Transforming Pressure Injury Detection: The Role of Vision AI and Virtual Nursing
In healthcare, AI is increasingly transforming the way we approach patient care. One of the key goals is making AI explainable and actionable. Imagine AI monitoring patients who have been in bed for extended periods without movement. Instead of being a “black box” spitting out complex data, AI can notify nurses, highlighting that specific patients haven’t moved and may need attention.
This makes the AI a powerful partner, not an overseer. Nurses aren’t told what to do but are provided with valuable context. The AI might suggest, “Check on Ms. Smith, who seems lethargic,” empowering the nurse to use their judgment and context, perhaps through checking medications or performing a video visit. This proactive approach enhances decision-making without overwhelming nurses with unnecessary data.
Pairing AI with virtual nursing brings even more opportunities. Virtual nurses can process large amounts of data that AI delivers, such as patient mobility trends, and collaborate with bedside staff to ensure optimal care. For example, AI might show a patient has been in bed continuously after surgery, signaling a possible setback, which the virtual nurse can flag for the care team.
The goal here isn’t to replace human decision-making but to support it. By analyzing data trends, AI becomes a tool that enhances critical thinking and helps nurses focus on the most pressing issues. AI functions like a “fastest intern,” always monitoring and providing helpful insights while leaving the ultimate judgment to the nurse.
This blending of AI and virtual nursing holds immense promise for more efficient care, better patient outcomes, and reduced stress for nursing teams. As healthcare continues to innovate, this partnership is just the beginning.
Listen to Holly Lorenz, Narinder Singh, and Tiffany Wyatt explore pressure injury detection revolutionized by Vision AI.
Video Transcript
Narinder Singh:
I want to shift gears because I think one of the things that has not existed in a lot of places inside the hospital is dealing with dramatic technology shifts. EHR was the 1 cent generation shift, but the real thing is that we’re seeing a rate of innovation with AI that’s very rapid in our personal lives, but trying to keep up with what that looks like inside of a setting where lives matter like a hospital is challenging. And so one of the things we’ve talked about, and I want to walk through a couple of examples, is how to make AI a partner to the nurse versus feeling like it’s somebody who’s telling me what to do. So a couple of examples that we think about is we think about kind of nudging action. So for example, being able to say here is a, here’s a situation where I see a list of patients, I’ve just sort of in the high, medium, low who we’re worried about for pressure injury and the AI is watching and saying, well, how do I explain this?
Well, we want to make the AI explainable. So it’s like these are the people that are in bed not moving very much and haven’t been turned for a couple of hours. So now the AI is not a black box. It’s like, oh, okay, yeah, I would want to pay attention to those. And for those maybe I’m going to look in the EHR and see what medications changed or look at their movement or even do a video visit with them to ask Ms. Smith why she’s been a little lethargic this morning. So this is an example for us where the nurse is not being told what to do in ai. It’s just that AI can be in every room all the time, whereas the nurse always has context. And I think there’s such a strong pairing with virtual nursing because we don’t want to overwhelm the bedside nurse with data.
We want somebody that can process through this. So by the time we get to a conversation, you’re talking to a peer who just seems like they know exactly what’s been happening on your patient’s room. So this pattern of AI that’s partnering with a virtual nurse to help support the bedside team is something that’s kind of paramount for us. And I want to just walk through two more examples and get your reaction to the whole group. Another example might be this is a chart of day by day. Is the patient in bed all the time? Are they starting to walk around the room? Are they sitting in the chair? This patient was in bed a lot, maybe they had a surgery in bed, in bed, looks like they were up and around and wait all of a sudden looks like they had a little bit of a setback.
They’re in bed a bunch again. And when you look at this data individually, it’s somewhat interesting when you can start to put it together and say, here’s everybody we’re thinking about discharging in the next 48 hours, but all of a sudden you can quickly see these three patients are all blue, meaning they’ve been in bed literally a hundred percent of the time. It gives you some skepticism that that’s true. So these are just examples of where we’re not taking decision making away, but we’re giving an opportunity for a virtual nurse to evaluate it with the information we’re showing, but also the whole context of that patient’s care. And so maybe I’ll pause for a second before I’ll get into more of the real time pieces. How do you think we go about instituting change in these areas to allow for that bedside nurse to feel like the AI is not my big brother, but it’s my guardian angel, especially when paired with virtual nursing.
