[music playing]

Speaker: From the campus of Harvard Medical School, this is Think Research, a podcast devoted to the stories behind clinical research. I'm Oby.

Brendan: And I'm Brendan. And we are your hosts. Think Research is brought to you by Harvard Catalyst, Harvard University's Clinical and Translational Science Center.

Speaker: --and by NCATS, the National Center for Advancing Translational Sciences.

[music playing]

Brendan: In health care settings, equal access to quality care plays a hugely important role for physicians and patients. The idea that access for any two patients is the same, though, isn't always true. When we think of disease and health as being impacted by a single cell or organ, we lose sight of many larger interwoven parts of a patient's life. Access to food or medicine, income level, and race, all compound to create different experiences for individuals in the health care system.

Using his clinical background and interest in business and bioethics, Dr. Junaid Nabi wants to provide patients with more than just equal access, but a truly equitable system for all. Dr. Junaid Nabi is a senior health care researcher at Harvard Business School. Dr. Nabi, thank you very much for joining us and welcome to Think Research.

Junaid Nabi: Thank you for having me.

Brendan: So tell us a little bit about your clinical training, and how that has informed your research on access to health care and quality of care.

Junaid Nabi: Thank you for that question. So I grew up in a very politically fragile environment in Kashmir, which is still undergoing a lot of changes in the last few years. And early on I've experienced the social health and economic inequities that our community in Kashmir has faced for a very long time. And I could sense from the very beginning, given that my father was a physician and my mother was working with the government, that it wasn't really always about just the pathophysiology, or what is wrong with the person's organ or a certain cell. But there were a lot of interrelated aspects to every single disease.

So for example, it's very easy to say malnutrition in children is because of lack of nutrients. But if we look more broadly, and if we understand the context, a lot of these children don't have access to proper food, they often don't live in communities that have resources where they can afford proper food, and a lot of times they belong to communities that don't have the power in the community to be able to provide for their children. So a lot of those reasons go into that malnutrition. It's not just about lack of protein or vitamins. It's about the broader socioeconomic determinants of health that do contribute towards these diseases.

So for me, early on looking at these different aspects in my own community, I was inspired to work towards these problems at a broader scale. And that's what got me involved in medical training, and eventually in Public Health training, and now in business strategy.

Brendan: And so you mentioned you grew up in Kashmir in India and you had a mentor-- you had a role model in your father, who was a physician. What made you want to pursue medicine? And what was your focus during your medical training?

Junaid Nabi: So looking at my father who was working at the front lines for a very long time in Kashmir, I could sense a couple of things. One was the fulfillment that he felt as a physician working for the community, the aspect of service-- that one was contributing some value to the society and could have a decent professional life. But the other aspect was how much value one can add just by-- not just the service aspect, but also helping people by providing them information, by providing them certain tools or even nuggets of wisdom that can really save them a lot of time and energy and resources down the line.

Given the community that we lived in, our resources were scarce from the very beginning. So he had to make a lot of clinical decisions that took that context into consideration and provided the prescriptions or other treatment modalities within that context. So that also provided me a lot of understanding of how to really take care of resources. It's not just a matter of looking at something in the textbook, but also how do we apply this information in a textbook in the context that we live in.

And so I trained in medical science at that point, was really inspired by my dad and a lot of my family members are also physicians. So I was inspired by that. In medical school, as I mentioned before, I sort of realized that it wasn't really about the pathophysiology, it was also about other elements of the society we live in. And then I thought I should get more training in public health, and that's how I got involved with the public health aspect, the broader health systems aspect of things. And I do continue to learn about clinical medicine and use that in my routine work, but all of this work is guided by the broader socioeconomic determinants that we have in front of us.

Brendan: Right. And so, as you said, your research examines the social determinants of health and one issue that you've looked at is racial differences in cancer outcomes. Could you tell us a little bit about that research?

Junaid Nabi: Yes. It was a fascinating research. So I've been involved with cancer care outcomes research for some time. And I've published some of the studies on, what are the socioeconomic determinants of these problems and where do they stem from. Because a lot of our information that we have while in medical training, or even afterwards, comes from textbooks and medical journals and the evolving body of literature.

And if you look at a lot of those resources, what we have seen is that a lot of times there is this sort of a blame towards certain communities. So for example, in the US if you look at a lot of the resources, it's very obvious that it seems that the outcomes for certain cancers are worse in the Black community. And our research wanted to really understand, was it really about the community or was something else going on. Because even now in textbooks and other forums of discussion there does seem to be a certain biologic determination for these elements. And I don't believe in that, and neither did our research group.

