How Can Pharma Navigate the Regulatory Maze of AI-Powered Interventions?

In this Pharmaceutical Executive video interview, Bill Grambley, CEO of AllazoHealth, delves into the potential of AI to revolutionize patient engagement and adherence. By utilizing patient-level data and predictive modeling, AI can identify individual needs and preferences, tailoring interventions to optimize treatment outcomes. This personalized approach addresses the common challenges of medication non-adherence and improves patient outcomes. The discussion highlights the importance of ethical considerations and data privacy when leveraging AI in healthcare. By focusing on patient well-being and using AI responsibly, healthcare providers can significantly enhance patient experiences and improve overall health outcomes.

How can pharmaceutical companies navigate the evolving regulatory landscape to ensure that AI-powered interventions comply with relevant regulations and guidelines? What are some key challenges and opportunities to doing so?

Within these programs, kind of using AI powered interventions, is not really a major focus of the regulatory landscape. Even that said, we are hyper focused on ensuring everything we do complies today and, in the future, and we keep an eye on that now, as AI becomes increasingly leveraged throughout the economy, it certainly can be a target for regulators to make policies which will impact pharmaceutical companies. When I think of the regulatory landscape in this world that we're talking about, there's really three areas that that I think of that will become important to keep an eye on. One, use of data. Two is kind of the type of algorithms, or the type of AI, and then three is the outcomes that you're impacting.

We've talked a little bit about patient data, and with the right permissions, the approvals, the safeguards, a wide variety of organizations can actually use that data to support the patient in their journey, the kinds of AI and the kinds of algorithms can get a little bit trickier. So, what we do? We use a number of widely available AI tools to make predictions. That includes some machine learning. It includes multi-dimensional test control studies. It includes a number of different, again, widely available tools that being said, you know, things that have come out over the last several years are really pushing the boundaries of how you're using data. You could think of large language models as an example.

Now, even in the world that we're in, again, using these tools that we use for our data. You know that that's not really been a focus, because we are supporting that patient, and we're using it very carefully, but we do have a responsibility to use them ethically, regardless of regulatory policy and so, you know, so things like, if you did use a large language model to come up with content, you're still going through kind of the overall review process, looking at is that content appropriate, is accurate, all of those things.

So, you know, as leaders in pharmaceutical companies and vendors, we have to continue doing that. Let me end though with the outcomes. And this is probably the area that I would say is where you'll see more focus what is appropriate to use as an outcome around this AI and patient data. So, if a prescriber believes the patient should be on a certain medication, there's a lot of justification to use AI to support that patient, to get and stay on therapy. And I think most people, most regulators, you know, most of the industry, would say that's perfectly acceptable. So, if you can continue focusing on that patient outcome, continue focusing on supporting that patient on their journey, use that as a guiding light. And if you can continue doing that, you avoid a lot of the concerns that come up with AI and that come up with these various news headlines you may see, because it is an area that we know there is an issue, and it's an area that we can actually improve, and the patient gets a better outcome from it as well.