Episode 33 Part 1: PPIRA In Action: Turning Data Into an Ally for Building Trust in the Fight Against the Opioid Crisis === [00:00:04] Michael Donovan: Welcome to the Evidence to Impact podcast, the podcast that brings together academics and their research partners to talk about insights and real-world policy solutions in Pennsylvania and beyond. I'm Michael Donovan, the associate director of the Evidence to Impact Collaborative here at Penn State. This is the first part of our two-part episode where we'll be discussing the PPIRA in action, turning data into an ally for building trust in the fight against the opioid crisis. My guests today are Dr. Nonna Sorokina, associate professor of finance at Penn State Scranton, Dr. Dusan Ramljak, assistant professor of software engineering at the School of Graduate Professional Studies at Penn State Great Valley, and two excellent student researchers, Tanuja Voruganti and Mahathir Mahbub. The Presidential Public Impact Research Award is a partnership between the office of President - Bendapudi at Penn State, the office of the senior vice president for research, and the Evidence to Impact Collaborative. The initiative brings research, student learning, and community expertise together to address challenges that - shape people's lives in communities they live. It builds on a national effort to use evidence more effectively in public decision-making, informed by engagement with the White House Office of Science and Technology Policy and the Office of Management and Budget, as well as our partners at The Pew Charitable Trusts and the broader research funding community. This work broadly is often called public impact research, using the framework advanced by the Association of Public and Land-Grant Universities. Penn State President Bendapudi co-founded the President and Chancellor's Council on Public Impact Research to better organize the whole sector around this mission. Across Pennsylvania, Penn State faculty and students are partnering with community organizations on locally identified priorities, from health access and food security to public safety, education and responsible- technology use. PPIRA supports research that combines Penn State expertise with the knowledge and experience of community partners. Our regular listeners will recall a recent episode discussing another important PPIRA team focused on microplastics in the Delaware River Watershed, and today we're gonna go into another really vital area of work that I'm excited to chat on today. Nona, would you be able to give us just a brief introduction to yourself, overview on your role? [00:02:27] Nonna Sorokina : Yes, I am an associate professor of business with concentration in finance and quantitative methods at Penn State Scranton. I'm a dual career professional, if you will. I had about 25 years of experience in financial services industry, some of that in overlap with my academic career, and my academic career is more than 15 years by now. I am an avid researcher, and I am focused on topics of social and economic sustainability. I'm very happy to be here. [00:03:02] Michael Donovan: Thank you so much for being here. Welcome. Dusan, how about yourself? [00:03:09] Dusan Ramljak: As for me, I'm a assistant professor of software engineering at Penn State Great Valley. And I do have multiple experiences as as Nona as well. I've been, I am a retired professional dancer and martial arts teacher. And I have been engaged in the leadership back in Serbia. Had eight years of experience as a leader, and we worked on substance use disorder prevention as one important part of the leadership. And my undergraduate degree was in electrical engineering with a focus on systems control, which then led to signal processing, digital communication, parallelization, optimization, finally to computer and information sciences. And my work my career in science or after undergraduate studies was in system administration supporting businesses, supporting research. And that role in this project I see as a support, data-driven decision-making support, artificial intelligence support [00:04:34] Michael Donovan: Wonderful. Welcome. Thank you. Tanuja, how about yourself? [00:04:39] Tanuja Voruganti: . Hi, everyone. I'm Tanuja Voruganti. I was a student researcher at Penn State Great Valley, where I did my Master of Science in Data Analytics. I worked on the PPRA project with Professor Dusan and Professor Sorokina, and I also had a chance to work with Professor Sorokina for a part of it. I'm really excited to be here and share a bit about that amazing experience. [00:05:05] Michael Donovan: Excellent. Welcome. And last but not least, Mahathir. [00:05:09] Mahathir Mahbub: Hello everybody. My name is Mahathir Mahbub. I was a two plus two major. I started off at Penn State Scranton, and finished my last two years at State College. I double majored in finance and IT. I personally believe that I it was me who asked Professor Sorokina, because I went up and searched her name on the Penn State directory, and I saw that she was an avid researcher, and I really wanted to be a part of it, and I really loved Professor Sorokina's class. I took her class and then got to know her a lot more, and I was a part of this amazing team. I was under Ahmed Suleiman, a PhD student and I was there vetting a lot of the research paper. I was working with Professor Sorokina for a bigger chunk of the timeline of the project, and I personally feel like that I have known so much even being an undergraduate student, because of how much research journals I would go out of my way to read for, about the, the project. This ultimately helped me a lot more when it came to writing the initial draft, and I personally feel like that I have had an impact on that as well [00:06:10] Michael Donovan: Wonderful. Excellent. It's so great to have student representation from the both graduate and - undergraduate experience here. This is really rich. Dusan? [00:06:20] Dusan Ramljak: As Mahathir introduced one name that's Ahmed