Top 7: How to work with a physician within Pharma to become a valuable partner

Benjamin Piske

Global Head of Biostatistics, PBS

Benjamin Piske is Head of Biostatistics at Cytel’s Project-Based Services with over 20 years of CRO experience. He leads global biostatistics teams and specializes in operational strategy, trial execution, and cross-functional collaboration.

Transcript

Top 7: How to work with a physician within Pharma to become a valuable partner

00:00
You are listening to the Effective Statistician Podcast, the weekly podcast with Alexander Schacht and Benjamin Piske designed to help you reach your potential, great science and serve patients while having a great work-life balance.

00:22
In addition to our premium courses on the Effective Statistician Academy, we also have lots of free resources for you across all kind of different topics within that academy. Head over to theeffectivestatistician.com and find the academy and much more for you to become an effective statistician.

00:49
I’m producing this podcast in association with PSI, a community dedicated to leading and promoting user statistics within the healthcare industry for the benefit of patients. Join PSI today to further develop your statistical capabilities with access to the ever-growing video-on-demand content library, free registration to all PSI webinars, and much

01:13
Head over to the PSI website at PSIweb.org to learn more about PSI activities and become a PSI member today.

01:28
Benjamin Piske is speaking from the podcast, the effective statistician. Today we’re talking about how to work with a physician within pharma. And I’m here with my co talker, Alexander. Hi, Alexander. Hi, Benjamin. Nice to speak again. Alexander, I think, as I said, we are today discussing how to work with a physician. And I think this is, this is something that where you have a lot of experience with.

01:58
a lot more than I do actually. So therefore maybe you can quickly introduce what this is or where this is coming from, the topic in itself, and maybe describe a little bit about how you worked in the past with this physician, your first experience, your experience in general, and little bits about how this is daily on a daily basis going. And maybe we can then start discussing and how this is different to my experience. Yeah. Okay.

02:27
Interactions with physicians, think is one of the key aspects of a statistician where we as statisticians can have a big influence and impact on the work and see the relationship with the physicians think for statisticians that work within the pharma industry or specifically within the sponsors. key. And maybe today I really want to focus on speaking about

02:56
the physician as a colleague, not the physician as a customer. So of course we have the physicians prescribe our therapies. We work with them as key opinion leaders or thought leaders. We write papers together with them and there’s lots of other engagement with external physicians, to say, external to the pharma companies. But today I really want to focus on how we can work effectively.

03:26
with physicians within the pharma companies. And they of course have lots of different responsibilities. We write together with them papers, we run the studies, we prepare all kinds of negotiations with payers or with regulators. So lots of the interactions is between statisticians and physicians. It’s probably the first people that work on the study.

03:55
as well as the last people that work on the study, for example. Together with the statistician. Yeah. Obviously together with the statistician. That’s why I think it’s really such a key relationship for us as biostatisticians. But isn’t a physician like a little bit of a statistician anyway? I remember that in the medic at university, usually have part of their lessons is about statistics. Isn’t this of great help for the

04:25
or the relations that we as a statistician have with them. Yes, that is of help. However, there’s a lot of variation in terms of stats knowledge between the different physicians. So there’s lots of variation about medical knowledge among the statisticians. Sure. But you speak to a very interesting point. I think one of the key things to have a successful relationship, successful working relationship is

04:53
that you try to speak each other’s language. And as the physicians learn to speak stats, it’s great that you learn to speak medic. So I think, and that was one of the key things that I learned over time is the more I learn their terms, their way of thinking, the easier it becomes. And of course, then you also have the possibility

05:21
to teach them about stats language. What is the p-value? I don’t know how Ofmar have explained that concept, but yeah, I’m probably not alone in that. Well, isn’t this like a of like a shared responsibility then that on the one side, we are asking them to learn what we are doing. And on the other hand, we expect or they should expect us to learn what they are doing. Yeah, I think that for me is an ideal scenario.

