The future of AI in long-term care with Allison Rainey and Daniel Zhu (Part 1)

Episode 74 October 31, 2023 00:15:45
The future of AI in long-term care with Allison Rainey and Daniel Zhu (Part 1)
The Post-Acute POV
The future of AI in long-term care with Allison Rainey and Daniel Zhu (Part 1)

Oct 31 2023 | 00:15:45


Show Notes


In this two-part series of the Post-Acute POV podcast, our host Lois Bowers, editor at McKnight’s Senior Living is joined by Allison Rainey, APRN, FNP-BC, head of nursing and clinical informatics at MatrixCare, and Daniel Zhu, VP of product management, data & AI/ML at MatrixCare.

Join Lois, Allison, and Daniel as they discuss what AI and machine learning are, the use of artificial intelligence (AI) in the long-term care industry, and its current applications in healthcare environments. The trio shares their experiences with implementing AI and how these technologies can improve patient outcomes and staff satisfaction. Listen to their conversation.

Topics discussed during today’s episode:

  1. [00:34 – 01:38]: Introduction to episode topic and speakers.
  2. [02:40– 04:04]: Daniel defines what artificial intelligence and machine learning are and how they work.
  3. [05:06 – 07:07]: Examples of AI in long-term care and how people feel about using AI in healthcare.
  4. [08:02 – 09:05]: Allison discusses AI use in care coordination and monitoring.
  5. [09:48 – 10:35]: How staff can use AI to help streamline patient assessments and documents.
  6. [11:32 – 12:13]: Allison and Daniel discuss AI technologies and tools that help improve patient care.
  7. [13:01 – 13:42]: Daniel explains how AI can help drive efficiency and asks Allison how she implemented AI technologies in her past roles.
  8. [14:04 – 14:46]: Allison gives examples of how AI positively impacts patient outcomes and nurse satisfaction.
  9. [15:15 – 15:30]: Conclusion.



The content in this presentation or materials is for informational purposes only and is provided “as-is.” Information and views expressed herein, may change without notice. We encourage you to seek as appropriate, regulatory and legal advice on any of the matters covered in this presentation or materials.