Holly Lorenz:
A really easy example to use is I’m old enough of a nurse that I remember when we saw a lab value.
Then miraculously you could see trends. You didn’t have to go back and look day after day in a paper chart that shows you my age or in an electronic record. It started to trend things for you. So when you trend things, it helps you really be able to use that trend line as someone who’s serving up some ability to do some critical thinking for you or even better to allow you to focus on something that looks concerning. So it isn’t telling you what to do, it’s not prioritizing your care, but it’s helping you triage through a lot of information, particularly with AI that we would never have the time to sit and discern. And it doesn’t mean we’re taking the thinking away. Because if you were to see all blue, it doesn’t tell you what to do.
Need to know what that means. You need to know and decide you’re going to intervene on patient A or patient B with what’s happening related to this. And that nudge I think is it’s like someone who’s tapping you on the shoulder,
Narinder Singh:
It was once described to me as like, we want to turn AI into the fastest intern you’ve ever seen, instantly checking on every patient in every room and telling you what’s happening when you need to know. And I think that’s kind of the pattern we’re looking for.
Tiffany Wyatt:
I was going to say real quick, nor I would take it one step further, Holly, in saying it’s not even information to sift through, it’s information that nurses have never had before. We have never had this level of information of how long have they been in the chair, how long have they been walking around the room? And I think it’s really powerful when you start to look into AI to understand how much information it can gather that we want to know that we’ve never even seen much less had to sift through like we used to. And in theory,
Holly Lorenz:
Some of this information doesn’t even need to be sifted to a nurse if indeed the protocol is someone’s been blue for three days or maybe it’s even a day that there is a team that’s a mobility team that manages that, so it doesn’t even have to be a nurse who does the interventions and some of these things.
Tiffany Wyatt:
Yeah, the opportunities are endless really.
Narinder Singh:
I think the thing that we’ve figured out is this pairing, which we’ll talk about later between the virtual nurse and AI or vision AI as we call it, is very powerful because it lets us start with saying, we’re not going to impact the bedside at all. We don’t have to show you this information. We’re just going to be there and say, Hey, by the way, your patient was restless all night last night. Okay, great. Now I have some information that I wouldn’t have had otherwise. And then maybe over time we say, actually every time you walk into the patient’s room and during a morning shift change, we’ll tell you how they slept overnight and we give that data out. So the virtual nurse gives so much flexibility for allowing innovations like AI to make it to the bedside without having a massive change management exercise when you’re still trying to figure out the right pattern.
So this is, to me, one big part of that and a place where we’ve already seen that is around virtual fall prevention. So this is a world where you have had the ability to change things in virtual sitting, but we still watch so few patients because it’s so hard for a person to watch more than 10 or 12, 15 patients. But with ai, we can set up a safety zone around the patient, the virtual fall prevention observer can adjust that, and then as long as the patient’s moving only inside that range, it’s fine, but if they start to put themselves in an unsafe situation, we nudge them. And so that’s a way that we’ve kind of dealt with fall prevention, which is a much more immediate feedback. Obviously they could fall. Then you take it one step further and say, the AI is going to watch every patient and now what we’re going to do is still we’re going to leave the final judgment to the human, to the person who’s saying, I think they’re just stretching versus this patient’s trying to get out of bed.
Let me talk to them. Let me get them to stay in bed. The AI is surfacing that, but the human decision making is still the humans. And so you get the best of AI being everywhere but still leaving the person, the virtual observer empowered. And so this pattern of what we call infinity watch, you can imagine for things that you said, infinity watch for potential violent situations where we nudge and sorry, that infinity watch for pressure injury, infinity watch for the blinds aren’t down in these rooms overnight.