In this episode, Raj and Faith chat with renowned biostatistician and public health researcher Bhramar Mukherjee following her Hansen Distinguished Lecture at the University of Iowa College of Public Health on April 29, 2026. Dr. Mukherjee shares her journey from studying mathematics in India to becoming a leading voice in biostatistics, epidemiology, and public communication during the COVID-19 pandemic. Together, they explore the “four quadrants” that shaped her lecture and career: ethics, community engagement, communication, and capacity building.

The views and opinions expressed in this podcast are solely those of the student hosts, guests, and contributors, and do not necessarily reflect the views or opinions of the University of Iowa or the College of Public Health.

Lauren Lavin:

Hey everybody, Lauren here. Before we jump into this week’s episode, Raj and Faith did a great job leading today’s conversation. So I’ll try to keep this brief, but I wanted to pop in because as I was editing this episode, I honestly found myself sitting with some of these ideas long after the conversation ended and I wanted to bring those up before you started listening to the episode. Dr. Mukherjee brings such an interesting perspective to public health, statistics, communication and leadership, but what really stayed with me were the final few minutes of this conversation. So you need to listen all the way through to get to these nuggets. But there’s a discussion towards the end about friction, intention, uncertainty, and the role that all of these things play in our growth and creativity as students and then professionals out in the field.

And honestly, I think that’s something that so many of us are wrestling right now. I know that I am and that many people in my generation are, is just this moving away from anything that causes friction, or tension, or disease. And so we live in a world that increasingly prioritizes efficiency and optimization and really like the smoothing out of every edge out of our process. But this conversation challenged that idea and I noticed it right away. Dr. Mukherjee talks about how friction is not necessarily something that we should avoid, but really something that refines us and that these mistakes and uncertainty and tension that we experience is where growth happens and I found that incredibly poignant.

So there are a lot of many beautiful, thoughtful moments throughout this episode about courage, and public health, and trust, and communication, and what it means to do meaningful work in public health. But I especially encourage you to listen really closely towards the end because I think there are some really great words of wisdom in there. So with that, I’ll turn it over to Raj, Faith, and Dr. Mukherjee. Let’s get Plugged in to Public Health.

Raj Daliboyina:

Hi everyone and welcome back to Plugged in to Public Health. Today we are bringing you a special episode from the [inaudible 00:01:54] distinguished lecturer for today, Dr. Mukherjee. Her talk today explored what it really means to be a public health statistician and not just from a technical standpoint but through four key dimensions, specifically ethics, community engagement, communication, and capacity building. I am Raj and I have Faith with me today and we welcome Dr. Mukherjee to our podcast today. It’s your first time with us. We are a student-run podcast from the University of Iowa College of Public Health and where we try to explore major public health issues and how they connect with real-world impact.

So to begin with, Dr. Mukherjee, thank you so much for joining us today and especially for such an insightful lecture that you gave and it’s great to have you here and we would love for you to introduce yourself for our listeners.

Bhramar Mukherjee:

Hi everyone and hi, Raj and Faith. Thank you for doing this podcast with me. I’m Bhramar Mukherjee, I am a professor of biostatistics, chronic disease epidemiology and statistics and data science at Yale University. And this is my first time to Iowa City. So I’m really enjoying my exploration of the beautiful campus of the University of Iowa.

Raj Daliboyina:

That is great. This is a good time to be here too with the kind of weather we’ve been having. So we are happy that you join us on one of our nicer days here in Iowa. To begin with, Dr. Mukherjee, what led you into biostatistics or public health?

Bhramar Mukherjee:

Yes. Thank you for that question. I grew up in India in Kolkata and I knew that I liked mathematics, but I wanted to do mathematics with the context and the purpose. And so it seemed like statistics will give me that opportunity. So I have been a statistics major for 35 years. That tells you how old I am, but I’m incredibly proud of my experience. So I was a statistics student from my undergraduate days doing my master’s and doing my PhD. The journey to biostatistics was somewhat of a serendipity. I was a professor of statistics in the College of Liberal Arts and Sciences at the University of Florida, but most of my interests were gravitating towards statistical modeling epidemiology.

