Insights on the Health Data Science concentration at UIC
Dr. Miriam Isola, Dr. Jacob Krive and Dr. Andrew Boyd share their insights on the Health Data Science concentration at the University of Illinois at Chicago.
This video will explore career opportunities in Health Data Science and how to prepare for those opportunities.
Dr. Miriam Isola: Welcome to this advising video for Health Informatics and Health Data Science students. Some of you may be considering pursuing the health data science concentration. Others of you may be nearing the end of your program here and looking ahead to graduation. So we wanted to provide some information, some advising. We get a lot of questions from students on these two things. And so that’s the purpose of this advising video, to get the faculty who are here today to give you some of their insights into these areas.
Dr. Miriam Isola: So today we have Dr. Jacob Krive, Dr. Andrew Boyd, and I am Dr. Miriam Isolaa. We’re all on the faculty in the Health Informatics Program and teach the Health Data Science courses. So to kick this off, the first question we thought we’d talk about today is why pursue the health data science concentration? We’ve had a lot of discussions about this in the past and some of our students come to us with an idea of why they might want to do it, but they still need more information about why they might do it.
Dr. Jacob Krive: This is a topic of interest that is very popular in the job market. Obviously healthcare, in terms of quality, in terms of some of the developments in genetics and science and medicine, really are increasingly relying more and more on analytics and the organizations in the healthcare domain are becoming more data driven to the point that data is really at the roots of the cultures of the most progressive organizations. What we typically get from students is that they look for jobs either when they consider HDS concentration or they’re just thinking about their careers, they often think of what will I do for a hospital? And that’s incorrect or more of a limited thinking because more and more analytics opens doors for others and old domains of care. And it used to be that a pharmaceutical world was a world of its own, very different applications.
Dr. Jacob Krive: It was difficult to be hired for someone who worked for a hospital for awhile by a pharmaceutical company and vice versa because those were different roles. Now, healthcare becomes more integrated. Data can be exchange and pharma teams are forming clinical informatics groups, data sciences everywhere. We’re talking often about the same types of data crossing multiple domains of care. It’s a very typical mistake that students make when they talk to me looking for information on how do I get a job at a hospital or “Oh my gosh, I live in the area where there’s only three hospitals and none of them are hiring today. What should I do?”
Dr. Jacob Krive: We don’t target only hospital or provider care. It’s a very big important area, but it’s not the only area and it’s not the only opportunity to get a job. Students should look into consulting into high tech rolled, into pharmaceuticals. Clinical pharmacists are getting involved with pharmacy informatics. There are all kinds of vendors looking and if you attend one of the bigger conferences such as HUMS or AMIA, you see huge representation from technology vendors and consultancy firms who are looking for graduates of the programs and/or experienced professionals from clinical informatics, specifically analytics and specifically those with analytics and healthcare data science skills.
Dr. Andrew Boyd: And besides those, some of the other smaller companies of the startups in the Chicago area, we have over 230 health technology startups and throughout the country, both in Boston and San Francisco and along the areas. Many of those individual firms are looking for people with health data science background to help drive the value, to help explain the value. So even some of the companies that don’t exist yet are looking for the students who don’t have a public presence. And so the neat and exciting part of this field is besides a hospital, you can look for individual companies or areas that you can grow into the position. You can help write your own job description, which can be scary for students at times.
Dr. Andrew Boyd: As they’re looking for individuals, this job or area of health data science has been around a short period of time so there isn’t a straight certification. There isn’t just a common job description and so, as you look across the industries, trying to figure out what they’re looking for and how your skill sets can be applied is one of those challenges.
Dr. Miriam Isola: Let’s talk about the… We’ve talked amongst ourselves, I know, about health data science versus just data science. There’s many industries that are getting into data science these days. So our students are particularly wondering what the health data science piece of it is and what they can bring. What’s so different about the health data science part of it?
