It’s surely not the norm, that a statistician from the pharma sector ends up working for BBC. But Robert Cuffe has taken this step and needs to explain statistics to non-statisticians daily.

In this episode, he shares the different methods he’s using. We also dive into common pitfalls, that we as statisticians trap into.

Finally, we also learn, how we can improve our publications to make sure, that the science is reported correctly in the broader news.

Robert Cuffe

Robert used to work in the pharmaceutical industry – at GSK and ViiV. During that time he was a chair of PSI. He now works at BBC News as the head of statistics. 

Transcript

What you can learn about communication of statistics from a statistician working for the BBC

00:00
You’re listening to the Effective Statistician Podcast, episode number 47. What you can learn about communication of statistics from a statistician working for the BBC. Welcome to the Effective Statistician with Alexander Schacht and Benjamin Pisgill. The weekly podcast for statisticians in the health sector designed to help you understand

00:24
improve your leadership skills, widen your business acumen and enhance your efficiency. The PSI conference in London is coming up. It happens from the 2nd to the 5th of June in London. The early bird deadline is only until March 20th of 2019. And if you haven’t already registered, do it now. I’m also there at the conference and will present there as well. So…

00:54
Come to London and then we can meet there. It’s really the best conference for statisticians that I can think of, especially for those that work in pharma or CRO. It’s really applicable to your day-to-day work. There’s no kind of ivory tower research in there. It’s all very, very relevant and practical. So just go to psivab.org, check out for the conference and register today.

01:21
In this interview today with Robert Kelff, who is actually one of the keynote presenters at the conference, we talk about what we can learn from him being a professional communicator and explainer of statistics to the lay community.

01:42
what we can learn in terms of training, non-statisticians, and what kind of things we need to take into account when we work on publications, when we communicate data from clinical trials, as well as what should be done.

02:03
You know, a little bit of a behind the scenes view in the BBC, which is also pretty interesting. So stay tuned and have fun listening. This podcast is created in association with PSI, a global member organization dedicated to leading and promoting best practice and industry initiatives.

02:24
Join PSI today to further develop your statistical capabilities with access to the special interest groups, the video on demand content library, free registration to all PSI webinars and much much more. Visit the PSI website at psieweb.org to learn more about PSI activities and become a PSI member today. By the way, if you register for the conference, you also become a PSI member.

02:58
Welcome to another episode of the Effective Statistician. And this time, I’m here with Robert Cuff. Hi, Robert. How are you doing? Very well. OK, very good. So Robert has taken on a very, let’s say, unusual career step as a statistician from the pharma sector and has now started to work for the BBC for quite some time.

03:25
How does that happen and what attracted you to this position? Well, I guess it had been something I’ve been dabbling with for a couple of years. So when I was involved in PSI, one of my first jobs there was to set up before the headlines. So that was a briefing service for journalists working in health, where we got a group of statisticians who would take a look at any of the papers that came out on their embargo.

03:55
describing some new interesting health story. And we’d look at whether the science supported the claim. So had they just for contenders properly, all those kinds of things. And the reason why we set that up is because almost every science story, almost every health story involves statistics. And there are experts who you can get in for quotes, you know, who know about the field. And of course, there are expert statisticians, but you can’t be calling on David Spiegelhalter for every single

04:25
science story and media. So we thought, because a lot of this is stats 101, you know, you’ve been working in pharma for a few years, you can probably do it quite well. We worked together with the science media center to train up a cohort of statisticians who would be able to respond to those calls. So that was one thing I did. A second thing I did was work with the Royal Statistical Society. They trained up a cohort of stats ambassadors to go out and talk about stats and do public talks and work with the media and train press officers and do those kinds of things.

04:54
And so I had all this stuff going on in the background. And I was leaning that way a little bit. So, you know, I was enjoying that work quite a lot. And when the BBC advertised for the new post of Head of Statistics, I thought I’d give it a go and the interviews went well and here I am. Okay, very good. And you have been the chair of PSI for some time. Did that play a role in your interview process? I think it was more the work that I’d already done with…

05:23
the science media center and the kind of the kind of the press facing stuff. It’s certainly true that being involved in PSI and being involved in the RSS and volunteering led to a lot of the the opportunities that got me the job. That’s unquestioned to be the case. In fact, I mean, almost every, every break I got both when I was working in pharma and with my move into into the media.

