Rising Stars, Hidden Jewels and Unicorn KOLs – Can We Identify Them and Is It Useful?
One of the great challenges in Medical Affairs is understanding our opinion and communication networks. Over the last few years, the diversity of stakeholders involved in communicating, advocating, treating and responding to our work has become more complex and dynamic. At the same time, communication routes have proliferated and opened, taking us away from our traditional channels. This wealth of data can be overwhelming.
In this podcast, Carlos Areia and Mike Taylor – two data specialists at Digital Science – discuss the ways in which Medical Affairs professionals can use this data to discover your next generation of researchers, those influencers whose voices are heard on Twitter, and how your KOLs are interacting with others online, and which publications they’re discussing.
Following is an automated transcription provided by otter.ai. Please excuse inaccuracies.
Garth Sundem 00:00
Welcome to this episode of the Medical Affairs Professional Society podcast series: “Elevate”. I’m your host, Garth Sundem, Communications Director at MAPS. And today we’re talking about using data to discover rising stars, hidden jewels, and unicorn KOLs. Joining us are Mike Taylor, Head of Data Insights at Digital Science. And Carlos Areia, Data Scientist, also a Digital Science. This episode is sponsored by Digital Science. So, Mike, it used to be that we would just go to ASCO and look at the stage and that was our KOL. Or we would do the same, digitally go to PubMed and see who was most listed on any given topic. So why now? Are we able to look beneath these obvious sources of KOLs, to these rising stars, hidden jewels and unicorn KOLs?
Mike Taylor 01:01
In short, gossips, don’t think to sites, as you say, as you say, you know, identifying the top 50, top 100 in a therapeutic area is is really quite straightforward. And certainly, my experience of working with people in Medical Affairs is, you know, if you prepare listening for that listener functions, and they’ll go, I know that person, I know that person, I know that person, they’ve retired, I know that person I know. But right, producing the top 50s is is is trivial. In several senses. Whenever you’ve got the data, you can you can find the top and the top 50 or, or were interested in supporting people to go beyond that top 50. The problem is, from a raw data, data current, if you lift off the top 50 or top 100 of us therapeutic areas, you might find 1000 people, 5000 people. The question for us is, how do we identify the rising stars or the people beneath that. And the way that Carlos and I decided to look at this was by looking at the relationship between those top 50 those those leading KOL cells to see how they’re reacting, how they’re interacting with the next 1000 The next year? So Carlos has this idea about seniorship in mentorship that, that we would love to talk about.
Garth Sundem 02:30
Well, let’s definitely do that. So that’s really interesting that, you know, it’s not the challenge of finding the next pool of 1000. The challenge is sorting that pool of 1000, to look at who you can interact with to truly drive your strategic actions. And you’re saying that that depends on their relationships with the top 50, Carlos? Is that right? Or are we just evaluating these next 1000? on their own merits?
Carlos Areia 03:06
Well, it’s a bit of both. So what we have this idea of like, first, how do we define who the leaders are inside, for example, a company or department or a whole therapeutic area. So within those top 1000, they’ll probably be a top 10, for sure. And those top 10 may influence the rest of the rest of the 1000, or may have may ever say on the overall picture of the therapeutic area or product. And so what we thought was, how to use the data that we have available at the moment to try and define and try and automate and calculate like some sort of score for senior shape, how do we define senior shape in terms of the world of publication and influencing inside a particular area or field overall? So yeah, so we’ve been tackling this challenge over the couple over the last few months. We still are, but we are making great progress on it and being able to establish this level of leadership. So we can then go back to what Mike just said in terms of mentorship, because once we know who are the leaders in the field, we can then try and see and identify who they are connecting and link in with in terms of not only as colleagues but also as mentors.
Garth Sundem 04:30
Because so so you’re taking the leaders and almost using them as a gatekeeper to tomorrow’s leaders with the understanding that today’s leaders have defined who will be tomorrow’s leaders and if you can look from who’s leading today, you can find out who will be leading tomorrow.