Dr. Andrew Boyd: One of the big things, the difference, between health data science and just straight data science, even health analytics is what is the definitions of the terms you’re actually using. From a programming point of view, when you talk to most computer scientists, you talk about how many beds are in the hospital. They’re expecting a single number. The number of beds in the hospital isn’t static. You can have beds in the hallway, you can have beds in the emergency room, you can have beds that come and go.
Dr. Andrew Boyd: So even something as simple that we all think of, “Well, how many beds are in our house?” And you can answer easily-
Dr. Miriam Isola: It’s pretty simple.
Dr. Andrew Boyd: It’s pretty simple. But even if you don’t understand the nuances of how we measure disease, what the concepts of disease are, just knowing a programming language isn’t going to help you provide the value. Applying a neural net to data without being able to explain what the value is and what the limitations of your method are is critical, especially in healthcare.
Dr. Miriam Isola: Yeah. You’ve talked about that a lot, I know Jacob, in your classes-
Dr. Jacob Krive: Yeah. And students often reach out to me, especially before courses begin. Which courses do I choose so that I learn our programming language or Python or a question of what are we learning? Are we learning Python? And I want to make sure that I learn Python. Well, it’s it’s very good and there are a number of opportunities at UIC, or outside of UIC, including online sources to just learn the code, including visiting your favorite bookstore and purchasing an inexpensive book. But that will not teach you what we focus on the most, which is the domain expertise.
Dr. Jacob Krive: Being that kind of an analyst who can go between those who are highly trained in programming and mathematics aspects and those who are highly trained in medicine, public health nursing, to help them identify opportunities for analytics, for quality improvement and translate data needs into medicine and medical needs into data, back and forth.
Dr. Jacob Krive: And with the changes in the industry such as automate the automation or artificialize the artificial intelligence, if I can invent the term. There are more and more companies, well known companies, out there who are creating the tools that will help address the coding issues, such as almost like hire your automated coder. But that automated coder will never be able to understand your domain. That that skill, not only will remain viable, but the importance of that skill will continue to grow.
Dr. Jacob Krive: And this is where we see the main bread and butter for our program, is prepare those people who can function in the domain of health care. There are not generalist coders or data scientists. There are other programs at the university who can teach you that without the healthcare component. But if you’re interested in health care, this is where we think we need to prepare you to be an effective analysts leader who can partner with others at a healthcare oriented organization to really drive the value of analytics in this domain.
Dr. Miriam Isola: So let’s build on that point a little bit further and take it into, it starts to lead us into our approach to the learning. So that’s another question that a lot of students ask. Some of them have been in some of the other courses in the program already and they start to wonder, “Well, am I going to be able to handle the statistics or am I going to be able to do the programming? What kind of assignments are in these courses?”
Dr. Miriam Isola: And so we’ve very specifically developed the health data science courses with a particular approach to learning that is applied. That is active learning. So it is not just theoretical and so, it ends up involving a combination of knowledge, skills and building your concept of what a health informatician actually is over time.
Dr. Jacob Krive: Right. And we will actually put you to work every week with hands on assignments and those assignments are typically either use cases that are real, taken out of most recent publications, most recent professional field experiences or, in some cases, we may ask you to develop your own data science program, program not being a tactical program where you write code, but your actual business and clinical program that will address the growing needs of an existing field or perhaps even work with an emerging field, such as cardiac oncology or precision medicine. Come and develop your own domain where nothing out there exists. You cannot go and just find on Google how to do it. It will be up to you.
Dr. Jacob Krive: And our approach, to many of this assignments, is leaving a lot of blank spaces. The instructions are purposely vague. Yes, I can tell you ABCD or one through 10 how to build this program, but that will be limiting the array of possibilities because there is more than one answer to the same question. There are more ways of going about it and there is really not a yes or no or correct or incorrect answer because we don’t really know what’s the right way to do this.
Dr. Jacob Krive: It’s all new, it’s all cutting edge and it actually makes it fun for us faculty on the other end to even grade these assignments because not a single assignment is a lookalike. So we’re not really grading and you’re saying multiple choice or correct or incorrect answer. Everybody comes up with their own unique approaches. We may comment what may have gone wrong or what pitfalls may be on the way in the feedback, but it’s not about did you provide a correct mathematical answer to this type of a problem or not.
