Erik van Zwet is an Associate Professor in the Department of Biomedical Data Sciences of the Leiden University Medical Center where he has been since 2009. He joined the school wanting to do more applied work in the areas of statistics and data analysis and has since published a paper in “Significance Magazine” that we covered a couple of weeks ago.
Episode Description
We’re pleased to have Leiden University Medical Center’s Erik van Zwet back to discuss his journey into the medical statistics field in this episode of Stats and Short Stories.
+Timestamps
What got you started on this path? (1:00), Advise for future medical statisticans (4:58), Other cool projects you've worked on (7:05)
+Full Transcript
John Bailer
Erik van Zwet is an associate professor in the department of biomedical data sciences at the Leiden University Medical Center, where he's been since 2009. He joined the school wanting to do more applied work in the area of statistics and data analysis. And he has since published a paper in Significance magazine that we covered a couple of weeks ago in a previous episode. But now we're pleased to have Eric back to discuss his journey into the field on this episode of Stats and Stories, where we explore the statistics behind the stories and the stories behind the statistics. I'm John Bailer. Stats and Stories is a production of Miami University's Department of Statistics and media, journalism and film, as well as the American Statistical Association. Joining me as my frequent partner in crime on the show, Rosemary Pennington, as well as our guest today, Eric van Zwet. Eric, thank you for joining us again. Eric, what was your trajectory as a professional? How did you get to where you are now and what kind of led you into this career?
1:00
Erik van Zwet
Yeah, so my father, who basically passed away, was a mathematical statistician. So he was a professor in Leiden, and he was a mathematical setState. It's very different from what I became in the end, but I didn't know that at the beginning. So when I came out of high school, I didn't really know what to do. And I thought I'd study maths. So I did that. And then ultimately, I really wasn't that good at math. And I became more interested in statistics. But for a long time, I thought I'd be a mathematical statistician, and be someone who proves theorems. And, you know, and, and really doesn't, doesn't do actual data analysis. And then, I became an assistant professor in Leiden. And after a while, I realized that I, you know, I couldn't really think of any good theorems to prove. And I know that there are real mathematical statisticians who somehow sit in their office and just think about theory. But if I tried that, nothing would sort of come up. So then I realized that this sort of stone's throw away from the mathematical Institute, in Ireland, there's a line University Medical Center, the University Hospital, and it has a very large statistics group. So this is unusual. In the Netherlands, where I said, stitches are typically sort of scattered in the different departments within a hospital. But at Leiden, there's just a large concentrated group of statisticians and most of them I already knew. So I, I, I approached them and asked if they had a nice job for me, and they did. They just happened to have one. So I moved to, I moved over to the, to the hospital and there. And that was, that was really good for me, because I found out that I like talking to medical researchers, because there on the one hand, they do important stuff. So that made a difference. For me, they're sort of happy with my help, because when I would use it if I had to have mathematical theorems and statistics, I didn't really have the impression that anybody cared, I was sort of more like now maybe to other people in the world. But these are so medical research and often busy students are, you know, they have to do sample size calculations, they're worried about if they can do the t test or the Wilcoxon test. And they, you know, and I can help them with maybe a research question and with some with, with some statistical analysis, and they're always very happy and said, Yeah, I feel very useful like that. Also, medical researchers are very passionate. Of course, there's a lot of pee hacking in the in statistics, or in science, or in research, where these significant results are so important that you know, the all sorts of fraudulent and shady business goes on to get a P value under point oh five, I within a hospital, I've never experienced that. They're all just really, they care for their research, they want to do it right to a fault. They're always you know, oh, dear, can I do this? Is it normal, and they're really worried about that, and I haven't experienced once, you know, where people would ask me if I could get the p value below point oh, five, so that never happens. They're, they're excited about their research, and that I feel genuinely useful. It's very inspiring. And then I'm also on the ethical review board. So I see a lot of trials that are being done. And then I say, I know, I started to see this pattern. And also, you know, when I will do the sample size calculations, I know it's a bit of a fake. And that, you know, they'll always say, well, we you know, we have two groups of 30 patients and what effect size Can we can we should we should we should put in the protocol, right? So, I realized that this is what happens, and I realized that this has consequences. So then, I got interested in this type of research, I ran into Simon who had this great database and the rest of history now, I'm talking to you.
Rosemary Pennington
Eric, queued you taught your path to where you are sort of seems almost serendipitous at times. And I wonder, given the work that you do now, if you have advice for people who want to do the work you're doing who who are who know they want to work in medical statistics would like to be in a in a situation like you, what things should they maybe be thinking about, or looking at, or, or making sure they're studying to be prepared to do the kind of work that you're doing now?
Erik van Zwet Um, yeah, so many people in my department have a math background, just like me. And I also had a sort of a fairly classical training in statistics, which is very messy. And a lot of my colleagues have this, have this background. And if, if we hire people for jobs, then we kind of know that they would have to speak that language. I am convinced, though, that anybody can do statistics, if you don't need a lot of math, because in the end, I never, in my work, I work with statistical software, and I don't do integrals. I don't take derivatives I, you know, I, I don't do a lot of math. In Leiden, there's a master program for statistical science, which sort of has this philosophy that's open to everyone who has a bachelor degree and has some exposure to statistics, but not a lot. And then we try to turn them into practical statisticians slash data scientists that can start work, maybe at a hospital or maybe at a business or something like that. Yeah, my own path is very, maybe old fashioned, or it's, it's something that my colleagues might also have done, but I don't think it's really necessary. So for students, anywhere near Holden, I would recommend that they do a nice master program at Leiden. And I'm not sure if similar programs, well, I'm sure that exists, but I wouldn't sort of be able to recommend them. And I also I, I'm really sort of torn if I would recommend studying math or not, I think it helps me. But for me, it's also difficult to know. I know, I don't do integrals, but maybe just mathematical thinking has helped me a lot. And I'm not really sure.
John Bailer
Well, you know, you've done this really fascinating work recently on these exaggerated effects. You know, this sort of this right, this inflation of effects from studies. I'm just curious, other than that, what's been one of the most interesting projects that you've really enjoyed working on, and in your current position?
Erik van Zwet Yeah, so from a purely methodological socio statistical point of view, this is really the the first big one that I've had since I changed jobs in, you know, 10 years ago, all the other stuff are consultation projects, with, with a real clear focus on on a medical question, and modeling a particular medical data set. So certain things I'm proud of, and where we had some sort of a complicated medical data set with repeated measurements with nonlinear effects and things like that. And so in the last 10 years, I've had to learn a lot about this surface specialist type of modeling. And that's another part that I like about my job is that whenever I'm involved in a new consultation it is always something different. There are some mainstays in logistic regression and linear regression, survival analysis, these things are almost always involved somehow, but there's also almost always something special about a particular project. And it's very exciting and yeah, and, and those are papers where I'm somewhere in the middle of a long list of authors and I, you know, I just, at some point, I tried to help them with something and and that's a lot of fun, but it's not a big methodological project where I'm first author, and I do something statistical and you know, the for that this is really a first for me.
John Bailer Well, this has been delightful. Eric, thank you so much for joining us for this quick conversation on Stats and Short Stories. Stats and Stories is a partnership between Miami University’s Departments of Statistics, and Media, Journalism and Film, and the American Statistical Association. You can follow us on Twitter, Apple podcasts, or other places you can find podcasts. If you’d like to share your thoughts on the program send your email to statsandstories@miamioh.edu or check us out at statsandstories.net, and be sure to listen for future editions of Stats and Stories, where we discuss the statistics behind the stories and the stories behind the statistics.