DRUID Talks Ep #3 The Secrets to Happy Patients and Medical Staff Unveiled with Joe Nieto

Episode #3 of the DRUID Talks webcast features Joe Nieto, Customer Engagement Director at Element Blue, and Subject Matter Expert Kieran Gilmurray. See the full episode and transcript below.

Kieran Gilmurray: Folks, welcome to episode three - The Secrets to Happy Patients and Medical Staff Unveiled with Joe Nieto, Customer Engagement Director at Element Blue. In this episode, we'll explore the current healthcare challenges our society faces and how conversational AI is helping solve a big part of these issues.

We will take you on an exploration journey through industry-specific case studies to discover how this technology is utilized for the healthcare system, how it’s going to improve patient care, how it streamlines communication and how it improves the efficiency as well. Our guest today will share with you some great advice and lessons learned from successfully implementing conversational AI projects in multiple organizations.

We will also look to the future trends and share some thoughts on how you best can leverage the power of conversational AI to enhance patients' and medical staff's experience alike. Welcome, Joe!

Joe Nieto: Hi, everybody! What a pleasure to see everybody here and just share a little bit about what we picked up. As Kieran said, my name is Joe Nieto. I am the Director of Customer Engagement here at Element Blue, and we have been working in healthcare for the past 20 years and bringing lots of improvements and innovation to healthcare. And I hope we can talk about some of these topics today.

Kieran Gilmurray: Joe, I suspect you, and I could talk for quite some time today, and that's the bit that I'm actually looking forward to in this show. I'm a little bit curious. Would you mind just telling us a little bit more detail about who your customers are and how you use conversational AI to create value for them? You don't have to mention customers, but just give us a little bit of a flavour for the audience so they understand where value can be created.

Joe Nieto: Sure. So, what are the important things that we focus on in our neck of the woods in our healthcare practice is that we work on the provider side. So, when folks come into the hospital or into the clinics, those are the people we’re interacting with. Not so much with the financial people that handle the claims and processing.

Now, don't get me wrong, we do a lot of that. But what we're talking about today, especially with the conversational stuff, is the part that faces the physician, the nurses, that human aspect that deals with the patients.

Kieran Gilmurray: Oh, fantastic! Well, little question then. What are the challenges they're experiencing, and what are you doing with conversational AI to help address some of those challenges for them?

Joe Nieto: Sure. Absolutely. So, you know, just kind of conservatively speaking, and I hear stats all over the place... We are roughly about 5 million physicians short on where we are and where we should be. I can't even tell you how many nurses we're missing in the field. And so all these are… we have to figure out a way either to produce more nurses and doctors, which would be great.

I can't help there, but what I can help with is the efficiency of these valuable resources and how they interact with patients. You know, we're getting to the point where conversational AI instruments, technology that we're building here, can really help handle the patients that are coming in, some of the patient load, so when the physician or the nurse is brought into the conversation, the conversation is handled.

A lot of that information, a lot of the interaction, sometimes becomes routine depending on the mission. We have different applications, have different missions. We have some that are… I believe in November, I was talking about our rooming process where we have sort of a funnel coming through where folks have telemedicine appointments coming up. And just like when you go to see a provider, there's somebody there to greet you, usually with a clipboard: you know, why you're visiting with us; what's your height and weight; you know, any pain; you know, you have these standard questions that go through. And those are things that conversational, conversational bots are really good at because you can ask those questions in different languages, ask those questions and collect that data. So, when the nurse or the provider comes online, they already have a leg up on the conversation.

Kieran Gilmurray: Oh! That's interesting there because you built a couple of things there. The figure of 5 million is truly shocking, you know.

Joe Nieto: Yes.

Kieran Gilmurray: You immediately think about, my goodness, there's not enough professionals to do the work, but the folks who are currently working must be under a tremendous amount of pressure as well. They're now covering the workload for 5 million colleagues who aren't actually there. That's kind of scary.

Joe Nieto: It is. And so, the industry is… we have folks that are getting burned out. You know, the top two reasons nurses leave is going to be: there's a staff shortage, you know, where they have to cover extra shifts, we have, they have more patients they can actually see. And it's getting into the second reason a lot of the folks are leaving the profession - is the work-life balance piece of it because they're working these 12-14-hour days. During COVID, we had so many nurses just and physicians, I mean, everybody in the front, I should say, everybody from nursing to physicians to janitorial, just, you know, there's a lot that goes on in the hospital systems and clinics that the general public only sees one piece of it, and that's the facing part.

