Episode Transcript
Speaker 1 (00:01):
Welcome back to the next installment in the latest series from the Post-Acute Point Of View podcast, our discussion hub for healthcare technology in the out-of-Hospital space. Here we talk about the latest news and views on trends and innovations that can impact the way post-acute care providers work. We'll also dive into how technology can make a difference in today's changing healthcare landscape for home and facility-based workers and the people they care for. Let's dive in. I'm McKnight Senior Living editor, Lois Bowers. I'd like to introduce today's speakers, Allison Rainey and Daniel Zhu. Allison Rainey has served as MatrixCare's head of Nursing and Clinical Informatics since May. She has experience in supporting the development, implementation and leveraging of healthcare technology for improved clinical and operational outcomes. And she also has experience in developing healthcare and healthcare technology strategy to meet corporate goals and objectives. Daniel Zhu recently joined the ResMed and MatrixCare team to lead their productization of AI and machine learning capabilities. His diverse background includes clinical and clinical research experience and technology entrepreneurship. And he also has spent time as a data consultant for large corporations such as Ford Co-op, and RBI. This is such a timely and interesting topic. So without further ado, I'm going to turn things over to Allison and Daniel.
Speaker 3 (01:33):
Thanks for our introduction Lois and hello everyone. Thanks for joining us today for our learning session on artificial intelligence in the long-term care space. I hope by the end of today's session everyone will be able to walk away with any key challenges and outcomes from implementing AI systems at your organization and what these intelligent systems hope to achieve in long-term care tomorrow. So Allison, did you implement any of these technologies? How did that affect your organization overall?
Speaker 4 (02:05):
Yes, yes, we did. We absolutely did. We use these technologies in many different ways to positively impact our patient outcomes and our nurse satisfaction. As every clinical person knows, managing negative outcomes, whether it be a fall or rehospitalization or anything like that, is extremely burdensome on the caregivers, on the cost of care, and certainly not at all what any of us are looking for for our patients, right? We get into this business and we do what we do to do the best thing for our patients and not have any negative outcomes. So those negative outcomes that are preventable are pretty heartfelt. With these active clinical monitoring tools, the clinical data is able to be synthesized to highlight those subclinical changes and non-subclinical changes right away that might need to be addressed. Pulling all of that information together, you can highlight various things in real time to be addressed before you have the rehospitalization or the fall. Those things that are changing.
(03:09):
Clinically, as we know, those of us that have taken care of the senior population and certainly in school and everywhere else, we know that this population, it's very, very important to identify changes early. Anything that is subclinical that we might not as humans recognize right away as a change, the machine can actually highlight those for us and pull different things together. So if certain things are happening in patterns in the background, one may not look like an overt change, but together the machine can actually say, "Hey, these four things or whatever are changing" and this altogether can make a difference, and does indicate that this patient is increasing for risk for hospitalization. They've got a change in condition, they've got potential for fall. So with those being floated to the top, it's so important that we take care of those things early.
(04:01):
As clinicians, we know that if we can address those things early before they've become too far gone, we can certainly change the trajectory of this change or what's going on with the patient much more successfully than if we were waiting till it was very obvious that there was a change in condition happening. So again, so much potential and so much that we can do for our patients. In a better way using this machine learning technology to impact the population that we serve. Beyond that, with the help of these systems, the clinical staff are also able to spend more time on other clinical items that are also identified by the platform. So we've got the big changes in condition that certainly we want to address, but we all know about antipsychotics and how important it is that we make sure that we are only prescribing those for the patients that really need them.
(04:56):
There have been great strides made in the appropriate use of antipsychotics in nursing homes, antipsychotic rates have gone down from 23.9% of long stay nursing home residents receiving an antipsychotic medication in Q4 of 2011, we're down to 14.6 in Q 4 of 2022, according to the DMS national partnership to improve dementia care in nursing homes reports. So a lot of good strides, but certainly still something to be focused on. There is appropriate use of antipsychotic medication in our senior population, so certainly there is appropriate use. But with the potential for adverse events like a potential for memory loss, there is potential for decreased function. There is potential for increased suppression and even mortality with these antipsychotic medications. And the potential for inappropriate use in the senior population is there, especially because dementia patients, I think is one subset that they do have behaviors that are not typically associated with maybe what's actually going on.
