Dr. Shilo McBurney is an Assistant Professor in the Department of Epidemiology at Brown University. Dr. McBurney is focused on teaching and developing new curriculum in the constantly evolving world of epidemiology, while also staying at the forefront of population-level healthcare research and policymaking.
Q: What kind of work and research do you and other epidemiologists engage in?
A: The thing about epidemiology is that it’s really just the broad study of health research in populations.
So any kind of health research you can think of, whether it’s clinical trials, drug studies, patient case reports, infectious disease outbreaks, all of that is sort of encompassed within epidemiology. We don’t actually have labs that often, Although we do sometimes. But we’re definitely not a bench science. All our analyses are done on computers. I wouldn’t say I particularly have a lab, but a lot of the research I do is what we call applied epidemiology. So meaning I work directly with public health units, public health data to try to use that to improve population health.
I have a history of doing this with sort of different topic areas. Infectious disease is one that’s really common. You can imagine with public health units and surveillance data, that would be common. But I’ve also done a lot of work around cancer screening and other sorts of interventions that we do in public health to see how well they’re working.
Q: What do you hope to achieve with what you’re doing or your work with these public health units?
A: I was looking at your questions earlier and I’m like, should I say something super lofty or something more realistic? Because I think all the time we have these like massive kinds of goals, but is it realistic to obtain them? Basically my main goal is to help serve the public by improving population health
Depending on which public health unit I’m working with, that need is very different for various times. So, for example during COVID 19, I was in my PhD program, so I didn’t get to work that closely with public health. But if I was at the time, I would imagine I’d be doing a lot of work with, you know, COVID 19 outbreak investigations and that sort of thing.
My main sort of focus that comes more from my academic perspective is that I really see the importance of good data for informing public health and then being able to do a very clear analysis with that data and then translate those findings into something that can be used for public health, either an intervention like a screening program or some policy like we saw a lot during COVID 19 like masking policies and that sort of thing
So I’m very much like a small part of a much larger system. But I think sort of a lofty goal that we could have is trying to, especially in the U. S. where data tends to be very centralized and siloed when we talk about health data, come up with better uses for the data that we have and better sources of data that we can use.
Q: A lot of your work centers around population level data, but I was wondering what issues come up with that, especially considering the fact that like you said, it’s very centralized. Are you and other epidemiologists looking at, or are you seeing more surveying methods coming up that are working better towards this goal of having better data? Or is it still like a work in progress?
A: I think it’s always going to be a work in progress. So I originally come from Canada: I actually moved to the U. S. about two years ago now, a little bit less than that. I’m used to being able to have very rich healthcare data because we do have a system there that’s paid for by the government. Everybody has a health card number, and we can link data from lots of sources in different ways there. And the U. S. is quite a bit more challenging. You can get through some of it through health insurance. You can get to some of it through specific hospitals and surveillance data.
But that’s why I say there’s, I think, a lot of room down here where we can improve in getting more centralized surveillance systems that use data kind of together. And so I do think there are a lot of challenges working with secondary data. So quality, like I said before, is one key challenge.
For example, if you go to the doctor and you go in because you have a cold, they record that down in some way. So they’ll record the billing for the visit. So just a general visit billing, plus they’ll record some diagnostic code. So maybe they think you have influenza. And so that’s the sort of thing that we’re collecting.
But as you can imagine, maybe at that time they do a swab and it turns out that you don’t have influenza, you have COVID 19. And so if we take that original influenza code, we have the wrong diagnosis for what you have. So really trying to validate the quality of the data is one of the main challenges, and then the second, as you kind of indicated, is access, privacy issues, ethics issues.
And then the final thing that might be of more interest to you when you’re talking about technology and that sort of thing, we are moving more and more into data science and being able to leverage big data sources using AI machine learning methods as well.
Q: In terms of upcoming technology, are you seeing anything that’s kind of promising to you that will help both your current work and other work you’re seeing in epidemiology?
A: I think there’s like two main areas where these sorts of methods would be useful for epidemiology and I’m really like, as I said, involved in curriculum development. So this is something I’m really passionate about bringing into our school here at Brown University.
The first is that it actually is very good for validation purposes. You can use things like natural language processing, et cetera, to try to get more information out of sources like electronic medical records. So instead of just having that code, like I said before, you could have information from all the notes that the physician wrote at your visit to get a better idea of whether this is a diagnosis or was that just sort of a box they checked to get further information?