But there's an army of people that make sure everything happens from, as I mentioned, janitorial - folks that go in and change out the rooms, change out the bedding, and all that stuff to food preparation, dieticians, pharmacy, all those folks make the whole thing work together.

Kieran Gilmurray: Wow, I wonder if we are kind to those individuals when they're at hospitals or possibly not…

Joe Nieto: We try to be, right? You know…

Kieran Gilmurray: We think as patients we should be, or as customers of the service, we should be. I don't know if we've always got the patience, from some of the stories I'm seeing in the press. But you don't realize, just beyond the front end, how many people are involved and the pressure that they're under. Because the healthcare staff they’ll enter the profession for a particular reason not to provide care. But that's something that's like an assembly line as opposed to an actual caring moment that you want to share with another human being who cares deeply about your health and well-being.

Joe Nieto: And that's a good point, Kieran, because it has come to the point where we have EMR - electronic medical records - you know, EMR systems that help manage for the financial side, right? We want to make sure that the appointment is documented, that the drugs that are administered are tracked, that all the symptoms, all that stuff… So, what ends up happening is that the physician or the nurse is sitting there keying this information, looking down at their keyboard, typing in instead of having a face-to-face conversation like we are and say, you know, tell me more about what's ailing you; you know, tell me about how long you've been feeling this pain; you know, and you become a little bit more tied into the patient, right, in their experience… they want to feel like you are there to see them. You're not there to fill out a keyboard, you know. And there's a long list of information that comes in.

And I mean, there are a lot of valid reasons that the information is collected. Financial, legal, medical, medical history, claim justification, all these things that come into play that there's a reason we have to do that. But if we can offload that part of it and let them focus on the patient, it's so much better for everybody.

Kieran Gilmurray: Just thinking there as well… It opens up the opportunity for lots of, I call them, inadvertent errors because no medical professional wants to make a clinical misdiagnosis or establish a…

Joe Nieto: Right!

Kieran Gilmurray: A patient's needs in a tablet that they don't need. So, Joe, give us some examples of some of the conversational AI applications or items that you have built to help alleviate at least some of this pressure.

Joe Nieto: Sure, sure. Absolutely. One of the ones that we were pretty proud of is our rooming bot that handles a conversation from the time that somebody joins into the telemedicine appointment. We know who that person is at the other end of the conversation. We've identified, we pulled up the medical records, and we know specifically which clinic they're there for: oncology, radiology, behavioural or whatever clinic it is. So, we walk these folks through the process. I mentioned earlier we collect some demographic information, just: height, weight, that kind of stuff. But depending on the clinic, we get specific into the questions. Yeah, we may ask questions about their extremities; we might ask questions about are their hands feeling cold.

We may ask questions about if they sleep well, we may ask questions about, oh, on some of the psychological stuff, we may ask, you know, are you interacting with folks on a regular basis? Are you living in a home? You know, there are 1001 types of questions that we go through. But what I think makes it really valuable is that it focuses on the type of planning that they're coming into so that when the nurse does come on the line, she has a list of all the different replies.

And the virtual assistant is smart enough to know that, oh, there are some trigger responses that I should look for, and it can flag it on a transcript that they're able to look at so that instead of rereading and regurgitating everything that was answered, they could go directly to the problem areas.

Kieran Gilmurray: I like that; I like that a lot because that sounds quite targeted.

Joe Nieto: Yes.

Kieran Gilmurray: Well, as opposed to just answering a laundry list that would make me feel as a patient or maybe the patient's family like, you know, they're part of a machine, not part of a healthcare system.

Joe Nieto: Exactly. And it helps both ways, right? When the patient comes online and, let's say, they have an appointment at 2 o’clock in the afternoon, that appointment is going to start at 2:00 in the afternoon. The virtual assistant is… they're willing and waiting, and… it starts. So, they know at 2 o’clock, their appointment is there, not 2:05, not 2:10, not waiting for somebody to get freed up.

They start to run on time, and they feel like, ok, I'm engaged, I'm on time. I know I'm going to get through this, through this procedure. And then on the provider side, because a lot of these questions are already done, they can cycle through people a lot more, a lot more than in an hour. You know, usually, it's what, 15-minute increments. We can maybe get, I don't know, 10 or 15, maybe 20% more folks through the queue than we could before because a lot of this stuff is already pre-answered, and they are reading those responses quickly as they come in.

Boom! You come in and know exactly... Especially if you have a surgical follow-up where it's just not a lot of this back and forth. Maybe you've had some surgery; you can do some follow-up. You come in; you have a conversation with the physician. They're going to ask the same questions: any fever, any redness, are you changing the bandages on a regular basis, do you need any care? We can send videos out to you to show you how to change the wound, we look… You know, we can ask those questions where we can get ahead of an infection or any kind of that problem and also highlight it for the physician so they can see and have some answers right away.