(06:03):
So some dementia patients are unable to really let their caregivers know that they're in pain effectively or that they're feeling discomfort or some other things that they might need. And so it's confused with behaviors that might need an antipsychotic because of some psychosis. So we really need to be careful and cautious that we're treating the patient as the unique patient they are and getting to the real root of those behaviors. But sometimes it just doesn't happen. And so those can accidentally be prescribed or misprescribed thinking that they're treating another cause. So they continue to be monitored, that's a long way to say they continue to be monitored by CMS. That's a part of the nursing home quality measures. And just to go into what those are very, very quickly. The short stay measure includes the percent of residents who newly received an anti-psychotic medication. Long stay measures are there for anti-psychotics as well. And they are for residents who received an anti-psychotic medication at all during the lookback period for their MDS.
(07:07):
So those are the two that CMS is monitoring on all those patients. So using these clinical monitoring tools, you not only can monitor for new orders, right? With these tools, they'll let you know a new order has come in. And so that can be taken a look at very quickly and for appropriateness by those that would be doing that. But clinical monitoring tools can also monitor for other things that might trigger an antipsychotic order from being initiated. So with some of these tools, behaviors related to treatable causes can be floated to the top and highlighted for clinicians so that they can address and look for the pain or the discomfort or really look at those things. So all of those things can be highlighted early and addressed appropriately before maybe a prescriber comes in and orders an antipsychotic not understanding all of the other things that are going on with these patients.
(08:10):
So again, very, very, very powerful tools that we have that can go into that record and filter through all of those items that are relevant in making those decisions. So what we saw with some of these clinical facilities or the clinical staff at the facilities have reduced their anti-psychotropic medication administration leveraging these monitoring tools specific to anti-psychotics. What we have found in studies already done are that in the short, say antipsychotic measure alone facilities using clinical monitoring tools as a part of their strategy actually decrease their rate by 20%. So that's big, right? Any patients that we can impact are big, but the fact that we are seeing these big trends is certainly powerful and noteworthy. Besides better patient care, these advanced technologies, they support your high quality staff and because your high quality staff can then do what they do, great, that translates to better health inspections and better quality metrics. So it's win-win. The fact that this is coming into our environment is just great for everybody.
Speaker 3 (09:19):
Yeah. Wow, that's some pretty amazing stats there. 20% decrease in some of these administration of these antipsychotics.
Speaker 4 (09:30):
Yeah. I mean, it really is. And it's simply from highlighting those key parts of the resident chart so that the skilled clinicians can take a look at it and do what they need to do appropriately. It helps the clinical staff sift through all the noise that I know, those of us that have [inaudible 00:09:46], there can be a lot of noise coming at us with all of the data and all the information, and they can focus on the important aspects of the clinical condition and then develop appropriate plans of care and really see those unique individuals that we take care of clearly for the best outcomes.
Speaker 3 (10:02):
That makes a lot of sense. So can you tell us a little bit more about the implementation process you and your team might have gone through?
Speaker 4 (10:11):
Sure. So I have to say the teams that I worked with were absolutely amazing. So I was very fortunate that I had very engaged teams that were really willing to do what they needed to do to get this done. But I think for most people, when you hear about a tech change or a new implementation, anybody that's been through some of these things especially in healthcare, you get very anxious and start thinking of all the what ifs. But on the bright side, these kinds of technologies that really are embedded in and automated in your EHR, because most of them are because that's the way they're designed, is to really pull that data out and be within the workflows knowing that there's so much to be seen out there and for these frontline caregivers to look at. So these automated clinical monitoring tools are not the kind of implementation you think of when you think about implementing a new technology.
(11:07):
They're fairly seamless because they're in the EHR, there is no data conversion to consider, so because it's all in the EHR, you don't have to look for it being pulled from here and transitioned over there and interfaces and all those kinds of things. It's right in the system. And because of that, there's no staff coming off your floor to take a look at the data and making sure that the data is appropriate coming from one system to another. No review along those lines, no additional URLs that they have to go to access these tools. The data, again, is coming from the EHR. So we were very lucky in the fact that that's where that was coming from. And so the implementation at base level was very seamless. It was just there. It was a matter of flipping it on. And then really the hard part came about when we were considering how to incorporate that into daily care.
