Please tune in to Episode 7 at Apple Podcast, Google Podcast, Amazon Music, Spotify, and Firstory

Transcripts:

Intro:

How will the biostatistics industry evolve in the coming years? Is SAS an important skill recruiters look at when hiring for a biostatistics-related job. In this episode, you will hear Dr. Xiaoling Wu answering these questions after being in the field for over 15 years.

Xiaoling holds a PhD degree in statistics from Johns Hopkins University. She has contributed to over 10 successful global submissions of drugs treating cancer and CNS disorder. She was a director and head of biostatistics in Legend Biotech in New Jersey and led the successful BLA submission of a breakthrough cell therapy treating multiple myeloma during her three-year tenure.

Currently, Xiaoling is the statistics technical lead of a game-changing ADC at a leading global oncology company. Xiaoling was the lead statistician of five successful IND clarins and one BLA MAA PMDA approval valued at $5 billion. She addressed statistics queries from FDA, SEC, and institution investors and co-authored over 150 abstracts published at ASCO, ASH, and other international medical congress. Let’s dive in this episode and see what Xiaoling is sharing with us.

Jocelyn: Welcome Xiaoling to our biostatistics podcast. It’s really nice to have you here with us. Thank you.

Xiaoling: Thank you.

Jocelyn: And guys, I want to talk about how I got to know about Xiaoling. So there is a really popular Chinese social media called Red, which is basically an integration of Twitter, obviously Twitter, Instagram, and Pinterest. And Xiaoling posted a lot of videos and posts about sharing her experience on career planning or some interview suggestions and how to be a better statistician in general. So I thought I would reach out and see if I could interview her and join our podcast. So Xiaoling, can you tell us a bit about your background and how you became interested in biostatistics in particular?

Xiaoling: Sure. And yeah, it’s my pleasure, Jocelyn, and thank you for this opportunity to talk with you about my statistic career and perhaps potentially sharing with other statisticians. The first time the concept of statistic came to me was when I was maybe in elementary school. I traveled with my mom and then we went to a university in Beijing. I think it was Beijing Normal University. And then she pointed to me to a building. She said that she got some training. She is a college professor. She’s already retired. And the name of the department is Department of Probability and Statistics. Yeah. So my mom chose the major of statistics for me. That’s how I know. Statistics, it looks like she maybe she did a good choice. We’ll see.

Jocelyn: I see. And after you chose this, I guess, this field, do you like it so far? I guess you do. You’ve been in the field for so long, but what’s your impression on it and why did you decide to stick with the topic?

Xiaoling: Yeah. So I went to college in Nankai University. Back then, statistics was part of the one major out of that math department. Now they have their own department. So I learned quite a lot, the foundation of both statistics and mathematics, including calculus, geometry and other things. And then the professors, most of them have U.S. experience. And so they said 20, now it’s 22nd century, is the century of biostatistics. And then so I wanted to be to study statistics and U.S. would be the place to be. And then and then luckily I received a full scholarship from Johns Hopkins University, the Department of Applied Mathematics and Statistics. That’s how I came to the United States in the in the beginning of this century. Yeah, and then I think of the test I when I became to know statistics better, it is actually a curriculum, a discipline that not only that you can use math to solve a real world problem and it actually requires a lot of communication between statisticians and between statistician and non statisticians. And of course, for several years, as the statistician was ranked as one of the best job in United States, like several consecutive year. And then maybe now it’s not like as hard as some of the tech, but I would say it’s still a good job.

Jocelyn: I see. And did you work on a lot of applications when you’re in school in terms of biostatistics or was it mostly mostly not statistical theory?

Xiaoling: Yeah, in the first for the at Hopkins, the first three years, we really focused on theory like probability, mathematical statistics, stochastic theory, geometry, algebra. And then and then we have to pass something called qualification exam like and then and then. But but I think because we had the advantage that we had a public health school and medical school and then so my professor, Dr. Professor Dan Neyman, and then he suggests I could look into some problem in statistic genetics. So and then I get to work on the back then we work at SNP. It’s a location over the whole DNA chain. And then so I work on problem of estimating the false discovery rate of the gene mutations for twins and triplets with certain genetic disease. So it’s both theoretical and have the potential to be applied.

