We recently spoke with Johnson & Johnson MedTech SVP and Global Head of Digital Shan Jegatheeswaran about the medtech developers’s new Polyphonic AI Fund for Surgery through Nvidia and Amazon Web Services (AWS).
At the same time as that interview, we asked for any advice he could offer to help medtech developers and manufacturers better understand device user needs, build trust in artificial intelligence, and advance digitization efforts within their organizations.
MDO Min-Vasive Medtech: Join us for live interviews and audience Q&As with minimally invasive engineers from Edwards Lifesciences, Jupiter Endo and Compremium
The following has been lightly edited for clarity and space.
MDO: What have you learned in your time in medtech about understanding user needs?
Johnson & Johnson MedTech SVP and Global Head of Digital Shan Jegatheeswaran [Photo courtesy of J&J]
What technology do we need in terms of infrastructure or next-generation components like sensors to achieve J&J’s vision for AI?
Jegatheeswaran: “Technology isn’t the limiting factor. The technology exists sufficiently enough where we can add value, whether it’s through AI or not. Ultimately, the end user doesn’t really care whether AI is involved. It makes it better. But they want an outcome. The limitations we’re working through and for which I think we’re uniquely positioned are in (no pun intended) the soft tissue around making technology work at scale. That’s how do you think about regulatory globally, not just at a hospital level, how you think about a trusted experience when it comes to things like AI, how you think about managing risk and contracts, and the change management for the folks who are ultimately going to use this software output in an OR, a very dynamic, human-first setting. Those are things that require time, patience, study, and that’s what we’ve done in the past with devices and human-centered design and human factors. We have to do the same with software and approach it in the same way with digital.”
Related: What J&J MedTech’s new Dualto says about the OR of the future — and Ottava
The term artificial intelligence is being thrown around a lot right now, everything from algorithms to large language models, and I’d like to know how you define AI and what kinds you think have the most potential for medtech?
Jegatheeswaran: “When we talk to customers, there are typically two flavors of value pools. One is clinical in nature, and so we’re hearing a lot of that from surgeons themselves and their teams. The other is more administrative on the business side of the hospital, and it’s more around efficiency. Both are valid value pools. Surgeons today are facing almost an impossible situation when they’re going to a procedure. Patients are living longer, comorbidities are more complex or exist, there’s a ton of new technology coming out, and the workforce is thinning out. It’s a perfect storm. Surgeons are asking for help, and AI can accelerate and augment and automate steps within the procedure process, at least initially, that make it simpler. Every surgeon, before they go into a procedure, is doing some sort of pre-thinking. They’re speaking with peers, looking at imaging reports, looking at patient health records and histories, and looking at their own notes on procedures with this patient or previous patients of a similar nature. All of that is manual and not recorded anywhere. Why can’t we collect that dataset and on top of that run an AI model to give them the salient outcome of the top three risks, things you can do preemptively, things that you want to make sure your patient does before they come into procedure, comments to your team in terms of how they can prep best the night before and the day of? That is something that can be done for surgeons with the tech that exists today, and that’s something that we’re working on. That’s a big outcome just in terms of preparing in the best way for a complex or normal procedure. On the efficiency side, we’ve seen this happen in the movement of people and the optimization of the movement of people and activities within a confined space. The OR is a dynamic experience: people coming in and out of the room, a lot of equipment working, a lot of sounds. The efficiency side is essentially how many procedures you can do in a day, and how you can decrease the level of complications coming out of procedures. You think about technology like ambient AI in the OR, laparoscopic video, and then connecting the dots with patient outcome and then the EMR coming in. That for me is an efficiency play that many companies do a good job of today. And AI has a role to play because you can optimize at an OR level, you can optimize at a hospital level, but with AI, you can optimize at a system-of-systems level, and those best practices can then be fed back to nurses, administrators, etc. That’s why I see the value of AI in the short term. In the long term, the jury’s out in terms of what AI can help surgery with. I actually think surgery is probably the most personal and sensitive use of AI. It’s literally within someone’s body. And so while speed is important, so is trust and quality. We want to approach this responsibly.”
Related: J&J MedTech arms its Monarch robot for futuristic lung cancer therapies
How do you build trust in AI?
Jegatheeswaran: “We don’t start with looking for a problem to solve with AI. It’s not a hammer looking for a nail. The first step is being authentic about what you’re trying to solve. We have very close, intimate relationships with surgeons and clinical teams. We’ve been in surgery for over 100 years, and we’re proud that we’re part of the craft. So the first step is starting with the problem we’re looking to solve. The second is meeting the user where they are. We can drop a Ferrari into the middle of a desert, but it’s absolutely useless. We need to build the infrastructure consistently and piecemeal and change along with our end users and the market and have that evolve. We’ve done that sort of curve going back to sterilization and laparoscopic. Coming forward to digital, we’re going to have to have that same change management curve. Net-net, it involves careful design and co-creation with our end users — which we’re pretty strong at — working backwards-in, understanding what is it we’re trying to solve and whether AI or digital is the actual path. In many cases the answer is yes. In some cases, no. And third is education and training, from new residents all the way up to very experienced surgeons and teams. There’s an element of education and training that often gets overlooked, and it’s really important and near and dear to us.”
Related: Philips offers recommendations for building trust in medtech AI
Can you tell us a little about the digital improvements you led at Baker Hughes before you joined J&J?
Jegatheeswaran: “I came from the oil and gas industry, which was great training in some ways for surgery. It’s a regulated environment, global business, the stakes are high, similar to surgery. … It was about efficiency: How do you make your production field more efficient, create the least environmental impact when you’re drilling, make sure you’re drilling in the right areas? It was a lot around workflow efficiency, training and safety for the workers, accuracy — whether you’re drilling or producing oil or gas — and doing it at scale. There’s no space for one-offs in oil and gas, similar to surgery. Being able to do that at scale in very remote places was our focus.”
Do you have tips for device developers and manufacturers trying to make similar moves toward digitization?
Jegatheeswaran: “Embrace it, because if it’s not here, it’s coming. Be humble enough to understand that no one company, no one entity, is going to figure it out. This is going to have to be a coalition of the willing. And that’s how we’re approaching it. Stay true to the starting point, which is patients come first. This is another technology wave. It’s not changing the fundamentals of what medtech needs to deliver, which is better patient outcomes, and that’s always going to hold true.”
Do you have a mantra or motto that your team would say you repeat all the time?
Jegatheeswaran: “Ultimately, the user comes first. That’s really the guiding principle. It could be a patient, it could be a surgeon, it could be anyone. That moment of truth when they’re engaging with the product we’ve built, and the susceptibility to use and reuse that product is what makes or breaks a good solution. Because we can have the best science in the world, but if it’s not adopted, it doesn’t really matter. Technology is one leg of the stool, but data and design are the other two legs. Data has to be quality compliant, etc., and then design is your technical component, which serves cyber and cost, and then your user component, which is your interface and your human design. If you don’t have those three legs of the stool, you don’t have a stool. The user comes first.”