So we tried to investigate this problem more deeply. And what we did, we analyzed the men with advanced prostate cancer to see what were the differences and where were those differences stemming from. And what we found out was access to care, treatment, and cancer characteristics, when we accounted for those, Black race was associated with better outcomes compared to other communities.

This was really revolutionary in a lot of ways, because so far it's always-- a lot of these discussions always seem to suggest that there is probably some kind of biological issue going on in a lot of these issues. But that's not the case. Our analysis shows that demographic factors, as well as socioeconomic factors such as access to care, the quality of care, these have a much more important impact on the outcomes for these cancers rather than what we-- look at the genes, for example. That's not the only thing at play here.

And these other demographic and societal factors play a much, much bigger role. And so our research was able to show that the incidence of prostate cancer is a little bit higher in the Black community. But the outcomes are not associated with some kind of biological background, but more about the demographics and access to care, and the quality of care that they receive at different places.

Brendan: What can policymakers learn from this kind of research?

Junaid Nabi: Yes. And a lot of that work, like I said, is an ongoing body of work. But all the signals that we received so far, all the papers that are coming out with the robust data, they are suggesting that that is where we should be focusing. And I think for health policy leaders these are really important considerations when they think about investments, when they think about allocation of care, when they think about what kind of centers should be promoted and what kind of centers should be invested more in.

I think we should take the lessons from this research towards enhancing access to care, enhancing quality of care, and really making sure that we are measuring not just the outcomes, but also how care is delivered. Because if you look broadly at the system right now, it's not so common to measure patient reported outcomes, for example. It's not that common to measure, how do these patients fare when they go outside the health system.

It's very easy-- I wouldn't say easy, but it's relatively straightforward to look at outcomes such as how many patients were re-admitted, how many patients had complications after certain procedures. But we should also be looking at what happens to a lot of these patients when they go home, three months out, six months out, 12 months out. What is happening?

Are they getting better? Are they able to go back to their functional state? Are they able to do the activities that they were doing before the diagnosis? And that requires measurement. And so I would say, broadly, greater investments, improving access to care, and as well as investing in modalities to improve outcomes measurement.

Brendan: And when we spoke previously, you talked about a program at Boston Medical Center that was doing some of those things. I wonder if you could talk about that example, because I think it's a nice, specific instance of how a large health system is doing this.

Junaid Nabi: Yes. Boston Medical Center has been a leader in the space of incorporating social determinants of health to its delivery of health care. And a lot of other institutions around the Boston area have invested in these programs. So, as an example, what a lot of these places do is, for example, we have elderly patients and they require a certain kind of care. Their needs, their patient needs are very different. So we must consider that context before making policies or developing protocols that think that every single patient is the same. That's not true.

So what a lot of these places have done is they have utilized information technology and other evolving technologies to improve patient care. And for elderly patients, that means, in a lot of these places, that they have provided them access to tools such as tablets, and information systems to contact their primary care physician in a much more robust way. So before, what used to happen for a lot of these places was patients would have to call the ambulance, they would have to go to the ER, or they would have to see a doctor who they're not familiar with for something-- it could be an ailment of a minor nature. It wouldn't necessarily require a visit.

And that also goes back to the cost issues. That increases the cost in the system, and that eventually burdens the system even more. So what they have done in a lot of the places is they have provided them access to nurses, at-home nurses, they have provided them access to care resources, care coordinators, to help them really navigate the space. And it goes back to understanding the system that's at play here, to really focus on the socioeconomic and, as well as, political determinants of health.

What is happening with a certain patient population? How are they-- what kind of care do they need? What are their needs? And addressing those needs where they are, rather than where the system wants them to be. So I think investing in those modalities is the right direction to pursue.

Brendan: As we've been talking about systems and structures, and that's sort of how your career-- that's what you've been focusing on in your career-- looking at systems structures, how they provide or prevent access to care. And one area that you're also looking at is artificial intelligence. Can you tell us about how artificial intelligence is used in health care, and what some of the problems are?

Junaid Nabi: Yes. That's my focus from the bioethics side. I was at Harvard Medical School for a year or so, training in bioethics, and my focus was artificial intelligence in medical devices and medical decision-making systems. And what I realized was-- so the first question is, what is the role of AI in medicine. And that's a really important question, something a lot of us have been thinking about.

