shutterstock_766970317.jpgThis is the final post in a three-part series on artificial intelligence in healthcare. 

Of all of the marketplace dynamics and advances currently posing a threat to pharma’s traditional methods and models, artificial intelligence and advanced data science are causing their fair share of consternation. I recently addressed this issue with my colleague John Piccone, a ZS principal and advanced data science expert who previously led IBM Watson Health’s life sciences offerings, and he succinctly stated the commercial transformation challenge that lies ahead: “Pharma’s role is going to change from educating people to educating algorithms.” How can pharma get there? What steps can companies start taking now to adapt? John has some ideas.

Q: We’re clearly looking at a significant upheaval to pharma’s commercial model, so let’s get down to how the industry can begin to change itself. A pharma company trying to invest in AI is embarking on a long journey that involves obtaining the right skills and then identifying and solving problems, all while understanding how the commercial model will be affected by other stakeholders using AI. What’s stopping companies in this industry from investing in AI?

A: I think that there are a few things—and this is personal opinion because I don’t have a lot of data or analysis to back it up. Number one, I think that for the past several decades, pharma has been working on improving “back office” items such as cost and efficiency, operating models and regulatory compliance, and their depth has been in a certain set of technologies, database, data warehousing and transactional application systems. But for pharmaceutical companies to transform with AI, they need to shift their focus from back-office stuff to the real bread-and-butter issues that pharma is in the business of doing: developing new science, translating new science from the bench to the clinic, engaging in patient care, and adopting new therapeutics in a way that generates true improvement and outcomes. That’s not the back office. It’s not compliance and efficiency and operating model stuff. It’s directly related to science and patients.

It’s a big shift in mindset, and it requires disruptive collaborations in both the R&D side of the house with academic and research organizations and innovators at the startup phase, and also in care, care delivery and reimbursement. It’s requiring very different collaborations, and that’s not in the wheelhouse of their traditional leadership and management cadre.

The second thing is that AI is a different set of technologies and requires a different mindset than traditional databases and transaction systems. Pharma organizations have a big learning curve there.

Third, pharma is a highly regulated industry and is careful about changing in ways where there could be sanctions or implications with the regulators. The good news is that the regulators have realized that these changes are happening and they’ve communicated that they’re open and supportive, preliminarily, to the changes that need to happen. I think we’ll see an acceleration of that.

Actually, I was at a conference not too long ago and one of the directors from the FDA came up and said: “We really don’t understand AI. We don’t understand advanced data sciences really well. We don’t have the competencies and we’re so far behind the curve that we don’t even understand, in many cases, the solutions that the industry is bringing in front of us.” It’s clear that regulators have a steep learning curve for AI as well.

Lastly, the industry needs to embrace open-ended inquiries. Pharma has largely been in a role of getting a molecule through the pipeline and creating all of the evidence required to justify a regulatory approval and reimbursement status for a drug, and open-ended inquiries can actually threaten that process. But pharma companies have been going into it from the early phases of the pipeline, saying: “All right, I have a hypothesis that this drug is going to work. How do I prove that hypothesis?” And it’s a close-ended approach all the way through.

With the new scientific advances and the need to generate better outcomes for patients and understand who is a responder, who isn’t a responder, who’s going to experience safety effects, who isn’t, etc., an open-ended inquiry is a much faster route to higher impact. Many pharma companies are very uncomfortable with that whole process because it’s not how they’ve been working, and that stance is contributing to AI’s slower adoption.

Q: What are a couple of critical moves that pharma companies can make to speed the adoption of AI, and to embrace it? One example that comes to mind—an analogy, if you will—is digital. Pharma companies went through a similar kind of upheaval trying to inject digital thinking and marketing into their processes several years ago. Digital was completely alien when it first came upon the scene and, outside of a few marketers, most people didn’t believe that it was needed. They saw it as an efficiency play and not an effectiveness play. Then they decided that it was time to embrace it and used the “center of excellence” approach, but it’s not integrated very well into the broad commercial sales and marketing processes at most companies.

I see a few parallels with the way that the industry’s reacting to AI, trying to embrace something that feels very big but quite alien to their way of thinking, where they haven’t quite been convinced yet that it’s going to change their whole world. What lessons for AI’s implementation can pharma pull from digital? What are a couple of critical moves that pharma companies can make over the next 12 to 18 months to adopt and embrace AI?

A: I think that piloting the technology and learning from the experience of the pilots is a key approach, and they need to be very clear on what they’re trying to accomplish with the pilots. Some companies have a laser focus. They know where the impact points are for insights and new technologies. You have to have a clear sense of what business you’re in, what your strategy is and where the impact points are, and then do pilots.

AI leaders in the technology, consumer products and financial industries recognize that there needs to be an agile and iterative approach. You need an operating model that allows you to introduce innovation, measure its impact, experiment with changes and adaptations, and continually refine it in tight time cycles. For all of that to work, pharma companies need to change their relationships with regulators, providers and payers, and incorporate those stakeholders’ objectives into what the pharma companies are trying to accomplish with AI.

As with the move toward digital, pharma companies preparing to implement AI systems need to ensure that they’ve defined their strategies and the business questions that they want to address before investing in any new technology—and they need to be careful not to let some aspects of their investments get ahead of others. Figure out what a worthwhile return on your AI investment would be first, rather than just trying to determine and demonstrate the ROI post-investment.

Pharma companies need to start figuring out how to reinvent their commercial models around their customers’ and stakeholders’ new AI-fueled approaches now, before pharma sales and marketing teams find themselves trying to tailor product value stories for a whole new category of customers—one whose servers and systems haven’t been calibrated to respond to emotional appeals.        

This is the final installment in a three-part series on artificial intelligence. Did you miss the first two posts? You can find them here.


RELATED CONTENT 

BLOG POST: Is AI in Pharma Ready to Move From Hype to Hope

BLOG POST: Paging Dr. Watson: Evolving Pharmaceutical Value Propositions in the Age of Artificial Intelligence

BLOG POST: AI Will Usher in Dramatic Changes to Pharma's Commercial Model


 

Topics: patient outcomes, outcomes, artificial intelligence, commercial strategy, pharma companies, artificial intelligence & pharma, educate algorithms, innovators, traditional model, AI blog series