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 innovative educational approaches for Nedical Affairs. Joining us are Tim Mikhelashvili, CEO and co-founder Amedea pharma, and Riaz Abbas, Learning and Performance Lead, JAPAC Medical with Amgen. So first of all, to both of you, thank you for joining us today. I’m looking forward to hearing about what is new in L&D. And, Riaz, I was hoping maybe you could get us started by taking us through some of the advantages of this information rich information age that we live in, what’s what’s going on, with our information, age, and learning and development.

Riaz Abbas  00:58

All Thank you, Garth, for having me. I would start by saying that we are so lucky, we are so blessed that we live in such an age well called the information age, where information is really available to us, within the palm of our hands. We are so well connected, we are so better connected than we were even a decade or so ago. And just to give you an example and put into perspective, I mean, and I promise we won’t talk about COVID After this, but if you imagine seventh of Jan 2020 is the day when the China CDC released and published the genomic sequence of this newly identified novel Coronavirus, and under 10th of Gen. There was an RT PCR tests being developed across by major labs in the US Europe sectors, the how fast the information travels, and how quickly we can get on top of it. And we can just imagine how lucky we are to be in that age. So that’s my first bit that we are just so well connected and so better connected to information.

Garth Sundem  02:04

Okay, so that’s some of the technological side of things that we have access things move so quickly. What about sort of the human side of things, Tim, how does L&D fit with this information age? What are the trends for L&D?

Tim Mikhelashvili  02:22

Yes, Garth, a pleasure to speak to you both love this topic. As you know, I think that somehow or another, we now have more opportunities to be smarter as a result of information age. So I agree with Riaz. Comments, information and our exposure to it determines the extent of education that we can also receive ourselves and share with others, which ultimately, also is plays an important factor in the quality of decisions that we make in health care. So born Polson data analytics expert and professor, he, for example, talks about the increase the exponential increase in the volume, the velocity and the variety of information, those three V’s you know, that we are all now experiencing and witnessing, and in terms of l&d and the human side of things, are there some encouraging trends, I think, Garth, because there’s data from culture, amp and ad LDAP. From L&D experts that I attended, that shows that there is an 11 100% increase in people choosing to join learning groups. So we’re not choosing to educate ourselves informally. And my last all out coach podcast, I had a chief learning officer Ruth cotton from Weill Cornell, who mentioned that informal learning is a key factor to success of these extreme high achievers, Nobel Prize laureates and, you know, what have you. So, yeah, yeah. And in healthcare, also in healthcare, there is there is a increase, you know, an increase by about 41% in us getting data gathering more data, more survey data from our employees, or, you know, colleagues, but nevertheless, nevertheless, when you when I think when you look at metal Medical Affairs, lnd or LMS systems versus let’s say, other functions, I think there is an increase, but the surveys show that there, is there still a, you know, remaining need in more link link learning and development? I think so. Yeah. So a lot of encouraging data. And I think I’m what I’m glad to see is that the data is moving from the lab into the public.

Garth Sundem  04:56

Yeah, you know, okay, well, so I hear a lot in there that I think we can unpack. One thing that I thought was really interesting that you just said is, it’s almost like decentralized learning where people are joining learning groups where it’s not just that, you know, your company is telling you to do this, or your university is telling you, but that people are seeking out learning groups on their own. But if I’m a person seeking out a learning group on my own, and I’m being bombarded by the opportunity, that this massively increasing lead velocity, volume and variety of data present, that also comes with with challenges. So Riaz it’s not only opportunity, but also challenge in the new learning landscape for medical affairs. Is that correct?