Soleiman. Another name was Mark Fahim. They worked on the PPIRI project directly. They were awardees the student awardees as well Mahathir and Tenuja. [00:06:40] Michael Donovan: Wonderful. Wonderful. Let's dig into it a little bit further here. I'd love to open it up about, telling our audience a little bit more about the project and really some of the, the research aims in trying to understand dimensions of the opioid crisis here. I'd open it up to Nona or Dusan. [00:06:58] Dusan Ramljak: I would like to continue with what I said, that I played a role of artificial intelligence and data-driven decision-making support and really understanding how we can trust the decisions that we make interventions that we make and everything else really I would like Nona to speak about as we were that support and trust while Nona was leading on understanding the economic impact and how to alleviate it. [00:07:35] Nonna Sorokina : I appreciate the significance that you assign to my role, Dusan. And I am at least I can't speak too much for that, but I can say that there's a history in my involvement in the topic, and this history dates back to Probably roughly 2017. That's been long time ago, and that's been when I visited Temple University for a seminar, and I saw a paper on the impact of opioid epidemic on labor force and activities of the companies who struggled with labor in the areas car- hit hard. And it was really striking how far opioid epidemic reached with its impact. At that time, I was not aware that it wasn't isolated to certain areas. The perception generally, it's an urban pro- problem, but it's not. I've learned about areas like Wilmington, North Carolina, for example, which I could never imagine being an epicenter of epidemic, and it was hit hard. So at that time, I met with my colleagues from Federal Reserve Bank of Philadelphia, Van Lee and Raluca Roman, and I told them about what I learned, and I said, "You know what? There must be other impacts." And that's when we started looking into impact of opioid epidemic on financial services system. And we have developed a paper that was presented probably at least a dozen times by now in many venues. Got lot of attention. It is under review at the journal now. That's a paper on opioid epidemic and consumer credit supply. That's how our research with Fed colleagues evolved. We also have a publication at Federal Reserve Bank of Philadelphia. They publish economic research, and the article came out in 2022. It was called Economic Impact of the Opioid Epidemic. We highlight how deep the impact is and how it affects diverse categories of population. Those from underrepresented backgrounds are affected disproportionately, we could see. We tap into how different legislative pieces affected the dynamics around the opioid epidemic. So there was really a long stream of research, and when it came to application for this program, Presidential Public Impact, this topic immediately came to my mind. One of the issues in our previous research- that was really prominent was an issue of identification as we did our statistical tests, and that's where I thought that more of support with sophisticated data processing, new data sets, and different methodology from data analytics perspective might help. And that's how we partners with Dusan, and that's where his area plugs in. We will speak later about what we developed within this project, but that's pretty much a background [00:11:00] Michael Donovan: Excellent. So a really organic and longstanding relationship with what became your community partner with the program the Federal Reserve Bank of Philadelphia. We'll talk about their specific contributions in terms of data and other pieces there. But anything more, Dushun? [00:11:17] Dusan Ramljak: I would like to address that history behind the involvement, what led us to focus on on this. And if that's okay, I would like to go a long time back, so not to 2017, but in the '50s when my grandfather formed, founded, and he was the first president of the association that dealt with prevention and treatment of opioid and alcohol addiction. So moving forward to that '81 when I started my engagement in leadership and when we learned about opioid addiction and how to really prevent that through that community engagement. And my grandfather would would add always that like he would like me to support that any way I can. And again, moving forward to throughout my involvement in martial arts and dancing, there was also involvement in really explaining the bad things that can happen with the addiction as I had, fortunately or unfortunately, the prime view into the center of addicts. My window was looking at at that where the friends gathered and used and abused the drugs. And I have seen people from like you and I happy in life, all of the sudden going and becoming the shells of themselves. And coming now back to future 2021 I started working at Penn State Great Valley. Immediately was work- explaining my work in support of research and company, Ignited, who has all the data for treatment, addiction treatment. They have all the data online. They asked, "Would you like to collaborate with us," right? "And support us in our quest to understand the engagement and satisfaction with the treatment, and does that lead to outcomes?". So that's how I started working on it here at at Penn State Great Valley with students who first worked with me and then interned at Ignited, and then we continued leveraging d- that data, improving the machine learning, then reinforcement learning. And and then when we just couldn't really help anything with with that how to quantify that help, how to quantify that trust that has been established Nonna. And the economic impact that we could really address came and that was what led us to use data and technology to study this issue [00:15:01] Michael Donovan: Excellent. A rich personal experience really drew you and ignited the initial i- interest in addressing this challenge. But then also the understanding that different