In Their Own Words: Participants Review the AI Summit

By Mario Stassen (Stassen Pharmaconsult BV), William Whitford (DPS Group), and AIO Team

​In February of 2022, the efforts of Xavier Health were assumed by the AFDO/RAPS Healthcare Products Collaborative. Because of the important work done before this transition, the Collaborative has chosen to retain documents that have Xavier branding and continue to provide them to the communities through this website.  If you have questions, please contact Timothy Hsu, Director of Health Technology Initiatives, at thsu@healthcareproducts.org.

Healthcare Products Collaborative (formerly Xavier Health) is committed to improving the progression of healthcare by supporting teams of representatives of academia, solution suppliers, and pharmaceutical product sponsors to discover and publish on the newest trends and technologies. 

In this blog entry by the AIO Team, we pursue questions regarding AI in pharma and the annual AI Summit in particular. Specifically, we looked to Mitch Tirea, Sr. Manager, Quality at Edwards Lifescience, and Zinatara (Zina) A. Manji, M.S., PharmD, Sr. Director Regulatory Affairs at GSK.


Q: Thanks so much for agreeing to provide your insights. Let’s start by learning a bit about your respective backgrounds.

Mitch: I’m the Sr. Manager for Quality at Edwards Lifesciences with an MBA degree from University of Southern. Edwards Lifesciences is a global leader in patient-focused medical innovations for structural heart disease, as well as critical care and surgical monitoring. The company collaborates with the world’s leading clinicians and researchers to address unmet healthcare needs, working to improve patient outcomes and enhance lives. 

Zina: I’m the Sr. Director of Regulatory Affairs for Innovation in Consumer Healthcare at GlaxoSmithKline (GSK). GSK has three global businesses that research, develop and manufacture innovative pharmaceutical medicines, vaccines and consumer healthcare products. In my role, I see myself as an innovation intrapreneur. I’m especially interested in emerging science areas, developing strategies for value driven health solutions and the newest steps in technology for advancing transformation in self care to improve health outcomes of consumers and patients. 

 

Q: Why is AI important for you or your business?

“AI is no longer a future possibility; It is currently becoming, or is already part of, many aspects of our daily lives

— Mitch Tirea

Mitch: There are many reasons. Foremost is AI’s potential to make rapid and precise decisions in so many areas of pharma and medical device industries. It is the next step in the digitalization of businesses, and we are now seeing its rapid adoption and application in, e.g., the electronics and automotive sectors. A major driver of this is that human capacity is simply inadequate to manage volume and disparate sources of data from modern sensing and analytics. AI is no longer a future possibility; It is currently becoming, or is already part of, many aspects of our daily lives. AI is providing a means of transitioning from the existing descriptive analytics to the up-and-coming advanced analytics. Comprehensive handling of the increasing volume of relevant data being generated is a key driver. 

Zina: The developing public/private partnerships are helping to define the future of public health in general. As the vision of healthcare is evolving, the value of volumes of relevant data becoming available is certainly important. But we might ask, precisely what type of data is needed for each initiative? The nature of the project determines the quality, quantity, timeliness, sources, and applicability of the data involved. Then, the project itself also determines the directions you will take, and answers to such questions as “What, specifically, do you want to accomplish with it?”, “What problem are you trying to solve and is the solution meaningful to the user and stakeholders”? It will be important to frame the innovation balance:  the benefits vs risks and how to mitigate for these risks, especially in a continuous AI model. This may be assessing benefit for personalization and early detection of disease, maintaining and improving health with generation of data that supports assurance of safety and efficacy—balanced with trustworthy data sources, system for continuous monitoring and risk mitigation strategies. 

“Explainability is a key aspect for AI adoption.

— Zina Manji

Then there is the nature of AI itself. Its power is demonstrable, but what about the need for deep understanding of how it works? Explainability is a key aspect for AI adoption. How much do we need to explain regarding the “black box”? It’s critical to frame the input, output and data sources for mitigating risks and how outputs will be used as continuous real world data to generate evidence for evolving applications.

From the highest perspective, the general driver for AI is simply that process and analytics data volume is exceeding the capabilities of human intellect and classical statistical methods. 

 

Q: Have you adopted AI in your company?

Mitch: We are working on adopting in many departments. I am personally working on advanced analytics for applications in the areas of quality for corrective action.  There is no doubt that “AI helps with precision in decision making.” Now, AI is not error-free, but it can be a significant step in enhancing our current capacities. 

Some human health sectors are lagging behind others in adopting AI. Therefore, contact with such peer companies like Sigma Health is important, as are such collaborations with academiaFor example, Edwards Lifesciences now partners with the University of California Irvine. I encourage all to find out how their master’s programs can do the same and become an AI innovator.

Zina: In general, my focus is product innovations—and where applicable—to evaluate how AI can be used in data generation and providing data to the user towards specific health outcomes. In understanding AI applications, I look to all of the different ways users and stakeholders may engage with  the application of AI to address healthcare challenges. Of interest is evaluation of AI as a tool that empowers continued development—not simply static “solutions”—but ongoing improvements and successes in a variety of continuous learning applications and product innovations—and, particularly, assessing the evolving regulatory framework.

 

Q: Why attend the AI Summit?

Mitch: I participate in the Xavier AI Summit to meet other industry peers, gain from their insight, and develop new perspectives. Such interaction not only helps to focus and benchmark our current initiatives, but also to collaborate and share in the path forward.  It’s all too common for us to regularly meet with the same colleagues or collaborators and develop “tunnel-think.” The workshops are especially big hits by presenting an opportunity to closely meet with regulatory people and better understand the specific checks and balances desired by the agencies.