Carlos Areia 04:55
Yes and no. So yes, because that’s one way to identify by them, but we are covering and I guess today we’ll talk about mainly three things will besides all the side care well, so if Well, no, like you mentioned, and Mike said, y’all know, okay wells work, where to where to find them in conference in ASCO etc, etc. This is like a data driven method, another method to try and identify not only KOL, but like leaders in certain areas and mentors in certain areas. So responding to your question. Yes, that’s one way of doing it. And we’ll also cover it today on the rising stars. But we also have other metrics to try and identify them, even without the mentors just through their own work. And then you can link it back to everything together like to like their supervisors, for example, in a PhD or to the lab to head off the lab, etc, etc. So there’s, there’s definitely, very, several ways to identify not only this SR KOLs, but also the early rising stars.
Garth Sundem 06:02
Okay, interesting. So we’re gonna look at these, then it sounds somewhat independently rising stars, hidden jewels and unicorn KOLs. I was thinking that there would be one way to find all three of these classes. But it sounds like we might be talking about how to find each of these individually, Mike, rising stars? How do we find why do we want them? And what do we do with them?
Mike Taylor 06:28
Well, why we want to find them is because, as we said, we all know the killer whales, it’s, and if we look at, say, if we look at it, if we if we calculate someone’s profile, so if we look at their entire publication history, now we can see that someone’s career might be 25 years long, 35 years long, in terms of their publication in their first paper, to, to the papers towards the end of their career. And during that time, we can look at the trajectory that they have, where they’re who they’re working with, what products, what molecules, they’re, they’re working with, who they choose to work with their relationship with institutes like Dana Farber, we can tell all sorts of things about someone’s profile as they’re developing it. But of course, if someone’s going to be publishing for five years, then you’ve got much less data. So we need to look at the kinds of characteristics that make up success in 2023 2018. And see if we can essentially make a prediction on behalf of people who are trying to build up relationships in the future. So we use all of these datasets in combination with each other. And, and as you say, I mean, all of these different things, rising stars and mentors, and, you know, unicorns, and DOL, all of these things are different. And all of these things are also related. You know, this is a case, particularly for younger people where their online profiles, what they do on Twitter, is intimately related to the work that they do. And that that act of choosing what you what you cite, and what you tweet about. It’s not the same, but they are related to each other. So the work that we do is all about building up these very complicated data models. But ultimately, putting something simple in front of you, you know, we’re not going to be saying, These are the next 50, or these are the next 20 People who are going to be, you know, the stars of this field. But you know, maybe it’s more a question of saying, this is the top 50, top 100 that we think look like the stars of the future?
Garth Sundem 08:40
Well, I’ll tell you what, if I was a young researcher, I would want to know if I were a rising star according to your metrics, or if I should look for another career in carpentry or something. So okay, so rising stars, we, I mean, do you give them a score? That this is sort of what I’m wondering, I mean, do you take these top 50 and say, you know, this person scored 78.2 on our metric, and so is predicting long career and you should be interacting with them is, is that works?
Mike Taylor 09:14
It’s always going to be a probability. Oh, sure. Yeah, of course. Right. It’s always we’re never going to say, you know, 78.21 means something other than we think that the signals indicate that they are more likely to turn that that kind of thing. I mean, typically, when we talk about senior ship, mentorship, we’re talking about here is the relationship between a junior researcher and the senior researcher. I know we’ve got quite a lot of data points in how we describe that relationship between Sr. And let’s see.
Garth Sundem 09:52
Well, okay, so, Carlos, let’s get into hidden jewels. How are these? First of all, how are these different than the rising star? hours? And maybe then how do we find them differently?
Carlos Areia 10:05
Oh, yes, and this is getting back to the to the price. This is somewhat related to the rising stars, it’s like those hidden jewels like because my like we mentioned, ladies, we also have like calculations between a relationship between a supervisor and a student or a mentor and a researcher in in a particular field. But there might also be some hidden jewels doing their own work or not as highlighted or, but the outputs are there, the publications are there, and the data is too. So that’s something that I’m still working at I’m I’m actually really literally worked on this just before I joined this call. So what we are doing is we’re trying to create a model to identify like a growth over the years, so still related with the rising stars, but in a more independent way, and not exploring so much the relationship without within their network, but like these, like lone wolves are not as exposed, and still being able to capture them. And this would be not in the total amount of like, research that they have, because they might be not enough different fields. But like the growth in over the past years, like not only on publications, but for example, the attention of this publication if they’ve been treated by by therapists, particular therapeutic specialists on Twitter, or if it’s been used on Wikipedia. So we’re using all sorts of data and the variability over the years.