And then I thought that maybe it is best to be closer to the data. And in 2006 from a statistics department, I transitioned into biostatistics and it was because of a combination of personal and professional choices. My family was in Michigan so I needed to move and biostatistics was the only job open. But when I ended up in Michigan in the School of Public Health, I really discovered myself. I realized that my strength is taking a public health problem and translating it into the grammar and language of statistics. So public health gave me the story and the anchor of the mathematical underpinnings of statistics.

Raj Daliboyina:

That is quite interesting how your whole journey just led you towards biostatistics from a math level, I must say. So building on that, your lecture introduced us to the four quadrants of being a public health statistician. What pushed you to think about the field in this broader, more holistic way?

Bhramar Mukherjee:

I think that it is my years of experience and when I was invited to give this very prestigious lecture, I wanted to come up with something which will be of interest to every student or colleagues and professionals in public health. And biostatistics, as you know, how awesome it is, but it is very specialized. So I wanted to step out of that biostatistical technical comfort zone, but still take the message that biostatistics transmits and show some snippets of my career how biostatisticians can go broader and can still collect data, train generations, can be in the public facing frontiers. So that was my goal to talk about biostatistics in a slightly unusual way.

Raj Daliboyina:

I must say it was a very interesting take on biostatistic because I know it’s stereotypical to just think biostats in the moment you think it’s coding, it’s about numbers. But it’s also about the people behind the numbers and what the numbers are affecting and what kind of change they bring in life of real people like the ASHA workers that you were talking about or the 300 plus people you have trained over these years. I think the whole college and our podcasters will also appreciate this holistic approach that you have brought to this concept of biostatistics, I must say.

Bhramar Mukherjee:

Yeah, thank you. And when I first moved to biostatistics, I was taking it in a very mathematical spirit. So for example, in my first collaborative meeting, people were talking about the data and I was saying, “I don’t want to know about the data. This is my X metrics and I am going to do X prime X and I am going to look at the IgAN values, but this is meaningless.” So one of my very kind mentors, but also very direct mentors took me aside and said, “Bhramar, we know that you are quite smart, but you must recognize that the rows of this data, the rows of your X matrix are my patients and the columns are biomarkers that I have spent my entire career in a lab discovering. So you better pay attention to what the rows and the columns of those X metrics are.”

That one singular conversation really shaped my career. And 10 years later, my mentor actually acknowledged that I have taken that conversation to my heart and really knew about the rows and the columns and myself were generating those rows and columns now.

Raj Daliboyina:

That just shows your passion for the science that you picked as a choice, right? So you touched on ethics in your talk today. Where do you think the most challenging ethical decisions actually show up in your work and beyond the formal review process?

Bhramar Mukherjee:

So that’s a great question and difficult question. So many times biostatisticians think that rigor and reproducibility is a part of a ethical pyramid that whether I’m doing this work in a rigorous way where I’m archiving all my codes so that my results are reproducible. But also when you are taking a broader look, you have to think about whether the data is actually collected in a culturally appropriated manner. Do we have social responsibility in terms of how the data are being used? Many times you feel that the PI who have collected the data, it’s their responsibility, but I see that as a responsibility of the profession as well to raise our voices when we feel that the data is not collected properly. That’s part of ethics. The data is intentionally or unintentionally leaving out populations or particular questions that we need to answer the scientific underlying truth.

I do think that biostatisticians and epidemiologists are uniquely trained to think about imperfections in data and to point that out when a study is being designed is the whole part of ethics. And then how are the outputs being used? For example, if you actually have a risk prediction model, which you have fitted to a large population, but you really do not know whether it works for particular marginalized subgroups. It’s your responsibility to really show that evaluation in the subgroups to show that the prediction exercise has been done in a fair way. A lot of people do not stratify by their data by saying that, “Oh, I do not have enough data from all possible ethnic groups or racial groups.” But that is your responsibility to be honest in terms of the models and the sensitive variables you are using in the model that may affect downstream policy consequences.

Raj Daliboyina:

Yes. And I’m sure the kind of work that you did in India with a thousand different kinds of ethnicities out there trying to figure out. And as my own personal experience during COVID, data was definitely a sensitive issue of how it was getting collected and the surveillance that was being happening during COVID and the apprehensions that the people or how much say that people actually had in the data they were sharing was causing fear in itself. Faith, what do you think? How was it in Nigeria or what do you think about what Dr. Mukherjee was talking about?