The reason why AI is making so much news, and a lot of progress as well, is for a couple of reasons. One, we have a lot of computational power currently than we ever had before. So creating certain models, creating certain algorithms and systems, are much more feasible today than they were ever before.

And the other issue is that of resources. If you look at how many patients we have in the US and how many physicians or other care providers we have, the match is-- there's a severe imbalance. It's very difficult to care for all the patients in the same way.

And what that has resulted is in fragmented care. That's why we have a lot of other care providers and care navigators in the system now. But even that would not be sufficient in the upcoming years because the population is aging, there's a lot of diseases that we're diagnosing at an early point now, and we do have-- we are witnessing a higher incidence of a lot of diseases than before.

And what AI can do is AI can automate a lot of these processes that were manual before, whether they are administrative tasks-- from scheduling meetings, or even prescriptions-- updating certain prescriptions, and other functions that a physician is routinely doing. They can automate a lot of those tasks. And so what that does is free up the time for a physician or a care provider to do what they are trained to do-- to talk to the patient, to counsel them, to spend time with them, and not spend their energies on just the administrative and documentation needs. Which are important, but they take up a lot of time. And there's a lot of burnout in care providers these days, and I don't know a lot of folks who are happy doing a lot of that work.

So if we can automate a lot of those processes, that would be really good. And it would save a lot of money as well. And that saving is going to translate in profit for a lot of companies. And that's why they are investing in those modalities.

For me, I was interested in these decision-making systems because they will be eventually used, whether they are for reading X-rays-- which a lot of the research shows artificial intelligence is getting better every year-- or even if they are pathological slides for, let's say, cancer diagnosis. They are getting better as well. So there is a momentum that is building towards integrating a lot of these tools and technologies into routine medical practice.

And my concern was what would this mean for the patient. Because a lot of times when we discuss these issues it is from a provider, or a physician point of view, or a health policy of view. But what would this mean for a patient, who often is not a player, or an active player in the development of a lot of these technologies.

And I was able to look at a lot of the literature, as well as a lot of the cases that are out there being used. And what I found was there was a massive problem of bias in the AI enabled devices. And I'm sure you must have heard about news in the criminal justice system, there has already been a lot of the investigative work from ProPublica and others how the AI enabled tools were punitive for minority communities because they were not taking the context into consideration.

And so there are similar concerns in the medical field as well. I think the training sets that are used for data collection, as well as building these models, they are not comprehensive enough. They are not broad enough. They are not diverse enough. They don't consider the population as they are, because a lot of this is voluntary, a lot of this depends on who the investigator is, how are they enrolling people.

So what that does is that creates a bias in the AI, which is created by us, the ones who are building it, to offer outcomes that are more in alignment with how we think about society rather than how the society is. So for example, there was this artificial enabled device which neurologists were using, and they were using speech patterns to decode what kind of degenerative brain disorders a person might have or might eventually develop. And so it's a really amazing tool.

If we could have that, one can imagine how much administrative burden this would reduce. Because then physicians and scientists, they can focus on actually treating the problem rather than worrying about, how do we screen patients, and spending time on those issues. So this had a lot of promise.

But what they discovered slowly was that it was really looking at only a certain type of accent from a certain community-- and I don't want to go into details right now, but it was biased towards what was the majority in a country. And it was not taking into consideration how people who are in the minority, how they speak, or how their accent-- especially in a diverse country with different populations from different backgrounds and different accents. It wasn't able to consider that. And so what it would do is it could give you an output saying, this person does not have a certain risk for neurodevelopmental problem down the line, but that person could eventually have it. It was just because they were--

Brendan: --or vice versa.

Junaid Nabi: Yeah. And it wasn't able to recognize that-- because the system was built on a certain accent. The system was built on a certain aspect of one community. But if you are living in a diverse community where there are millions of people from thousands of different backgrounds, we may not be able to capture all of that if we're realistic. But we have to capture at least all the major communities. And we have to capture all the major characteristics of different communities, so that we can actually create a device that addresses those problems in a way that we would have wanted them in real life.

If a patient walks inside a clinic and speaks in a certain dialect or a certain accent, no physician would misdiagnose them, or diagnose or under-diagnosed them just because they don't speak the language in a certain accent. And that's where my work has been going on. And a lot of my research is looking at, what are the factors that we should consider, and how can we develop more diverse training data sets, and who are the people that should be involved in development of these data sets.

