 Hey everyone, welcome to theCUBE's live coverage of the Women in Data Science Worldwide Conference 2022. I'm your host, Lisa Martin, coming to you live from Stanford University. I'm pleased to welcome, fresh from the main stage, Tina Hernandez-Boussard, the associate professor of medicine here at Stanford. Tina, it's great to have you on the program. Thank you so much for this opportunity. I love being here, and I've been coming to WIDS for many years. So it's exciting to be part of this and participate. It is exciting. It's one of my favorite events since I was telling you before we went live, and if you think about it, they only started back in 2015, and how it's now, it was a one-day technical conference, and now it's this worldwide phenomenon in 60 countries, and just 200 different local events. Talk to me about, I caught part of the panel that you were on this morning, and one of the things that I loved that you were talking about was mentors and sponsors. Talk to me about what you guys were talking about on the panel overall, and some of your mentors were, as you came up in your career. Yeah, so mentorship is so important, and really it makes a difference in people's careers. So I come from first-generation family. No one in my family has had any higher education. So having a mentor, an academic mentor, was just made all the difference in the world for me. So I started undergraduate, and I was immediately paired with somebody, a mentor because I was first-generation. And this person, he's no longer with us today, but he believed in me, and he opened doors for me, and he opened my eyes to all of these different opportunities. And having somebody who believes in you and really can help you pursue these other ideas is just means, it's so important. And so we talked about, in the panel, we talked about the importance of having a mentor, but we also talked about the importance of being a mentor, and helping people and students coming into the field find their place and develop the confidence that this is for everybody. There's something for everybody here, and you've got to try, you've got to put your name out there, and having support is really important. Oh, it's critical. Even some interviews I was doing last week for an International Women's Day, which is tomorrow, I was surprised at the number of women that I talked to who said, well, I was told no. No, you can't study computer science. No, you can't study physics. And talking to- This is a really difficult field. Are you sure you want to pursue this? Well, yeah, I, you know. Yeah, instead of so having those mentors and that encouragement to help build that confidence from within is a game changer. Absolutely, absolutely. Tell me a little bit about your research. Yeah, so we do, we use electronic health data, all types of different health data to really define and predict, prognose healthcare outcomes. We develop AI algorithms, tools to analyze the data, and we really try, and one, incorporate the patient's voice in the tools we develop, and two, we try and get those back to point of care. So I think a lot of emphasis has been about model development and model performance, and we really focus on, okay, that's great, but it's only useful if we can get it back to the hands of the clinician, back to the patients to really improve health outcomes. And so that's a big piece of what we do, and as part of that, understanding patient values and patient preferences is really a big important aspect of developing optimal treatment and optimal models. That's good, involving them in the process. How can data science promote health equity? First of all, what is health equity, and how can data science help drive it? So health equity is really an important topic, and there's a lot of different definitions about what is health equity. Health equity, what we want is we want equal outcomes, and that's not equal resources. And so a lot of times, there's this contingency behind, are we trying to have equal resources for every patient, or is it equal outcomes? And so we really focus on equal outcomes. We want everyone to have the opportunity to have the same outcomes. So health equity requires that we really think about different populations and their different needs, different preferences, et cetera. So that's what we focus on. And so to your question about how data science can promote health equity, one of the things we've been working on is really thinking about the gold standard of clinical care, which is the clinical trial. So a clinical trial is our gold standard for what treatments work, in what situations, and for what populations. However, a lot of the clinical trials are developed in a non-representative population. So we have to have patients who can come into the care setting at multiple times during a period. They can't have particular diseases. They can't, for example, in one particular trial, we were looking at how they had to have a specific BMI. They couldn't have diabetes. They couldn't have all of these other healthcare conditions, which at the end of the day, it doesn't really represent the community at risk, right? And so when we develop these models using clinical trial data, a lot of times it's not generalizable to routine care. And so AI can really help that. AI can help us understand how we can better identify patients to include in trials, what patients are going to be more likely to complete the