Jocelyn: Right. And then if that’s the case, how did you make a decision to go into the industry after you graduate?

Xiaoling: Yeah, that is a relevant question, because back then the most of the graduate students in my department will choose academic and the professor when they talk about from the industry, they will say she or he went to the dark side. And I well, because my parents were both college professor, I was just kind of tired of. The campus life. That was one reason. And then and then for my first job, like I get something like a summer intern, but it’s pretty it last several months that the company was actually focused on statistic genetics. So I got to apply some of my understanding and knowledge from the thesis work to the to the job. So and so that’s so even my professors thought that that fits my interest.

Jocelyn: I see. I guess we’re not talking about exactly we’re doing your current job, but can you give us a brief I guess back a brief introduction of in general, what do you do?

Xiaoling: Yeah. So since I’m still new to the company, so I like to, when I when I’m ready, I can share. So I’m currently the technical lead of an ADC compound treating breast cancer in a top oncology company. So then as as the technical lead, the understanding is that I’m like the gatekeeping of all the statistic delivery of those specific projects. So the work not only include like ensuring the statistic validity, but also timeline. And then and then the regulatory submission, the entire strategy, the execution strategy and submission of that compound regarding to the statistic aspect.

Jocelyn: I see. That’s interesting. As you mentioned, this is your new job. So I’m wondering, is this sort of similar to whatever job you were doing before this?

Xiaoling: Yeah. So I yeah, I since the pandemic, I did change job twice. Like and then before the pandemic, as you can see in my LinkedIn profile, I was the director of statistic at Legend Biotech sort of start up focus on treating cancer with cell therapy. And then and then I work for another top pharmaceutical company in that. But I see that although it’s it’s like three different jobs, but I see I learn different aspects and then leading to my next ideal job, which is to become a statistical leader in in a company of reasonable size. So I think it’s I learned different aspect of that kind of job.

Jocelyn: I see. So when you’re talking about statistical gatekeeping, was it more of a technical gatekeeping or was it was it more of a regulatory side of the gatekeeping?

Xiaoling: Yeah, it’s it’s first make sure everything we deliver is. Right in terms of statistics, right, and then. Whether first, like there are four for statisticians in the pharmaceutical industry, we have planned analysis, which is described in the protocol of a clinical trial. So then the planning and then including the endpoint, the statistic methodology, and then how to report that. How to report these data, like in what kind of table listing or figures. So like make sure we we meet all the standards, the industry standard and the regulatory standards. So and they’re they are correct. It’s validation is a very important part of a statistician’s job. You always have to either validate other people’s work or validate your own work. And then, yeah. But oh, and then there’s also determine the strategy of the compound that I’m working on, along with the other functions.

Jocelyn: I see. And I guess I’m also wondering if it is true that people say in the pharmaceutical industry, it is SAS is the main language to use in terms of clinical submission and everything.

Xiaoling: In the past 30 years, yes. So in the past 30 years, SAS is the almost the only statistical software that’s used to produce table listing figures. And there’s something this is something like we didn’t know at school is that for regulatory submission, the sponsor not only actually not only required to submit the table listing figures of the result, but actually there’s things called Adam and SDT. They’re actually SAS data set as well. But FDA required to receive them. So their statistician can use those ready to run analysis type of data set that’s Adam and SDT to replicate the sponsor’s result. So then all those table listing figures and the ready to analysis data set are required to be delivered in the format of SAS. It’s only in the past five, maybe in some European company, they started to utilize R as well because R has other advantages. But it’s only in the past even five years that there is so the FDA is also getting more flexible. And then in addition, well, but then at the design stage of a clinical study, like when statistician design the study, for example, for sample size calculation, power calculation, that’s the most important part from statistic aspect when designing study. So FDA, I think they start to accept simulation that’s generated in R. But I think in my past experience, they prefer commercial software. Commercial software means not only SAS, like statistician can calculate sample size and power using some commercial software. I think they prefer commercial software than open source, which means R. It’s changing.