Dr. Andrew Boyd: And talking about the correct mathematical answer, one of the challenges of the field and one of the challenges you’re getting at in the coursework is just getting a numeric answer isn’t sufficient for healthcare organization to make the informed decision, to help drive the innovation, to help make better outcomes. And so building a program, figuring out, we can measure this, we can identify this, but how do you put it in the hospital? Does a clinical decision support where you’re helping the physicians and nurses and pharmacists? Are you engaging the patients? Are you engaging the team? We have all these powerful tools, but how does it, for your organization, or how do you make it generalizable across lots of organizations? The mathematics will be solid, but how you actually drive the innovation and drive the change in health data science across the number of different industries you can be working in is one of the challenges that we see the students with the health data science background being able to address.
Dr. Jacob Krive: This is where programming and mathematics are insufficient. And in fact, some of the publications, such as Harvard Business Review, have identified the main reasons behind failures of a health scare or any data science program, is inability to tell the story with the data. That storytelling ability is something that we stress. We ask students to present on their projects or create at least PowerPoints based on some of the technical work that they do because it’s inability to understand the environment and translate the needs into tangible outcomes, meaningful outcomes, relevant outcomes. Or sometimes the outcomes may be relevant, but there is nobody on the team who can translate it from the language of data science to somebody who is not a data scientist. And most of the people in the room will not be data scientists.
Dr. Jacob Krive: So, you may have a really smart team and nobody understands the value of that team. So for all intensive purposes, they’re labeled as useless because they’re not relevant to the organization.
Dr. Miriam Isola: So we, in the end, health data science helps them to be critical thinkers. The students will be using all health care examples. They’re using healthcare data to look at these problems. So it’s a combination of some of these business skills that we’re talking about and being able to present relevant information, useful information, for healthcare executives to be able to make decisions. And so the business side and the quantitative side.
Dr. Miriam Isola: So there may be an expert data scientist in the room that’s got the big skills on quantitative, but you’ll be able to work on a team with someone like that. You’ll be able to participate on a team with someone like that and be able to take it to the people who are making decisions and rolling out programs and solutions.
Dr. Jacob Krive: Yeah, and in fact we do routinely have people who are technically strong, mathematically strong in the classes and they frequently comment on how much value they get from applying it to the domains and use cases they never thought about. So maybe executing the technical part is easy for them. But application of that, to the specific use case and telling a story with that data and presenting that data in the format consumable outside of their immediate domain of expertise, is a revelation for them and wow, they’ve never thought about such things.
Dr. Jacob Krive: And by the way, we will employ our Minitab and Excel and interoperability formats. Jason H07 will do some of the assignments in the Unix shell and teach you the command line. All of it comes along. It’s not like we are disregarding any of the technical aspects. We are incorporating them, but yet, we don’t stress those by themselves. It’s really the combination of understanding the domain, being comfortable in a domain, and applying those technical skills along the way that we target in this program.
Dr. Miriam Isola: Okay, so we’ve covered a lot of ground on the approach to learning. Let’s talk some more about some of the typical questions that we get from students. We’ve touched on this a little bit, but the background. What kind of background do they need to be successful? That’s a question we get a lot. Have you had students ask you about that?
Dr. Andrew Boyd: The background, because we’re blending the fields together, can be quite diverse. Both from the clinical background, are you a licensed provider or from a business background, are you from a startup? Are you from a Fortune 100 healthcare company? Also from the technical backgrounds and realizing our students come from a diverse background is you can learn from each other as well as from the individual faculty members, but there isn’t a prerequisite where you need a specific degree in order to be able to pursue this.
Dr. Miriam Isola: I just, this week, had a student at the end of a very simple data analysis exercise, say in the in the anonymous comments at the end. “I didn’t know if I could do this or not, but I’m really glad I tried it.” So I think that people that are willing to jump in, if they have an interest in this area, they can definitely consider. And not self eliminate just because they don’t have some kind of a background they think is needed.