(11:59):
So simple orientation tools. I think that it was very important that we talked early about how they were designed at a high level so the clinicians felt very comfortable with what they saw before we even got started, because there's always going to be that skepticism with any new technology. So there was that, we really had to talk through some of that. And then the expectations, because it is a little bit of a change, but there was no launching other websites. There was no additional usernames, no different screen layouts because it was all typical to what we already saw. I would suggest if anybody out there is looking for effective clinical monitoring tools, just make sure that you're choosing some that navigate that careful balance of having the right amount of information, I should say, at your fingertips versus too much information creating noise. Because there's always that balance. So as you're evaluating your tools and what you're considering for your facility, just look out for that because it certainly, it's critical if we have too much noise coming at us, it's hard to make decisions. The right information at the right time is critical.
Speaker 3 (13:05):
Yeah. That absolutely resound with that for sure. Because I think we have to on a high level, stop imagining AI as something that's added to your work, but the systems that are being built today, they're being built right inside. Like you said, they're being built right inside the EHRs, right inside systems that you're used to using every single day. One good example is, I'm not sure if everyone here remembers when auto complete first came out. When you're texting your friends, it started out a little bit clunky, but it didn't really impact the people who didn't use it. Eventually, as it got better and more efficient, more accurate, you've gotten used to using it and it starts to fit really well in the way you text, the way you search in Google as well. And now it fits seamlessly in your day-to-day interactions with all these technologies. So we talked about some of these, the implementation process, but Allison on your side, were there any challenges during the implementation process?
Speaker 4 (14:06):
Well sure there were. Anything new there's a challenge with. Again I have to say I have to give credit to the teams that I worked with were amazing, and I had some great leaders that I worked with. But even though tools were easy and fit nicely into the current workflow because it was an EHR and those types of things, there's not going to be automatic adoption because it's new. So again, I go back to more explanations so that folks feel comfortable with the conclusions that it's coming. It's hard to change patterns. So clinicians typically have a routine for how they review the EHR as they're doing a task or taking care of a patient or access pattern. So they access this and they access that. And so those are really, if you do that a lot during the day or a lot during the week, those are stuck.
(14:57):
And so to change those are difficult. Even though we talked about, "Hey, if you use these technologies, it might be able to take some steps out of what you do in other areas, or it improves outcomes. It's very, very valuable". Changing longstanding existing patterns can be hard. And facilities, especially nowadays, they've got so many competing priorities, especially those frontline folks that are really at the bedside taking care of patients. There's a lot. And so again, I go back to changing those patterns can be hard at times. And so getting groups to adopt was a little challenging in some areas.
Speaker 3 (15:37):
Got it. So knowing that, would you have done anything differently?
Speaker 4 (15:41):
Yes. Yes, I absolutely would. In hindsight, there are a few things that I would've done differently, especially now that I recognize the impact that these tools have. I think as a leader, I would've been much more diligent in monitoring and producing the usage metrics that aligned with the KPI that we were impacting, or the key performance metrics that we were impacting. Really, I just wish I did that more frequent and got it to the folks that were really managing the systems and the workflows. I think if I were more consistent with that and speaking directly to the value and the changes that we were seeing, I know we would've been more successful with general adoption much earlier in the process. Certainly that came, but I think we would've been more successful earlier.
Speaker 3 (16:25):
Got it. Earlier you described a falls program. Can you tell us a little bit more about that?
Speaker 4 (16:31):
Sure, sure. So we also in our monitoring tools have falls, has a fall tool, and even with something as intuitively positive as falls predictor tool or a falls risk tool, knowing how important that would be in our environment, change is still hard. Again, we found that clear consistent messaging was important, expectations that it's there and what's going on. But with that one very specifically, there were a lot of questions because it's important. So we really changed our approach to make sure that we gave a very high level information education around what AI and ML was as we were doing our presenting of what we were trying to do with these tools. And I have to say this framework is a great framework to launch from because I think this would've helped me as we were trying to explain that a little bit better. So your slides are perfect.
(17:31):
So initially they went through the skepticism and then there was a fear of use. Because we're dealing with humans, right? If there's mistakes, it's critical that the tools that we rely on are accurate. And so again, that descriptor and really trying to explain it early so that people understood and felt like they could rely on the information was very, very important. And then I think with the falls program specifically, we talked a lot about the benefit to the patient because at the heart of every healthcare worker that spends any length of time in a SNF or a long-term care facility is the patient. I mean that's where our hearts are, it's the patient. And so the benefit to the patient spoke volumes to those dedicated nursing home caregivers. And then we talked a lot about time because the more time that we can give back to our clinicians and our caregivers, the better outcomes we're going to have and the better field they're going to have for what they do.