Kieran Gilmurray: I kind of love that. I've done a lot of work with contact centres… and just to get that in my head. So, for the first 5 minutes, we're collecting data. We spent, you know, in the hospital context, healthcare context… That is technically a waste of time. It's vital, but if that can be done by technology, let the tech do it! So, then that healthcare professional, as you say, well, can put more people into an hour. But rather than doing the admin piece, which is automatable, then they’ve more time to do the care piece, which is what the patient and they want to do as well. It sounds like a win-win.

Joe Nieto: Absolutely. And one thing that I don't touch enough on is that the virtual assistants are multilingual, right? So, if we know the patient is… English is not the primary language; maybe it's Spanish or French, and we can start that conversation in their language. And you’ve already broken down some barriers, some cultural barriers when I'm like, oh, you know, I mean, read this English, what it's asking me and try to determine what's the right answer, you know?

And so, if I can present that in their native language and have those prompts and responses, so there's clear as possible, it’s either a yes or a no, a number or date kind of response, right? So, we're not going to ask a long, you know, question that's going to generate a paragraph or a Rembrandt, right? We want a very clear and precise answer.

Kieran Gilmurray: I like that. It's interesting because I went for… it just reminded me of something else. You forget how valuable that is. And in a moment, I’m going to ask you another question because that's prompted something else in my head… I went to surgery recently for a COVID injection.

Joe Nieto: Oh goodness.

Kieran Gilmurray: And the patient couldn't speak English. And I think they were new to Ireland. And what happened was that it took about 20 minutes for the doctor using Google or whatever it was on their phone to explain that they needed to get a translator, that they needed to rebook an appointment. So, it ended up the nurse couldn't actually spend the 5 minutes that she wanted with that patient.

It was 20, and everybody else got delayed. The meeting had to be reset, and therefore the person who was there to get their COVID injection had to go and not get it and go with their family and organize another meeting. So, you can start to see how all those inefficiencies appear by not having a multilingual interface.

Joe Nieto: Absolutely. And that is, I tell you, that's been one of the biggest responses we've gotten from folks because they're, you know... at first, it's kind of novel, they’re like, wow, look at this, it's speaking, you know, whatever language… Yeah... And then you come in, and it's, after a while, they really latch on to this, right?

Because there's no vagueness, you know, it's like you depend on the patient's mastery of their English language to understand what we're asking. So, we just ask them in the, you know, their native language and get a better, clearer response.

Kieran Gilmurray: But it must be more personal as well, as opposed to you coming in and someone insisting you speak English.

Joe Nieto: Yes.

Kieran Gilmurray: Now that starts with the care part, taking care of the experience of the patient right from the very beginning. Joe, you must have learned a lot as you've rolled out conversational AI. My next question: what are the three or four biggest lessons that you have learned? Having implemented lots and lots of healthcare conversational AI solutions.

Joe Nieto: Sure. I think looking at the end product instead of how we're necessarily going to get there, you know, I generally go into these meetings… tell me where you want to be, you know, leave how I’m going to solve the problem to my team and myself. Just tell me what you want to see, give me this big visualization, this big picture… because I'm very upfront. I think you always have to be very upfront with the clinical folks. They're top in their industry; they're on top of their game, they're operating on top of their license. And you know, we're coming in, you know, as IT developers and everything. Right? And I'm, you know, I always tell them, look, I work with a lot of folks like you, but, you know, I'm just a visitor.

I am just temporarily walking in your shoes. After I leave, you're going to have to live in this world with what we built. So, I want to look at this big picture along with you and go: where do we want to be? What do you want to see? What does that feel like? What does that experience going to be like?

You know, for you as a clinician, for you as a patient, for you as a doctor or provider, you know, what is your role in these things and where are we getting to? And I won't back from what we have, right? Because if I tell you, look, I've got five pieces of data, and this is all I can give you, but that's not going to

I mean, it may solve that problem, but it won't get you on board with owning the solution. Because I feel if we start at the end and I'm, like I said, a very upfront look, I am not an M.D., but I'll walk with you. Right? And I think if you have that kind of partnership early on in the process where we all know we're in this together, then I think that’ll boost your success rate way up.

Kieran Gilmurray: I might refer to it as the M.D. for the rest of this show because, by the sound of the things you actually do, you’ve got a lot of experience, doctor Nieto! By the way, those folks listening in, Joe is not a medical professional, so please don't reach out for medical advice. Hey, but do the doctors believe what you're saying? Because you're walking in as a technology professional and saying: I can give you everything that you've ever dreamed of!

