Harnessing AI and Data Modernization: Transforming Public Health for the Future

Artificial Intelligence (AI) and data modernization are having profound effects on the public health sector as they have in the health care and other sector.  But the exact nature of the impact on public health is just beginning to be fully understood.  Preliminary areas with great promise include assisting in communications (producing materials in multiple languages and literacy levels), in efficiently simplifying tedious processes (like gathering necessary public health reportable data from health care records in a fraction of the current time), and identifying early disease outbreaks and other health concerns as early as possible so they can be nipped in the bud (using such tools as syndromic surveillance in emergency room settings and Wastewater surveillance screenings).  ICF is actively involved in assisting CDC and other public health agencies in understanding and tapping the possibilities as well as understanding and helping to overcome the challenges Two of ICF’s AI leaders, Senior Vice President, Health, John Auerbach and Principal Data Scientist Eddie Kirkland, discuss these matters.

John Auerbach
John Auerbach is the Senior Vice President for Public Health at ICF where he serves as a thought leader and liaison with federal agencies and non-profit public health organizations. Over the course of a thirty-year career, John Auerbach has held senior public health positions at the federal, state, and local levels in both governmental and non-governmental organizations.

Eddie Kirkland
Eddie is ICF’s Principal Data Scientist, a statistics and data science expert at ICF with more than 20 years of experience in data and software engineering. He specializes in guiding data-rich projects from concept to delivery, working directly with clients to identify areas of need, developing custom solutions in an agile framework, and delivering clear and meaningful results.

ICF is a global consulting and technology services company with approximately 9,000 employees, but we are not your typical consultants. At ICF, business analysts and policy specialists work together with digital strategists, data scientists and creatives. We combine unmatched industry expertise with cutting-edge engagement capabilities to help organizations solve their most complex challenges. Since 1969, public and private sector clients have worked with ICF to navigate change and shape the future. Learn more at icf.com.

To learn more about this topic, check out this webinar: https://www.icf.com/insights/health/role-of-ai-advancing-public-health 

Thanks for listening! Would you like to know more about us? Follow our LinkedIn page!