Jocelyn: I see. I just think in a lot of school trainings, I don’t think we are really familiar with SAS as a statistical software. And also the SDTM and Adam you mentioned, I think I’ve only seen them in the job descriptions when I started looking for jobs. And then I look at them, they said, oh, so you’re familiar with this. And I don’t even know where to get those resources because it seems like you have to take courses. And then so I guess another question would be, where do you usually get knowledge on about SDTM and Adam?

Xiaoling: Yeah, I understand. Like in the past five years, because I’ve screened like the resume of hundreds and hundreds of students, PhD or master’s students, I understand that.

Jocelyn: And so how do you like, and then if you go to a new job, then how do you transfer from an R user to a SAS user?

Xiaoling: Well, I think in big pharma, like if you have a programming team, they are a statistical programmer or some company call them statistical analysts, then it’s probably okay that you can import Adam SDTM to your R environment and then you validate other people’s work or run modeling and simulation with using your R code. But then your R code is not probably not fit for regular submission. But then when you receive like dozens of SDTM and Adam SAS data set, then what do you do with them? I may, yeah. So usually a company of certain size, they will constantly provide you training, including SAS and then SDTM and Adam. And then you actually get to review and actually have to validate them for your job. So then otherwise you can look out for trainings provided by programmers from the industry. And then there’s also guideline by FDA. There’s actually very detailed guideline by FDA.

Jocelyn: So if you’re saying if we go to the FDA website, we might be able to find some resources regarding the…

Xiaoling: Yeah, in the past, I’ve trained fresh PhD students on this request. So I understand that there is a gap.

Jocelyn: And do you think this gap is very universal as in most of the students who are fresh graduates, they don’t have this kind of knowledge?

Xiaoling: It is like in the past, at least the students are, they would say, I know both SAS and R. And now it’s like, I only know R. I don’t know why is that. Your question is how do you get to that level, right?

Jocelyn: I think my main question is, is it common for freshmen?

Xiaoling: It’s quite common, but when we talking about fresh graduate, they may be completely fresh to a company, but nowadays a lot of the fresh graduate student, they actually have intern experience. They even had intern experience working as a SAS programmer. I don’t know how they get to that job if they don’t. But that’s… Not everyone came to the pharma industry completely naive about.

Jocelyn: So yeah. That makes sense. I guess regarding that, how do you think the industry will evolve in the coming years? And what are some new development or innovations do you see on the horizon? I guess, including how will the regulatory be changing, for example?

Xiaoling: Yeah, that’s a big question. The first thing I think about is I think the pharma, biostat, the society, how to say, the community is getting more and more competitive. Things are changing almost year to year, especially after the pandemic. Before the pandemic, I think things were not that competitive. And then ever since the tech industry is grabbing some of the talent, statistic talent, people will feel like there is a safety net. But then now with a wave of layoff happening in tech industry, or just… And also because the pharmaceutical industry is getting more and more global. You are based in Canada, right? And you and I are talking. And don’t forget there’s talent in Europe and Asia as well. They can do a job with a fraction, maybe not a fraction in Europe, maybe less, but in Asia, wouldn’t that be a fraction of what you… But they can do pretty good job. And some of the talent, statistic talent in say, India and China, or even Japan or Korea, they’re pretty good. They can be good at communication, English and statistic and programming. So I would just say the trend is that the pharmaceutical statisticians are getting more and more competitive. And we should learn from statisticians from other industry. Of course, the data scientists from tech, but also from different section of government, like FDA, there’s also other government that they employ statistics.

Jocelyn: I see. I guess with that being said, do you have any advice for students or people who are looking for a statistician job on how to make ourselves more competitive under this big environment?

Xiaoling: Yeah, I think that’s something we should think about at least weekly. How do I become a better statistician? So first, you need to notice the trend. Not only, of course, first is if I’m already in the industry, you would know what are the major companies out there, what are their product, and of course, your own company’s product. So you know what skills you need and what’s expected out of you. And it’s also because you get yourself prepared for new opportunity. Otherwise, if a new opportunity approaches you, you may not even know that’s something that’s worthwhile. Or you may not have the skill set at the time the opportunity approaches you then. And then definitely, that’s what you guys maybe have already been doing, attend the industry conference. And now, nowadays, there’s a lot of virtual training as well. And then I read something like people in the tech industry, they spend about 100 hours every year in learning new skill. Because well, maybe their skill is changing more often. But then nowadays, the statistics is also that requires a lot of learning.