At the summit, we share technical experiences on exploring the role of data as a valuable tool in the decision-making process. But, maybe more important is hearing of approaches on how to bridge the gaps between the academic and business components, and how to get leadership engaged to fully embrace AI.

Zina: One way professionals have interacted and expanded their awareness is through industry and professional associations. They can provide significant value but tend to have memberships from a specific area of healthcare. However, consortia such as the Xavier AIO initiative in general, and the summit specifically, is successfully bridging the gaps and breaking down walls between such disparate groups as regulators, industry professionals, academics, providers, payers, vendors, process and product developers, and production engineers in this very unique platform. In bringing everyone to the same table, they encourage not only the exchange of perceived benefits and challenges of AI, but reveal ways to establish cross-functional teams to actually implement paths to practical solutions.  The initial proposals of GMLP (Good Machine Learning Practice) and further incorporation into FDA’s thoughts on an AI regulatory framework is a great example. 

Xavier has developed a program that is very open and inviting and has attracted an eclectic group of professionals representing diverse backgrounds and functions.  It provides a venue to interact with inter- and intra- company (or organizational) talents who you otherwise might not have occasion to meet.  

 

Q: What do you expect from the Summit?

Mitch: This year, I’m looking forward to a couple of activities within the working groups. Besides the expectation to share the improved tools to implement ‘change process’ of adapting AI in organizations, my favorite subject will be the transition from CSA to CSV approaches to validation. CSA represents a step-change in 21 CFR Part 820 relevant computer system validation—introducing critical thinking into the CSV process, as opposed to the traditional, more prescriptive, approach.

“I expect this year’s AI Summit is a means of bringing people of various backgrounds, positions and specialties together with the singular goal of successfully achieving AI at scale.

— Mitch Tirea

The different perspectives observed in the rather dynamic environment of the summit help us to discover new ways to prepare and harness AI as a tool for creating agile organizations. The National Progression Award (NPA) in data qualification also promises to be controversial and exciting.  

Elements of the resistance to new ideas can be diverse, so vigorous debate and exchange prior to presentation within our respective organizations is also quite valuable. This is especially true when presenting to non-data professionals. A robust dialogue, with clear metaphors, helps to elucidate AI to those not specifically trained in data manipulation. In other words, the diverse and dynamic dialogue at the summit help us to be able to “sell” to others what we personally know to be true. I expect this year’s AI Summit is a means of bringing people of various backgrounds, positions and specialties together with the singular goal of successfully achieving AI at scale. And, of course, the accredited regulatory personnel who attend are an invaluable feature.

Zina: I’m looking forward to hearing the latest developments on AI applications and shaping of new regulations as well as discussions on how the pandemic has brought forward different tools and applications. Some personal goals of mine at the summit this year are to further understand the (1) distinctions in the types and sources of the data behind AI implementations, (2) latest thinking on standards and aspects of bias and (3) requirements to develop AI as instrumental in supporting the claims of product value and controlling risk. Additionally, I always look forward to updates from the working groups and plans for further activities.

 

Q: Have you attended one of our summits before, and If so, what did you appreciate about it?

Mitch: I’ve attended two AI summits in the past: One was in-person, and the last, virtual. Last time I missed the face-to-face interaction, although the remote presentations do allow for an easier to follow and focused experience, with fewer distractions. The technical operation and success of break-out groups varied, depending upon the provider software and quality of internet connections. While I personally enjoy an in-person venue, I see the decision between in-person and virtual is a balance between such factors as features, budget, travel, and COVID imperatives.

I particularly appreciate the openness and transparency of the participants within the summit. It seems to attract not only subject matter experts and certified officials, but more significantly, those interested in helping to both share and gain advice and guidance.  

“It’s instructive to hear what professional AI application practitioners have to say about what is actually happening—not simply what is technically possible.

— Zina Manji

Zina: My first AI Summit experience was in 2018 and I’ve attended each year since. I have been impressed by participation of AI thought leaders, including FDA. I’m passionate about the value of bringing creative experts from diverse backgrounds to the endeavor. This brings not only technical depth, but legitimacy to the dialogue on topics addressed. I particularly appreciate the individual focus of the summit roundtables and sub-teams, and the approaches being taken to help enable AI applications in pharma operations. I see Xavier’s board and summit organizers ensuring that we are moving in meaningful directions. Xavier’s Good Machine Learning Practices (GMLP) team, which I’m most familiar with, is an excellent example of the value of the collaboration in the initiative. It’s instructive to hear what professional AI application practitioners have to say about what is actually happening, what problems need to be solved (and the challenges)—not simply what is technically possible.


Note for readers: Hopefully, you will be as enthusiastic as Zina and Mitch at the next AI Summit. For more information:

  1.  AI Working Teams: These teams of industry members work collaboratively to create deliverables that help advance the industry (such as the white papers that FDA cites and uses, or the AI Maturity Model, etc.). Members have all different kinds of backgrounds and experience levels. We just ask that you have the time to actively participate. These teams have formal co-leaders and administrators.  
  2.  Healthcare Products AI Summit: We have an annual summit that brings thought-leaders and practitioners to the stage to specifically address AI related to the pharmaceutical and medical device industries. FDA’s Bakul Patel (Director of FDA’s Digital Health Center of Excellence) is on our strategic committee.