Finding a biostatistician to work with

0:00 Welcome to the Clinician Researcher podcast, where academic clinicians learn the skills to build their own research program, whether or not they have a mentor. As clinicians, we spend a decade or more as trainees learning to take care of patients. When we finally start our careers, we want to build research programs, but then we find that our years of clinical training did not adequately prepare us to lead our research program. Through no fault of our own, we struggle to find mentors, and when we can't, we quit. However, clinicians hold the keys to the greatest research breakthroughs.

0:40 For this reason, the Clinician Researcher podcast exists to give academic clinicians the tools to build their own research program, whether or not they have a mentor. Now introducing your host, Toyosi Onwuemene. All right everyone, welcome to today's episode. I am Toyosi Onwuemene, your host on the Clinician Researcher podcast. What a pleasure to be here today. Thank you for tuning in. Thank you for listening. Today, I'm talking about how to find a biostatistician to work with. And before I start, I want to share with you that we are opening up slots for clinicians who are interested in Academic Negotiation Academy, and this is to help you negotiate your research career so that you can lead the research program that you really want to lead and thrive while you do it.

1:35 So if you're interested, sign up on our website, clinicianresearcherpodcast.com. All right, we're talking about how to find a biostatistician to work with. And I want to tell you, it kind of depends whether you're looking for a biostatistician or not. Depends on the kind of research you do. So I'm a health services researcher. I do a lot of statistics heavy type analyses. And finding a biostatistician to work with can be difficult. So it does depend from institution to institution. I'm going to share with you some insights that I have that I want you to think about as general principles. But I've always known that I needed a statistician, at least kind of theoretically everybody tells you you do. And you're like, well, I need a statistician, OK.

2:18 But one of the ways that I finally really understood that was when I finally had submitted a grant. This was a career development award that I had submitted. And when the award was not discussed, I had the opportunity after receiving the summary statements to talk with the program officer. And the very first thing the program officer said to me when we started talking was, oh, you're the one who didn't have a statistician. And I think I was kind of surprised by that because I was like, but how could you tell I didn't have a statistician? And you know, it's interesting. Of course they could tell. So on that panel, they usually will have a statistician represented to answer questions of does the statistics seem appropriate?

3:00 Are the analyses appropriate for the particular hypothesis? And so yes, there will be a statistician reviewing your grant. And if you wrote the grant without a statistician, they can usually tell. Now maybe you have statistics background and you can write the methodology for statistics. Sure, good for you. But if you don't have that background and you're going to write a grant that involves any statistics, it is wise to get the involvement of a statistician. And many times, clinician researchers don't necessarily have access to statistical help. And so I want to just shed some light on a couple of things I think are helpful for finding a biostatistician collaborator.

3:41 And actually, later on the show, I'm going to have a statistician come talk to us about how clinician researchers can work with statisticians for more effective collaboration. So stay tuned for that. Okay, so the first thing that I want to share with you is that a biostatistician is a collaborator. I want to say that again. Biostatistician is a collaborator. Because in research, we have this sense that we're doing the work and then the statistician is like the person who just comes and does like the magic analysis at the end. And that's the end of the story. Like, oh, great, we finished the analysis, come and do the statistics. And it doesn't actually work. It doesn't work that way.

4:21 I mean, you can do it that way. It's not very effective. And the reason it's not very effective is because statistics is not just about doing math. It's about setting up the right equations, so to speak, to be able to do the math well, setting up the room well. Like, maybe if I would correlate it to like podcasting, you can make audio sound better post-production. But if you were there at the beginning, like, you know, making sure that the space was quiet and that you could record the best possible audio that you have at the beginning, then you don't need people doctoring much at the end. And so I think of a statistician in that way, where it's like if you're setting up the study, having a statistician work with you to say, okay, these are the parameters that you need to make sure that you have considered.

5:11 These are the variables you need to collect. This is the systematic fashion in which you need to collect these variables. And then at the end, I'll come and do my magic math. It'll be so awesome. That's how you collaborate with a statistician. So from the very beginning, a statistician should really be part of your formulating your research question. And that's why it's important not to think of a biostatistician as an afterthought. This is not like a, I finished my analysis, who's going to analyze it for me? And to be honest, many of us, that's where we are. It's like, well, the project's been done. Now we need some help. Okay, great. And biostatisticians are awesome people, at least the ones I've worked with are. And they're usually able to take things and make it better.

5:51 But if you really want optimal strategy, it really does help to meet with a biostatistician upfront to be clear about what question you're trying to answer. And one of the things that they do well is they understand the research methodology where it's like, hmm, that's a question you want to answer. This way that you're going about doing it will not answer that question. This is the question it will answer. And that's helpful to know upfront. So first of all, figure out that the biostatistician is your collaborator. And for that reason, think about how do you incorporate the biostatistician early on in the whole process and not just waiting until it's time for the analysis. So biostatistician is your collaborator point number one. Number two is to develop your research plan.

6:33 So for many of us who may not have had research training, sometimes there's a sense that you are just kind of doing research on the fly. But research is not really successful when it's done on the fly. You could get lucky and sometimes we do. But really being very systematic about what question you want to answer through this project and being very systematic about the variables you're collecting and how those variables are being collected is pretty important. Sometimes as clinicians, we have the tendency to just say, let's just gather as much data as we can and then let's see what shakes out. And really, it's not a very effective strategy.

7:14 It's important to figure out what is the question we hope to answer at the end and what is the data that we need to collect that helps us do that. And sometimes there's a sense of like, we've got to collect all the data we can now. We can't come back. And that's kind of like, I would say it's a scarcity type mindset. You will always be able to come back. Always be clear about what you're trying to answer and then be focused. And it helps you to be able to accomplish things a little bit more effectively than when you're paralyzed trying to collect all possible data. And at the end, you miss the data that you actually should have collected to be able to answer the question you wanted to answer. And so recognize that you need a plan. And this is why your statistical collaborator can help you at the very beginning of the plan to figure out how to pull it together and what variables you're going to need to make the plan a reality.