And natural language processing is really good because it takes into the full context of the sentence, the paragraph, etc. The main other way we use that sort of technology is through prediction. So taking all the information that we can from you using things like genomics data or gene based data, all your sort of characteristics, health records, drugs, and then being able to decide, okay, you present with this condition.
What is the next step for you that’s going to give you the best outcomes? And I think that’s a really exciting kind of field too. Somebody was just talking about this in my department recently. They do pharmacoepidemiology, so drug-based research. And their lofty goal is being able to put in all your information and then be like, this is the best drug for you right now with your condition and history.
Which, like I said, is talking about lofty goals. But can you imagine that? That would be such a useful patient tool to have.
Q: You mentioned developing practices for communicating complex scientific topics and driving public health and policy change, and you mentioned how you wanted to do this on a larger scale with companies, and obviously bigger policy changes, so what does this kind of look like?
A: Whether I’m developing, teaching curriculum, or talking to public health units, policymakers, physicians, those are all very different target audiences and I need to be able to communicate what I’m saying in very different ways.
And so I think sometimes research gets stuck because we go. Here’s the data. Here’s what I found. This makes so much sense. It’s so rational. But then it doesn’t get picked up down the line, and that’s because A, we’re often seeing it in a very complicated way that nobody really understands, and B, we’re not putting the values on it that other people would put on it.
That’s the key when trying to translate research into policy, even though sometimes it might be an icky thing and we don’t like to talk about it, like cost effectiveness. In reality this is what often matters to policy makers. So finding a way to still get kind of the result that, you know, is the best for the population by translating your information into something that’s going to be something that they’re interested in hearing about is really, really key.
Tech Talks Insight: In almost every one of the past interviews arises the fact that there is a lot lost in translation. What researchers work on now, as at least from the words of somebody else, sometimes takes another 10 years to come by.
As I was actually talking to one of Dr. McBurney’s colleagues at Brown, Dr. Kesari, he was saying, “I’m a mathematician, I look at my work, and I know it will work, I know how it works, and it may seem impossible to other people, but I know it, but then I have to show other people about 10 different models of the same thing in reality for it to kind of come across.”
It’s really interesting to hear about the kind of presentation of Dr. McBurney’s work and presentation of facts to other people.
Q: How is your work typically received and what pushback do you face?
A: I think the audience I typically have to work hardest with to send my message through is physicians. They’re very much entrenched in like, this is how we always did it. And so trying to find ways to appeal to them. So a lot of them might be very kind of set in their ways, but they’re really interested in improving patient health. So sort of couching it in that this is something that can really go a long way for patients can be helpful.
Q: How have you seen yourself changing curriculum or trying to craft new curriculum based on the circumstances and what do you think students should be emphasizing right now?
A: I’m always really interested in, I think, what would serve people best and so I mostly teach graduate students but I’ve given talks to high schools. Some of my students are focused on epidemiology, other ones are focusing on other areas of public health like social and behavioral health and that sort of thing.
So I do give a lot of thought to what these different groups of people need. And also despite me just saying everything’s changing, there are certain skills that I think over time are going to be important. So I think critical thinking and scientific literacy are sort of the first things that I focus on.
I think we need that at every level. In the global day, we get so much information thrown at us, and we need to be able to assess, is this useful, not useful, wrong, intentionally wrong? And so I think that can kind of serve every population. I also try to have my students focus more on developing skills than focusing on topic areas.
So I think a lot of times in research people get very fixated on like, oh you’re an infectious disease epidemiologist, you’re a cancer epidemiologist, you do clinical epidemiology, but really a lot of epidemiologists truly at the heart have a methodological skill set that we can apply to lots of different areas.
And so one skill that’s very popular is learning how to program in R, or we have other statistical software. Some people are brave enough to try Python. So just developing those skills, I think is key and then relating to what I talked about a bit earlier like getting communication skills is just essential. It doesn’t really matter where you go, having good communication skills is going to get you really really far. And so that’s sort of my overarching thing and then I do consult with public health units, etc. to get more specific skills that they feel that we’re missing in the workforce and that sort of thing. But I think those aren’t as broadly useful as some of the other things.
Tech Talks Insight: Dr. McBurney’s work lies at the extremely interesting intersection of research, policy, and education. From redefining the preconceived notion of epidemiology at the beginning of the interview to explaining the roles of different individuals and practices towards the goal of improved population health, Dr. McBurney provides a new perspective of technology and research’s service towards society.