Transcript

0:00

Welcome to the Fed Fed podcast where we dive into the dynamic world of government technology. In this podcast series, we'll be joined by current and former federal leaders and industry trailblazers who are at the forefront of innovation. Here, we speak openly and honestly about the challenges and opportunities facing the federal government and the Department of Defense and its partners in the modern age, driving innovation and the incredible capabilities of technology along the way. Whether you're a federal leader, a tech industry professional, or simply fascinated by it modernization. Just like us, this podcast is for you. And we're so happy to have you tuning in. We'd like to welcome John Auerbach and Eddie Kirkland. John Auerbach is the senior vice president for public health at Ikea, where he serves as a thought leader and liaison with federal agencies and nonprofit public health organizations. Over the course of a 30 year career, John Auerbach has held senior public health positions at the federal, state, and local levels in both governmental and non-governmental organizations. And I'd like to introduce Eddie Kirkham. Eddie is Ike's principal data scientist, a statistics and data science expert at Ikea with more than 20 years of experience in data and software engineering. He specializes in guiding data rich projects from concept to delivery, to working directly with clients to identify areas of need, developing custom solutions in an agile framework, and delivering clear and meaningful results. Welcome. Well, John and Eddie, thank you so much for joining us today. Thank you. It's a pleasure. Glad to be here. Wonderful. Well, why don't we dive right in. I'm so excited about having this opportunity to talk with you all about AI and other really important topics that are, within that public health sector focus. So let's go ahead and start with the first question. I has become of great interest to all of us. We've heard of some of the most promising applications in health care and other sectors as well, but we haven't heard as much about its applicability to the public health sector. How is it applicable? Well, maybe I'll start and then Eddie will probably have some concrete examples. I think it's very applicable. I think is the is the answer. But part of the challenge in the public health sector is it's a very uneven sector. We have the CDC and other federal agencies. Then we have the state public health organizations, and then we have 3000 local communities with their own public health departments. So when we think about AI, we need to think about its potential applicability for a wide range of different sizes. And, types of challenges, in order for it to be useful. But I'd say it is useful, particularly when it is in response to the identification of some concrete existing problems that already exist. I think that's great. And it it's interesting because I think the applicability has changed in the last few years. And it's just grown. So much more. If you look back, you know, 3 to 5 years ago, I was largely limited to, you know, only people who were working with massive amounts of quantitative data and they had to have a lot of expertise to be able to leverage it and leverage it appropriately. With the advent of newer approaches to it, particularly generative AI. It's opened AI up to operate even just in big quantities of text data. And I have not met anyone in public health that doesn't have a junk drawer full of text data that they wish they could process, they wish they could gain insights from, but they haven't been able to. Three years ago that wasn't possible. And it is now. So it's I would say it's much more applicable today than it was, you know, a few years ago. And it's growing by the day. That's awesome. Great perspective. So ICF has been directly involved in this work with the CDC and other agencies and organizations. Can you give us some specific examples and lessons learned? Sure, John, I can I can jump in and, give some specifics and then if you have more, please feel free to toss those in as well. In our work with CDC, we've seen a lot of different applications of AI. We, particularly saw a really powerful application of traditional AI to help with, document legal epidemiology documentation. And it was helping to analyze those, to process them quickly. They, a team had a huge backlog of papers that needed to be reviewed. We were able to eliminate that backlog and create a much broader and more useful data set for their team to track, legislation over time as it related to, in this case, it was Covid, mitigation mandates. So that was one application of traditional AI. Now we're seeing a lot of different applications of generative AI, where we're seeing it being used to tag and classify or code. A lot of people, reference, coding different data points, different documents in a much more efficient manner. We've seen it used in a lot of ways to consolidate and summarize reports, to summarize anecdotal data, to understand topics and themes, and be able to notice patterns across text data. There's a lot of very specific applications we've been able to champion, specifically at CDC. That's awesome. Maybe I, I just a couple of other examples. You know, again, some of these are very practical. Like, for example, NIH has asked us to help them with, literature searches that are very complicated and in the past took a long time when they're very complicated, specific literature reviews to identify, what has been learned, over the years about, certain health issues. We've also we have a tool called Comment Works, which helps when, federal agencies are changing their regulations and they need to have public input, that sometimes they get thousands of members of the public that will have a comment about a proposed regulation. And they have to, summarize that data. They have to sometimes reply to each of the people who has or offers input. And again, we have a tool that helps them in terms of, getting that, that work done. And then of course, we'll talk more about this. I suspect the heart and soul of public health is often, surveillance or epidemiology that is tracking diseases and AI is increasingly becoming helpful in terms of not just capturing the most accurate information about, diseases, and where they're occurring, but also subpopulations. And it offers increasingly the possibility of forecasting in a, in a pretty accurate way what might be coming. So we don't just know what has occurred, but we are able to estimate what will occur under certain circumstances. AI is, a valuable tool in terms of, making that kind of forecasting effort. That's awesome. And just to go a little bit deeper here internationally, there's got to be a lot of talk about how you will how the CDC is working with as an example of an agency working with agencies that are similar overseas or internationally. Is AI being used across the board? How is that being managed? Well, we you know, one specific example is we've been working on a pilot project for a team that is providing guidance to public health departments in developing nations. And like John was just saying, one of their biggest needs is the ability to do very rapid literature review. To make good, qualified decisions with the best in class research, on given topics. And in many cases, they are trying to contextualize that research as much as possible to their region globally. And so we're creating a tool on behalf of the CDC that will potentially help those people quickly search through the documentation to find the best