Faith Ibitoye:

In Nigeria specifically, I would say that the government itself did not take into… They didn’t take the situation as serious as many countries like India, you talked about it and so many other things like that. But specifically in Nigeria, the lockdown was serious that the state itself, especially how the states in Nigeria or some of the states in Nigeria didn’t take the information as serious. But I would say that we tend to follow the trend of what other countries were doing because there was not this big understanding about what the public health or what the physicians should do in terms of how to protect the citizens. The one thing that was standard or one thing that was put in place in Nigeria was that everybody should stay at home. There was no going out. There was not all those other things. My mom was a nurse, but she still had to go out to treat her patients sometimes. But the pandemic also affected her work in that place where she has to stay at home most of the time for that.

But in terms of other aspects, the Nigerian government itself in the aspect of taking the information about lockdown, using your mask, about doing all those things, they took that serious for a certain period of time. But to also certain period of time, many of it were not accounted for. I think I remember every day you hear the news and you’d be like, some number of people died and there was no that accountability of how these things are going on. So there was always a gap in data and I think that’s something Nigerian government itself suffers in terms of public health. The data is scarce. We don’t get to hear the communication about these things that are even happening among the people itself. So it’s very scary for you to even believe the news you see every day.

So even though we see the data, even though we see the number of things that are going on, sometimes you still question yourself, is this true or is this actually going on or what are they actually doing in the state itself? So it was a lot of uncertainty for many of us as Nigerians. I know a lot of people will have better information on that, but from what I saw every day, there was just that uncertainty about, I don’t think this number is true. And people are just dying or the government is not taking this thing seriously.

Bhramar Mukherjee:

Data is very political, right? Data is the modern weapon and governments should have a very transparent system of data collection and data dissemination, which is independent of the political arm of the government, but that’s not really true, right? So this is not like a justice arm, but we need more data justice and independent collection of data. For India, I think when we are working on the pandemic, the paucity of data that you can have models on various different compartments, but you did not have data that can afford the models and the opacity of the data, for example, undercounting of deaths became a huge issue. And if you don’t have the correct data, it’s like flying blind, right? You cannot really project and people do not believe you. If there is a death undercounting, people may think it is not serious. And so accurate reporting of data is really a critical part of policy and interventions and any decision that we take in modern world.

Raj Daliboyina:

It could also be to essentially [inaudible 00:14:19] the state of democracy at this point with data being the most valuable asset that’s being traded in the modern world. So that’s a profoundly impactful thing that you’ve said, Dr. Mukherjee. Coming from that, so can you share an example where statistically correct answer didn’t fully align with what felt responsible from a public health perspective?

Bhramar Mukherjee:

So this is a great question. I will tell you a couple of examples that I actually am these days teaching in my ethics and equity class. One is an example of a prediction algorithm which was used to decide which patients are going to get referrals for certain diseases. And it was a machine learning algorithm which predicted a score in an electronic health records and then based on that score, you crossed a certain threshold and then you were being referred to physicians. And then somebody tried to validate that prediction model against actual chronic conditions these patients have. And it turned out that for every level of the risk score, Black patients had higher chronic burden than white patients, but they have the same risk score. So the underlying health condition for Blacks with the same risk score was actually much worse than the white patients. So how did this happen? Was the algorithm biased?

And this actually determined referral patterns for more than 200 million people. And when they looked at the algorithm, what they discovered, this paper by Ziad Obermeyer in Science discovered is that the algorithm was predicting cost of care and cost was taken as a proxy for underlying health. But we know from past data and it has been shown that for the same kind of conditions, the cost of healthcare is less for Black patients because they don’t get the proper interventions and referrals. So optimizing cost flipped the assumption that the chronic health condition. So this was not anybody’s fault but lack of contextual knowledge. So it’s not a problem with the model. You cannot blame the statistician, but it’s really that contextual knowledge and learning about the question.

So I think that what I practice as a statistician before going into the problem with a preconceived notion of a model, spending a lot of time with my collaborators wrestling with the idea as a team before we jump into models. Because I do not think that there is an omnibus solution or there is the model, you really have to look at the question wrapped in reflexivity, wrapped in domain science. So that’s why I think collaboration becomes very important so that even if you’re making a statistical conclusion, which is right, but then the context tells you that you should not have done so and so. And that has happened many times in my life and my career.

