From Company Culture to Data-Driven Decisions

From Company Culture To Data Driven Decisions

[00:00:00] Alexander: Welcome to another episode of The Effective Statistician and today I’m super happy to talk with [00:00:10] Matt Davies from Jazz Pharmaceuticals. When I first learned about Jazz Pharmaceuticals, I thought about Jess, that’s an [00:00:20] interesting name for a pharma company, but we’ll touch base on that a little bit later.

[00:00:26] Alexander: So first over to you, Matt, maybe you can introduce [00:00:30] yourself to the listener. 

[00:00:32] Mat: Absolutely, Alexander. Thank you so much for having me. It’s an, it’s an honor to be here. So I’m excited to be able to talk to you. My name is Matt. I run the data science organization [00:00:40] here at Jazz Pharmaceuticals. I’ve had kind of an interesting career journey to get here.

[00:00:44] Mat: I actually started off my career as a middle school math teacher, so I used to teach 7th, [00:00:50] 8th and 9th grade. So 12 through 14 year olds were my, were my sweet spot and did that for a couple of years, decided that I wanted to leave and get into the pharmaceutical industry because I was getting my [00:01:00] master’s in applied statistics at the time, ended up working at a CRO for about seven years.

[00:01:05] Mat: When I realized how much I love this job, but I did not really enjoy the [00:01:10] consulting side of it. I preferred at the time to be on the pharma side. And the only way I knew how to do that was to go get my PhD. So I got my PhD in biostatistics from UPenn [00:01:20] at the same time as I worked full time. So I did school full time and worked full time, which was a nightmare.

[00:01:25] Mat: And while you see all this gray hair on my head, and I finished that up, I [00:01:30] got hired as an individual contributor at Teva Pharmaceuticals. And in a span of a number of years went from a statistician to a therapeutic area head, to the head of [00:01:40] statistics, to the head of biometrics at Teva, before leaving there to join a small company called GW Pharmaceuticals.

[00:01:46] Mat: GW was acquired by Jazz and I was named as the head of biometrics [00:01:50] at Jazz about three years ago, and then became the head of data science now. And just because data science means statistics, different things to different people. We’re responsible for [00:02:00] bioinformatics, clinical statistics, statistical programming, medical affairs and value statistics, real world evidence, and pharmacoepidemiology.

[00:02:06] Mat: So that’s the background and that’s the team. 

[00:02:09] Alexander: [00:02:10] That is pretty awesome that you have all these quantitative scientists all together. Which in many other companies Work quite [00:02:20] isolated from each other and sometimes more against each other than rather than together. 

[00:02:26] Mat: I would share that’s one of the hardest parts that I’ve noticed even in previous jobs [00:02:30] is when you have so many quantitative individuals that aren’t actually working together.

[00:02:34] Mat: It’s actually a huge deal. Disaster and nightmare. You might have two or three different quantitative groups [00:02:40] working on the exact same problem, but they do using different definitions, and they all get an answer that’s conflicting. And then nobody knows who to believe. And it’s not very efficient. So we’re trying to solve [00:02:50] that by harmonizing them under one shot.

[00:02:52] Alexander: Yeah, I’ve seen that before and where, whereas in kind of the different departments had kind of [00:03:00] fight each other, says that well my message is wrong and is right and your is wrong, and you should only believe me, but not these guys over there. And which of course [00:03:10] kind of doesn’t help anybody. Yeah.

[00:03:14] Alexander: So that is actually very closely related to the topic that we want to talk about [00:03:20] today. I recently stepped over an article on LinkedIn that you also highlighted. It is a Forbes article by your CEO [00:03:30] about jazz culture. And I read it was A lot of interest because there’s a lot of [00:03:40] talking about culture and trading culture.

[00:03:43] Alexander: And in the end, very often culture is a product of decades of [00:03:50] hiring, firing and all kinds of other decisions that happen within the company. And for Jess, that’s a little bit different. Can you [00:04:00] tell us a little bit about how Jess. Got its culture. And why is that so important for the company?

[00:04:07] Mat: Yeah. So I’ll start by saying, you [00:04:10] know I respected him. I, or Bruce Cozette, our CEO very much. He says that he started this company based on culture, that he wanted to have a company that was culture first, [00:04:20] and he is a man that has lived up to that. Aspiration. I’ve gotten a chance to observe him publicly and privately and can tell you that this is a man that walks that walk, [00:04:30] which I’ve seen other people that can’t and he and he can and it permeates itself through all of the organization.