05:51
However, I think when I see across the industry, that ideal scenario is very often not reached. And I think that is for variety of reasons. The first is I think when physicians come as industry, and maybe they were already very well established within their medical society. So maybe they were head of a department or head of a clinic.

06:21
And then they joined the industry.

06:26
They actually have a very hard time at the beginning to work in this very different culture. Imagine you’re in physician and you work in day in, day out in an emergency room setting where, you know, every minute or every second counts and what you do has a direct impact. And so everything that you say and tell really needs to be followed.

06:55
So it’s really in crisis situation and very quick decisions need to be made. then there’s no room for discussion or lots of additional explanations. If you come from that setting and then you come into a study team where people push back and people want to know the background and people want to argue with you. It might feel very different.

07:25
probably will feel very different. And lots of physicians, think, at the beginning struggle with this cultural change. So I think…

07:37
Lots of physicians come with them into the industry and have this natural thing of thinking, okay, I’m the lead. I tell what it’s done and then you as a statistician, go and implement. And so that is, think very often something that’s quite naturally coming from them because that’s how they have done their job in the past. And also isn’t it, it seems that

08:06
people or physicians coming from hospitals might then also think rather on the short term. Isn’t it? Then if they have a problem, if they have a patient, if they have a problem to solve, they need to solve it as soon as possible in order to find surgery or whatever the requirement is. However, if they work in pharma, they are about to plan a strategy for a product, for a study.

08:35
on a long term, so years in advance. So they have to plan more than just the next two weeks or how long it takes to get a patient up and running. But it might take years to get a product on the market or some, so there’s some other aspects of the whole thinking that might be difficult on, we might need to anticipate. And because it is the same for statistician, the statistician is always, is also on the one side is asked to solve problems. We have data, we have

09:05
We have a problem, we need to get a result. However, if we work with a medic on the planning for a study, it is more than this. It’s really like a partnership with the medic in understanding each other’s thinking and each other’s strategies in order to get the results in a two years, three years. Exactly. And as I said, sometimes the physicians.

09:34
to come into this kind of partnership mode, but likewise, I see many statisticians struggle to get into this partnership mode, just maybe more from the opposite side. I think as a function, as a statisticians overall, we have a tendency to be rather introvert. And if we then kind of, if you’re more an introverted person and then you

10:03
to work with maybe very extroverted physician that maybe has a little bit of a belief he knows it all and it’s for he has treated these patients that you are developing a new therapy for, then that might not lead to this partnership. So it might lead to a situation

10:29
Whereas the situation is just on the kind of receiving end or feels he’s just on the receiving end and just implement what he’s told to implement. And I think that’s not how you can have an impact as a statistician. And what I want to see statistician doing is going into these kind of discussions and meetings with a mindset of

10:59
playing on the same level.

11:04
I that is very important. Both in this teamwork, you need to have all the different pieces. You need to have a very good statistician and you need to have a very good physician on the other hand. Of course, there’s lots of other functions that we’ll not speak about today, but I think if you work on the same level, then you can really drive the project forward in completely different ways.

11:33
Another point that I think is important is here to really think about this as a partnership. I remember one incident earlier in my career where I went with a couple of other statisticians to lunch. And there was also this statistician that was coming fresh out of university and he was hearing all the other kind of statisticians complaining about their medical counterparts. And…

12:02
In a discussion with him he mentioned, yeah, I’ve learned a couple of things in the first weeks here in the farm industry. One of the things is the physicians are the enemies. Is it? Wow. Okay. No, I think only because lots of people are complaining about physicians as counterparts doesn’t mean that they are the enemies.

12:29
And, but I still, this, that struck me really is that even because there is this misunderstanding, that can even be perceived in exact negative ways. No, there are definitely not there are our enemies. As you said, they are the partners. And even though it might be difficult to get along with them for several reasons, I agree that the success in any of the tasks that we have as a statistician.