Faith Ibitoye:

Talking about collaboration, I truly value the… Shifting into the part of community engagement, I truly value the way you bring community into your work and I think as an upcoming public health researcher, I rarely see that in other departments than that. But I want to ask from your perspective, what does meaningful community engagement look like in statistical research and where do researchers tend to fall short?

Bhramar Mukherjee:

Yeah. I think that collaboration, if you go back to the root of the word collaboration, it’s co-labor. So you have to find somebody with whom you are going to like to work together. That’s the first thing. And I think that I have worked on projects just because I had a good person to co-labor with. Gotten into questions that were beyond my immediate interest, but I really liked my co-laborer. So I think people, again, are very important and then people finding questions that connect themselves. I see academics statistics and you probably saw it in my talk that I see each part of my existence is really relating to people. Public, public is a big part of the public health world. It goes back to the root word of the word public, publicus, of the people. And so it has always been about finding the right people and then we find questions that interest us. So that’s how I think about collaboration, not in a strategic way, but much more of an organic way, which bottom up, not top down.

Faith Ibitoye:

That’s powerful because that’s something I also learned in my class. There’s this class I take called community engaged research and it talks about collaborating with community members, those that have been affected with what you’re actually researching about. But I want to hear your perspective. There was one thing we spoke about like working with difficult people. This might be what you want to research and you find the people you actually want to work with. You have the same research interests, you have the same, everything aligns. But when it comes to that area, because for community engaged research, it takes a while, there’s conflicts, you have to resolve those things. So how does it look like in your homework when you have the data and you’re trying to communicate it and how do you balance that?

Bhramar Mukherjee:

Yeah, I think it’s really important for a statistician to understand the population from which the data are coming from. One thing as we were working on building a massive biobanking cohort at Michigan, University of Michigan Health System, we did was to form a community engagement studio. And we even asked those, our study participants that, why did you consent? Why did you consent to giving your electronic health records with genetics, with Social Security numbers so that we can link with other variables? Why? Because there is a lot of things that you hear about privacy and confidentiality concerns in the public about giving their data. And I thought that they’re going to talk about science, they’re public good, but most of the people said that we trust University of Michigan researchers. That trust. And why do we trust? Because someone in my family has gone to the health system or to school. It has changed the outcome, health or educational outcome of my family, my community.

So I think that relationship of researchers and the participants is totally cemented by trust and the trust does not end with just collecting data. It’s constant iteration of going back to the community with your results and then refining and redefining your question based on the needs of the community. So I think this is an iterative process and that is why I really love public health because it gives you a framework to do that.

Faith Ibitoye:

That’s powerful. Going back to the community, showing them the results, the integrative process, bringing the aspect to the communication aspect. During COVID-19, statisticians were suddenly communicating directly with the public. So what did these moments reveal about how we handle uncertainty and trust?

Bhramar Mukherjee:

Yeah. We like mathematics and statistics because we do not like people, right? We are introverts and that’s a trait of our profession. But what you can do is even if you’re not a natural communicator, you can get a lot of training. There are lots of resources available and one thing is that you are speaking your science that gives you strength. One reason I did the communication, the scientific communication work in India and it took a toll because with the time difference, I had to get up in all kinds of hours to be on Indian television, right? Because I was living in Ann Arbor, Michigan. But two of the reasons I did it was that there is no statistician in the media. Modeling is our lane. Economists were talking about it, physicists were talking about it, medical doctors were talking about it. Statisticians who were largely absent in the public eye.

The second thing was there are very few women in the public because as soon as you start talking about undercounting of deaths or public health interventions, you know that you are going to be trolled and grilled in social media because it is so political. So you have to be somewhat robust and desensitized to those brutal attacks if you want to be in the public communicating. With 1.4 billion people live in India, even a fraction of them reacting to your comment is going to flood your mailbox. So I do think that was another reason that I felt that I am speaking my science and if I make mistakes, they’re genuine mistakes and I can own up to them and it is not driven by any agenda or motivation. So that give me the strength along with a little bit of data feminism.

Faith Ibitoye:

That’s powerful and I think that’s also inspiring for upcoming statisticians and also female leaders. That’s important for we looking up to you, that inspire us to see you how they’re saying the truths. Coming to the trade offs of communication, how do you balance accuracy with clarity when data are complex to your audience and how do you make that as simple as possible?