Riaz Abbas 05:50

Absolutely. And I completely concur with Tim, I think I mentioned about the advantage of information age, I think the limitation or the challenge, as you say, is, we almost feel like drowning in that data or the level information. So just to again, give you an example, just in the space of non small cell lung cancer. Yeah, they were close to seven and a half, 1000 publications, seven and a half 1000 publications, in the last 12 months. In the last 45 days, since the beginning of the year to now Akoto, according to PubMed, there are 520 publications in just a space of Norseman lung cancer. So if you put our x, you know, us at the center of it, as well, as more importantly, I will say, the stakeholders, the physician, the patients that we we serve, it’s like some mounting amount of data, some I’m on information, that’s a challenge for us the information this is just the challenge is not information, or can we get the information, the challenge right now is too much information and people feeling almost drowning in what they call data, or as some people refer to as the data dilemma that we live in. Yeah,

Garth Sundem  07:00

You know, it’s almost like HCPs used to need MSLs to give them the Cliff Notes for the 50 studies that were out in the last six months. Does Medical Affairs in general. Now, do we need another layer of gatekeeping? To help us make sense of what did you say 7000 studies or something? Tim, how do we make sense of all this, of all this data? What What can we do? Or equip ourselves with as medical affairs professionals to make sense of this?

Tim Mikhelashvili 07:38

Yeah, I think you bring up a good point, Garth, that, we’d be probably remiss not to talk about, you know, responsible AI and AI accountability, right, because, because of just how fast that data is moving, right? Like both of you just remarked. And the way this plays out in Medical Affairs is, for example, if you’re conducting a study, and you have now completed your manuscript, and it’s going goes through a rounds of reviews, you know what, during that time in the process, there’s a lot of people who may or may not be qualified to discuss health care and make decisions, who are already who may be speaking about that same topic. You know, and the data shows that in dermatology, for example, dermatology related hashtags across social media, if only 5% of them are being discussed by board certified dermatologists. So, patients are demanding care for which they may or may not be appropriate. And so there. So I think there are a lot of the innovative companies are that are employing these exciting technology, you know, AI are now also looking for ways to standardize and validate, you know, and the government also the HA HA as a human health services. You know, last year, they announced their AI ambition, and the government as well in the United States. In Europe, they’re following suit as well. There’s a association for healthcare, social media, a trend across social media called hashtag verify healthcare, which are also kind of providing some balance in order to make sure that, you know, AI and technology and education and medical education and health care is used for the right reasons.

Garth Sundem  09:32

You say that only 5% of the conversation about dermatology on Twitter is by dermatologists. Yeah, before we get back to making sense of all this data using technology tools, does this mean that the conversation has gotten away from us as as medical affairs lost? What control or leadership of the conversation? What do you think?

Riaz Abbas 09:56

I wouldn’t I agree with to an extent with Tim but I don’t think we We have maybe at the cusp of I think we have got a huge opportunity by using technology actually, in my opinion, the technology let the technology solve the technological dilemma. Right. And you will say, Lester, let the virus be dealt with virologist and the politicians can worry about something else like politics, right. So I we are, for instance, heavily involved in looking at how tech can help us solve this. So we are doing core got a lot of work in l&d at the moment. at Amgen, we are really increasing the digital literacy of our colleagues. And we also looking at options and solutions potentially they were AI and NLP can really help us solve this, the solution. So when I say seven and a half, 1000 publications, I know most of the audience, most of my colleagues here, they’ll be thinking, well, that’s beyond comprehension. But Ray has come on get real, not all those papers, not all those publications are relevant. And are front and center. Absolutely. But how do you find that information? Which is you need to be front and center. And as you say, golf? How can we then serve our physicians better? So I think AI NLP can really help us and just to kind of say, in my opinion, what NLP really is it NLP really puts a structure to unstructured set of data. It’s a bit like to me like a, if you could imagine a brick of a piece of Lego bricks. And you can get a Batmobile out of it. Or you can get a you know, another structure like a Lego city type, you know, thing out of it. And LP, you’ve literally makes that decision, that information of unstructured piece of Lego, and visualize that for you in a matter of seconds. So I think there are some very exciting advancements happening at the moment where NLP can really help us, in my opinion, at least get to the first step of resolving this data dilemma.

Garth Sundem  12:00

So follow up on that real quick. So natural language processing, everybody’s listening to this knows that. But anyway, NLP. So you’ve got this tub of Lego bricks, and you can build a Batmobile, or you can build a cityscape. Does that mean that when we are using NLP in unstructured data, we need to know the shape of what we’re trying to build? You know? So you’re saying that this type of Lego bricks, it can go one way it can go the other? Is there any single meaning in the type of Lego bricks? Or do we need to tell NLP? What it needs to tell us?