disciplines can provide different lens to look at for this challenge. Nonna, did you wanna come in here? [00:15:20] Nonna Sorokina : I just want to add, and you mentioned that different disciplines. This is true interdisciplinary research, and it interdisciplinary became a buzzword recently b- but it does get to add value when applied correctly. It is amazing how much we can extract leveraging expertise from different fields. In our case, economists generally live in their own bubble, if you will, and they look only at economic impacts, and they struggle with their own struggles. But at the same time, this is a health crisis, and when we don't talk to those who are experienced in health issues, we really miss out. And then there's a lot of data involved, and if we only focus on the specific methods that are customary for economics or finance or subfields of economics we really miss out when we don't talk to data analytics professionals who may have very different view on how you can process and visualize and highlight the information. And then finally, and most importantly that's that speaks to the presidential initiative itself. Once we know something about the process, what do we do about it? How do we make sure that we provide knowledge back to the community and action is taken? So in, in that sense, we were able to tap, to some degree, in every area with this project, and it was very important. [00:17:08] Michael Donovan: An almost , a kind of a pervasive all of society challenge, the opioid epidemic really requires a inclusive approach understanding that many different disciplines and perspectives can contribute. And that's also really valuable , for students who are part of this as well, to understand how this can be synthesized into, maybe their future opportunities as well. Dusan? [00:17:31] Dusan Ramljak: Yeah. I wanted to mention like how we can then give back to the community. 15 years ago now more or less the Start Talking Science initiative has been formed at Drexel. Kristi Love also a PhD and a professor there. And we gathered around the idea how to help the researchers speak to the layman audience. And that is also an important part of how we get how we get to give back, right? And our students have presented at Start Talking Science event, and we'll speak more about that through how did the project develop through the program, as it, it really helped us to give back and in various ways. [00:18:29] Michael Donovan: Great. And I am very keen to include our student partners here as well. But before we go to them and hear their experience, I'd love to just get a little more granular on some of the types of health and social and economic factors that you were really examining , through this project. That would be really helpful, I think, for our audience to know what the type of data you were pulling from and the various sources, and also the complexity of working with such data how that was solved, and lastly, the challenge of working across multiple campuses. You had students across different campuses as well, so any insights on that would be lovely. [00:19:06] Nonna Sorokina : As a result of the project, we had several presentations at various regional and pre-interdisciplinary events. And we developed two research papers that are in under review stage. One of them was addiction as a banking risk factor, where we looked at various areas of exposure for banks, and we measured several factors looking at different aspects of bank risk in search for specific impact. And the other paper on the paper we partnered with Se-Jong Park from , Penn State Harrisburg. She has a public health focus, and she looked at change in OxyContin formulation as it impacted deposits of banks. So that's quite an unexpected twist as you can see. It was even more specific not simply an exposure to overall epidemic addiction, but a s- very specific event in the timeline of epidemic. And it turns out that it did impact small deposits of banks. And this is the story to tell because small deposits is really the population , that may not be as economically established who was affected the most by the opioid epidemic twists and turns. What we used as a data came from CDC. CDC has a database of opioid-related death, and they have a database of opioid related prescriptions. If you know about the history of the epidemic, it started with prescription epidemic when opioids were improperly prescribed and over-prescribed. Since then epidemic evolved, prescription crisis was tackled quite some time ago, but prescription data still tells us about where the epicenters are they remain a lot of times, even when you talk about illicit drugs. So we used exposure, opioid exposure data from CDC, and we used publicly available data on banking. Banking area is highly regulated, and it's good for us researchers. Banks have to submit not only their major financials quarterly, they have to submit really pretty much any metrics related to their activities quarterly, and it's all publicly available. So there's a lot of room for studies. So we tapped into that call reports database that is made available to us by the regulators. So what we found that risk in banking increases as well as deposits that are made by smaller depositors are affected when opioid epidemic hit. [00:22:20] Michael Donovan: Thank you. And I wonder if you could just touch on the multi-campus approach because it's such a rich opportunity to bring in both multiple disciplines, but also multiple campuses of students and investigators, and that adds complexity to the project, right? In terms of how to organize the work. Dusan, did you want to note on that? [00:22:41] Dusan Ramljak: Yeah, and I would extend a little bit that note to first recognize more names on associated with this project and introduce what they meant. Tayo Basani he worked on establishing what data really we need and , what are the cybersecurity implications and privacy implications when using the data for this. Then another name Janil Kaka who essentially made the