Bhramar Mukherjee:

That is a wonderful question because as data are becoming more complex, for example, I often work with electronic health records where data used to look like square matrices. Now data is in the form of text or clinical notes. It’s in the form of lab results, the values that you get when you do a blood test. It is in the form of images, a CT scan, or an x-ray, or an MRI and now all of these are your data. So how do you actually take all of this structured data and unstructured data and blend it into a model and learn about human health in terms of prediction, in terms of prevention, in terms of treatment response? So this is a massive maze. And often with big data, your original intuitions of cross-examining the data through a scatterplot or through some univariate analysis, those do not work anymore.

So you have to adapt to modern visualization tools in terms of looking at the data, but also accept that your traditional checking things, model checking tools may not work. So what grounds it is really replication, that what you found actually holds in another population, in another study, do lots and lots of sensitivity analysis that how flaky is the result? If I change and tweak a little bit of something, is the significance going away or my finding is appearing to be lost? So we can apply different tools of sensitivity analysis and also replication in diverse population to get to the ultimate scientific truth.

Raj Daliboyina:

I find it so funny that I had had this conversation with my father during COVID. We were looking at those graphs that you have created on the YouTube or on the website and I had to explain to him what the peaks were, what the dips were, what was happening. And he comes from a pretty mediocre education. He’s an educated man, but it was so hard to communicate with him going over his biases against the data that he was seeing or the ideas of policy that he believed in. And I think that was a very big pattern in India specifically during COVID. And I must say you were very courageous to tackle what you did at that point of time.

Bhramar Mukherjee:

As I said in my talk that the four Cs that are important to me is computation, communication, collaboration, and then courage. Courage actually makes you stand apart because you take risks that you truly deserve.

Raj Daliboyina:

To call back to that, has anyone apologized for their… Over the years, after all of this time, I think it’s been six years. I want to ask if there was a hint of apology or humility from the people who have wrongfully told you or questioned your data over the years.

Bhramar Mukherjee:

So I think that there is science is science and I am okay with people. I’m not doing science for receiving apology. And so what was really full circle moment though when the original actually official death registration data from India from 2021 and that was released in 2025. And we immediately wrote a paper showing that the actual data is very much in alignment with what we had predicted. And the reported number of COVID deaths is much, much lower than the actual number of deaths that were reported and the excess deaths that were reported. So you can say that this did not happen because of COVID, it was delay in care, it could be institutional collapse, but we know that such a large massive change in mortality does not happen without COVID. So COVID will be a large fraction of those excess deaths. So that gave me some sanity check and also I can never forget the faces of the people or the families of the people whose death were not counted.

Raj Daliboyina:

And it’s almost ironic that I love SRK, but if they had to make Jawan and had to talk about the deaths more than they believed us, scientific community, it was very sad watching that movie because people were like, “Oh, it really happened?” And we were like, “We were telling you for five years that it happened and all it took was a Bollywood movie for you to consider us correct.” But I guess that doesn’t change anything, that doesn’t change what we need to do or what we’re supposed to do.

Bhramar Mukherjee:

But that also reinforces that science has to be more accessible to the mass and to the public. So this underscores an important point that we should think about other modalities of communication so that we can get our message across and the public… Trust has to be earned. Trust is not given to you. And I would say that I did write a paper comparing our projections with what was released just as a closure. It was almost my therapy moment that, okay, this was not a mistake and you went to, yes, you showed your courage, you stood ground on your science and your models, but there is some method in this madness.

Raj Daliboyina:

Totally. I must say we all applaud you for doing it, because at any given point of crisis, it’s not… These days there’s a thing going on in India, so I’m just going to flip that. Someone said recently that you don’t fight in a fight when fight is unfair, but that’s not what real people do. You do fight even when the fight is unfair, even when things are not playing towards you or for you. That’s when your real courage shows up, right?

Bhramar Mukherjee:

I do not know if it’s a broader fight for an agenda or an advocacy or a political fight, but when it’s a fight with data and concrete numbers and you know your stuff, what you’re doing and you can reproduce your analysis, then I think that you just have to find the courage to go up and say it. That’s the biggest part. I think there are many people who are running very credible analysis. Very few people have the ability or the interest or training to go out in the public and talk about it. My biggest advantage is that my family is all of actors and performers. So they’re very courageous. Live theater is one of the riskiest profession. You make one mistake and that show is done. And so my father at age of 86 can get up on stage and act. So I think I carried a little bit of those genes, which made me a little bit more fearless than an average statistician.