Riaz Abbas 12:39

Absolutely. This is exactly the key keep it here. So we should not be worried about tech. And I know Tim, you’re passionate, very passionate about this. You’ve spoken about this topic at length. There is some of our colleagues that get worried about oh, is NLP going to take my job? I would we need to MSLs. And well, I actually believe that is the result and solve our issue. So yes, exactly. To NLP, Dudley’s give you structure, but you have to give it some parameters. Yeah, okay. So we defined in no small cell lung cancer, if I take as an example as to, which is the area, which are the parameters, we really need to take out front and center piece of groundbreaking data, bring it forward, and gives it that structure to say, well, this is these are the key list of publications or papers, which based on your parameters, you’ve done it. And it literally just processes that information at a much faster speed, speed and beyond us, however, the contextualizing our cognition is far more smarter if I could use that word. So we need to use that human skill, our expertise as subject matter expertise are still paramount to make NLP work.

Tim Mikhelashvili 13:47

Yeah, so you know, years ago, I was somewhat of a skeptic. And somebody taught me the Ctrl F function, for example, which allows you to sift through lots of data, right? And give it a particular command and search for that term right through a large heaps of data. Well, I look at NLP and natural language understanding as that control for f function, but to the parts of the exponential power, if you will, right, which which also the which depends on the ontologies and the lexicon, so that you, you’re able to ingest various different formats of data, which were exposed to a lot. You know, I started in medical education, but I’ve spent most of my career in the field Medical Affairs role. And we’re bombarded with different spreadsheets, the CSR reports, right PDFs. So I actually even though I used to be a skeptic, I advise a natural language understanding company called Source zero, as well. Because I’ve I’ve been convinced of the uses like Riyaz, just also alluded to, and so I think deals it addresses that third V of the variety, the variety of data because not all the data that we need is all organized well, in rows and columns, a lot of the data is unstructured to raise this point. Yeah. And that’s, that’s what it allows you to just organize that data. So ultimately, if you’re a medic in medical affairs today, you can either choose to be confused by all of this data, or you can choose to be more organized using novel technologies like like this one that really has just highlighted?

Garth Sundem  15:27

Well, so we’ve been talking about making meaning from data, let’s bring this back to education and helping people don’t know, make meaning of data. So all of these tools, are these going to eventually improve the quality and outcomes of medical education programs? Or is it? Or is it more of an internal understanding? You know, we’re going to use NLP and AI to say, internally understand the volume of data.

Tim Mikhelashvili 16:00

Yeah, well, you know, you mentioned I can I can take a crack at this. First, I guess. So I think in medical education, we there are various different ways now that the education is being delivered. And people have various different ways that they learn whether informally or formally, right, so personalizing that medical education is a key factor, I think, in first of all, making sure that you can trace that activity of a medical education program to the outcome that you’re seeking in terms of patient outcomes, or, or physician outcomes, because I think that’s what connects all of us in medical affairs, we’re trying to improve the quality of health care and its continuity, right. But if you look at metrics, for example of a lot of the social media, these innovative medical education programs, I recently did a review. And you’ll find that about 70% of all of such metrics are are reporting the acceptance, which which we already know is there, but very few, only about 3% or so are reporting on the direct, you know, the outcome has a impact on the outcome. So it has to do with the metrics that does that How, how we look at metrics of these innovative programs, you know, is it just the views? Is it the number of attendees to virtual center conference? No. And we need to I think more from attention to context and more deep understanding of metrics and designing metrics more deliberately, I think,

Garth Sundem  17:33

and maybe more personalized way, just to make sure I understand what you’re saying. So when you survey people who are taking part in innovative medical education programs, they accept the format of the program, the innovative format, but we’re not necessarily measuring it in the right way to understand how innovative medical programs are influencing outcomes. Is that Yeah, there’s Yeah, data