trials. And so there's a lot of opportunity there to think about how we diversify clinical trials and also how we can start thinking about stimulating and doing pragmatic clinical trials if we don't have enough data to represent certain populations. Right, one of the exciting things about data science is all of the things that it's informing. It's also, it can be, there's pros and cons, right? But when we talk about AI, we always talk about ethics. How is it being used in healthcare? How are you seeing it being used ethically and effectively in healthcare to really turn the table on some of those biases that to your point weren't representing some of the most vulnerable parts of the community? Right, and I think we've been taking this holistic approach of the AI life cycle. So not just focusing on the data we capture, but the whole life cycle. So what does that mean? That means where's the data coming from? Who's capturing it and in what setting, right? If we're only looking at the healthcare setting, well, we're missing another large population. If it's only collected via a mobile device, there's another big population we're missing. So thinking about where the data is coming from and then thinking about who it represents and who's missing from that. The next step is really thinking about the questions we're asking. I'll give you a good example. We can ask, can I use AI to predict a no-show appointment? Or can I use AI to identify barriers for this patient to access care? So really even thinking about how we can flip that question to make it more equitable, make it more diverse is really important. And then there's been a lot of work in model development and algorithmic fairness and so there's a lot of research on that. But then there's another piece that we don't really see a lot on and that's model deployment. What are the biases when you introduce this into the healthcare system? The clinicians, how do they understand the data we're giving them, the tools? How do they use that to make clinical decisions? And then also what systems can actually deploy these AI algorithms because they're very resource intensive. So we think about the AI and equity along all of those aspects of the AI lifecycle and it really helps us get a more holistic view because each of these components intersects. They do, you're right. Tell me, I'm curious a little bit about your background. You are associate professor of medicine here at Stanford but give the audience an overview of the path that you took to get where you are. Yeah, and so not a straight path which is often typical that we're hearing today. So I started getting a master's in epidemiology and public health. And from there I was like, I wasn't really sure what I wanted to do. I applied to medical school. But I'm like, I'm not sure that's what I want to do. So I went and got a PhD in computational biology. I'm very, I'm only data savvy. And so thinking about how I could use data and then I was interested in healthcare. And so I got my PhD in computational biology. And from there I was thinking about, well I was really interested in the application of data science to the healthcare field. So then I got a master's in health services research. So it's the combination of all these different degrees that make me really have, I think a diverse view. I really understand the need for multidisciplinary teams and how we need opinions and viewpoints from so many different disciplines to really create something that's equitable and fair and something that is feasible and usable. Thought diversity is so important in every aspect of life, whether we're talking about business life, personal life and without it, there's bias. Absolutely, absolutely. And so we see this, we'll have a clinician maybe come to us with a question and then we'll have the health economist think about it. We'll have the other people think about it. And we kind of work it and we massage it to get to something that's meaningful and something we can really use that's gonna change care for patients or particular patients. And so it's really important to have that diversity. Absolutely, talk to me about your team. Yeah, so my team is very diverse. And I'm very proud of that. We have diversity across every aspect. So we have racial ethnic diversity. We're probably about 80% female on my team. And so interestingly, one of my members was like, wow, I didn't realize I'm like such a minority on your team, it was a male. And so I'm like, and I'm very proud of that. But we also have very diverse disciplines. So we have a lot of medical students, medical faculty. We have computer scientists, engineers, epidemiologists, health policy experts. And so it's very, very diverse. And what I like to do is I like to pair people up in teams. So I might put a health economist with a computer scientist and watch them go. And it's just amazing how they can learn from each other and the directions they go in. It's just, it's really incredible. Well, the opportunities that that interdisciplinary relationship builds, I mean, opportunities and possibilities must be endless. Yeah, and it also allows students to understand how to speak to different groups because we don't speak the same language. We really don't. And equity is a good example. So equity to me might have a certain meaning, but equity to the health policy expert might have a different meaning. And so even understanding how we speak to other