Raj Daliboyina:

The arts are just as important as science. That’s what you have said. And I must say I agree with that. So all the attributes that you have described, I think that ties it up to capacity building that you were talking about. So do you think there’s anything currently missing or how do you think we should train future public health professionals or statisticians in particular?

Bhramar Mukherjee:

I do think that it is very important for our students to believe that their work has impact. You cannot go a day without making impact and the small ways you choose to make impact is in your education every day, in your classroom, in helping a peer, in making a point which has never been made before. It does not have to be huge every day, but those small gestures of impact that you do every day, the small gestures of kindness that you do every day, that really adds up. And when you graduate and you are going into the workforce, it makes you a better scientist and a better statistician. But also I think keeping up with the skillsets, right? What does education mean to a person? Why are we investing our time and our money and our resources into education? If you asked me five years ago that one of my students is going to be hired as a prompt engineer, I’ll say no, because this job title did not exist.

You can see that data science, data scientist as a job title did not exist until the early 2000. So how can I predict the future? What I can teach my students is the ability to think critically and have some foundational skill sets to pick up future skillsets. Our students have to be future ready. Anything specific taught in their classroom may not be relevant, but the ways that we teach them how to learn, how to approach a poor problem, how to ask a question and from the question get to the answer and interpret it is what makes a huge difference. So I think that if you follow your inner instinct of developing foundational skills, that is eternal, right? That’s cross cutting. That’s going to stay with you regardless whether the buzzword of the day is data science, or AI, or quantum computing, your skills are going to stay with you. So try to build those foundational skills as opposed to just focus on the specificity.

Faith Ibitoye:

For your students and those that you currently teach or maybe the summer program itself, do you incorporate the later parts of this four quadrants as early as possible?

Bhramar Mukherjee:

That’s a great question. And yes, I do. The way I structured my new course at Yale because also I could actually start from zero zero. So that was brilliant because I did not have any carryover. I did not have any mortgage of slides and so I really created it from scratch. And so the way communication is a huge part of it because with AI and homeworks being done by GPT, you can never trust your learning unless you are standing up and presenting. So that you will see that when you have to explain to a group of people, that’s when you dig deep into something and learn the best. And so most of the class was based on discussion and communication. The final exam, which is going to happen next week is an oral exam. And so where they have submitted their projects, but I’m going to ask them some questions and they’re going to explain.

And this is exactly what I want my students to learn. I want my students to learn how to solve problems and how to explain their reasoning behind solving the problem and where they’re going to take this work, the next step. And the alternatives they considered and eliminated in the process. Science really progresses much more by elimination of other alternative paths. I do incorporate that. Community is something that the students, most of my students are engaged in terms of projects involving real data and ongoing studies. I try to take them to the field as much as I can. Ethics and equity is a critical pillar and then the fourth quadrant of training and capacity building, I hope my graduate students are mentoring undergraduate students, my postdocs are mentoring that other PhD students in the lab. So I’m hoping that I’m creating a circle and orbit where mentoring is valued, appreciated and is sought after.

Faith Ibitoye:

That’s really great and powerful. And thank you for doing that and instilling that in your students to really have a great future in public health itself. There was one thing you said towards the end of your lecture about sometimes when the quadrants doesn’t work, you have to go back to the origin. Has there ever been a time there is actually a tension between those quadrants and how do you navigate that?

Bhramar Mukherjee:

Yeah. The tension is really to be lived for. Tension in life cannot be eliminated. As a statistician, as I told you that uncertainty and tension are very poorly understood and processed, but they should be really embraced. Because the tension between viewpoints is how science progresses, right? But if you’re talking about tension between how much time should I give to each quadrant, that’s really a complicated question. And I’m sure as students you face the same challenge that how much time am I going to give to your four quadrants of passion? And sometimes if you do one, you cannot do the other. So my approach has always been to try to find projects which touches each quadrant, at least two of them, so that they’re not disjoint for things I have to do.

For example, the Big Data Summer Institute, it’s really community building, it’s really capacity building. Students learn how to communicate about statistics. They have scientific writing workshops and so on. And then they get strong training in foundationals of responsible conduct of research in science. I think that you have to touch all four quadrants through your projects, at least some of them.

