PAP 2024: Embracing the Power of Artificial Intelligence

News
Article

Fireside chat uncovers ways to leverage this technology in order to boost medication adherence, engagement, and outcomes.

PAP 2024 session, "Medication Adherence—Embracing the Power of AI to Improve Medication Adherence, Engagement, and Outcomes.” March 20, 2024. PAP 2024, Philadelphia, Pa. Image Credit: Nicholas Saraceno.

PAP 2024 session, "Medication Adherence—Embracing the Power of AI to Improve Medication Adherence, Engagement, and Outcomes.” March 20, 2024. PAP 2024, Philadelphia, Pa. Image Credit: Nicholas Saraceno.

Medication adherence—the process of following a regimen that oftentimes requires patients taking their medications regularly—is easier said than done. After all, we’re only human.

However, artificial intelligence (AI) in this space is continuously evolving. Here on Day 2 of Patient Assistance & Access Programs (PAP), in a fireside chat titled “Medication Adherence—

Embracing the Power of AI to Improve Medication Adherence, Engagement, and Outcomes,” Jordan Armstrong, VP of business development with AssistRx, a patient support solution provider, and William Grambley, CEO of AllazoHealth, a tech company that delivers AI-powered healthcare experiences, sat down to discuss the manners in which this technology can be used to increase adherence, along what the future holds in this space.

This chat has been edited for length and clarity.

Jordan Armstrong: What exactly can AI do to support the therapy journey? Will there be initiation and engagement, and what are some best practices?

William Grambley: AI covers a lot of things. ChatGPT is one version of AI, but there's also predictive modeling, predictive analytics, machine learning, and neural networks—all sorts of things can be defined as AI. When we think about the types of things that you can use AI for, it does cover a lot of different applications. Now, obviously, we have one specific way that we've approached that, which is about predicting patient behavior, and then driving the appropriate intervention to support that patient. But there are a lot of other ways you can do that, such as looking for different patterns in your data. When we end up using a bunch of external data, as part of our algorithms, we can enable those kinds of insights as well.

AI is almost a category as opposed to a thing. That category can cover a lot of different parts. I would also highlight that you can use it operationally, you can use it within your field teams, you can use it to support initiation, and you can use it to support ongoing adherence, The question should be, if I'm not using AI, what could I be using to make my programs more effective? It’s alsoso broad that you should be asking why you're not using AI, instead of how to find the right AI.

Grambley: What types of programs are you seeing that are getting the biggest impact on adopting AI, and what are you seeing from the AssistRx side?

Armstrong: There’s different categories of AI that they have, where everyone just lumps it all together. Whereas you have machine learning, you have all of the different complexities of what AI can actually be. We typically break down exactly what one means by how we are leveraging AI, because a lot of what AssistRx is consists of pulling in specific data at that exact instance, so that we're not necessarily using predictive models to understand what benefits look like, we're not using predictive models to really go through any type of reimbursement access activities. At the forefront, it’s more transactional. Where we've seen it really be impactful is from a workforce management, as well providing our manufacturing partners with insights from the day-to-day interactions between our frontline individuals, case managers, universal specialists, with payers with patients with healthcare providers (HCPs), so we leverage technology internal to our platform that's actually able to pull out specific key phrases, as well as understanding exactly how tone recognition is, and understanding exactly how the patient is actually feeling through the process.

Armstrong: What are some pitfalls of this technology, and for life sciences organizations who really leverage AI, how can they internally navigate everything?

Grambley: There are lot of areas that we need to be very thought out about how to leverage AI, and we'll just talk about them for a moment. I'd be amazed if anybody here used ChatGPT or something similar to it, to create patient and/or HCP-focused communications without going through your normal medical legal regulatory (MLR) submission process. AI is still developing. When we think about it, we start with a lot of the statistical models, a lot of predictive algorithms, and we're looking at variables. When you think about how to adopt AI—access to data, reliability of the data, the size of the data—all of those things are areas you really need to think carefully about.

One of the one of the benefits of using an outsourced company is that usually they have access to data that you're not going to have inside your company. Often times, they'll be able to get more data about the patients. In our case, we actually use things like consumer behavior data. But you obviously need to have consents, you need to go through your privacy, you need to go through legal reviews, and those can take time. But they're all things that if you do those appropriately, can really increase the power of how you can use something like AI. I think as manufacturers, there's a very real kind of change management process.

Grambley: When you see programs, you know, how do you think about this from the execution side?

Armstrong: Execution generally can be very difficult. I think to your point on technology, and leveraging as best as possible, that's where the differences are, because it's truly a manual process. This gives you a cue for what's happening; there are nimble factors that really come into play where you're working with an organization that can actually leverage the technology to make those changes on the fly.

A big part of what we do is part of the implementation process, part of the “go live” process, where there are unknowns. When you’re talking about utilization of technology and understanding the call out, I think that consensus is very important. With only your first name, last name, date of birth, gender, zip code and your consent, you can basically find out anything about an individual. understand exactly everything about them. But this is where you need to be able to take the program so that rather than having you think about that experience, you can think about the hope that this brings to people, in terms of helping them live a more normal, active, healthy lifestyle. As part of that, you need to think about how can I move from point A to point B—which is a clinical decision to product dispense in hand—as quickly as possible?

Reference

Armstrong J, Grambley W. Medication Adherence—Embracing the Power of AI to Improve Medication Adherence, Engagement, and Outcomes. March 20, 2024. PAP 2024. Philadelphia.

Recent Videos
© 2024 MJH Life Sciences

All rights reserved.