[00:04:35] Mat: I’ll just give an example of how that I mean, the Forbes article kind [00:04:40] walks through a number of the things that he thinks are important in the culture, which admittedly is putting patients first. Our purpose is to innovate, to transform the lives of patients and [00:04:50] their families. I’ve memorized that because it’s central to everything that we do, but they also talk about wanting to make jazz the best experience of the career of every jazzician, which [00:05:00] is what we affectionately refer to as anyone that works at jazz.

[00:05:04] Mat: But there was actually a story a couple of years ago. I actually applied for a job at Jazz before [00:05:10] my career led me here. And in the application process, they sent another article that Bruce had written for another magazine a couple of years ago. And it outlined his [00:05:20] expectations in the culture and basically right up front, they said, if your values don’t align with this publication, this probably isn’t the job for you.

[00:05:28] Mat: Because we’re a culture first [00:05:30] company. And I was blown away by that, especially walking out of a culture that I didn’t love to say that it was culture first was just transformative in terms of how I thought of a workplace. [00:05:40] 

[00:05:40] Alexander: Yeah. And I think it’s really, really interesting to have both the the employee focus and the patient focused on it.

[00:05:48] Alexander: Most [00:05:50] companies, especially the bigger ones focus first on the shareholders. Yep. And talk mostly about them. Of course, they also talk [00:06:00] about the, the patients. But when push comes to shoals, it’s usually let’s sell our drugs. That’s right. 

[00:06:09] Alexander: [00:06:10] And so What really kind of stood out for me is that he was, for example, speaking about the the role of the sales [00:06:20] representatives and said their role is predominantly to be educators and help the physicians.

[00:06:29] Alexander: They [00:06:30] support the HCPs. It’s a kind of typical pharma abbreviation to make the right decisions, not to just, you know, [00:06:40] prescribe our drug. And I think this is really very much aligned with what my vision is to provide [00:06:50] the right evidence at the right time to make the right decision. And how is that reflected with Your organization said supports [00:07:00] also data for, for that.

[00:07:03] Mat: Yeah. Well, I think that it’s kind of like the shining beacon that we all work towards. [00:07:10] Right. And it’s a filter that we view every problem that we come against as what is best for the patient at the end of the day. And as data scientists, [00:07:20] statisticians, programmers. The whole umbrella. I think we’re faced with a lot of those problems often, especially if you’re on the more medical affairs value side [00:07:30] of things, because you know that when you’re trying to present data, you’re trying to interpret a clinical trial.

[00:07:35] Mat: You can interpret it through a lot of different lenses. You can interpret it through the lens of what’s best for the [00:07:40] company, what’s best for yourself, what’s best for the product, or what’s best for patients. And you can tell different stories based on that. And we live this [00:07:50] mantra right up front that says we do what’s best for the patients.

[00:07:52] Mat: And if that means that we interpret a post hoc analysis or a clinical trial readout or real world evidence study as [00:08:00] not being favorable to us, but it’s very informative to the patient, that’s really important for us to strongly consider because it’s patient first. So [00:08:10] it, And I’d say the other side of that, maybe sometimes, you know, that there’s always a temptation for us as data scientists to put that little spin or just [00:08:20] that little extra piece of the analysis that would make it more favorable to us to get that e value under 0.

[00:08:26] Mat: 05, for example, when we know that’s not the best way to do it. We don’t do [00:08:30] that. Because it’s not what’s best for the patients. And so I think at least as much as we can try to do that, it really makes us live. Just view all of our data [00:08:40] outputs and communication about our data as being patient focused first.

[00:08:45] Alexander: That’s a pretty cool thing about it because [00:08:50] I know that lots of, lots of people, especially statisticians in that space very much love that perspective yet still [00:09:00] struggle very often to, you know, Fight against company, internal culture. Although of course the official policies always say that [00:09:10] your, our publications are well balanced and so on.

[00:09:14] Alexander: Yeah. So speaking about it and doing it. Are two very, very different [00:09:20] things. So how does that kind of materialize in the everyday doing? Can you give an example? 

[00:09:28] Mat: So, so a couple of [00:09:30] things, I mean, the biggest example I can give is Not that I can give a whole lot of specifics about it, but being in higher level governance situations where you’re discussing [00:09:40] one of these different outputs and there’s a lot of opinions around the table on what we should and what we should do with it.