Garth Sundem 11:39
Mike, go ahead.
Mike Taylor 11:41
Yeah, sorry, I, a couple of years ago, I was working with a client in Switzerland. And they were interested, they are interested in a very, very narrow area where there’s not been a breakthrough for years. And I, we I came up with this notion of citation ship, which is the slightly odd one, because what we’re trying to do is to identify people who were working in a space that weren’t getting a great deal of attention, but where they weren’t getting the attention. They were getting it from very influential people. Oh, yeah. And what we did, what I identified there was to identify a group of Chinese researchers, and I’m not gonna say what area they’re working in. But there are a group of Chinese researchers that just didn’t pop in the data when you look at it. But when you looked at who was citing them, it was very important people who have been linking out their data. Now they had done anything like CO authorship with them or anything like that. But for me, that’s the kind of really interesting signal that makes you look at it and go, what’s going on there. Now, I’m not saying that that group are solving a problem, or they’re getting anywhere, what what my argument would be is to say, if you’re out there, and you’re working in this field, you ought to look at these chaps, these these people, because it looks as though other people are looking at them, see what they’re doing. And that’s the kind of signal that we’re trying to surface. When the crowd all goes, oh, you know, all heads turn, like, oh, did someone famous just come into the restaurant?
Garth Sundem 13:20
Well, and in this analogy, you know, it seems like it’s not all heads turning. It’s just a good point. Yeah. It’s the heads in the know, you know, who are recognizing the people who aren’t celebrities, but who they respect. And so again, you’re looking at these, these influence networks, in which the the KOL is the ones that we would identify easily, or almost vouching for people in different ways. You know, they’re vouching for the people that they’re directly connected with. These are the rising stars kind of, and they’re also in a way vouching for the people who they’re not necessarily connected with, but who they’re citing who they respect. And that would be hidden jewels. Is that, oh cool!
Mike Taylor 14:13
Okay, that’s a way of finding hidden jewels and, and also, of course, what we’re doing on Twitter. Now, one of the nice things we mentioned unicorns, right, so we got a unicorn, a unicorn. This case is someone who is both a KOL and a dol. So if we just unstripped that as well as being key opinion leader, in other words, preeminent finishing or academic, they’re also a digital opinion leader. And we see this. We see this across our data sets. So Carlos did some did some work a year ago, and we identified with really high probability about two thirds of a million KOLs who are also active on Twitter. So if you like We can we can find Dr. Areia. And then we can map Dr. Areia over to Carlos Areia on Twitter. And that means that we can track between the two profiles. So we can we know what that they’re working with. We’re not they’re publishing, we know how people are talking about their publishing, we know who their co author with. But we also know what they recommend who they’re recommending. So we’ve previously talked about sentiment analysis. So we can see whether that cancer writer who we noted the doctor around on dimensions, we know that he is promoting a piece of research that was he’s, perhaps I shouldn’t use the word promoting but recommending a piece of research and who he’s recommending. And this is another way of identifying those all important heads turning, if you like, heads turning on Twitter, in this case.
Garth Sundem 15:55
So that’s so interesting that you know, it’s almost like a bottom up and a top down approach. You’ve got rising stars hidden jewels, then you’ve got these massive unicorn tail wells. I mean, Carlos, the unicorn KOL seem like they would be fairly easy to identify just because they’re high profile in both spaces, you know, the research space, and maybe in the social space or the digital space. But, you know, are we interacting with these unicorn kala wells differently? Or are we asking of them different things than we would have in the past? Asked of, you know, nondigital KOLs? Or I’m wondering, what is the use of these unicorn KOLs and does it provide opportunities, but also some strange challenges for Medical Affairs teams having to dip their toes are more into the digital space?