12:59
is only there if we do work with a physician together. I think sometimes, as I said, I think it might be sometimes that the history of the physician, so the career steps of the physician may be part of the reasons why they’re acting as if they were the boss. Sometimes I experienced that they had, they were actually the budget holders. So in a way they were the boss because they could make decisions that others couldn’t, especially not the statistician.

13:29
And nevertheless, I think in my, from my experience, it is always that the success or the smoothest studies that we were running were if the medic was involving the statistician before making any decision. Not any decision, but any like relevant decision, for example, regarding protocol or changes to the protocol.

13:59
Partnering with the statistician from the medic side was one and vice versa obviously was one of the key experiences which led to successful and smooth working on studies in the past. Completely agree. I think if there’s a really good relationship, you can…

14:20
create something that is much bigger than if people work in isolation. Just as an example, when the data of the study are presented, it’s sometimes really helpful if that is done by the physician and the statistician together. I have very good experience with that, that you can then speak to the, let’s say, medical specialties quite well.

14:49
because you have the medic there, but then if it comes to the, let’s say the core data, then hand over to the statistician and explain where the strengths and where the limitations of the data. Because that is really where the statistician usually can speak much better to the and you have a much better presentation with two presenters that can throw the ball over to the other side.

15:18
backwards and forwards and have a much more lively presentations. And if you try to train the physician to speak about a stats methodology that he’s not really familiar with. what would you say is then a good strategy for convincing the medic to, for example, not do a decision or to not do this? There are some, there are quite often reasons.

15:45
from a statistician, let’s say the data quality, the data, just the extent of the data that is available that would lead the statistician to say, I wouldn’t recommend to run the same in a specific subgroup, for example, because, but the medic would say that this subgroup is the one that the audience will be interested in. So we need to present something. So what is your strategy then to convince or to discuss?

16:14
the point or the issue with the medic? I think there’s two different strategies here. The first strategies that I’ve seen is that people say, really push for it in terms of that they say, oh, I’m the statistician of the study, so I have the ownership and I have, this is my play field, so I will present that. That’s one way of doing it. And sometimes it works.

16:44
What I found is works actually better is that you try to sell it more of helpful for the physician as well. For example, on the stat side to give him a couple of challenge questions. And if you see that he is uncomfortable with that or that he gets awake about the answers. Then just what about if I present that and you.

17:14
It makes the introduction and the conclusion and these kind of difficult parts, the steps, that’s heavy parts. I do present that. And then you’re relieved from the pain to answer these kind of nagging questions that you’re not fully prepared for. So I think coming from this being helpful is usually a better position. Of course, that gets you a little bit in an inferior position, but over time.

17:42
people will really acknowledge that. And I think that brings you in a better position to come into this, to present together, for example. But sometimes it also helps to be a little bit pushy. So I’ve seen that, for example, with first and senior authorships on publications, where very often there’s this tendency that by default the physician or the physicians take these

18:11
positions and sometimes it’s really, it’s necessary to put a stake into their grounds and say, okay, the majority for this study or for this paper was actually done by a statistician and therefore that statistician should be first or last author on that. Yeah, that is probably a topic that the pharma statistician is facing more often.

18:39
And for example, the statisticians working in the CRO business. Yeah. I guess so. But on the, but just one more thought on the, working together with the medic and making decisions together. What I also, what I’ve also done in the past is to provide the medic with data to support his ideas, but not to provide everything that he’s asking for. So finding a compromise.

19:08
For example, physicians love p-values. The p-value is a thing that they have a p-value, so they have a, let’s say like a proof, so some kind of significance. And so they are happy with it, but sometimes the data doesn’t allow to provide p-values. And I’ve had many discussions about p-values, providing p-values, not to provide p-values. And I think this is one of the things, for example, where you can find a good compromise saying, okay, this is the data we…

19:36
provide the data, we show what it is, we can explain it, but we don’t give a statistical significance with the p-value saying, now everyone preventing the physicians to prove to the world something which is not reliable from the statisticians. Yeah. The p-value is actually a very good point. So I think that is also where in this relationship you need to push back. Sometimes you really need to…

20:05
stand the ground and don’t just do what you’re told. So as you speak about P values, one thing is for example, P values for baseline tables for randomized studies. And of course the physician will come with some kind of paper where says for each baseline covariate like age and gender.