(18:26):
So we spoke a lot to our healthcare workers about how these tools could decrease the time spent preparing for meetings because the reports are already built for the most part for them, or at least highlighting the patients that maybe needed to be brought to certain meetings or talked about in certain meetings. We also could use some of these tools for shift change reports, putting the falls risk, here's the patients to focus on and here's the why. And then orienting transient staff, knowing that agency, there's a lot of agency out there, it's getting better, but it's still out there. And those nurses need to know too about the patients that they're taking care of and their patient loads. So this is very powerful in being able to help with those folks too, to make sure that they're administering the best care that they can. In general, it's very important to highlight how these tools prevent falls, rehospitalizations, or other outcomes that can take away from the clinician and caregiver day so that they can definitely take care of those patients, prevent falls. And I think that was very, very helpful as we implemented this product to our teams.
(19:29):
We needed to share what was in it for them. And I think that was something that maybe we missed a little bit at the beginning too. Because as leaders, we want our caregivers to know that we're doing this for them as well. The patient's outcomes, absolutely. But we all care about our caregivers. That's where the rubber meets the road. And so it was very important to share how this is going to be helpful to them too. And that's what I think got our falls program up and going as they leveraged the falls tools in their processes. Okay, sorry. So we did see amazing clinical outcomes by proactive clinical monitoring directly related to the falls. So as we all know, risk for falls increases with age and so do negative outcomes related to falls.
(20:15):
I mean, those are known facts, especially in our industry and the population that we serve. And we don't want falls to happen. So preventable falls we want to address and we want to get that whatever it is taken care of so they don't fall. So not only do we not want our patients to be impacted that way, but with that comes a quality of care measures. So falls impact provider metrics related to falls, but they also can impact our length of stays, our re-hospitalizations and other publicly reported quality metrics and ultimately overall cost of care as well. So in this value-based care world that we live in, that's certainly very important. Using the tool that we used, those groups that implemented it so well, we saw long study measures including the need for more help with daily activities, independent movement worsened and experienced falls with major injury, all decreased by 25%, 26%, and 8% respectively in a study of the sites using the machine learning supported advanced clinical monitoring tools for falls risks.
(21:17):
So these types of tools are very effective in the population that we serve. Not only that, but some of those same facilities also experienced a 7% actual increase in their short stay measure around residents making improvements in function. So good machine learning. Again, it's good for the patient and I think it's only going to do more great things as we understand it more. Every care in this industry that's been in this industry for any significant length of time, [inaudible 00:21:43] to take excellent care for seniors. How could we not use this technology to improve what we do? It makes a difference.
Speaker 3 (21:50):
I absolutely agree. And it's a great way to highlight that these technologies are only here to empower and drive efficiency and effectiveness. In fact, I think if anything, we're just going to see more and more of this technology transitioning into the next slide in the future of long-term care. So personally, I believe AI is going to become more and more ubiquitous in long-term care, just like how it's already in a lot of our other day-to-day life scenarios like auto complete, like the way that detectors work here and there. And I think it's going to start empowering every form of documentation, not just in long-term care, but everywhere and making our clinical decisions just so much more informed and just so much more faster. Right?
Speaker 4 (22:39):
Absolutely. AI is, Daniel, it is going to be a game changer for our industry. Our industry is going to be transformed by this technology, there's absolutely no doubt. So I think I'm just going to speak to a few major areas that I see changing in our industry. I think it's going to improve the care that we deliver. I think it's going to decrease spend. It's going to improve healthcare satisfaction across the board. So I see it helping our capacity to really care for each person as a patient, as a unique person. Using data we can proactively, dynamically adjust our care in response to a person's unique clinical status history, and preferences. I mean, how amazing is that, right? Timely, being able to do that to take care of that unique person. I also see technology driving improved patient and family experience. Secondary to what I just spoke to, but they're going to have increased comprehensive care, higher quality of care and improved outcomes. That's what we get into this for.