05:50
It’s kind of as a result of something extra that I did and a lot of that is, is I put down to PSI. Yeah. That’s actually a very, very nice learning for all the listeners that, um, there’s lots of opportunities that just open up through this volunteering work. And, um, you don’t know exactly what kind of opportunities you get, but, um, it’s for sure quite a nice investment into PSI. You may be biased in saying, I think it’s true.

06:20
course it’s true, you know, when you’re working on the projects, you get to see a certain level of responsibility and it’s only when you join these committees that you’re building all those soft skills, you know, getting people to kind of care around what we want to do. That will be useful to you when you step up to the project role or you’re stepping into management. So you do get kind of training wheels way of building those skills. So yes, and you never know what things will flow from that, so you just kind of do what you’re interested in and see where it goes.

06:50
For sure, you get much more kind of external exposure through that. So that’s for sure. So as an employer, you might not want that, but it’s not one. On balance, it’s better that your staff are well trained and learn that thing. So when I was in PSI, I found that most employers were willing to, you know, take the risk that someone would get poached in order to ensure that they’re better employees.

07:16
Yeah, and of course, there’s also a benefit for the employers that they get more visibility through that as well. So, yeah, I think. Yeah. In terms of your new job at the BBC, what are you doing there from a day to day basis now? Helping journalists find and tell better stories with numbers. And how does that work? Is then these journalists, do they, you know, just…

07:45
step by your desk and then ask you questions or how does that work? It’s a couple of different things. So one is yes, I will absolutely help journalists on individual questions. They have a survey has come in or they’re doing a freedom of information request and they want to understand what they can say about that or how they can explain the numbers appropriately. I’d also sit in on the editorial meetings. So the big stories that are running through the day or the week, they’ll be…

08:14
of a conference that will happen at nine in the morning and then again at three in the afternoon where all of the different programs in the BBC and on all of the different specialists groups get together and say, you know, what, how are we treating this story? Are we making sure we’re doing it right? And that’s my opportunity to come in and say, well, do you know what, I’m not sure we can go on this or we need to put this caveat in. Okay.

08:42
of culture and skills building part as well that you’d be familiar with. And, you know, it is part of any large organization. So the BBC is a massive employer of journalists. There are some people I’ve come across who were just amazing on numbers. I mean, they, they, you know, they don’t need a statistician to help them understand what, what they’re doing. And of course, there’s a variety of, of confidence and expertise around numbers. And some people do need a lot of help, even with the very basics of calculating a percentage. And so.

09:10
It’s my job to make those people more comfortable. And obviously I can’t reach everyone when there are 2000 journalists in the company. So it’s about creating networks of people who can help each other, building training materials, all those kinds of things to help to kind of up the base level. So the kind of training non-statisticians on statistics kind of things that you would do in pharma quite a lot as well, yeah? Yeah, yeah.

09:38
And, you know, it’s, it’s already the case at the BBC would train their journalists on how to think about surveys, how to think about numbers, how to, to be confident. It’s kind of one of the early things I did when I joined the BBC was go on their, their week long training program for, uh, for new young journalists. And they did spend quite a bit of time thinking about numbers on that. Um, but it’s making sure that you kind of ingrain all those lessons again and again and again, to make sure that they really, they really take hold. Okay. Very good.

10:08
In terms of your job now, you also sometimes appear kind of live or recorded on the news yourself, not just through journalists. Did you get any special training for that? Oh, God, yeah. Yeah. So even as part of my interviews, in fact, they brought me in. So the first interview was a normal kind of panel. The second one, I sat down in the main

10:37
BBC News Studio where they put out the six o’clock or the 10 o’clock news. And we did a mock interview with one of the presenters. And I have to say, first of all, obviously absolutely terrifying are scary enough, but if you, you know, your, your dog prospects, hang on it. It’s, it’s even worse. And secondly, it was amazing. The skill of the presenter, uh, she was phenomenal because she was so charming and so good at putting me at my ease.