Joe Nieto: Yeah, I, you know, I think from that perspective, I think there're two things that really work in our favour quite a bit. One is we listen intently. We listen, even though we may have heard it before, we don't want to make any assumptions because, you know, there's a reason hospital A is better than hospital B which is way above hospital C.

And it's, you know, they’re all seeing patients, they’re all treating different diseases and stuff. It comes down to the execution and operations piece of it: how efficient they are with getting results, their communication back to the patient, is the patient engaged, you know, that makes this hospital better than the other. So, we focus on that as well.

Right? What are we doing to take care of the patient? What are we doing to help facilitate this as efficiently as possible? And so…

Kieran Gilmurray: Thank you, you've a wonderful bedside manner as a tech, as an IT tech...

Joe Nieto: I, well, try to be. I try to be. I think the other thing that is helpful, I think, is that we ask for their opinions. We do… they’re very involved in the design, and they’re very involved in the delivery piece of this. And to a tea, everyone I’ve ever worked with has gone… I don't know the answer, but I can find it, or I can… let's get together.

There's just some whiteboarding. We do a ton of whiteboarding, not only from me educating them, you know, because, you know, you draw the lines and circles, here's where things are happening. But on their side, they're going, oh! You know, that's great…your computer. But I'm dealing with people, you know. Especially in paediatrics, you know, paediatrics is, you know, it's like other hospitals, but everything is smaller. Everything's more complicated because the patients are smaller. You have mixed units, mixed family units. You have mom and dad. Then maybe one mom, maybe two dads. You have different types of situations that you have to deal with and allocate, you know, to be able to work with. So, when you're working in telemedicine, it's extremely important you identify who is on the other side of the conversation and is that person who's on the other side of the conversation the patient authorized to speak from the patient. So that brings those big challenges that, you know, aren't as common in other hospitals.

Kieran Gilmurray: See, that's interesting for me because very often, sometimes, folks consider IT as a cold technology. That sounds, Joe like technology's very much an empathetic part of everybody's experience.

Joe Nieto: Yeah, yeah, absolutely. You know, and I think that has to… the solutions that reflect… you know, we constantly have to remind ourselves, why are we there? You know, we're trying to heal patients. We're, we're trying to make as efficient as possible, you know, the processes… For somebody who's there on the worst day of their life, right? Mom and dad are coming through the worst day in their life: their child is injured, and they’re coming to the hospital. They're going to the emergency room. You know, how can we get them through that process as quickly and as smoothly as possible? If it's virtual assistants that we have kiosks when they walk into the emergency room where they can enter their information so that it's up to the front as quickly as possible.

And we ask those questions that can triage those people so that they move up in the line ahead of somebody with a… that has a cold, you know, maybe not as yeah, it's important, but maybe not as much as, you know, your child is just you know, unconscious, barely breathing, you know, bleeding out. You know, those things need to get attention as quickly as possible.

Kieran Gilmurray: See… having sat with young kids in hospitals, not having any communication, that's not a criticism of the doctors. They are so busy! So actually, having that thing to do and feeling like you're contributing to solving the medical issue, but that as a parent, I know to me, with conversational AI in that process, I would have been delighted because there's nothing, as you say, scarier than your child being ill. And you not knowing what to do, not knowing who to talk to or understanding any part of the process or being able to contribute to it.

Joe Nieto: Yep. The other use case, you know… we’re talking about interesting use cases that we're working on. Another one that we're working on has to do with new moms, you know, mothers. And, you know, it's the first baby you’re delivering. You know, we're still on the board with this thing. The technology is there, not even worried about the technology, not even worried about the conversational part pieces of it… It’s like, what is the content we're going to deal with? And the mission is, you know, you're a brand-new mom, and you're a brand-new family, right? You brought your child home, your baby’s home, and there's no operating manual. You're literally, what do I do now? Right?

Who would you reach at 2 o’clock in the morning when the baby can't sleep? You know, who would you reach? Who would you reach out to? So, you know, there's a need for a virtual assistant that's available to say: let me log in, let me tell him who I am. I've got my baby record here. I can see I'm dealing with I don't know…

You know, a child that's pretty normal. Maybe it's having some kind of other issues, nutrition or something, you know, and you could start a conversation to and say: hey, the baby is not sleeping well, you know, the last time I fed, and the bot can go back and forth and get some information to help you decide whether, hey, maybe we should be coming into the emergency room.

Maybe the bot can go ahead - you know, this sounds pretty important based on what you've told me. Let me connect you with somebody live to address that situation. Other times… look, this is perfectly normal, you know… These are the kind of things to expect, you know. So those are the ones we're working on now, too.