groups is so important. And being able to translate something in a simple language that other people get is really key. Absolutely, here we are. Tomorrow is International Women's Day. Exciting. And it is exciting in Women's History Month. We get a whole month to ourselves, which is fantastic. But one of the things, you know, when we look at the, at the data, 50, you know, the workforce, 50, 50, males and females, but the STEM positions are still so low. Right, right. Below 25%, are you seeing, obviously WIDS is a positive step in that direction to start shifting that. But what do you tell the younger set in terms of, yeah, it is a challenge. Yeah, it is a challenge to really, and this is the example I always give, as a woman, we've all walked into these rooms that are all male. Oh, yes. We've all walked into these rooms where you're sitting at the table. Oh, can you take notes? And it's hard. It's really hard, but you know what? It takes courage. So again, that mentorship, being able to speak up, being able to set your place at that table, I think is really important. And we're doing better. We're doing better, but it really is through consistent mentorship, consistent confidence building, et cetera. It is. Yeah. And this event is fantastic for that. It's going to reach about 100,000 people annually. Amazing. Men and women of all ages, of all different career backgrounds, which is fantastic. But the International Women's Day theme tomorrow is breaking the bias, hashtag breaking the bias. Yes. What do you think we are on that? I think we have a lot of ways to go. Yeah. And so there's bias in our teams. There's bias in the way we think. There's bias in our data. And there's been a lot of publicity and hype about the bias in the data. And it is so true. And certainly in healthcare systems. And it's important to understand that when we're developing AI and all these machine learning or data-driven models, they learn from the data we give it. So if we're giving it a biased data sample or an unrepresented data sample, it's going to learn those characteristics. And so I think it's important that we think about, how do we do a better job at capturing data, diversity, voices from different populations? And it's not just using the same tools and technology we have today and going to another community and saying, okay, here's what I have. It's not working that way. So I think we need to think outside of the box to think how do these people want to communicate with us? How do they want to share our data? It's about trust too, because there is trust is a big issue with that. So I think there's a lot of opportunities there to just further develop that. Do you think there really is going to be such a thing one of these days of a non-biased, unbiased set of data? I don't think so. I don't think so because the more we dig, the more biases we find. And so while we're making great strides in race and ethnicity, diversity in our data sets, there's other biases, male, female, age biases, disease biases, et cetera. So just the more we dig into this, the more we identify, but it's great because when we find these gaps in our data or gaps, we take steps to address that and to mitigate those biases. So we're moving in the right direction for sure and putting a spotlight on it and being transparent about it I think is key to move forward. I agree, that transparency is critical. Yes, absolutely. And we often say she can't be what she can't see. Right. And so from a transparency perspective in the data and also in the visibility of the leaders and the mentors and the sponsors, that transparency is table stakes. Absolutely, absolutely. What are some of the things that you're looking forward to as we hopefully move out of the pandemic into the end of it? I can imagine with the masters in public health you have an MPH. Yes. Your perspective must be so interesting on living through a global pandemic. Right, it is. And it's interesting because we've gone through the pandemic and now it's turning into this endemic, right? And so how do we deal with that? And one of the things I think that is really important is we find a way to still meet and collaborate face to face, share ideas. This conference is amazing where we can share ideas, we can meet new people, we can learn new perspectives and being able to continue to do that is so important. I think that during the pandemic we really took a big hit in the transfer of learning in our labs and in our teams. And now it's funny because my team, they're like, let's go to lunch, let's do a happy hour. Let's, you know, they just want that social interaction. And it's more to better understand the perspectives of where they're coming from with their questions, better understanding the skills they bring to the table. But it's just this wonderful opportunity to think about how we move forward now in our new world, right? Yes, we're getting there slowly but surely. Well Tina, thank you for joining me, talking about your role, what you're doing, the importance of mentors and sponsors and the opportunity for data science and healthcare. We appreciate your insights. Absolutely, thank you for having me. It's been a wonderful, wonderful thing. My pleasure, excellent. Thank you. For Tina Hernandez-Boussard, I'm Lisa Martin. You're watching theCUBE's live coverage of Women in Data Science Worldwide Conference 2022. Stick around, my next guest will join me shortly.