Carlos Areia 16:52
I think it’s a bit of both, but I think it opens definitely opens a multitude of doors. And just what we’ve been saying, for example, if you find unicorn KOL, and is actually really active on Twitter, you might help in circling back to what we just discussed in my help finding this, even Jules because if he’s tweeting, maybe not citing, but tweeting about particular paper, from a very new researcher in the field, that might mean something. And might that might be, that might be a signal on its own. And getting back because then we can circle back if both are in our data set. So for example, imagine that the very senior leader KOL just tweets about a particular piece of research. And it’s retweeted by like their followers, et cetera, et cetera, on this very new, early career researcher getting to this field. And then we managed to find that research on Twitter, link it back to his research and see all the papers and all the research activity that he has. And we keep, like within this circle covering a multitude of data that can open several doors and when we can very easily get into a rabbit hole here on both ends either on either on the senior ship, your senior ship side or on the early career research side and but but the data is there and we have it and it’s just a matter of like linking the dots
Mike Taylor 18:24
We should push a little back a little bit. It’s not easy because you know, you might have a Dr. Burton on in dimensions or in PubMed. But Dr. Burton might easily be oak doc on Twitter. And identifying those two entities are the same with hyperbility is it took a little bit tricky but together we had over a cup of coffee, we had a bit of a brainstorm about the data underlying the data and some of the trends that we see on Twitter and I’m not going to talk about the the magic the magic meets there, well, the Quickie corner whatever it is, but that was how we got to two thirds of a million and we we think we’re probably going to be up push it up to a million which would be quite a breakthrough. Because you know, we are identifying people who are very active on Twitter and not yet that productive in terms of research. So you do get people on Twitter who are doing doctorates for example, so, you know, they’re not yet sort of even in that top 1000 But we’re very engaged
Garth Sundem 19:28
And it seems like a real innovation you know, Carlos, I thought you were gonna say okay, the thing to do with a unicorn KOL is to you know, equip them with the context of science and understanding that they need to drive the clinical and scientific conversation, you know, in a in a disease space. And where you went immediately was, you can use and use is probably not the right word, but but you can work with unicorn KOLs to identify other que LLS and, and so you’re using your KOL is not only to amplify your your messages and to drive the scientific and clinical conversation, but as part of the network to then you know, have a door into find other cables, you know is that the innovation of your data science is using connected networks to find the people that you need to talk to?
Carlos Areia 20:36
Well, like I said, like we can get into very different rabbit holes here. So for me in because I’m actively working on this almost on a daily basis, I’m interested in using this data to find more they find more, they tend to find, like new singles that I’m looking for. But as you correctly said, it is very meaningful. It’s like having these unicorns and see what they do, not only what they published what they research in, but they also talk about and what they and the sentiment behind it can also drive the conversation forward. And I think that’s that’s I think that’s you rightly said. So that’s one of the main things and main benefits of this data.
Garth Sundem 21:15
Go ahead, Mike.
Mike Taylor 21:17
Yeah, hey, we’ve come such a long way. As just reflecting four years ago with the beginning of COVID. You know, we were all using zoom, was it four years ago, or three years ago, I’ve completely forgotten. I had less gray hair anyway. But so So three years ago, people weren’t really weren’t talking about what was going on in Twitter. Even though we were seeing an explosion, we saw a near doubling of the amount of people who were linking to research and triangles on Twitter. In that March, a number by the way, which has been maintained to this day through all of its rambunctious changes that we’ve seen on Twitter the last few months hasn’t hasn’t blinked in terms of the volume of tweets that we’re getting. But you know, back then, three years ago, people weren’t talking about this at all, even my most advanced contacts, were sort of saying things like, well, we’re looking at it, and you can’t move now. I mean, you you and I were at MAPS and Lisbon, and everybody is talking about this, and we’re talking about tick tock, they’re talking about all of these different platforms where people are engaging. And of course, we’re hearing the Omni work, right. And that’s another huge change in the last three years that omni channel is everywhere. And omni channel means Twitter and apart from everything else. It means all of these different platforms coming together.
Garth Sundem 22:36
Oh, man, that’s a red flag. You say omni channel and I see another six podcasts in our future. But for today, I think we should we should leave it at that with using. I mean, how do we want to sum it up like using networks of connected publishing and social spaces and influence and citations and interactions and mentorship to dig down through these things to predict who is going to be the KOL of tomorrow, so All right, Mike, Carlos, thanks for joining us about how your organization can partner with digital science. Visit digital-science.com. MAPS members don’t forget to subscribe. And we hope you enjoyed this episode of the Medical Affairs Professional Society podcast series: “Elevate”.