[00:09:46] Mat: And then having the voice of Bruce or somebody else that steps up and says, [00:09:50] Hey, everyone, We are here to do what’s right for the patients first. So how do we make this decision in terms of what’s right for the patients first? And I think [00:10:00] that’s what’s so unique here. It’s as soon as that sentiment comes out as we’re debating on what to do with this, it kind of reshapes everyone’s [00:10:10] focus.

[00:10:10] Mat: And I think Because we all believe this so much in our hearts, you know, sometimes even if your head isn’t in the right space while you’re making that decision, as soon as [00:10:20] someone says that, you’re like, right, of course, patient first, let’s focus on patient first. And I think it’s just the common vocabulary, vernacular, and also shared [00:10:30] values that say that that’s where we, that should be the guiding light for our decision making process. So that’s just one way that I see it. 

[00:10:37] Alexander: Yeah, I think that is actually a very, very [00:10:40] powerful way. If people always speak up when things don’t go along with this culture, and they are acknowledged, and [00:10:50] then everybody appreciates it, and comes back on the ship. That is really, really important. Yeah, that actually I think makes a big difference.

[00:10:59] Alexander: Yeah, [00:11:00] because there’s always going to be someone that, you know, violates goes off, off track, you know, comes from another company, doesn’t know the culture [00:11:10] yet has been measured differently. 

[00:11:12] Mat: Yeah, and it’s also interesting when that happens. I think the other way that you can determine how well a company is doing with this is what happens in those situations, [00:11:20] right?

[00:11:20] Mat: Is it just swept under the rug? It’s like, oh, that’s just how this person operates. Or, or is there something really done about it? Do people really care? Does compliance step in? Are there [00:11:30] other ways that happens? And it’s like the other side of that coin. It’s so when invariably, There’s a bad actor or something like that happens.

[00:11:37] Mat: What does the company do to help fix that [00:11:40] culture? And I think that speaks volumes on the other side as well, because we all know that’s going to happen at some point. 

[00:11:45] Alexander: Yeah. And I think the, you know, unfortunately we all have [00:11:50] these kind of compliance departments but the more you don’t need them.

[00:11:55] Alexander: Yeah. That’s right. Yeah. So it’s kind of, if. [00:12:00] People within the teams directly correct it and don’t kind of call the compliance number anonymously. That is a much, much better [00:12:10] way and more powerful way to, to correct it. 

[00:12:13] Mat: Yeah, because if you really think about what that means from a cultural perspective, I mean, we want to build a cultural culture where people [00:12:20] feel safe to speak up, that they have the right relationships, that they have the autonomy within their teams to be able to solve their problems and self correct, so to speak.

[00:12:28] Mat: And when you have team members [00:12:30] that respect each other enough to do that with one another, there’s not a whole lot of compliance phone calls in that case, because you’re actually able to solve it. Person to person and be more autonomous to be [00:12:40] successful, which is, which is some of the things that we strive to do.

[00:12:43] Alexander: Yeah. And I think that is a great fundamental thing for you to also have a [00:12:50] place where people strive. Psychological safety is so important for all the different things that we do. Yeah. To speak up when there [00:13:00] is a weird signal and the safety data or when there’s a. Certain parts of the population that look like they get more harmed and then [00:13:10] benefit from being struck or all these kinds of different things.

[00:13:13] Alexander: So there’s conflicting data, yeah, that people can bring that up irrespective of where [00:13:20] they come from and so on. That makes a huge difference. 

[00:13:23] Mat: Absolutely. There’s something that I learned when I was teaching that I use a lot as I think about setting [00:13:30] the culture and leadership and everything, at least what we try to do for our team.

[00:13:33] Mat: And that’s Maslow’s hierarchy of needs. I don’t know if you ever heard of this, but, you know, that concept [00:13:40] of, you know, If for anyone that isn’t familiar with this that’s a gentleman in Maslow that put this out as terms of like, what do you need for somebody to actually be their best self? And we used it in [00:13:50] teaching when you have a child that wasn’t performing at their best.

[00:13:52] Mat: It was like, why wouldn’t they be performing at their best? And the very lowest level of the hierarchy of needs is your physical logical leads like Okay. [00:14:00] You can’t expect a child to do well on a math test if they don’t have somewhere to sleep at night, or if they don’t have somewhere to eat. Like, that’s not a fair thing for them to expect.

