re in a very different place right now. It’s hard to say that the vaccines have significant positive externalities now. They mostly protect the individual to a variable degree. And so it’d be very hard to argue that a Covid vaccine mandate is necessary. But if you look at childhood vaccinations, the parallels are very strong. There are very large negative externalities as associated with being unvaccinated, positive externalities with being vaccinated, and we should do everything we can to continue to have mandates, particularly on children, in order to prevent major outbreaks for vaccine preventive disease.
Harlan Krumholz: Let me just, let me just jump in one second because I think, but while very useful information, you’re taking that as kind of like today, what should we do about vaccines? But I’m just wondering what the threshold is for society to mandate that somebody does something. What’s that level of benefit? So take for example seat belts. Seat belts save lives, but at the margin, the truth is the absolute benefit for anyone who seat-belts themself is really, really small. Now, by the way, if we banned cigarettes, we would save millions and millions and millions of lives, but no one’s prepared to tell people that if we banned alcohol, we would save millions and millions of lives, but we’re not prepared to do that. Now you’re saying we should formally tell people, “No, but you should wear helmets.” Now, where does the threshold tip?
Howard Forman: Yeah, no—
Harlan Krumholz: … when the benefit is enough for us to do that.
Howard Forman: There’s no perfect answer to that. And by the way, there’s so many examples of that, speed limits on highways, another example of that. We impose limits on certain populations; we don’t on others. I would say seat belts is one of the greatest public health successes, but one could argue that most of the benefits flow to the individual. And the reason why we started to impose those laws was in some ways to get manufacturers to put better seat belts in. But it’s harder to make those arguments about why we do certain things when most of the gains are internalized, not externalized.
Harlan Krumholz: Yeah. I think that our society’s going to continue to struggle with this. And for example, the vaccine’s large net positive, but lots of people saying, I should be able to make my own choices. And then there is this question like you raised, how does my behavior affect other people? And where do you, but I don’t think we’re being very consistent right now in society at a certain threshold of net benefit. Anyway, lots more on this, but thanks so much for sharing it’s really such an important topic and very relevant here in Connecticut, given the helmet laws or lack thereof. You’ve been listening to Health & Veritas with Harlan Krumholz and Howie Forman.
Howard Forman: So how did we do to give us your feedback or keep the conversation going? You can find us on Twitter.
Harlan Krumholz: I’m @hmkyale, that’s HMKYALE.
Howard Forman: And I’m @TheHowie, that’s @T-H-E-H-O-W-I-E. You can also email us at email@example.com. Aside from Twitter on our podcast, I’m fortunate to be the faculty director of the healthcare track and founder of the MBA for Executives program at the Yale School of Management. Feel free to reach out via email from more information on our innovative programs, or you can check out our website at som.yale.edu/emba.
Harlan Krumholz: Health & Veritas is produced for the Yale School of Management and the Yale School of Public Health. And what makes this podcast great, Howie?
Howard Forman: I do.
Harlan Krumholz: It’s Miranda Shafer, our producer. Inez Gilles and Sophia Stumpf, our student research assistants. They’re the ones who make this—
Howard Forman: 100%.
Harlan Krumholz: Yeah, for sure. Thanks so much, Howie. Talk to you soon.
Howard Forman: Thanks, Harlan. Talk to you soon.