11:07
You know, when somebody is really listening to you, you know, as you’re getting ready and you’re sitting, you’re sitting in the seat, getting ready to be interviewed. And someone just, you know, ask you questions. Are you two working in the pharma industry, work in health? What do you think about how we, we cover stats or, you know, what kind of things do you do? Um, and then really listening to what you’re saying and then asking you another question, that’s a, you know, a follow on, they’ve, they’ve thought about what you said, so really kind of, you know, you feel like you’re in the spotlight of, of their attention, but at the same time.

11:34
You know, she’s threading her microphone through her sleeve, hooking it onto her ear, talking to the people in the gallery who were controlling the program and the editor. So while I’m talking, she’s saying to somebody, okay, so we’re gonna do it into this camera, the lights are coming down from here, we’re gonna cover that question first, and then turns back to me, and gives me this quite insightful question about what I’ve just said. It was amazing that she just had so much charm to burn that she was able to charm me while doing something else. And that’s actually sometimes almost

12:04
the opposite of what I need to think about. Because for me as a statistician, most of the numbers in news, it’s not marginal structural models. We’re not doing these, you know, inverse probability weighted analyses. It’s about making sure that numbers are communicated simply and clearly. And we can talk about that a bit more in a minute. So even after I had a baby and I came into work two weeks later, having had maybe three hours sleep over the last fortnight, I could do the numbers fine. That’s not a problem. But the hard part.

12:34
is communicating it right in a way that will land with an audience. And that’s where I’ve had to learn an awful lot in the course of my first year or two here. But you asked me about training for, for, for interest. Yeah, yeah. Yeah. So yes. So one, as soon as I arrived, they sent me on courses writing for TV and radio, how to do two ways, two ways, like an interview where you’re sitting with somebody and they’re just asking you questions. So

13:03
how to do that well on television, how to dress for telly. We had some fun where I was standing in the, in the main newsroom again, you know, waving at numbers on a, on a screen and guiding people through it and still looking at the camera. And there was another exercise where they, they got my auto queue to break halfway through. Okay. And I had to keep on going. So lots of things where they just test you a little bit and push you outside of you, you know, this is what you’re supposed to do. This is what you do when things go wrong. Yeah.

13:33
In terms of that, you mentioned at the first thing you did some training on writing. What was kind of your key takeaways from this training on writing? I think the main one is, it depends on what you’re writing for. Okay, so having your audience in mind is the first one. Yeah, always have your audience in mind, but not just your audience, your medium.

14:00
So if you’re writing for radio, it’s different to if you’re writing for television or for online. So for online, you can get all the facts in in quite a slow way. If you’re writing for TV, you need to really hit people very quickly. So there are different ways of doing that. And the thing I would say for statisticians that, you know, working in pharma, the most common situation when you’re writing a script is when you’re doing a presentation.

14:30
The most useful thing to take from it is write out the words that you’re going to say, but not on the slides. Yeah. And then read them out. Make sure it sounds natural because it’s only when you stand up, book yourself a meeting room to yourself where you can stand up and read it out loud because it’s only when it’s out loud that you start hearing, oh, that’s not really how I talk. That doesn’t actually sound natural. So that you then have your script for your talk and your slides can focus on showing people the data rather than giving them the bullet points that you speak to.

15:01
Yeah, yeah. I always think that for presentations, the slides should support the speaker and not the other way around. So that’s just one important thing in terms of the other thing is we usually see far too many words on slides and I’d like to have, you know, that condensed, condensed, condensed and then condensed again and then it’s usually right. So, um, Yeah, but you still need to support yourself if you’re not super confident speaking.

