MAPS   00:06

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,

Jennifer Riggins   00:33

Welcome back to elevate the Medical Affairs Professional Society’s podcast. As a series within this podcast, we focus on digital first communications how digital is transforming Medical Affairs. In these podcasts, we speak with experts in the field of Medical Affairs and discuss how digital transformation is opening opportunities for Medical Affairs communicators. I’m Jennifer Riggins, a co host of this podcast. I currently serve as a member of the Digital Focus Area, Working Group I’ve worked in pharma for more than 30 years, focusing on medical information, scientific communications and medical digital and I currently work for phactMI, a nonprofit consortium of medical information leaders. I’m joined by my co host, Steve Casey of Omni Healthcare Communications, or Omni HC, for short. Steve is even older than I am, and has been in pharma for over 35 years, and has led Omni HC for the last 10 years to become a leader in digital, first, medical communications. We have a great topic in store for you today, as we’re delving into an area that I think will be beneficial to our listeners, we’re going to be discussing AI use cases in medical information.

Steve Casey   01:43

So most of you are probably experimenting with AI for personal use, and maybe even are having some success with the use of AI in your day to day work. We want to share some exciting learning with you from some proof of concept work that has been done in medical information. Joining us today is Sebastian Lewis-Saravalli, Medical Information. Gen AI, Content Lead for Pfizer. Thank you for joining us today, Sebastian.

Sebastian Lewis-Saravall   02:08

Thank you. Thanks for inviting me.

Steve Casey   02:10

It’s great to have you on the podcast. Could you give us a brief overview of your background and introduce us to your current role at Pfizer? You know going to working with AI and generative AI?

Sebastian Lewis-Saravall   02:21

Absolutely. My name is Sebastian. I’m currently on a second man as the medical information Gen AI content lead advisor. And in this role, I focus on using generative AI to enhance our medical information processes and to improve how we deliver information to healthcare professionals and patients. A bit about my background. I have a strong foundation in medical information science, with a master’s degree in pharmacology. I’ve been in the industry for over 10 years, working on various projects that integrate technologies into our workflows. And at Pfizer, my team and I are dedicated to exploring innovative Gen AI solutions that can streamline our operations and provide more accurate and timely information to our customers. So that’s a bit about me and what my team does. I’m happy to share more about our journey and the insights we’ve gained along the way.

Jennifer Riggins   03:18

Yeah, absolutely. Thanks for providing your background and giving us a glimpse into the work that you’re doing. I’d like to move us now into a bit more of the details of that work and some of the use cases you’ve explored for AI and Gen AI.

Sebastian Lewis-Saravall   03:33

Yes, so earlier this year, we explored several exciting use cases for AI in the realm of medical information. First, we focused on medical writing. We explored how Gen AI could accelerate our content creation process while maintaining high quality. We’ve shown that by automating parts of the medical writing process, we were able to reduce the time it took to produce high quality medical information materials. This improved our efficiency, allowing our team to focus more on critical tasks that require human expertise. Another use case we looked into is using AI to help our frontline specialists during customer interactions. The idea is to help them find information faster and more efficiently by integrating AI tools into our customer service workflows, this would allow our specialists to access relevant information in real time, and it’s something that we are still exploring. And lastly, we used AI to analyze large data sets and extract valuable insights. This helped us identify trends and patterns that can improve our services. For example, by analyzing customer inquiries and feedback more efficiently, we can identify areas for improvement, refine our content and tailor our services to. Better meet the needs of healthcare professionals and patients.

Steve Casey   05:04

Wow, that’s really fascinating. Those are three great examples, where have you seen most of the success with your proof of concept work that you’ve completed thus far? And can you tell us a little bit more about that success?

Sebastian Lewis-Saravall   05:16

Well, there’s a lot of potential and still plenty to explore in each of these use cases, the area where we’ve made the most progress is in content generation. As you know, in medical information, we handle a high volume of customer inquiries, which means we need to produce a lot of content so having a robust process for content creation and responding to inquiries is crucial. Earlier this year, we conducted a proof of concept to evaluate Pfizer’s Gen AI platform capabilities in summarizing scientific data for this experiment, we’ve summarized 100 different documents, including posters, abstracts and published articles. We had two main objectives, first to achieve a 25% increase in efficiency through time saved compared to human control group, and second, to attain a quality score of at least 50% for the first draft generated by the Gen AI platform. The quality scores were based on six dimensions, accuracy, completeness, language, flow, reliability and paraphrasing. So the results have been quite promising. Using Gen AI, we were 55% more efficient at summarizing scientific documents than without using Gen AI, that means we cut down the time by half. That’s significant. And in terms of quality, Gen AI achieved an average score of over 70% when combining the six different quality dimensions. And that was only at the first draft,

Steve Casey   06:59

70% success on a proof of concept? That’s pretty exciting.

Sebastian Lewis-Saravall   07:04

Absolutely, this proof of concept really showed us that Gen AI can efficiently summarize scientific content, adding significant value to our organization. It proved that we can rely on Gen AI for those time consuming tasks, freeing up our team to focus on more impactful activities. So building on these promising results, we decided to expand this use case, and we kicked up several other use cases to further explore the impact of Gen AI across the content lifecycle.

Steve Casey   07:36

Interesting. Can you tell us more about those content related use cases?