Dr. Jacob Krive: Yeah. And we do go through interesting discussions with are over the phone, live or over the email, doesn’t matter the format, back and forth. And at the end of the week, I get something from a student, “Wow, I never thought that I would get this far with this assignment or, overall.” And it’s, “Yes, you can actually do more than you may be thinking that you can do.” And if you don’t, we put you on teams, like Dr. Boyd said, you may be in yours, who will be working with somebody who is more technical in the team. Those teams resemble the real environment. You will not be alone by yourself in any healthcare organization, whether small or large, figuring out the clinical concepts and the technical concepts and the programming and the data and interoperability.
Dr. Jacob Krive: That will be done by a team of individuals who represent those areas. We want to get you ready as much as possible for a number of those jobs, but you will never be alone, struggling. And you don’t have to think about this program as being, “Wow. If I don’t click all the buttons and I can do this, this and that, that I can only be successful in this way.”
Dr. Miriam Isola: So let’s talk a little bit about this group of students that are looking ahead to graduation and they’re now really seriously starting to think about how do I transition from being a student and take this education that I’ve got, and build a successful career in informatics. Sometimes taking those first steps, people are a little bit unsure about exactly how to approach that. What advice have you given students about that?
Dr. Andrew Boyd: I’ve routinely recommended students both go to the national meetings, HIMSS, American Medical Informatics Association meeting, as well as the local chapter meetings. We’re not talking about flying across the country. HIMSS has local chapters all around the US, going to the monthly or semi monthly meetings, getting to know the community, getting to know the individuals. As we said, this is a young field. Making yourself known is more than just submitting a resume or CV for a job board. But as you get known, as your work gets known, the opportunities that are both publicly posted and privately posted will reveal themselves.
Dr. Miriam Isola: Yeah, I think you bring up a very important point that I always stress when I get the question. A lot of times students are thinking about where they’re very successful at this point in being a student. They are finishing their degree, they get how you pass classes and how you do things, but it’s a little more. So the next thing that they’re considering is getting another credential and you can always get more credentials. But I always emphasize, get out there and get known. Get a network, start building your network, whether it’s through LinkedIn, these professional association meetings and get out there. I think that is so important.
Dr. Jacob Krive: And analytics is a hot field, let’s remember that. And it’s projected to remain hot for awhile and that means employers have a hard time finding the right candidates and that means some of them, if not, and that there’s an increasing number of them are becoming more flexible. No, they will not hire somebody who is not qualified, but they’re becoming flexible and seeing somebody who may be the right cultural fit, who may be the right learner. In fact, I was just working with a few students by answering their questions and coaching them in terms of careers. They applied with smaller high tech firms that are more nimble. The jobs that they ended up getting were not necessarily the title that was posted. They interviewed, they came in, they talked. The company liked them, they saw potential in them, but perhaps not in the job that they interviewed for.
Dr. Jacob Krive: So guess what? They went and rewrote job descriptions or perhaps that they kept that job open but they hired into a different slot that was also open, but they haven’t thought of. There are smaller companies that can easily rewrite the job description and then, there are established companies that may be looking for a more senior person, but seeing that somebody more junior, they’re willing to drop the level down for that person to allow him or her a chance. And that means that even though you’ve got a more junior job, you already know upfront that there is a more senior job that was posted that’s available for you. So come and grow, it’s a matter of being brave to step into something where you can develop farther.
Dr. Miriam Isola: Yeah. Those dynamic kind of situations that you’re describing require you, as the person applying, to be able to scope it out and to be able to recognize potential opportunity and how it fits to you. So you have to have a pretty good understanding of your own personal narrative. What is it that is of interest to you as a health informatician? What is it that you have some experience doing and, because of the courses that you’ve taken, what have you actually done that you can share with this person to articulate to them how you can fit into their organization. So students have to be able to make that transition to be able to talk to these potential employers.
