00;00;05;04 

MAPS 

Oh. Welcome to this episode of the Medical Affairs Professional Society podcast “Elevate”. The views expressed in this recording are those of the individuals, and do not necessarily reflect on the opinions of MAPS or the companies with which they are affiliated. This presentation is for informational purposes only and is not intended as legal or regulatory advice. And now for today’s “Elevate” episode. 

00;00;33;09 

Adeola Davis  

Welcome to episode two of the SciTech Critical Evaluation of Literature and AI series. Literature never sleeps. Staying ahead without burning out.  

Sarah Snyder 

Hey, everyone. My name is Sarah Snyder from MSL Mastery, and I’m here with experts Adeola and Ruth. Adeola, would you mind giving us a quick introduction and then we’ll turn it over to Ruth? 

00;00;55;28 

Adeola Davis 

Hi, I’m Adeola Davis, Senior Director of R&D, Learning and Development at a global pharmaceutical company. And I’ll hand it over to Ruth. 

00;01;05;15 

Ruth Nicholson 

Thanks, Adeola. I’m Ruth Nicholson and I am a Global Medical Affairs and Medical Excellence Lead and super excited to be discussing this with you today. 

00;01;14;19 

Sarah Snyder 

Thank you both. I know that you’ll agree with me that staying up to date with literature is harder than ever. If you’ve opened PubMed lately, LinkedIn, your inbox and felt overwhelmed, this short episode is for you. Today we’re going to talk about how Medical Affairs can stay current without burning out or losing credibility. So Adeola, why do you think keeping up feels so impossible? 

00;01;44;00 

Adeola Davis 

Thank you for that question, Sarah. I’m going to have to say volume, velocity and variability. Preprints, abstracts, social media amplification. Evidence is moving faster than our ability to appraise it. 

00;01;56;25 

Ruth Nicholson 

And not only that Adeola, but not all publications are created equal. We’ve got preprints, predatory journals and they all blur the signal. 

00;02;07;17 

Sarah Snyder 

Wow, it’s a lot to take in. Can you guys help me define the risks? 

00;02;13;18 

Ruth Nicholson 

Thanks for that. And I think it’s so important to be aware of the potential risks that come with the volume, the pace and the quality of the information that comes at us. You mentioned preprints as an example. So these, are early access information, but there’s no associated peer review. So it can be very useful for awareness but not for conclusions. 

00;02;38;02 

Adeola Davis 

No, I completely agree with you, Ruth. In addition to preprints, there’s great literature like Congress posters that may never be published but still influence thinking in the field. 

00;02;47;03 

Sarah Snyder 

Interesting. What about predatory journals? 

00;02;51;03 

Ruth Nicholson 

So predatory journals again, they often lack peer review or even rigorous peer review, but they look legitimate. It’s very easy to look at a predatory journal and not realize that, perhaps the authors that have even paid for that article to be published in that journal. So Medical Affairs teams really need to be aware of this type of journal. You need to check the indexing , check, really dig deep. Look at the editorial boards and even the publication processes of these journals to make sure that what we’re looking at is a credible journal. 

00;03;30;16 

Sarah Snyder 

Good point. We assume that peer review. But that wouldn’t always be the case. Let’s switch over to preprints. So, do either of you have a learning moment where a preprint looked convincing? But then a further review that didn’t hold up? 

00;03;46;26 

Adeola Davis 

Oh, I have an excellent example of this. So during Covid 19, as you can imagine, preprints and repurposed therapies such as hydroxychloroquine and other small preliminary studies were circling rapidly on Med RXIV and RXIV. If you’re familiar with those where a lot of preprints are published. And the traction was huge, right? There was this huge need out in the community to have an understanding what was going on with Covid 19 and clinicians and even some public health guidance discussions were influenced by those early results before formal peer review. 

00;04;21;03 

Ruth Nicholson 

I’m sure we all remember the press coverage of this at the time. But when the larger peer reviewed randomized control studies such as Recovery or Solidarity came out, they didn’t confirm that early optimism for hydroxychloroquine these formal trials actually show there was no meaningful clinical benefits. And regulatory bodies had to withdraw their earlier emergency advice due to these recommendations. 

00;04;48;06 

Adeola Davis 

And that created a very interesting situation. And it put Medical Affairs specifically in a tricky spot. So everyone’s talking about this potential game changer. But the evidence quality was uncertain. It was important to stay aware and responsive, but without endorsing premature conclusions. 