(23:36):
So our healthcare industry unfortunately has the highest spend currently on healthcare per person, and yet we still have some of the worst outcome rates in the United States. So not our industry, our healthcare industry in the United States overall has some of the highest spend and not the best outcome rates across the globe. So we know that value-based payment models are expanding to try to adjust for that. They are expanding in their adoption and their levels of care being incorporated into the equation. So I see machine learning and AI helping to ensure that patients are being cared for in the right level of care. I think they will help us transition to the lowest appropriate level of care at the right time for these patients with the right plan of care that supports that unique patient's best outcome.
(24:25):
I see with all this. Of course, we're going to see improvements in length of stay rehospitalizations, total cost of care, through these AI supported technologies. I also see technologies through collaborative discussions with our regulators. I think AI supported technologies might be able to fill some of the voids we are currently seeing in healthcare. So I think there's an opportunity there for growth. I also feel that healthcare workers' experience being transformed from one where there's a lot of mundane and repetitive tasks that take up a lot of time, hoping to move that to where they have time to perform at their highest level of licensure certification or skill by augmenting or automating some of the work that they have to do today. So I think there's going to be a lot of opportunity there too. I see a future where healthcare workers have time for empathy, comprehensive clinical judgment and true patient connection.
(25:16):
So I think that's going to, hopefully with some of these technologies, give time back to our healthcare workers to be able to do those sorts of things. I also see a future of highly fulfilled, highly empowered, and appropriately informed and highly effective healthcare providers, being MDs, PAs and NPs, clinicians and other healthcare workers. I think this is really going to improve satisfaction for what they do. I see a future with great things for a healthcare system where providers, clinicians, caregivers, and patient interactions, experiences, outcomes, and satisfaction are all at their highest level. These technologies are going to do great things.
Speaker 3 (25:52):
Yeah, I couldn't agree more. I think at the end of the day, it really kind of gives the power of care back to the care providers. And in addition to working with all this great documentation that's already in the EHR, AI is only going to accelerate, like you said, as more devices are introduced in long-term care. From helping with vitals to patient monitoring, there's going to be much better and more consistent ways of monitoring, and that's much less intrusive for our residents in all these long-term care facilities. It helps us keep up to date on their condition at any time right?
Speaker 4 (26:27):
Yes. Which is needed, especially like you said this time with the healthcare worker shortage. As an industry, we are trying to tackle this issue from so many different directions, and one of the biggest opportunities I see is the expanded use of appropriately leveraged technology. That nurses, they're simply not out there. And so nurses and other clinical staff need to be functioning most efficiently and at their highest level to provide good outcomes. I mean, imagine a world using AI technology where we have intelligent sensors and AI assisted care, intelligent comprehensive coding support based on data within the record voice-driven documentation, automatic data-driven falls prevention support, plan of care suggestions and recommendations in home, in room, or telehealth, and basic needs support, all automated. Or kind of basic things like comprehensive circle of care inclusive admissions and discharge planning meetings, right? Bring all of that information in to make sure that we are, and all those people in, to make sure that we are taking comprehensive care of the patient at critical times in their healthcare journey. The possibilities are absolutely endless.
Speaker 3 (27:42):
Yes, absolutely. The key takeaways here are AI systems are a variety of technology and tools that are utilized and use data to generate patterns to help facilitate a specific task. We've described quite a few today, so definitely have another listen. They are particularly effective with large amounts of data that may be difficult traditionally for people to monitor and for people to continuously sift through, and therefore we use AI systems to help us with that. AI is also designed to compliment the work you do. And when considering implementing AI systems in your organization, they shouldn't feel like a chore, but rather look for systems that feel natural alongside the tools that you already use today. These systems will still require some adoption and implementation considerations that Allison highlighted, but make sure your teams are fully aware of the overall benefits and do thorough communication and you will be able to achieve successful implementation in the long run. And finally, the takeaway I think is most important is that AI and these systems are only going to increase in long-term care. So taking the first step to learn how to better work with these systems can greatly improve capabilities as caregivers.
Speaker 1 (29:01):
That concludes the latest episode of the Post-Acute Point Of View podcast. We have a lot of guests and topics coming up that you won't want to miss, so be sure to subscribe. To learn more about MatrixCare and our solutions and services visit matrixcare.com. You can also follow us on LinkedIn, Twitter, and Facebook. Thank you for listening. Be well and we'll see you next time.