©2023 by MatrixCare

View Full Transcript

Episode Transcript

Speaker 1 (00:02): Welcome to 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. Lois Bowers (00:34): I'm McKnight's Senior Living editor, Lois Bowers. Our speakers will be discussing the current state of AI implementation in the long-term care industry. And now, 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. (01:08): Daniel Zhu recently joined the ResMed and MatrixCare team to lead the 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. Daniel Zhu (01:38): 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 general awareness of what AI is and how it's currently being applied in the long-term care space today. In order to achieve our objectives for today, this is our agenda. We'll quickly go through a review of what artificial intelligence is. Immediately after, we'll dive into how it's being used today. And Allison can walk us through her experiences implementing some intelligent systems at her previous organization. So to start us off, what is AI? The original definition of AI is anything built by people to mimic human intelligence. Unfortunately, this is a very abstract definition, so I'm going to try to simplify it and put it in terms that might be a little bit more relatable. (02:40): So what is AI? It's any type of system that makes use of available information to complete a specific task. So what does that mean? Most people often associate AI with things like self-driving cars or ChatGPT where there's a large amount of information that comes from a variety of mediums such as video and text and voice, and it's able to complete some specific tasks like driving the car or answering your question. But with the original definition of AI, it actually includes some very simple everyday systems such as calculators, where you're providing the information as numbers and operations and the system actually completes it for you. That being said, modern day AI is often associated with a subcategory called machine learning. Machine learning refers to a platform that utilizes data to generate its own set of rules to accomplish its task. So AI is any set of rules, either people could tell the system to apply those rules. Machine learning is a specific subset of AI where the system generates its own rules to complete the task. (04:04): One very common machine learning system that you might have experience with is Google Search. It uses all the historical searches as data to generate potential search terms as you slowly type. This is a really commonly associated form of machine learning. So how does machine learning work? I'm going to quickly go into this just on a high level. What machine learning does is it generates mathematical patterns to predict the possible outcomes, and in order to successfully generate accurate and robust results, you need to have good data. And very quickly, I'm going to go into what large language models are because with the recent popularity of ChatGPT, these are specialized machine learning models that work with language data to accomplish all those cool things you've seen ChatGPT do. (05:06): Now that you have a good understanding of what AI and machine learning is, we have to emphasize that they're not always the right tool. That being said, AI is good at a number of very important types of tasks. For example, they're good at taking large amounts of information and performing specific actions when the rules are very well-defined. Also, when there's no clear procedures, but an abundance of data, you can count on machine learning to pick up patterns and generate predictions as we mentioned earlier. It's always important to start our session with this brief section on AI because now we can better imagine and understand how AI is being applied in long-term care industry. In fact, we conducted a survey with McKnight's earlier this year asking healthcare providers like yourself, if you believed AI could be used to improve patient outcomes, and the resounding number of respondents agreed they believe so. (06:12): Similarly, when we asked respondents if AI systems were accurate, a lot of people agreed that they're fairly confident in the accuracy of these systems. Yet, when we asked the same group of people if they have adopted or if they plan to adopt AI in the near future, a majority of people said they haven't and they aren't sure if they will in the future. The main concerns highlighted by the respondents had been, "Are the systems really complex? Is it really complex to implement these systems?" And some legal and regulatory concerns and uncertainty around these systems in general. Hopefully, today we can tackle and go over all these concerns through the different use cases that are available today. So believe it or not, modern AI is already used in a lot of different ways in the long-term care space. (07:07): Just off the top of my head, for example, in wound care, there is some very advanced technology out there already that's able to measure the area and the depth of a wound from an image just taken with your phone. That really provides consistency amongst measurements because if you can imagine different wound nurses might measure length and width slightly differently, resulting in slightly different measurements each time. Not only that, there's also some pretty cool technology I've seen in long-term care where they're able to detect and identify the people in a room and determine if any of these people and any actions taken by these people might result in falls. So as you can imagine, there's huge applications to that in all the different spaces in long-term care. Allison, can you name a couple of good examples for AI in long-term care? Allison Rainey (08:02): Like you said, Daniel, there are so many examples already that are in long-term care. So they're just transforming what we do currently. So I guess to start, we know care coordination is such a critical piece of optimal patient care. We really need to make sure that everybody in that care circle, which is typically big in our population, has all the information that they need and that we've got it all consolidated for appropriate care. So there are many technologies out there to support care coordination, pulling information from various healthcare settings across the continuum, and not just pulling it into one consolidated place, but also elevating that to identify a patient risk based on all of that information coming together, best suited levels of care. So pulling from that information, where would the best place for these patients be? And then even to the point of identifying maybe the location for the best care for this patient's specific condition, so a lot of information coming together to really help us make the best decisions or help the patient and the family make the best decisions for themselves. (09:05): We also have monitoring systems like Daniel already mentioned, that can recognize activities in a room. So ambient technology can identify whether patients are lying or standing or whether they're sitting, or even if they've fallen, which is a little later than what we'd like to know, but it can certainly do that. With that information though, these technologies can identify if