[00:14:09] Mat: I think the same thing [00:14:10] happens at work. I mean, you’ve got the physiological needs, which many of our people have that are covered just due to the nature of the industry we work in. But then you build off of that [00:14:20] safety needs, which is where that psychological safety comes in. Right. It’s like, I don’t feel safe to be able to speak up or talk to another employee or be able to work something out within my team.

[00:14:29] Mat: I’m [00:14:30] never going to be successful and be innovative and speak up about the safety signal. And then you build on that love and belonging and then esteem. And then at the very top [00:14:40] after safety, love, belonging, esteem is the ability to innovate. And that’s it. You can never expect people to truly innovate unless you’re [00:14:50] building the whole pyramid up, right?

[00:14:51] Mat: And so that’s where, like you mentioned, trying to really be intentional about how we’re building the culture by making people feel cared for, supported, [00:15:00] belonging, that they have all the tools they need to have the support from us as management. Now you go and change the world. And I think that’s how we really can build a culture that [00:15:10] supports this purpose of innovating.

[00:15:12] Alexander: Yeah, innovation is also mentioned quite a lot in this Forbes article, and your [00:15:20] CEO speaks about taking risks and learning quickly from failure so that you can. Move forward and have success. [00:15:30] How does that work within your organization? When it comes to learning from failure, learning from data, learning from from evidence, [00:15:40] how does that, you know, help you to make better decisions about your pipeline?

[00:15:48] Mat: So one of the things I [00:15:50] love about jazz is the size of the company it is. I think at this point jazz is somewhere between 3, 000 and 3, 500 employees and it’s, it’s well funded. So I think the [00:16:00] good, the good part about that is that we have. The ability and the resources to do a lot of exciting things, but it’s not so big that it’s this huge ship that we have to be [00:16:10] able to steer.

[00:16:10] Mat: So I think one of the ways that we do that is by maintaining agility, not only in our programs, but in all of the different ways that we work. So, for [00:16:20] example, you know, there’s this concept of double blind sponsor open studies. I don’t know how familiar we are with those, but it’s where we try to keep a blinding at the site.[00:16:30] 

[00:16:30] Mat: And the investigator. But then we at jazz keep a number of internal people unblinded to it, like in a phase one or a phase two, a study. So we might be able to look at the [00:16:40] data more often and make a decision about where we want to go. And other companies that concept might take. Years to be able to come up and running and how do we get that straightened out at [00:16:50] jazz?

[00:16:50] Mat: I think it took us about two months to be able to say yes that fits right in with what we want to do We want to make these decisions as we go. Let’s let’s make it happen So I think [00:17:00] that it’s a company that just maintains agility at its core to look at data on an ongoing basis in Obviously scientifically rigorous compliant ways, but [00:17:10] that is not afraid to pivot from a strategy when it makes the most sense for the compound.

[00:17:15] Alexander: Agility for you means making decisions fast, [00:17:20] sticking to it, and then reevaluating as you get more data. 

[00:17:25] Mat: Yeah, I think, I think the biggest focus is the speed of, of the [00:17:30] speed of data and the speed of decision making. The sticking to the decisions is key. is always something that we’re continuing to grow in.

[00:17:38] Mat: But I think definitely the speed [00:17:40] and how we make the decisions is really important, which lets us to use to not only progress our compounds the fastest, but also allocate our resources to different things. Because obviously [00:17:50] we’re not You know what? We have a lot of resources. They’re not infinite. So in some cases, a decision on one program means that we have to shift our resources toward a completely different program.

[00:17:59] Mat: [00:18:00] That’s something we’re doing well now and able to shift things across as needed, and the only way to do it is through these more rapid decision making processes that the data scientists [00:18:10] actually are leaned on quite heavily to make sure it has that rigor and has the right information that’s interpreted correctly so we can make those larger portfolio decisions.

[00:18:19] Alexander: So [00:18:20] that’s really interesting what I know from lots of other companies said to make a decision. You usually [00:18:30] go through many different layers of governance. Yeah. How does that work at Chess? 

[00:18:36] Mat: We have two layers of governance. So [00:18:40] that that’s your answer. So we have a layer of governance that focuses more on the on the portfolio level of governance in terms of what are the [00:18:50] large decisions that we need to make for this compound.