papers. That align with a natural language question instead of having to come up with some kind of, you know, really complicated, librarian search string. For many people in these public health departments, English isn't their first language. And so we want to give them the tools to just ask a simple question, get a wealth of information back, and then using AI as well to narrow those things down to their unique contexts and their unique scope. So that's one example of a way that we're, you know, working and collaborating. Across, you know, international agencies. But maybe I'll mention just one other quick one, which I think builds on what Eddie was saying. And that is AI offers the ability to translate materials into other languages very quickly, and you can even customize the way that you want to communicate so that you can, factor in literacy levels or weight, whether you want to have pictures instead of words. A lot of that kind of translation and communication interpretation, is very resource intensive historically. You need to have a lot of people on staff, or you need to have contracts with experts. You there always needs to be a person in the in the loop that someone always has to check that that translation is accurate. But the ability to translate into multiple languages and customize communication tools to specific subpopulations is much, much easier with AI. It's very it's a. One other little tidbit is, as we work with ICF, Europe and Asia team, they're involved in AI as well, and they're very well versed in that. And one of the benefits of being connected internationally is we're all holding each other to the highest possible standard of safety ethics. And in many ways, Europe leads the charge on that. Some of their safety and stability laws are very restrictive. And so finding ways to partner and, allowing ourselves to think through how would this be implemented in a place like the European Union? It just it makes everyone safer, makes it a better, outcome. I love it so and that was my follow on question to you all is so there's so much promise for AI across the board, both here in our country and internationally. What are some of the obstacles to that might impede such progress, and how are they overcoming those obstacles? And, you know, everybody's risk averse, right? Everybody's there. Everybody wants to make sure that there's guardrails so forth and so on. So could you tell us a little bit more about that? Well, I'll mention 1 or 2 and then and then turn things over to Eddie. One of them goes back to what I was mentioning earlier that we don't have a uniform public health system. We have a very uneven public health system. People who live, for instance, in well resourced large cities get a different level of public health services than people who live in rural areas where the resources are limited and things state by state. There are differences as well. And so when you have potentially valuable new tools in AI, you we really want to avoid the possibility that will reinforce, a two class system. Some will have it and some won't. The public deserves to have a uniform, high quality public health system in place. And so, a concern is, the importance of making AI available for all. And that partly to answer your question about how do we grapple with that? It means recognizing it and acknowledging that different kinds of technical assistance and support from federal agencies and others is necessary to ensure that the lower resourced areas or those that have, some obstacles in terms of their, connection to the internet or having the latest equipment, gets support so that they benefit as well. Yeah, I think that's absolutely true. And it I think on a broad scale, the, the things that make AI, that we should be cautious about as we, you know, embrace and integrate AI into everything that we do. There's two big ones that come to mind. The first is, just the security aspect of using this AI model. And some of that comes from potential biases and models that we need to be aware of. We need to work against. We need to assume that models have bias and work to mitigate that. But some of it also comes because these these models currently are so large and so complex that they're have to they have to run on other people's machinery. And so we're all sharing time on the mainframe of another person's, infrastructure in order to use these. There's definitely secure ways to do that. There are ways to kind of fence that off. But that's that's going to be a concern for some people, especially in public health. If you have to, you know, deal with AI and, you know, data that could be very dangerous if it were to be released. The other, aspect to think about with this is sustainability. With AI, a lot of these AI models are very energy hungry, and, they require a lot of electricity. And so obviously there are issues with that. And there is even financial issues with that. And there's a question as to whether or not that's sustainable in the long term. But I think with both of those challenges, I'm seeing a trend in this industry that's not very publicized, but it is happening where these models are actually, getting better at being smaller and still being very, very powerful. And so as that trend continues within the next few years, my hope and what I, what I want to see happen in this industry is that the models become small enough to be used locally on your own device, on your own phone, on your own computer. And that helps answer both of those questions. It help to answer the security question and the sustainability. So there are definitely pitfalls we need to be concerned with. But I'm excited about the promise of what's coming. And I think the more we can, you know, lean in that direction and encourage it in the industry, the better will all be. That's really great, John. Anything. I'll stay out there. I think that we have also seen in certain parts of the country, a bit of, politicization of AI. There's some locations where, some states that have thought, I, threatens people's privacy. Maybe, maybe medical records will be too accessible to the public. And so there are locations around the country where, elected officials have said to the health departments, you can't do it. You can't use AI. And in those instances, I think the the key is building trust by demonstrating that you can, in fact, ensure that the AI systems that you're using have, guardrails that prevent any possibility of or take the extraordinary steps to take them, to prevent the possibility that, that inappropriate leakage of, say, personal information would ever occur. There are definitely ways to do that. That's one of the things that ICF works with federal agencies and others on to, to ensure that they're built into any work that's going on, because over time, I think the more that it becomes clear that those safeguards are possible, it will, address the concerns that people have raised that have led to, some of the, the policies, against the use of AI. That's excellent. Gary, I have one more question for you. So I hear from so many organizations. Well, we know we need AI. We know we need help. How do we get started? Yeah, that's a fantastic question. And we hear it all the time. Because I think all of us are in this boat. Part of it is, you know, this is a unique moment where a new technology has blown up in the public domain in a way that others haven't. So, you know, when when Microsoft Excel first came around, there weren't conversations in Congress about it. But all of a sudden, this technology has come around and