Brendan: So who should be involved? I mean you mentioned looking into different communities and taking into account the diversity of a population. So how do you create these more equitable systems? How do you involve more of the population of a country?

Junaid Nabi: It's all about diversity. It's all about considering different perspectives. And it's about enrolling individuals from different communities, and being intentional, that we must focus on these elements when we are building these models.

It's not just about, how do we create a certain model that can save us money. How do we create a certain model that can save us resources. How do we create a model that can save us time. But a person has to be intentional about, how do we bring the voice of the patients to this model. And that requires having a very broad data set, being very intentional about enrolling different communities, and making sure the proportion of people who are enrolled in these data sets are proportional to how they represent in the society that we live in.

And the other thing that we can do is we can bring different experts to the table, not just engineers. We can bring physicians, we can bring sociologists, we can bring economists, we can bring up public policy experts to advise on these projects. They shouldn't be an engineering project where we're just focusing on the technical issues that are involved. They shouldn't be just about the policy issues that we're trying to address.

But they should also be about how does society look, how does society function outside of our model, and how do we bring that perspective to the model. It may not be perfect, but that's the whole goal of any machine learning system-- to learn, to develop these models. And slowly they'll get to a point where they will represent the communities we live in.

Brendan: Yeah. It just goes back to who's creating these systems. And I think it leads in to your current focus on health care strategy, where you're studying at Harvard Business School looking at value based health care. So we just have a few minutes left. I just wanted to wrap up with that.

So you're looking at health care from the business side. Why did you think that was important? And how can leaders make changes that promote what we've been talking about in universal and equitable access to care?

Junaid Nabi: And that's a really important consideration right now. I think given what has happened globally in the last decade, I would say, it's very clear that leadership matters, and leadership matters in every single field, not just politics, but medical science, health care, and public policy as well. And so, for me, there are a couple of things that I have noticed in the last few years that motivated me to pursue more investigations into looking at, how can we use business strategy to address some of these questions on equity.

And my conclusion so far is, we don't have a lot of buy-in from leadership at different places. When people talk about equity, when they talk about improving access or improving quality of care, a lot of times they're talking in these abstract terms, and they're not investing in resources that would make these dreams a reality. And for that we need leadership.

For that we need leadership that understands these issues, and leadership that initiates projects to get this work done. And unfortunately a lot of those considerations to come down to the financial aspect of a lot of these projects. And, for me, understanding that, understanding how does strategy, how does finance, how does organizational behavior, how do these elements contribute to a lot of these problems was a really important question.

So, as I mentioned, a lot of my work so far has been looking at health systems and public health, even the-- whether it's AI or cancer outcomes, research, or even other global health work that I'm involved in. So many of these projects cannot move forward because we don't have resources, because we're not willing to invest in people, in infrastructure, in models, in facilities to make these dreams a reality. And as unfortunate as that is, I also realize that what we often need to make is financial case-- is a case of, how does this help an organization grow.

And I believe, personally, that all of these issues, if they were addressed, they would lead to not just profit, which a lot of financial leaders are concerned about, but they would also lead to better performance. I personally believe that. And I've seen a lot of emerging data on that. But, as I said, the buy-in is not clear for a lot of the individuals, a lot of stakeholders in the process.

And, for me, what is the current task is, how do I make a case to business leaders, to nonprofit leaders, to political leaders of investing in these resources, of investing in these ideas. And what elements of business strategy, finance, organizational behavior, or other business concepts do I bring to the table to make this case. Because I already have training in clinical medicine and public health and bioethics. But what is often missing is making a case through, let's say a balance sheet, or some other financial tool.

To some individuals it matters, it does matter out there to a lot of individuals. Return on investment, how does that influence these ideas? I think that is a consideration for a lot of leaders, and if I could close that gap, if I could make that case using some of these concepts, I would say that that would be mission accomplished.

Brendan: Well, Dr. Nabi, thank you very much for joining us. It was a pleasure to have this conversation with you.

Junaid Nabi: The pleasure was mine. Thank you so much.

[music playing]

Brendan: Thank you for listening. If you've enjoyed this podcast, please rate us on iTunes and help us spread the word about the amazing research taking place across the Harvard community.

Speaker: To learn more about the guests on this episode, visit our website, catalyst.harvard.edu/thinkresearch.

[music playing]