Raj Daliboyina:

What you said about tension was so beautiful because it reminded me of what my mom used to tell me through med school. She used to say that if there’s no tension in the string, there’s no music. So it’s all about having that balanced tension between any aspect of life. You need it, but you need to know how to balance it out for you to achieve what you’re trying to achieve in life.

Bhramar Mukherjee:

Yeah. I actually listened to a presentation by a student at Yale yesterday and they were talking about that with use of GPT or generative AI tools, everything is becoming very smooth because the writings are smooth. People are very polished, but maybe we do not want that. If you think about friction of ideas, mistakes that you make, that teaches you much more. And I’m really concerned about this, everybody using five tools in order to come up with their answers that is leading to a epistemic narrowing of ideas that we have never seen before. If the distribution of human mistakes shrinks and becomes narrow, we are not going to be as creative. If you think about how you solve a math problem, you actually probably go through multiple routes and you say that, “Oh, this is not working out. This is not working out.” That’s your learning process. If your distribution of mistakes are in one direction guided by one platform, I think that has massive consequences on human cognition and I’m really worried about that.

Raj Daliboyina:

It’s a constant struggle between trying to be the most efficient both in the product and also in the process that is where the contention seems to be arising from because you have a great product by the ChatGPT, but the process seems to be extremely compromised at this point.

Bhramar Mukherjee:

So the question for us to answer is that what point to introduce any of these tools into our teaching or research workflow? So the way I have been using it, and that may not be perfect, is that I never get my original idea from these tools. I think about the original idea myself and then in order to finesse it, I partner with dialogues in terms of some of these deep research functions. Claude is amazing in terms of doing literature review and telling me some synthesis of what my fundamental flaws in my ideas could be, but I never generate that foundational idea. So that question part is very important to me. And then I think more and more we have to be prepared to do end-to-end science. We find five molecules that could really bind with this protein. Let us take it to wet lab.

If we actually find something really important for community, let us actually put it to test. See that if it can improve the health in terms of the outcomes that we care about. So the pipeline of building machine learning models and risk prediction models will probably be very streamlined, and shortened, and automated, but then what do you do with it? Maybe we should think about, we should really engage the white space that is left because this time is now shorter to think about those broader issues. So I think that how to strategically delegate your tasks from which you do not learn anything, but really don’t delegate your understanding because that will be the end of it.

And I know that as students at the University of Iowa or anywhere in the world, students are there because they want to learn. They just do not want a degree, a checkbox given to them. And it is the onus is on us to figure out how to still keep challenging them and stimulating them.

Raj Daliboyina:

Yeah. I think as a community of education, both the professors, the faculty and the students have to discuss this, but on a lighter note, I think the emails have been really smooth these days. And I think all of them who are in the education sector would agree with the ton of emails that we had to deal with.

Bhramar Mukherjee:

Not just the emails, but there are some amazing uses and efficiency gains as well. For example, like just editing papers took me a very long time, right? Even if the ideas were in place. So I think editing papers have become much easier because I always tell my students that do not write the paper, just put the logical structure in place. After that, involve any editor or any GenAI tools, but put the logical structure, how you are thinking about the problem first in place. So I think another good use of GenAI, I’m going to give you an example that I was writing a paper on we really need a biobank collecting South Asian data, right? Like the point that I tried to articulate that we need genetics data more from South Asia. How can we form a consortium in South Asia?

So I actually use the deep research function to identify and create a table which has all of the leading biobanks, has all of the PI’s names, where can you request data? What are some of the data use agreements in terms of biosamples into one table? You should have taken my graduate student two months to do that and now that table is created. So then this is done, this is of course not a paper. Then what did we do? We actually contacted all of the PIs of the biobank and think about what would it take to put together a consortium of South Asian biobanks. So I do think that you can use it gainfully and use your empty space or the newly gained time to do the bigger thing, the next step. And I think that’s very profound in terms of public health researchers that what do we do with that finding in terms of improving human health?

Raj Daliboyina:

Especially we should stop sharing personal stuff with ChatGPT.