everyone's talking about it. And so it does put leaders in an interesting squeeze where they're being squeezed from the top down, saying, we need to use this technology. They're also being squeezed from the bottom up with fears of, are you replacing my job? Are you, you know, outsourcing, you know, the critical things that I can do. And so we try to partner as much as we can with leaders in those positions to help them walk through the process of very small, safe, approach steps to using AI. And so for someone who, you know, isn't working with us or they're just, you know, on their own, maybe a small budget, it's first of all, making sure that you have worked through what the guidance is in your local health department or your agency to make sure you're not violating any policies. That's really important. But secondly, it's, you know, leveraging tools like ChatGPT, like, you know, llama models, these things you can find for free in the public, leveraging them in ways that make you, that open up time in your day to do what only you can do. And many times that's really simple, mundane things. So it's, you know, using generative AI to help draft an email. It's using generative AI to help analyze a document that you typically would take two hours to read. And you can read it in ten minutes. They seem like very trivial use cases. And again, I would encourage, if you're using public tools, only use them with publicly available data. Don't use them with anything private or anything. You know, that's proprietary. But if there's a publicly available document and you need to read it, try it, try it with generative AI and see what the results are. Because what we've found is it's not a tool for replacing jobs. That's not where it's most powerful used, especially in public health. What's powerful about it is it frees up your time from all the monotonous and mundane, so that you can do the things that only you can do for public health. And that's personal conversations and deep thinking. And, you know, using that meta analysis that you can do with your life skills and experience, that an AI model just will never really be able to do. That's excellent. Thank you so much, Edie. So, John, in January, we'll have a new presidential administration, new party and the majority in the Senate. Is it likely we will see a change in the policies or practices with these political changes? You know, of course, time will tell. But my sense now is that there will continue to be support for data modernization and the incorporation of AI into the public health sector. Because, there's a recognition by both political parties that this is essential. We need in order to protect the public. We need the most accurate information we can gather. We need it to be timely, and we need to be have it be granular, broken down into, subpopulations, for example. And that's a recognition, I think, that is very much appreciated, by the incoming administration as well as the current administration. You know what what I think we're going to, need to pay attention to is will there be the necessary amount of funding to actually ensure that all of the changes that are necessary can take place not just at the federal agencies, but also at the state local level as well? Unfortunately, AI and data modernization is not a one time purchase. You don't just buy the hardware and then you've got it. The main, expense in this work is people. You have to have people who are, good at gathering the data, good at inputting the data, good at analyzing it, designing systems, updating them, and then helping in terms of their practical applications. So, so investing in the expertise of people on an ongoing basis, does cost money and so that there will, there will certainly be discussions about what's the proper a funding amount for the federal level as well as the levels at state and local communities. Excellent. Eddie, your thoughts around that? Yeah, I think it's a really great point. And I think, you know, one of the biggest strength of what I can bring, especially recent developments in AI and generative AI, is if it's leveraged the right way, it can be leveraged in a way that's incredibly transparent, where you can help demonstrate exactly why decisions were made and why, you know, specific analyzes and outcomes were advocated for. It's becoming less and less of a black box and when you think about a world where, you know, we now know the next four years, but we don't know eight years, we don't know 12 years, there's a possibility that things may shift back and forth. And so in a landscape where things are very evenly divided and there's a lot of opportunity for contention on both sides, having tools that can provide that transparency. And that can always provide documentation to say, here's what we did and here's how how that works. I don't think that's going away anytime soon, no matter who, is kind of setting the mandates. Thank you for sharing that. Or thank you both for sharing your your perspectives. So finally, if you were offering any final thoughts about how to approach the use of AI in the public health sector, what would you offer our listeners? I would say, you always need to have people in the loop. You always need to ensure that there's a lot of attention being paid to, the data that's coming in, the data that's being analyzed, the reports that come out, AI is not foolproof. It is a tool. And what has to be done is you've got to have the right people checking it, making sure it's accurate, questioning whether or not, interpretations, need to be reconsidered. So keeping a person in the loop at all times is critically important. And I think. That's that's huge. I mean, that's definitely the most important thing is to keep that person, that human centered approach, is massive. I think the other thing I would encourage is to lean towards, an openness in understanding and incorporating AI rather than a restrictive ness. And so that means a few different things for different departments and agencies. But one of the biggest is just education, just helping people understand what I can do and what it can't do. A lot of the times we find that when we get a lot of pushback, it's because people don't really understand what it's doing and they're, you know, imagining all kinds of different scenarios that just aren't realistic. So finding ways to let people understand what it can do, what it's really great at and what it's not really great at, hopefully will continue to open those doors and make it a more valuable experience for people. That's great. Well, thank you both so much for joining us. This was such a fantastic discussion, and we're looking forward to working with you and learning more about what we're seeing over the next. It seems like everything's moving so fast. Six months, 12 months, 18 months, just to see how everything accelerates and the work that you all are doing. So thank you so much for your leadership. And thank you so much for your time. Thank you Susan. That's great. Thank you very much. This concludes today's episode of the 33rd podcast. If you enjoyed this episode, please don't forget to subscribe, rate and leave a review. Your feedback helps us continue bringing you thought provoking conversations with the brightest minds in government, technology. Stay tuned for our next episode, where we will continue to explore opportunities to harness the power of technology and explore what's next. Developing a more innovative and efficient government. Until then, this is the Fed to Fed podcast by GovTech Connects. Thank you for joining us.