Sebastian Lewis-Saravall   07:41

Sure, when we think about the process of responding to a medical inquiry within medical information, it can be broken down into five steps, defining the question, researching the topic, evaluating the evidence, synthesizing a response and Sharing the answer. We looked at all of these areas to see where we could make improvements. So when it comes to researching the topic, we found that Gen AI can significantly simplify the literature search process and cut down evaluation time. For example, it can automatically build search strings filter out their relevant publications and even assist in analyzing the relevant ones. Next in terms of summarizing the response, we’ve shown that Gen AI can quickly identify relevant information from various scientific documents and use it to generate high quality summaries. This allows us to create first draft content more efficiently. And finally, when it comes to sharing the response, we’ve demonstrated that Gen AI can effectively customize content for different channels or personalize it for different audiences. So for example, it can generate patient response documents, ensuring that the information we provide is relevant and useful to our customers.

Jennifer Riggins   09:07

So I love that you’re seeing such great success in your initial proof of concepts, and then it’s really benefiting. It sounds like so many areas of your work, from researching the topic to summarizing the response to generating patient, friendly versions of your responses, but, and we always know there’s a but, I’m sure you’ve run into some challenges as well. What challenges have you been addressing as part of these POCs?

Sebastian Lewis-Saravall   09:34

Good question. One of the significant challenges we’ve been addressing is ensuring the accuracy and reliability of the AI generated content. While generative AI has shown great potential, it’s crucial that the information it produces is precise and trustworthy, especially in the medical field, we’ve implemented rigorous validation processes to review and verify the content. And before it’s used or shared. Another challenge is integrating AI tools seamlessly into our existing workflows. This involves not only technical adjustments, but also training our team to effectively use these new tools. Ensuring that everyone is comfortable and proficient with the technology is essential for maximizing its benefits. Lastly, respecting intellectual property rights while using generative AI is crucial. We need to be very careful about the sources of data we use and the content that we generate. We’ve established strict guidelines for sourcing data, ensuring we only use data from authorized sources to comply with copyright laws.

Steve Casey   10:43

Those are some important challenges that definitely need to be faced head on. Sebastian, what other information do you think is important for our listeners to know as they embark upon their own pilots and proof of concepts with Gen AI?

Sebastian Lewis-Saravall   10:57

One important piece of advice I would give to anyone starting their own pilots and proof of concept with AI is to have clear objectives. Knowing exactly what you want to achieve will help you design your experiments more effectively and measure success accurately. Another key point is to be prepared for an iterative process. As you know, generative AI is a rapidly evolving field, and what works today might need adjustments tomorrow. So embrace the learning curve and be flexible in adapting your your approach based on the insights you gain along the way. I think it’s also important to address some common fears associated with using AI. Many people worry that AI might replace human jobs. So it’s crucial to highlight how AI can actually augment human capabilities rather than replace them. For instance, AI can take over administrative tasks, allowing employees to focus on more strategic and creative aspects of their work. So I think that by addressing these points and by being mindful of the challenges and concerns, you’ll be better equipped to navigate the complexities of Gen AI and maximize the benefits it can bring to your projects.

Jennifer Riggins   12:16

So, that’s really good information, and I think it’s really important for any team wanting to use AI and Gen AI in their business model. I love the advice about being prepared for an iterative process. Gen AI is definitely evolving rapidly. So what are the next steps for you and your team?

Sebastian Lewis-Saravall   12:34

For our next steps, we’re focusing on a few key areas. First, we’re upskilling our colleagues to ensure that they can fully benefit from the learnings we’ve gained, and effectively use Gen AI in real life settings with actual products and questions. This involves training and hands on experience to build confidence and proficiency with the technology. We’re also refining our processes in technology to better integrate Gen AI into our workflows, aiming for seamless and efficient operation. Additionally, we’re exploring new use cases for Gen AI and content generation and customer interactions to enhance productivity and deliver high quality information quickly. So overall, our goal is to stay at the forefront of AI innovation in medical information and continuously improve how we serve healthcare professionals and patients.

Steve Casey   13:32

One final question for you, if, if you had to give three takeaways from our discussion today for our listeners, what would they be?

Sebastian Lewis-Saravall   13:39

First off, AI, especially Gen AI, has incredible potential to transform medical information. It’s not just about speeding things up, it’s about making our work more effective, from summarizing documents to helping specialists find the information they need. Second, AI tools are really enhancing our customer interactions by giving our specialists real time access to relevant information, we can respond more quickly and accurately, which leads to a much better customer experience. And finally, embracing AI means we need to keep learning and adapting. It’s an ongoing journey, and being flexible and open to new insights is key, this way we can stay ahead in innovation and continuously improve our services.

Jennifer Riggins   14:32

So, those are great takeaways. You know, I hope everyone out there noted those key takeaways, the transformative potential of AI and improved customer experience and the need to embrace continuous learning and adaptation. So Sebastian Steve and I want to thank you for providing an overview of the use cases and outcomes from your POC work with AI and Gen AI to our listeners, we hope you’ve enjoyed this discussion and received some great takeaways. If you. Appreciated this podcast. Please make sure to like it and feel free to comment to us. You can find our contact information on LinkedIn. Thank you for joining us today and listening to our podcast series, Digital First Medical Affairs, a podcast production of the Digital Focus Area Working Group of the Medical Affairs Professional Society.

Steve Casey   15:19

If you’re an old MAPS member or maybe a younger MAPS member like Jen, thank you for your support. If you’re not a MAPS member or an old MAPS member like me, I want to encourage you to join so you can access additional resources. Visit the MAPS website today at MedicalAffairs.org/membership.