8:04 Because what you don't want to do is get to the end of the study and it's like, oh, we can't answer the question because we didn't collect variables correctly. So develop your research plan, ideally with a statistical collaborator as part of the early development process. Okay. Number three is to determine what kind of statistical support you need. This is really important because I started first of all talking about the need for a statistical collaborator. There are many different kinds of statistical collaborators. There are many kinds of statisticians. And the experience that they have is myriad. So recently I started working with a statistician who's like, well, my expertise is trajectory class analysis.

8:48 And I'm like, well, what? What? But every statistician has an air of expertise. And especially when you think about statistical collaborators, you're really talking about faculty. Faculty who have statistical methodology that they've honed and actually they're probably innovating and creating new types of statistical methodology. So your collaborator is really someone who's able to help think big picture about the project. But they may not be running their own statistical analyses. That's important to note. And sometimes clinician researchers are just like, well, we're just looking for someone to run the numbers. Well, you're just looking for someone to run the numbers, maybe a master's level statistician.

9:33 But your faculty bio statistician, usually a PhD faculty statistician, is really like the big thinker. And so you want to be clear about what kind of statistical support you need. And hopefully you're not just looking for someone to crunch the numbers, because that's not necessarily what the collaborator does. They can crunch numbers, but you're better served with them really helping you think big picture rather than with them doing the number crunching. So you might need number crunching help and the statistician kind of like as a collaborator to really think through the data. But definitely be clear about what kind of statistical support you need. OK. Number four is to find out where the statisticians hide. So at every institution, every institution is different, but statisticians kind of exist independently.

10:19 So sometimes there are statisticians who are associated with divisions or associated with departments. And sometimes statisticians have their own department where you kind of have to go to a hub to be able to kind of recruit a statistician. But whatever happens, it's important for you to understand how statistical support works at your institution. What are the norms? How have people succeeded? So it's knowing what the general norms are and then understanding what's outside of the norms. So for example, perhaps the only way you can get access to a biostatistician at your institution is for them to be part of an established departmental budget or something.

11:00 So to have access to the statistician, then you kind of have to go through a process where your project is approved and then you can talk to the statistician. Maybe that's the way it works. Or it could be that just individual people hook up with biostatistical faculty and they're like off to the races. So you do need to understand the structures within your own institution and then how other people have succeeded. Because it may be that you don't have the bandwidth to wait for the couple of months that's necessary to be next in line for the faculty biostatistician that's affiliated with your program. It may be that you are able to figure out a way to find a statistical student to work with you and a faculty member who's willing to look over their shoulder and just make sure they're doing things right.

11:43 There are many, many, many, many ways to make this work. But understanding what the range of possibilities is at your institution allows you to know how to meet the needs that you have within the confines of the resources that you currently have available to you. And so that's really important. Know where the statisticians hide, aka figure out what the processes are, what it takes to work with a statistical collaborator, and what are the margins of what that could look like. Okay. Number five is to pitch your statistical collaborator. And I say pitch them. And so it's interesting because some people work within groups where this is just a statistician for the group. It doesn't matter what your project is. This is just a statistician everybody works with.

12:27 And I will tell you that an engaged statistician will always be better than the statistician you just have to work with. And so helping statisticians that you work with understand your project. Why does that matter? Why is it important? Why is the question that you want to have answered important to be answered? Why have you decided to answer it in this way? And so it's important for you to help your statistical collaborator come on board and really understand the premise behind the work. Help them understand what's going on in the background because it does help them kind of better understand where the numbers are coming from and what the numbers are supposed to do for you.

13:07 And ultimately, your biostatistical collaborator is likely not going to be a clinician. So there's so many things they don't understand clinically. Your clinical expertise is so important and you framing the project in light of the clinical problem is going to be super, super important as well. So you want to pitch your statistical collaborator and help them have a great reason to work with you. Okay. So there are five things I talked about. Understand that the biostatistician really is a collaborator. Make sure that you develop your research plan probably in concert with a statistical collaborator. Find out what kind of statistical support you actually need. Find out where your statisticians hide.

13:49 So understand the structure of your institution, how you can work around them, and then pitch your statistical collaborator. Actually sell them on your project. Help them to really be a contributor to your project as well. So that's definitely important for finding a biostatistical collaborator. And if you're struggling and you're like, I'm trying to move work forward. I can't even find a statistician. Yes, there's someone in my institution who's supposed to be helping me and I can't make it work. It's possible that coaching will help you move it forward because what you really are doing is negotiating for resources. See it all comes back to negotiation. So if you are interested in learning to negotiate better for resources, including negotiating access to biostatistical support, we can help you in the Academic Negotiation Academy.

14:36 For more information, you should visit our website, clinicianresearcherpodcast.com. Okay. Somebody else needs to hear this podcast. I'm going to ask you to please, please, please share it with someone else. And if you've been a listener for these last couple of episodes, please subscribe and rate the show so that other people can find us as well. All right. It's been a pleasure talking with you today. I look forward to talking with you again the next time. Thanks for listening to this episode of the Clinician Researcher Podcast, where academic clinicians learn the skills to build their own research program, whether or not they have a mentor.

15:20 If you found the information in this episode to be helpful, don't keep it all to yourself. Someone else needs to hear it. So take a minute right now and share it. As you share this episode, you become part of our mission to help launch a new generation of clinician researchers who make transformative discoveries that change the way we do healthcare.