15:31
you know, have the things that you’re going to say. And so thinking about how you write that, maybe you do start off with that very wordy set of slides, speak it out loud, make sure that works. And then you can condense the slides to make sure they support you, as you said. And maybe also condense your notes so that you’re more confident and you can, you know, you can appear more fluent rather than holding a page up in front of you. Yeah. In terms of that, yeah, it’s about practicing things and

16:00
repeating it and repeating it. Okay. And yeah, repeating maybe also for the unforeseen things. So preparing for that is also very, very nice takeaway. So like, you know, if there’s some technical hiccup in the recording process, like at the BBC or wherever it can happen, things like that can happen at on stage as well, if you’re presenting. So that’s a good thing. Sweetheart.

16:29
Even in my time here in the BBC, we’ve recently introduced a new system that supports all the running orders and scripts for the shows. And within the first couple of weeks of that, it crashed at about three o’clock in the afternoon. And so all the team preparing for the 6pm news had to leave the building from which it was going to be broadcast and moved down to our other studios in Westminster and start from scratch and build the program there. Oh my God.

16:58
thing. Do you know what it is? It is amazing. To me, this is something I kind of communicate about people who about the stuff that happens in the news. It is amazing that any live broadcast ever happens. Because there are so many things that can go wrong at any one time. You know, just you know, somebody needs to kick one plug, or somebody somebody needs to be smoking in a toilet and start a fire evacuation from the building. Yeah, yeah. And you know, the whole thing can fall over so that anything ever

17:27
gets broadcast is amazing. And it’s because people are working their backsides off to manage all of the different pieces that need to come together. And it’s important to understand that when you’re dealing with the media, because when they say they’re on a deadline, they really are. They’re kind of up against it for live broadcasts a lot of the time. Yeah, that’s a completely different deadline set we usually work to.

17:52
Oh, yeah. So that was a big realization for me when I came here. So I’m used to, you know, you’re working on a drug project where you set up your trial and a year and a half later, it’ll report out. So you prep your tables, you get everything ready to go on, and then you kind of crunch it for a couple of weeks when the data come in. So you’re talking about years for your deadlines, really, or at least months, whereas here it is today. And it’s not even a deadline. You know, you’re going on air at six o’clock and six o’clock doesn’t really care if you’re ready or not.

18:21
that’s when you’re going on air. And those very fast turns around and non movable deadlines, they change the way that people think about time. Yeah, absolutely. And if you can hear someone crying the background, well, that’s my little one just waking up. Okay, so in terms of training others, that’s important, but…

18:50
It’s also about kind of statisticians that write up the source material. What can statisticians do to better communicate clearly and in an easy way in their presentations and especially also in the publications, in the manuscripts? Okay, there are two different things there. So in manuscripts…

19:16
When I think about this, I think about communication to the lay media. And the work that statisticians do go through a series of filters before it makes it into a news article. So you make a contribution to the paper. There’s a press release written around that. There may be some work the journalists do to find out more about the data.

19:42
and some questions they have to ask about it. And at each of those stages, the statisticians can have an important contribution. Let’s take one example. The commonest problem that we have in a lot of reporting of health stories is that focus on the relative instead of the absolute difference between two interventions or two conditions. Whenever I speak to journalists about it, science journalists who’ve been on the

20:11
know, within about three months, they know, even sooner than that, they know they should be looking for the absolute risk difference and they should quote it. Fine, you might put the bigger number in the headline, but they want to quote the baseline risk in the story. And this probably isn’t an issue for us in clinical trials as much as it is in maybe in epidemiological studies. But they ask, they ask the press officer or they ask the lead author of the study, you know, what is the base rate in

20:40
for the population under study, or what is the absolute risk difference? And they don’t say, they won’t tell them or they say it can’t be calculated. Now as statisticians, we know, unless it’s a case control study, you can work out from the population you’ve studied what the base rate is. And so there is where a statistician can be enormously useful and influential by being available on the day. When the press release goes out by, you know, making sure that they get that information into the manuscript.