there's changes in patient patterns that maybe we wouldn't recognize as humans, but the machine does and can let us know as caregivers, as clinicians can send these notifications out. So maybe we can do a little more analysis of what's going on with the patient to prevent those potential negative outcomes and get in there early and take care of those things. (09:48): From monitoring technologies, we can also combine patient preferences. So if we've got patient preferences in one area and we've got all of this information in another area, it's coming together so that not only is it recommending actions, some of these technologies, but they're also recommending actions for the unique person pulling into those person-centered approaches to make sure that, again, we're following and we're recommending things that make sense for this unique individual. Flipping it over to providers and physicians, MDs and nurse practitioners, there are technologies that actually prep charts so they can go through and sift through all of the data and all of the information that might be in a chart that might be critical for a clinician to look at to make good analysis and make good decisions, and really help guide those decisions. (10:35): Instead of having to go through all of those pages and clicks, it assists what's important to the top, and again, allows those providers to really focus on what's important, which is what the interactions with the patients, the things that they can do, the assessments, all of those very specific tasks that a machine can't do that a provider needs to do so that they can have really good care and that they can document it appropriately. But even with that, there's technology out there that can synthesize those verbal interactions, so the provider is in there taking care of patients and doing what they need to do, and the technology can synthesize those interactions into an appropriate healthcare note, certainly would have to be reviewed and checked off on. But there are technologies that actually can do that to really efficiency providers and others so that they can do what they need to do with the patient and focus on what's important, which is not documenting. It's really taking care of that patient in a fantastic way. (11:32): The technologies that I am most familiar with are most associated with risk management being in the business that I'm in. And so there are tools out there that can do continuous scanning of the patient record for data that together with other existing data could be an early identifier for condition changes, risk for falls, rehospitalizations, depression, all of those kinds of things. So the technologies, as data is entered into systems, it can really highlight information that might be relevant for a clinician to review or an action to be taken. So it's really that timely notification of things that are happening as they're happening, so you can have appropriate comprehensive adjustment of the patient care. (12:13): You can change the clinical plan of care as needed or make, this is on another level, but you can make clinical staffing adjustments for improved care. So really matching up those clinicians with the patients that are appropriate which ultimately results in what improved clinical care and then hopefully would translate into staff retention. So you've got the right patient with the right clinicians and caregivers, and everything rises higher. AI certainly will not replace clinical staff. We need empathy, we need judgment, but it will empower the existing staff to be able to do their job even better, and it's already significantly impacting clinical efficiencies in our environment and ultimately improving patient outcomes. Daniel Zhu (13:01): Yeah, definitely. I think AI powered clinical surveillance is an extremely, extremely useful tool to drive, as you said, efficiency. If you can imagine nursing staff nowadays, like they have a staff of over a dozen care providers entering clinical information into the EHR every single day, and care providers can be working to manage three or more dozens of residents at a time. So I'm sure getting some help keeping an eye on all those residents and all your staff members, what they're entering would be always highly appreciated. So Allison, did you implement any of these technologies that you mentioned? How did that affect your organization overall? Allison Rainey (13:42): Yes. Yes, we did. We absolutely did. We used these technologies in many different ways to positively impact our patient outcomes and our nurse satisfaction. So we used some technologies that afforded the clinicians to know the patient acuity level at the time, so any information that was entered now would actually impact this patient acuity score, which would allow for not only knowing where the patient is in their clinical journey, but also allow on a big larger scheme for those that were admitting folks to facilities to really be able to triage those patients to units that might have appropriate staffing ratios and skillset to meet those patients' needs or to spread the patients out to make sure that we aren't caring for the patient in the right way and with the right amount of people for good satisfaction and for good care. So with those acuity scores, we really were able to enhance what we do and take care of those patients. Those centers that were working were amazing at leveraging that tool to do those sorts of things. (14:46): By matching the patient to the staff, what they also found was it also promoted some professional fulfillment. It encouraged a positive work environment, and ultimately supported their goals for retention and recruitment. So if you've got those staff members that are really able to positively impact patients feeling good about what they do, we have the right staffing ratios to be able to take care of those patients, then certainly that is one of those things that makes a big difference to the caregivers that are caring for them. Speaker 1 (15:15): That concludes this portion of our latest series from 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 You can also follow us on LinkedIn, Twitter and Facebook. Thank you for listening. Be well and we'll see you next time.

Other Episodes

Episode 76

December 12, 2023 00:40:48
Episode Cover

Getting prepared for new trends in at-home models of care with Rob Stoltz, senior business development director, and Kathy Piette, co-founder and CEO of Corstrata

Introduction In this episode of the Post-Acute POV podcast, our host, Diana Eastabrook, staff writer for McKnight’s Home Care, is joined by Rob Stoltz,...


Episode 28

November 05, 2020 00:34:28
Episode Cover

The 4 pillars of a successful COVID-19 response with Dr. Dheeraj Mahajan, President and CEO of Chicago Internal Medicine Practice & Research (CIMPAR)

In this episode, Kevin Whitehurst, SVP of the Skilled Nursing Solutions at MatrixCare, sits down with Dr. Dheeraj Mahajan (Dr.Raj), MD, to discuss the...


Episode 23

October 08, 2020 00:29:24
Episode Cover

Information security: the non-negotiables for digital healthcare

In this episode of the MatrixCare Podcast, Navin Gupta, SVP of the Home and Hospice Division at MatrixCare, sits down with Todd Friedman, Chief...