Riaz Abbas 17:58

is yeah, the data is a little bit, you know, it’s missing, you know, it’s lacking. And, yeah, and, you know, another area that I recently learned about where the data is lacking is in drug discovery, for example, where the databases for drug discovery are, do not have a lot of data in terms of the activity between a structure, the structure and activity relationship of a drug, you know, how a drug interacts with particular, you know, proteins or molecules, right, the drug target interactions. And that’s where AI, that’s one of the areas that I’m now starting to be fascinated with, in terms of drug because now we have a first AI discovered AI discovered drug for idiopathic pulmonary fibrosis that’s already in human clinical trials, where AI is not only allowing you to, to analyze a lot of data and organize it, but actually imagine new molecules using software, these new new software programs that are called generative adversarial networks, where there are two software programs that are essentially competing against each other, they’re pitted against each other when one generates generates them generate potential molecules with preferential qualities, let’s say solubility or protein binding, and the other one is trying to identify them. But over time, it learns to fool the other ones such that it develops a new, new, new product that could discover new drugs and I didn’t know that until recently.

Garth Sundem  19:37

Okay, so innovative, technological uses of artificial intelligence go beyond creating meaning from unstructured data. Well, well, I guess that’s another example of creating a kind of meaning like a useful drug would be the meaning in that case, and it emerges from this primordial soup through sort of Like evolutionary competition with other molecules, and I guess that is sort of like creating meaning. But yeah,

Tim Mikhelashvili 20:08

sorry, just so sorry, God, just so I know React is going to kind of but just to finish my thought on the meaning, which you just, which I probably didn’t completely address is that I think what I’m interested in now is to actually see how accurate some of these software programs are in predicting, you know, these these molecules and in augmenting those databases. So I’m actually presenting my overview of that in the new drug delivery conference taking place in May in Scotland. So yeah, that’s yeah, that’s what I’m interested in as well. And then the meaning of those new general adversarial networks systems.

Garth Sundem  20:51

That’s near Yes. What do you think?

Tim Mikhelashvili 20:54

Yeah, I think when it comes to meaning, I think I would ask all of us and who listen to the podcast today to think about what is the meaning of education? Okay, and it may vary from listener to listener, what’s the purpose that we’re trying to achieve? So I look at education as a holistic piece, which is trying to get as a skill set, overcome the knowledge barrier, but then translate that knowledge into actual meaningful action. So when it comes to l&d, we are trying to bridge haste and make it simple for people to understand complex science and data so that they can then translate, contextualize it, or when they’re discussing with HCPs. And likewise, in medical education, we should we have to start with that, again, what’s the outcome we’re looking for? What’s the purpose, and then work the modality not the other way around? So that’s my first bit about what’s the meaning. My second piece would be when we are working at Tech, or digital solutions, in my opinion, are function agnostic. They are enablers. Okay. But we need to understand what the problem we’re trying to solve, and then look at Tech to help us solve it.

Garth Sundem  22:06

That’s sort of those two things are a bit similar in that they are both outcomes based approaches Correct. You’re talking about education, designing education from your desired outcome. So is it is it that you know, future medical affairs, education will be designed with a with what, like the measured outcome in mind?

Tim Mikhelashvili 22:33

Oh, it has to be I mean, it should be we now say it should be designed with an outcome in mind. I mean, I think that’s where the challenge is that we sometimes are too bogged down. And I’ve been guilty of it myself too bogged down about what, what’s the data? And how can we quickly and how rapidly can we expose and share that with with physicians and healthcare professionals, whereas the approach the right approach should be, what’s the right outcome, we want to get what good looks like and then work backwards, including what we need to communicate and how we communicate in the most effective way. So that we can really overcome and not create this drowning in data. There’s some physicians and clinicians and opinion leaders that I know of. So yes, absolutely should be even today. But I think tech can really help that haste and those connecting those dots that can be really accelerated by using the right kind of AI and technological solutions.

Garth Sundem  23:29

All right, well, we’ve got tech as the nice problem to have too much data and tech, also, maybe the solution to having too much data in that it can help us make meaning from it. I know that this is the first of a brainstorm three part podcast series. So let’s leave some for next time. Thank you, Tim. And Riaz for joining us today. Thank you so learn more about modern approaches to external education. Before the next episode comes out, check out the maps Content Hub and sought for external education. You’ll find a bunch of good stuff there. Don’t forget to subscribe to this podcast. And we hope you enjoyed this episode of the maps podcast series. Elevate