21:10
by leaning on their lead investigator to make sure they’re willing to give that information out if asked. So there are a couple of ways in which a statistician can really help on those main things that people don’t understand. So working with the first author of the paper together to help them kind of answer such questions in a good way, that’s one of the things to think about? Yeah, absolutely. Be available. Yeah. Yeah. And that takes all that kind of…

21:39
you know, developing that relationship with the lead author or with the press team. You know, to show them that you’re useful, you’re somebody offering solutions, you’re adding to the output and the likelihood that this thing is going to get taken up because, you know, it’s the message I’m giving. And in fact, the health journalists who were here long before I ever arrived are giving people is we want to know what that information is, and you’ve got to include it. So I think that’s a yeah.

22:08
That’s one of the ways in which statisticians can be very helpful in the communication of, of data in the, in papers. Okay. Very good. Yeah. I said, I’m always a big fan of reporting all this together because when you just hear, well, the relative risk is two or something like that, I don’t really get personally a lot from that unless I’m an expert in the field and I know what the base rate is.

22:37
But otherwise, it’s really useful for that. So another topic, and just to go on that a little bit more. So that’s one thing that’s people important for everyone to understand, you know, exactly how big a problem is this. But the difference between the relative risk and the odds, you know, I wouldn’t obsess about that or the hazard ratio. When it may be important in a technical context, but when you’re talking about the lay audience, unless it’s one of those situations where you’ve got like a very, very frequently occurring.

23:07
condition, as long as the odds is a reasonable approximation of the risk to within, you know, if it’s 2.3 instead of 2.4, who cares? You know, you’ve got to focus on the stuff that will drive the most important part of the story as well. So we have sometimes have to be a little bit flexible in terms of what we think about the audience and the important caveats and numbers to get across.

23:37
precision and communication in such a way that it’s easy understood. Yeah. Yeah, exactly. Okay. And the last point that I would like to speak with you about is actually your presentation at the PSI conference. So you are one of the keynote speakers for this year’s PSI conference in London. Anything you would like to…

24:05
talk to our listeners about that, what kind of any kind of things that would encourage them to listen to it.

24:18
going to be amazing. Actually, if you have listened to Robert on a presentation at a PSI conference, he’s a very, very good presenter. So it will be pretty surely amazing from a delivery point of view. But from a content point of view, what will you talk about? Well, so the first thing to say is we’re going to be turning off the recording. It’s not something that’s going to be available online

24:44
This is an interactive experience where I’m going to be putting the audience in the newsroom and asking them to write up a breaking news story on the basis of some stats. So they’ll get that real feeling of what it’s like to be a journalist and what it’s like to work in the media. And there’s not much point to just watching people do that on a video because you’re watching people just sit there and panic for five minutes. Doesn’t make very good audio or television. And the second thing is we want to talk in detail about some of the kind of the more entertaining examples that we’ve…

25:14
we’ve discussed in the BBC. Awesome. I’m really looking forward to seeing you in London and of course, I’ll be there as well. And thanks so much for the interviews. That was a very, very special kind of situation for me to actually interview someone that is professionally working for Broadcoast Organization.

25:40
while dangling a baby on one arm. Actually, while dangling a baby on one arm, Zed didn’t sleep long enough for this interview. Oh, no. Was she supposed to do her two hours? OK, thanks so much. I’ve been there. This show was created in association with PSI. Next week, you’ll learn about a survey on data science, so stay tuned for that one. Thanks for listening.

26:09
Just visit thee to find the show notes and learn more about our podcast Boost Your Career as a Statistician in the Health Sector.

Join The Effective Statistician LinkedIn group

I want to help the community of statisticians, data scientists, programmers and other quantitative scientists to be more influential, innovative, and effective. I believe that as a community we can help our research, our regulatory and payer systems, and ultimately physicians and patients take better decisions based on better evidence.

I work to achieve a future in which everyone can access the right evidence in the right format at the right time to make sound decisions.

When my kids are sick, I want to have good evidence to discuss with the physician about the different therapy choices.

When my mother is sick, I want her to understand the evidence and being able to understand it.

When I get sick, I want to find evidence that I can trust and that helps me to have meaningful discussions with my healthcare professionals.

I want to live in a world, where the media reports correctly about medical evidence and in which society distinguishes between fake evidence and real evidence.

Let’s work together to achieve this.