00;05;03;27 

Sarah Snyder 

This is a perfect example of why speed doesn’t equal certainty. We all want those, and those early preprints can help us be aware of the emerging hypotheses, but they also need careful appraisal, context and confirmation before we can base any decisions or create messaging on them. Both of you know, I love AI, and I think that it can really be an amplifier for Medical Affairs. So there’s places where AI can help us in this, and there’s places where it can’t. So, Adeola, let’s start with you. Where do you think AI can fit into this? 

00;05;47;06 

Adeola Davis 

Oh, thanks. Sarah, I think AI can definitely help, you know, scanning hundreds of abstracts, tracking updates, identifying emerging themes. 

00;05;55;27  

Ruth Nicholson 

But AI isn’t the perfect solution though. It can’t for example understand context clinical plausibility or impact. It’s restricted. It can summarize what’s written not whether it’s reliable or applicable. 

00;06;10;09 

Sarah Snyder 

So what I’m taking away from you both is that AI can support awareness, but not judgment, Lets pivot to confidentiality, because this is where teams can get into trouble, and we want to avoid that. Ruth, what do you think about the confidentiality aspect with AI? 

00;06;28;16 

Ruth Nicholson 

And such an important topic and often is the main topic that’s discussed by medical affairs when they’re looking at introducing AI into their workflows. You know, medical affairs handles so much sensitive and confidential, even preliminary data, internal strategy documents, pre-approval materials, and that content must be kept confidential. It’s commercially sensitive and it should never go into an open or unsecured AI environment. 

00;07;01;19 

Adeola Davis 

Now, I completely agree, Ruth. And as I think about it sometimes, too, is that Medical Affairs or anyone in industries, we are stewards of the data, right? Not only we evaluating it, but at the same time, we also have to make sure that we’re leveraging it appropriately. And so one simple rule that I use is that if you wouldn’t email it externally, don’t paste it into AI. 

00;07;22;15 

Sarah Snyder 

Nice. 

00;07;23;26 

Ruth Nicholson 

Yeah, I completely agree. I would recommend that you stick to it to enterprise approved tools or if need be, to de-identify content anonymized your data or even use public data when you experiment and you try to set up these workflows. 

00;07;42;19 

Sarah Snyder 

I love those last two tips that you both provided. When we think about putting those into an actual framework, even with the approved tools, the outputs still need validation. You can’t just take what AI spits out at you and then put it into anything. How can we do that efficiency efficiently? What do you think? Adeola. 

00;08;06;28 

Adeola Davis 

I like a simple four check approach. The first one is can I trace this back to the original source? That’s very key. And keep in mind, even when you’re asking AI questions, you can ask for to include the original source and you can go back and check it. The second one is did the AI go beyond the data? Is it starting to try to make sense of it? As Ruth mentioned, it may or may not had the context that it needs. So that is that you have to be aware of. The third one is, is this applicable to the patient population or use case. And then lastly, the fourth part of the checklist is are limitations and uncertainty clearly stated? I think when you run through all of those that it really allows for you to help with the validation. And I would also say, as you think about this, if you’re on a team, you may want to have someone else take a look at it as well. Ruth, what are your thoughts? 

00;09;01;04  

Ruth Nicholson 

So I completely agree, Adeola. I think AI can be an incredibly powerful and beneficial tool to Medical Affairs, but ultimately it’s the people. It’s the humans in Medical Affairs that must make the decisions and the judgments. So the judgments in Medical Affairs is not optional. So this mindset, I think, reduces our risk while still optimizing the use of our resources and avoiding medical affairs burnout. 

00;09;33;22  

Sarah Snyder 

Yeah. I think that it’s going to be necessary to know how to use it and when to use it to avoid that burnout with so much information. Thank you both for all of your valuable input here. Adeola, can you kick us off with a final takeaway here? 

00;09;53;13 

Adeola Davis 

Yeah. Thanks, Sarah. I would say my my final thought is AI can help us keep up, but only if we stay critical. 

00;10;01;29 

Sarah Snyder 

What about you, Ruth? 

00;10;03;17 

Ruth Nicholson 

So I would expand on that. And I’d say only if we protect confidentiality and scientific rigor. 

00;10;11;21 

Sarah Snyder 

Wonderful. That wraps up our episode about how literature never sleeps. And we know that in Medical Affairs, credibility is everything. So we’ve got to keep up with the literature. We are going to have more on this topic in episode three, which is going to be humans plus AI making sense of scientific papers at scale. So we’ll talk to you then.