SDSU professor wins Fulbright award to study AI in health care

S1: Welcome in San Diego , it's Jade Hindman. On today's show , we talked to an Sdsu professor who was awarded a prestigious grant to study AI in public health. This is KPBS Midday Edition. Connecting our communities through conversation. So think back to your most recent doctor's visit. Did you notice them using AI ? Well , it's becoming more and more common for health care workers and researchers to use artificial intelligence. So some public health experts are now diving deeper into the use of it. That includes San Diego State Professor Susan Qin. She was recently awarded a prestigious Fulbright scholarship , and later this week , she'll be going to Spain to look at how AI can be used in public health around the world and locally. She joins me live. Professor Qin , Kean , welcome to the show.

S2: Thank you so much. It's such a pleasure to be here.

S1: So glad to have you. And first of all , congratulations. Thank you. That's so great.

S2: And I , I think that the the mission of the Fulbright is really a mission that is very central to me in terms of the the bridge between cultural divides and fostering understanding between different nations and , and their people. And I think certainly right now that's such a crucial thought to have in our minds. So it's a it's a huge honor for me.

S1: Oh , that is great. Well , talk a bit more about how AI is used in health care , in public health.

S2: And , you know , I don't want to go too deep into it , but I think there's several kind of what we would call use cases for AI in , in healthcare and public health. And kind of there's kind of four main buckets we might put those into. So first would be like AI to guide population health. So we might think of that as like predictive analytics as to who might develop or develop a particular chronic disease or the progression of chronic diseases. Or it could forecast disease trends. So like when we might have an outbreak. So kind of like speeding up analytics of like how can we sort of do population health better. Then the second thing would be patient virtual health assistance. So this might be what we think about like chatbots. So things that might help you in your health as like um one use case for this that I've I'm familiar with is around mental health. That might be a chatbot that assists you with , um , you know , you have a you want to chat about something around mental health , around maybe depression or something , though. We need to be careful around those types of things. Then another one is the frontline health worker virtual assistants. So these might be things for nurses to help them with their work , to speed it up , maybe help with if they're doing any type of diagnostics to help with that. And then finally physician clinical support decision support systems , which would also include radiology. So I think the biggest area where it's really being used , like day to day , where people might be familiar with is with radiology and helping to actually determine like , is this actually a fracture ? And it's actually quite good at that. And it has very good predictive ability with detecting fractures. Actually , I think better than humans. We maybe not. We maybe don't like that , but it's actually really useful in that area. So those are kind of some main areas where AI is being used within healthcare and public health.

S1: The possibilities are really amazing. So it has a lot of uses.

S2: And so thinking about how does AI help enable that ? Um , and for me in particular , how does that how does that look like for populations maybe in hard to reach areas in rural areas , populations that traditionally have , um , harder , more challenges in accessing health care. So that's where to me , I think the potential is , is bridging some of those gaps.

S1: Do you see the the I don't know the the current push against diversity , equity and inclusion impacting AI's A's ability to to to be of use to various populations.

S2: Um , I mean , I think possibly I mean , I think the , I think AI also has some challenges in that way. And that's actually one of the things I'm kind of interested in , in going to Spain , because I think they're trying to address that biases around language. So I think the way that AI is developed can also create biases. Um , and that's kind of how I'm interested in what Spain is doing around language biases and trying to address that within their own country , because they have a lot of co-official languages and how they're working to address that and really localize things. Interesting.

S1: Interesting.

S2: And so I you know , I think with anything , we had to be cautious about kind of thinking about how can we make something really serve the needs of all people. Um , you know , we are a nation with many , many people.

S1: I mean , you mentioned that there's all these different languages , but tell me a bit more about that. Yeah.