Dr. Andrew Boyd: And another potential network is the students you’ve been going to class with for the last several months or years. They, many of them, are working in the industry or in tangential industries. And one of the neat things of working in a course or program like this is meeting your fellow students. There are obviously some students you get along with better. There are others, for personality reasons, you don’t. Stay in touch with them. This is a long career path and you never know when one of your classmates will end up either looking for a job or you’ll be on the hiring table and one of your classmates shows up and you’re going to hire them. So please, keep in touch and please, make sure that you leverage that as well.
Dr. Miriam Isola: And we do have and we encourage all of our students to join the LinkedIn group, Health Informatics LinkedIn group, and we post job opportunities there sometimes. And once we do, I often hear from students saying, “Thanks for posting this.” And they are looking at that. So I encourage everybody to do that.
Dr. Jacob Krive: Yeah, and in fact, I was just attending some of the first attempts to bring together the local healthcare providers here in Chicago, who are interested in exploring more analytics and specifically artificial intelligence as a sub domain of analytics. And one common theme I hear is it’s difficult to find people and it’s difficult to retain those people and keep them happy. So the market is really moving, but there is that disconnect between the job posted and the person who may be well qualified for it, or maybe a candidate who is promising. So that that kind of networking is really important to make the connection between a job that may have been posted for the last nine months and hasn’t been filled in and yourself.
Dr. Andrew Boyd: Well, and hospitals, well not all jobs are in hospitals. Hospitals are used to hiring credentialed individuals. Nurses, doctors, pharmacists, they’re all licensed by the state. There’s all these tests and so, HR, for many of these organizations, are used to checking boxes and in these dynamic fields, trying to get a description of who you really need through a Human Resource Department that’s used to seeing check, check, check. It’s really hard.
Dr. Jacob Krive: And as a leader also involved in the industry, I have that experience working with recruiters who are used to that type of hiring. You either have an RN or you don’t, but how do I hire for an informatics position and what about array of all this required and nice to have and maybe have. And oftentimes, I end up saying, “Just give me access to the system to see everything because you might be taken out that best qualified person of the pool and you’re just not seeing it.” Maybe the person didn’t highlight his or her skills very well. Maybe you look by keywords, but it required reading into it a little bit more or recognizing the kind of degree that the person has.
Dr. Jacob Krive: That whole promise of culturally fitting into the position gets missing when you just go and try to hire by skill. But like Dr. Boyd said, this is what HR is used to doing in, not only provider organizations, in fact pharma is guilty of that as well. And whether it’s retail or the innovation side, insurance companies to some degree are doing the same. There are much easier about hiring an actuary or a mathematician of another kind than someone for a new job.
Dr. Jacob Krive: And oftentimes the job description is very vague because they don’t know how to write it. It’s more like they hire somebody and rewrite the job description for that person.
Dr. Miriam Isola: I think this whole industry, we’ve been talking for a long time about how it’s maturing. And I’ve done work for many, many years in many different positions, but nobody ever once called it Health Informatics, but it was. So we’re getting to the point now where I think we’re starting to see some employers recognizing that terminology, but there’s still a lot of work to do along those lines, as we’ve been talking about.
Dr. Miriam Isola: So students need to be the one to fill that gap, to not only focus on passing your classes, but also do some thinking about who you’re going to be out in the job market and how you’ll fit into some of these jobs that you want going forward after you graduate and as you’re building your career. I think we’ve talked through some of the things we said we wanted to cover. The only thing I guess we would add is, as far as next step for students, they can always talk to their student advisor and ask them some more questions about requirements. We didn’t want to take the time in this video to go through the details of requirements, but your student advisor can go through all of that for you and talk with them about whether you’re ready to enroll in any of the health data science courses.
Dr. Jacob Krive: And as far as helping choose which courses, sometimes you can follow your passions and in other cases, when you don’t know where your passions are, take some of the introductory courses because in those lectures and assignments we actually address other more advanced domains of healthcare data science. So maybe through those courses you will be able to figure out, do you want to take a clinical decision support course? Do you want to take a visualization or artificial intelligence class? Maybe you need a little bit of that introductory time to develop your farther passions.