Dr. Edmondo Robinson Outlines Hospital Readiness to Receive Large Amounts of Patient Data

Hospitals and health systems are largely not equipped to handle all the patient data they could receive and turn it into actionable information, explained Edmondo Robinson, MD, MBA, MS, FACP, senior vice president and chief digital officer at Moffitt Cancer Center.

Hospitals and health systems are largely not equipped to handle all the patient data they could receive and turn it into actionable information, explained Edmondo Robinson, MD, MBA, MS, FACP, senior vice president and chief digital officer at Moffitt Cancer Center.

Transcript

Oh, absolutely not they have they don't know what to do with those data. Especially we start talking about wearables, right? So, you think about it, a patient says, "Hey, I, I'm averaging 10 more steps this week than I was last week." What does that mean? You know, other than saying, "Yes, keep going. That's wonderful," it's hard to know what that means.

And so I think there's a process around structuring your approach to understanding, in this case, remote patient monitoring, wearable types of data, to say, are there criteria whereby a certain change — because it because some of those data are even continuous, right, continuous monitoring — so are the criteria whereby a change would trigger an intervention of some sort? Are there summaries? You know, a weekly summary, a monthly summary, whereby understanding those data at that level across over time is important for many intervention perspective. It depends by the data. And that's not even including all the research you want to do, which is going to ask the questions differently and attack them differently, depending on the hypothesis being built in that particular research study.

So, you know, it could be everything from "I have a patient on a particular chemotherapeutic intervention; they're now at home after getting their infusion in the infusion center, and I want to monitor adverse drug reactions or events." Well, I could use wearables for that. But what does it mean, if the heart rate went up a little bit? Or if the heart rate went down a little bit? Is that a 88 [heart rate]? Or is that just normal variation in a heart rate?

You've got to set those parameters and you've got to understand what it means to have those data; but just looking at it constantly, it's not gonna be helpful. So, again, you really got to be disciplined about your approach to those data.

Will health systems and hospitals need to implement things they don't currently have, like artificial intelligence (AI) or different algorithms?

It could. That could absolutely accelerate your ability to sift through those data and get to...separate the signal from the noise, right? You absolutely could leverage machine learning to do that. You need a ton of data to train those models really well. But, ultimately, I think that's where it's going to go, whereby what happens is, the AI surfaces up, "hey, this, look at that, that's important."

But even short of that, you can just set some hard parameters around some of these things, right? Where, you know, they're not always perfect, but they at least signal that you should pay attention. So, for example, if your heart rates over 100 [bpm], I'm going to pay attention. Now, it could just be that you were out for a run — not a big deal. Or it could be that you are having some kind of reaction. And I'm thinking about even for patients who are on the inpatient side. It could be that I caused that high heart rate or I know the reason for their high heart rate. Or it could be that I don't know the reason. And I better go figure out what the reason is, right? So though that's just a hard, it's not even AI driven. That's just a hard line, that 100 right. So, there's some of that.

The reason why machine learning might be really critical is often it can pull signal from noise that we can't see normally. Right? And it might do it sooner. And that's actually really important.