The Biggest Impacts of AI in Pharma
During this panel, industry experts (showed above) discussed the biggest impacts of AI in Pharma. We’ve included a short transcription of the panel, beginning at 18:41 of the webinar.
Sandeep Burugupalli, Pfizer: From preclinical studies to drug discovery to manufacturing to commercial, where is AI making the biggest impact? Where are you guys seeing the most potential for impact as well? So Tatiana, I’ll start with you here.
Tatiana Sorokina, Novartis: So it’s interesting to see how different divisions within such large organizations such as Novartis adopt AI and what point in the journey they are. Our team is part of the digital offices as I mentioned before and I have this unique opportunity to work with divisions pretty much across the company. So especially when we were hit by COVID-19, I realized how unprepared the whole industry is with respect to finding the right treatment for a disease that just appeared. And on top of that, how unprepared we are to quickly find solutions and treatments for the diseases that have been around for decades. We know they exist, we can diagnose them, but how come we haven’t found treatments for many of them yet?
I think part of the reason could be attributed to how long it takes for the drug to go from a molecule and discovering that molecule, all the way to the clinical trials and through the R&D process then to the approval and the commercial use and the launch and how expensive it is. Overtime, the time to get the treatment to Market hasn’t decreased. It’s actually increased and the cost has increased drastically. The question is, what are the areas of that development journey that could really be affected by AI in a positive way?
We have multiple different AI initiatives in different parts of the organization. Starting from research and development where you are looking at the studies in the discovery process and potentially, in addition to using your own data and using your own processes that you have in place, you also want to apply someone else’s from the real world data that’s out there already. You can look at it retroactively and get some insights from.
It helps because oftentimes you don’t run controlled experiments in the drug discovery or in the drug development phase. Oftentimes you’re looking to confirm or disprove a certain hypothesis, you’re not looking for different things that are out there. That’s what the experiments are for. Now if you look at the real world evidence and use real-world data, a lot of it is exploratory, so you’re letting a lot of insights emerge before your eyes.
Structurally, how do you then use those insights and turn them into hypotheses to maybe streamline your clinical trials or to streamline your hypothesis and therefore cut down on that time to market. So that would be one use case.
There is very interesting work going on in the manufacturing site. So as the drugs and the products become more complex in terms of manufacturing, can you automate some of your manufacturing processes that are in place to improve the yield of the product, for example? And if you can improve yield by just a few percentage points, the impact on the business is huge. So the idea is to look into those areas of the business and identify those opportunities to then see if we can improve the process and have an impact of even one percent two percent. Whatever kind of impact it’s going to have on patients and on their well-being and then finally the commercial, so there are a lot of commercial use cases, of course.
I mentioned NLP before but then you have all this plethora of data from digital marketing. How do you market to Physicians? How do you market to patients? And because all this is digital, in addition to the field force activities that we’ve been doing in the field. Now have all this data that we’ve collected from many different products. The question is, how can we utilize that data to do instead of just marketing to doctors? Can we actually start a conversation with the doctors and personalize that conversation to the extent that the doctor doesn’t feel that he or she is being bombarded by the information, but rather having an intelligent conversation with the company. Overall, being really a partner to that company as opposed to just an object or the target for the advertising.
Sandeep Burugupalli, Pfizer: Yeah, and what I really find fascinating about what you said, and I kind of echo the sentiment here, is that there are a lot of use cases across the organization and they all very dependent on whether it’s R&D or commercial or any of the spectrum in between and so I think it’s really interesting. When you say AI and pharma, there are no clear-cut use cases. It’s really about the technology enabling the use cases in a certain way. So I really appreciated a lot of the insights that you provided.
Learn more and watch the full video on YouTube: https://youtu.be/j3cxWQLcPe0