Jocelyn: I see. That’s pretty good advice. I guess, to be more specific, what are some skills that exactly that you think?

Xiaoling: Yeah, when I used to manage and mentor fresh or intern students, I think there are several aspects of statistician skill set. The first is, of course, the statistic foundation. That’s the classes maybe you learn at school, survival analysis, regression, hypothesis test. But you not only should get good grade, but you should also learn to explain statistic concept to non-statistician. And you will try to be able to translate their question into a statistic problem. Yeah, sometimes they might not know what they want, and you need to move them to the… You need to let them know what you are talking about, and you also need to understand what they want. Sometimes they don’t know what they want, they need your help to define what they want. And then that goes to a second part is the ability. So that includes oral communication, but also written, like you are able to specify what analysis you plan. And so that’s the statistic foundation and communication. And then definitely the programming skill. But I think nowadays the younger generation is pretty good at the program and foundation, but we talk about most of the students is much better at R than SAS. So it definitely helps if you can learn some SAS because it’s actually a statistical software. So it involves a lot of statistics, and it’s different from R. So I guess to summarize, to become a more successful statistician, first we need to have a really solid statistical foundation. And we need to be able to communicate those statistical concepts with the non-statisticians. And then we need to solidify our SAS programming, because that’s what’s needed in the industry if we were to go to the industry. Is that right?

Jocelyn: I see. So is there something that you wish that you knew when you first started your career in biostats?

Xiaoling: Yeah, like the statisticians always talk about doing a good job in a company in the industry is very different from school, because there is no test. And then or someone grades your test, and then you know you pass or fail or you get a very good score. And there’s no textbook. There’s no actually, and no one is there to be your teacher. Everyone else has their own job. So I think the soft skills, for example, well, in Chinese, we also say emotional intelligence is for one’s success, but translate to corporate life, how you can become successful in the corporate setting. Emotional intelligence means that you can understand, you can sense, comprehend other people’s emotion. And then you can manage your own emotion and utilize these skills to build relationship with people and then even getting the job done by because you build a good relationship with people.

Jocelyn: I see. That’s definitely really useful. And I guess we have a time constraint. So to conclude this episode, the last question is, what is one question that you wish I could have, I would have asked and you have answered it, or it could just be something that you want to share personally to the audience that you think would be helpful.

Xiaoling: Yeah, I think what so yeah, we briefly touch about it. I think the whole pharmaceutical statistics community changed since the pandemic. Pandemic is just an important change factor in many sense. Like I think that one, we talk about like the changing environment in the tech industry because they bloom, they prosper during the pandemic, but also they change how we work. Before the pandemic, like most of the old school leaders were completely were not acceptable to work from home. Now it’s like, so we take things for granted, but before that, so that just changed not only the way we work with our colleague, but it’s also how we can build a work life balance. Because at home you can work 24 hours, then how do you make sure yourself is healthy physically and mentally and not getting burned out. And then how for, this is not just for women, maybe for men and women, like how do you maintain a work life balance when you have kids? This changed, because when I was at graduate school, I was like, if I study hard, I will get good grade, then if I work, if I have a job, if I work overtime, then maybe I’ll get better. But when I have kids, maybe I cannot work 24 hours. So yeah, you can, so as when you are at the stage of your, the beginning of your career, you can think about not only like, I want to be VP of this company in 10 years, but you also, you can also think about, do I ever want a family or not? Either way is fine, but you can plan your life and your career around it.

Jocelyn: I see. I guess that’s really good advice, because sometimes I think about like personal future and the academic future, and I guess the career future, when I said personal future, it’s like personal plans in terms of family and stuff. I guess it is something to think about when you start your career, because it needs a long term plan. Yeah. Well, thank you, Xiaoling, for sharing all your experience and insights. And it was really nice to have you on this episode. Thank you.

Xiaoling: Thank you so much, Jocelyn.