Big Data, AI, and Advancements for More Personalized Medicine

Geoffrey Gurtner, MD, FACS:

I am Geoff Gurtner. I'm the chair of surgery at the University of Arizona. I am a general surgeon and plastic surgeon, and I run our wound care programs and have an NIH-funded laboratory that looks at the basic science of wound care.

Yeah. So I think wound care and wound healing problems are so common that we really struggle with being able to take care of the volume of patients that have these problems. And so the promise of big data and automation and machine learning and artificial intelligence and sensors and wearables are that we can somehow be able to pick out the patients that are very likely to do poorly so we can devote more resources to them and then understand which patients would be predicted to do well so that we can use less resources potentially on them since they are likely to heal without much of an intervention. And so really figuring out at the very beginning which patients are at high risk for bad outcomes such as amputation, sepsis, admission to the hospital, all of those things that cause problems for patients and for healthcare systems.

Well, we've been working in this space for probably 10 years now, and we started off working with the electronic medical record, and we've evolved into using sensors and wearables to predict bad outcomes. And so it's just been a big focus of my laboratory and my clinical trials for the past 10 years. And so when I was invited to speak on this topic, I jumped at the opportunity.

So we originally thought that just from features in the electronic medical record, that would give us enough resolution to actually decide who was going to do poorly. And although it is very useful, it's probably more useful as a population-based predictive model. And so with that as our background, we decided to start working in sensors and we at Stanford actually developed a wearable that could monitor the wound, the wound conditions and then also potentially intervene so the patient could have a treatment delivered without ever seeing a doctor through kind of an autonomous closed loop system.

And so that is really, I think, where we eventually will get to with personalization. We will have enough data that we know if we see this constellation of indicators that something bad is going on and that could either alert the doctor, tell the patient to go to the emergency room or go to the wound center, or potentially even deliver a treatment, which might be very useful in low resource settings or in austere environments in developing world where there's not a lot of doctors around. So that I think is the ultimate goal.

Well, I mean, for me, even though usually in wound care center we see the patients every week, there's always a patient or 2 in any given clinic who deteriorates in that week-long interval. And if you could see them when they just began to deteriorate and potentially could intervene, you could prevent many of these patients need to go to the emergency room then, or go get admitted to the hospital or have an operation. If you could intervene between doctor visits, that might again theoretically decrease the complications.

I think that these technologies will be useful in the near term that I think that sensors, wearables, AI will be able to be an early warning system for our patients so that they can hopefully get treated earlier and have better outcomes. Again, we haven't proven that yet, but I think that would be kind of the holy grail. If we can prove that, then there are multiple pathways where you can envision these technologies being reimbursed. And that's the real problem right now is that there really isn't any reimbursement for many of these technologies, and so they just don't get adopted, they don't get used other than as investigational tools. And I think if we can actually prove that using sensor technology and artificial intelligence and machine learning, we can prevent bad complications or bad outcomes, we can save the system money. And then I think there will be an easy path to reimbursement and wider adoption of these technologies.