Bhramar Mukherjee:

And that again really brings us to a point of public understanding of privacy and confidentiality of data. I have asked many of my family members that, do you know that your prompts are being seen by the company, are being kept for 90 days? Or do you know that your data is being used to retrain the models? 90% people say no. And that is because it’s an opt-in consent. It’s not an opt-in consent, it’s an opt-out consent, right? So basically you just by default you will be in and you’ll have to opt out. But to me, I think you should specifically ask for opt-in consent. And then when I tell them actually that all your data is going back, they tell me that it’s out there anyway. I can see that anytime I do something that all of the ads come up into every device so this is really out there anyway. So I think we really need to have more privacy, confidentiality, security, data security related lessons and literacy for the public.

Raj Daliboyina:

Yeah. And I think the European Union is I think one of the strong proponents of that, like trying to get-

Bhramar Mukherjee:

They have GDPR, but then again, I think that you have to have policy at an upper level in order to make that happen, right? Because otherwise if it’s not influenced at a higher meta level, it’s not going to be the individual’s responsibility.

Raj Daliboyina:

It was amazing, Dr. Mukherjee. So taking away from all of that we have talked today, if the students are listening to you today had to take one thing from your work, what should that be or what kind of a change that you would say they should take away from their approach toward public health from this conversation?

Bhramar Mukherjee:

So I personally feel that I’m a human being with very modest abilities, but where I stand out is my courage and the willingness of taking risks. Even every move, I recently moved from University of Michigan to Yale University and after being at a place for 18 years with a lot of friends and community and research capital to let go of all of that and look for a new future is not an easy decision. But I did that because I think that you have to experience growth pains to stand taller. And so the immediate growth pains have to be endured in order to stand taller. So this is one thing about taking risks and if you’re not happy with the current situation or if you are feeling like you need a change, please embrace that change, look for the right change. That’s one advice I’m going to give.

And the second thing is this skill building, right? That we are not really solving particular problems or picking up particular skill sets. We really are toning our intellectual muscles to run a long marathon. Never forget that. I tell all my students that do an hour of coding and hour of math every day because that’s training your muscles. That could be for any reason. Your PhD problem may not be related to what you’re doing, but ultimately it really prepares you for the long run and long haul and whichever career you choose, be it in academia, be it in government, be it in industry. I believe that all of you want to have a long legacy and durable platform of creativity, so that’s all.

Raj Daliboyina:

That was such an amazing thing to say and end this conversation, Dr. Mukherjee. We are so happy to have you here. We were amazed by your lecture and we were even more highly impacted by… We are so much more thankful that you agreed to do this podcast with us and shared your ideas and your visions. And I think it’ll be a great asset for future public health people to listen this and find the courage that really is acquired in this world today. Thank you.

Bhramar Mukherjee:

Thank you so much for having me and we wish you both and the student community here all the very best.

Faith Ibitoye:

Thank you so much, Dr. Mukherjee. And for everything shared, I’ll share the last advice you gave about growth pains and about taking that risk. I think it’s very ever inspiring to me and to the students that is going to hear that. And I’m actually thankful this is actually something I get to test personally. So thank you so much for the things you’ve shared and for the stories you have brought to bring light to that, to your journey as well.

Raj Daliboyina:

So listeners, that’s so nice of you to say that, Faith, this is our episode this week. Big thanks to Dr. Mukherjee for joining us today. This episode was hosted by Raj and Faith and was produced by the Plugged in to Public Health Team at the University of Iowa College of Public Health.

Lauren Lavin:

So that’s all for the episode this week. A really big thank you to Dr. Mukherjee for joining the podcast and for such a thoughtful and deeply reflective conversation. I think one of my biggest takeaways from this episode is the reminder that public health is not just about technical skill or data analysis. It’s also about courage and communication, trust, collaboration, and the willingness to engage with complexity, uncertainty, and even sometimes discomfort. Dr. Mukherjee spoke so powerfully about the importance of tension and friction in both science and life, and I think that message will stay with many of us, or at least I know it will with me.

This episode was hosted and written by Raj and Faith and was produced by Lauren Lavin at the University of Iowa College of Public Health. You can learn more about the University of Iowa College of Public Health on Facebook, and our podcast is available on Spotify, Apple Podcast, and SoundCloud. If you enjoyed this episode, please consider sharing it with colleagues, classmates, friends, or anyone interested in public health and the future of science and communication. Until next week, stay healthy, stay curious and as always, take care.