S2: Yeah. So the reason I one of the reasons I chose Spain was because Spain actually partnered with IBM to develop what we call small and large language models in Spanish. So a lot of AI is developed using English and not so much , so much in other languages. And so they partnered with IBM to develop these , what we call small and large language models in Spanish and some of their other co-official languages , which really kind of helps to address this language bias that exists when AI is developed in primarily in English. And then they're also implementing obviously AI I in various parts of the country , so I'm really curious as to see how have they , as they've scaled things up , how have they then kind of localized and tailored things in different regions ? So I think that's really the one of the important things is you can't sort of say one size fits all. Like we have this. Yeah , we have this great tool and we're just going to say , okay , all hospitals or all clinics have to use it. And it's like it's just one word. It's going to roll it out and everyone has to use it as is. But it's thinking about , well , how in this area do we need to adjust it , how do we need to tailor it , and what's really the process for that and the work that I'm doing , I'm not really on the technical side. I'm more on the what we call implementation side is studying. How do we do a better job of both adapting things that are evidence based ? And then also how do we what are the best kind of implementation models ? So how do we do the best job of implementing something and like , what's the systematic way of doing that ? So then we can kind of reproduce that way of implementing something. So that's really kind of the angle I'm taking on this. And that's really the big gap is that there's lots of AI happening , but we're not really studying what's the best way to actually implement it. So that it's most useful to people , to providers , to patients , to everyone. And it has kind of high usability. Right.

S1: Right.

S2: They have a masters for AI in healthcare program there , so I'll be doing a little bit of lecturing within that. And it's a program that has students from all over the world. So that'll be really interesting to learn , both from the students as well as doing a little bit of teaching.

S1: That's cool. So there's still a lot of , you know , concern around the use of AI in health care and more , you know , more broadly , for example , like research suggests that doctors might actually become too reliant on AI assistants.

S2: I mean , I think like any AI , even like students using AI , right ? If certainly it can sort of it can give us wrong information. And I think that's one of the biggest concerns , you know , anyone has about using AI of like. It can be it can give you something that's completely wrong. And if you don't as a human then interpret that information. I like to I like to use the word interpret of like we need human interpreters of AI or we need to filter it. There are also some kind of additional programs that are used with AI to kind of do that filtering of , um , so it's sort of like you have the initial AI and then you have something else to do to make it to improve the accuracy of it. So I think those are very important concerns. And so there either needs to be a significant human element in deciding if this information is actually accurate and useful or not , or there needs to be kind of that additional , um , program that supplements it to ensure that the information is accurate. But I think that that's really on everyone's mind as far as we're not just going to use the information that's provided , we're going to actually interpret it or screen it before we use any of that information.

S1: All right. Well , you've studied public health all across the world , particularly in , you know , HIV prevention and reproductive health. And , of course , we're seeing a lot of upheavals in international aid programs due to funding cuts there.

S2: And , you know , I in a way , sometimes that breeds more innovation and it breeds more collaboration. So , you know , I think it'll be some challenging years ahead. But , you know , I think everyone's very resilient. And certainly the people that I work with were all very resilient. And I think that , you know , we'll figure out how to do , you know , the best work , the best science and the best things for the the people that we work with and the patients and populations that we really care about. And we want to make sure that we're advancing health for them. And so I think for us , it's really our heart is in it , and we still want to do the work and we'll figure out a way to do that work , whatever it takes.

S1: Yeah , yeah. Is that pretty much your hope for the future of AI and other related technologies ? Yeah , yeah.

S2: Yeah , yeah. I mean , certainly the my interest in AI is stems from kind of , as I mentioned , kind of build bridging some of those gaps , figuring out how do we , um , like , improve health care for people who kind of need access to health care , whether it be , you know , locally in San Diego County or in this region or globally. And so that's that's kind of my hope.

S1: All right. Susan Keene is a professor of global health at San Diego State School of Public Health. Professor Keene , congratulations once again and safe travels.

S2: Thank you so much.

S1: Thank you. That's our show for today. I'm your host , Jade Hindman. Thanks for tuning.

S3: In to Midday Edition. Be sure to have a great day on purpose , everyone.