Leveraging AI for the Benefit of Patients
In an interview with Pharma Commerce Editor Nicholas Saraceno, Matt Hawkins, CaryHealth’s Chief Technology Officer, describes how artificial intelligence can not only diagnose patients, but also deliver accurate medical advice.
PC: How can AI accurately diagnose patients and provide reliable medical advice?
Hawkins: I think the benefit of AI—especially in 2024 as we've seen the proliferation of generative AI over the last three to four years—is its capability to be able to look at a huge amount of data. I would look at that data in context, and that's typically an operation that would that for an individual would take a long time to do, so reading through medical history, looking through medical images, looking through things like lab results. Bringing all of that information together takes a lot of concentration. It takes a lot of time from a human in that situation. AI provides the ability for it to look at this full data corpus very quickly and to find the patterns in that data and give the human on the other end a very quick answer that’s very much in context to the question that has been asked, and to do that at a scale which has not previously been possible before.
When you talk about the reliability as well, I think we all know that data can be analyzed in such a vast capability right now. How do we know that that's reliable? I think it's very important when we're presenting things like medical information, is that we're able to back up the decisions that we've made with referenceable answers, so that when you're getting a response—whether in a patient context or in a healthcare provider context—you get instant feedback on why you're getting that answer. Where's the reference, where's the clinical guidelines, where's this piece of information come from? How does that relate to the specifics within the patient profile? That serves a couple of purposes.
One is that it's giving you confidence that you're getting an answer that is based within truth. And two, it's giving you that ability to instantly verify that information as well. Context on the way in, and context on the way out. I think from that point, there's several places that you can go. We can talk about personalized treatment plans. We can talk about simulating medical scenarios. There's such a broad application of this that we could about lots of different applications.