20:34
treatment and what have you says there’s a p-value and so yes, the other colleagues has done this as well. And we need to have these p-values in our paper as well. No, you don’t. But then really explain it and explain him why it makes no sense and why it may be even perceived as being, yeah.

21:04
being unscientific, I think that is the key thing. So I completely agree from time to time, you actually need to push back. The other thing that you said, I think is also important to not just provide the tables and then think that job done, but provide the tables together with a good interpretation. But of course, for that, you need to really understand what the physician

21:32
needs and to really need to translate it into his language. So I think for that initial listening for what is really the problem is really key. I think this is what you mentioned before with trying to learn the language of a physician and understanding where he’s coming from. And probably before delivering any results or any interpretation of the results to the physician, I think it’s a good idea.

22:01
always spend a lot of time with the physician to understand what he’s looking for. Because just because there’s a sentence in the SAP, let’s say, or somewhere in an email asking you to provide this and this, it doesn’t mean that you understand what the thematic is asking you to do. This is sometimes completely two different languages, two different meaning. So really about going back to a point, which is key in many things is really communication. So really get down.

22:30
with the, with that person, talk to them, understand, maybe repeat whatever he’s saying in your language. So to understand what this is the same, this is the same or has the same meaning as what he’s asking for and provide to the physician what the statistician should do, the data, the outputs, the results and the interpretation for the medic. Yeah, completely agree. It’s really about early checking, common understanding. And sometimes physicians

22:59
may only be able to see it if you present them with a table, even with maybe fake numbers in it at the beginning or made up numbers, just to get him a feeling for how that will look like. Because the vast majority of physicians I have worked with have really big trouble to read the SAP. Primarily because the SAP is not a document that is in

23:29
targeted towards physicians, it’s more targeted towards other statisticians or to programmers. If you just provide them with a 50 or 500 word document of specifications, he will not really understand it. So I think they’re putting it into a format that he can understand it and that he is used to will help a lot to clarify that there’s a common understanding of what needs to be done.

24:00
And if you provide the data to like a dummy data or shells to the physician, explain it to him. Exercise explaining the results to the physician. Because whatever you present in this output, if the medic doesn’t understand, who else should understand it? These guys are clever. These guys are smart. So if you provide an explanation or an output to them without explanation or with the incorrect explanation, and he doesn’t understand, nobody else will understand.

24:28
the outputs, except other statisticians, but other statisticians are not writing the clinical study report. Other statisticians are not writing the publications or presenting the results. These are the physicians. So they need to understand what the statistician is trying to put into a tape. Completely agree. I think of it as the statistician explaining it to the study physician is the first step of getting the results into the medical community.

24:55
So if you think about it from that, it goes into the publication. It goes to key opinion leaders and then it trickles down in the medical society. So if you fail to explain what these data actually mean in the first step, this whole cascade cannot work. So that is really important. Maybe one idea of better understanding physicians is actually to have a physician as a mentor.

25:26
Because that helps you to get a completely different view on the industry and also to get a better understanding of their way of thinking as well as a better understanding of their language. That’s a good idea. If the setting is allowing for it. And I think if you’re working together with the medic day to day, day, over months or years, I think this is a really good idea. Yeah, completely agree. Okay.

25:54
I think we are coming here to the end of our half an hour podcast already. So thanks a lot for listening and talk to you next time. Talk to you next time.

26:11
This show was created in association with PSI. Thanks to Reine and her team at VVS for help with the show in the background and thank you for listening. Read your potential, read great signs and serve the patients. Just be an effective statistician.