introduction between us and Abhay Haridas, who established that first first machine learning development and then reinforcement learning at IgnitEd. So they're all from Penn State Great Valley. We need to first establish Penn State Great Valley as a school of graduate professionals. So those are not your average Joe students. Those are the students with experience, Like three years up to 30 years and more experience, and they come to Penn State Great Valley to sharpen their skills in a particular area, right? Emerging technologies or, So we have programs in artificial intelligence and data analytics systems engineering, software engineering management and then on the management side as well. And all of that adds the complexity and the students that I mentioned were from all the mentioned programs. And they-- then we also included the students who are from World Campus as well. The undergraduate students like Mahathir and, We did incorporate the projects, the topics for students. Direct results are those two papers that Nona mentioned. Indirect results the whole team of students has been working on establishing the promise of artificial intelligence in substance use disorder. So understanding the data that comes in and what the research has been doing on, on, on the data itself. What artificial intelligence models frameworks were applied and how, and in this aspect, what is the future of applying that? Understanding what publications, what research disseminations were made. All of that were students really working on. And then again, remember, those are not your average Joe students. Those are CEOs, CFOs, CIOs of the companies the business owners workers everywhere, and hundreds of those students were introduced to this project. At around hundred students were involved in this promise of AI in, in substance use disorder. So more than just, research paper model or this, it is that information about this is the problem and there is a solution And then that information is provided to the leaders. That's at least 100 leaders that can spread that information further. And companies like Vanguard, Comcast ex-- and then others that we have students from. So that, that's that value and the complexity of it's not just campuses, it's skills, it is levels companies. So all of that, I think it is worth mentioning, and we are grateful to you for facilitating this great opportunity. [00:27:20] Michael Donovan: Thank you, Dusan. That's a wonderful segue to hear a little bit from our student partners themselves not just from their experiences, but representing the larger core of students that were exposed to this project. Before we do that, Nonna. [00:27:36] Nonna Sorokina : I'll try to be brief, but I absolutely have to add to this part. And I would say that impact on students' experiences is one of the major achievements in this project. For me I'm less global than Dusan. Dusan you should see his research meetings. He has dozens of students there. He really spreads the word out. He gets everyone involved, even those who are not directly on the project and this is great. I run it slightly different. I have a lab. It's called Sustainable Communities Lab. And in that lab, it's a fairly small group of students up to, probably up to eight different times, depending on our funding. But what's interesting that the level of students is very different and ranges from undergraduate to postdoc level. And so you have undergraduates, you have master's students with a practical focus. You have PhD students who are very theoretical, right? And we had postdocs a couple of times. And they all work together, and that really brings a perspective. On top of that, the way I run the lab comes from my industry experience. So I really run it as a research team at a commercial organization rather than academic. And although I naturally mention research products as our output, the way we get there provides practical training. I see Tanuja is nodding there. Yeah, so there's a lot of expectations of independency, of setting goals, meeting expectations, meeting deadlines communication on different level. Students learn a lot [00:29:21] Michael Donovan: Wonderful. Wonderful. So lots of complexity as well about how you engage different students in different ways at different levels. Dusan? [00:29:30] Dusan Ramljak: Yeah, I want to follow up on that. So we do after this experience with PPI RA we have established meetings every Monday through Thursday when the students are in classes. And those are regular meetings at one thirty every, as I said, Monday through Thursday. And I would like to also mention that the professors here at Penn State Great Valley were involved and are involved in those meetings. Professor Satish Srinivasan and then Natalia Nascimento and Professor Everton Guimaraes. So they are actively involved supporting us. Also Professor Robin Q. and Sally Richmond, always our director of engineering programs, Raghu Sangwan, and our dean Colin Neal. So those are all the names. And I wouldn't like to miss anyone but everyone was so supportive and all the teams and staff were really always supportive and always willing to help us. And yes my meetings are always with a, a lot of people. And that is alumni they come back. So they they come back and volunteer their time to help the students that are currently there. M-my colleagues from my PhD then back from Serbia, Austria they're all coming to those meetings. [00:31:10] Michael Donovan: That's a wrap for part one of our two-part episode on the PPIRA in action, turning data into an ally for building trust in the fight against the opioid crisis. My guests today were Dr. Nonna Sorokina, associate professor of finance at Penn State Scranton, Dr. Dusan Ramljak, assistant professor of software engineering at the School of Graduate Professional Studies at Penn State Great Valley, and our two student researchers, Tanuja Voruganti and Mahathir Mahbub. Thank you so much for everything today, and please join us for part two, which will be coming out soon.