And we can imagine, you know, as I said, the bots are going to be available 24 hours, 365, seven days a week. And to be able to have these conversations and help the new families get through their early years might be really important for them.

Kieran Gilmurray: Joe, where were you 18 years ago when my son was born? By the way, if I had a bot to work out my thinking about him at the moment, that wouldn’t be useful. I don't think your need for advice ever changes. Let me ask you the next question if you don't mind. If you were advising other companies or healthcare providers that were looking at implementing conversational AI, what would your advice be to them?

Joe Nieto: I think the central point of everything, especially the design, is to focus on the conversation, focus on how the user, patient on the other end - in my case, is patients-, but maybe it's a new customer. Visualize how their conversation journey will take place. And that's, that's really… I mean, we have whiteboards and sticky notes, and we go, here's where the conversation comes in; here are the multiple branches that they'll be able to take. And where are these eventually going to go, right? So, spend your time on the design piece of it. And then, the follow-up to that would be focused on the metrics. You know, I can't say enough to go... where are my customers and patients spending their time when they come in through the virtual agent, and they have these engagements going on with everybody going, and where's my time being spent?

Monitor the metrics that tell you somebody put an answer that was unexpected, and we couldn't resolve it. You know, you know, they give you an answer you wouldn’t expect. We track those, right? We know when somebody puts something in, and you kind of go, huh? The next time I see this kind of response, I want them to go on this journey.

I want them to go to this pathway or maybe this. Maybe it's something that, you know, more people are asking about this that I don't have a response for. So, I should really build something that the next time they ask, I have an answer. So, focus, focus on the conversational design piece of it. Technology is going to take care of itself.

We've got really wonderful products these days and making sure that the product implementation is smooth, of course, from an AI perspective. Still, from a design perspective, the conversational impact, it's probably where I would spend my time.

Kieran Gilmurray: I love that advice! My final question then… Are you seeing any trends within conversational AI or the conversational AI industry that could have a really big impact in the years to come?

Joe Nieto: I think, you know, this ChatGPT business that... you know, for clinical, I think it's going to have a large impact, right? And I think the way that we look at it is… it's the source, you know, so, you know ChatGPT has a lot of different products that can actually hold that conversation, and it'd be contextual, right? But from a clinical perspective, you know, if you come in and you go: Oh, I have a patient, age 2, yellowing eyes, running temperature, you give all these symptoms, right?

And it says: oh, you've got jaundice, you know, something like that… That’s great, but it doesn't… you don't know where it got the answer from. You know you don't know that the answer is correct. You know… and at some point, that has to, that's going to have to be weighed out.

You know, I like, I like where it's going. I think it's pretty monumental in the way that it's going to be handling conversations, right? Still has ways to go... So, I don't want to say that, you know, don't let perfect be the enemy of good enough to let's continue building on this thing. But I really see that, and I really see the conversational piece of it taking less and less on the keyboards. Because you and I can have a conversation, and we're not going to be typing back and forth the responses. And in the clinical setting, if I have an ambient microphone hanging down that can say, you know, hey, whatever the wake-up word is, you know, hey, Joe, hey, Joe, schedule me a blood test for my patient. Right? And contextually, it knows, it knows who you are, the provider. It knows your patient because it knows it's in this room.

It knows who the patient is and who all the attendees are. And it can say, I can just tell it, you know, order a blood test and it would be able to run automation to generate that test, write all the documentation, fill all the paperwork that usually a medical transcriber is sitting there on their little keyboard righting it.

So, more efficient, faster, and more accurate. And I know, yeah, and we have that working in the lab and… but you know, we've got to work. We've got to get it out. Yeah.

Kieran Gilmurray: Okay. Tell me, is that the Star Trek lab? Because it sounds like this is literally way ahead of its time.

Joe Nieto: I think when we get closer to clinical research diagnoses, those kinds of things... I think we're going to see some really big improvements in that area. And I'll tell you, that's going to be extremely helpful because, as a human being, physician, researcher, and nurse, it is absolutely almost impossible to keep up with all the latest publications, right? Because you're spending your time seeing people.

You don't have time to read the research, the latest research from New England, you know, Journal of Medicine and go, oh, wow, they got a new treatment for COVID! Oh, they've got a new treatment for, you know, for a heart condition. You know, it's a, it's impossible just about… So if you are able to take conversationally, and I think this is where the skill is going to be coming in for the physician, I kind of see, you know, how in developers we deal with “if then else”, you know, programmers….

I think that skill is moving over to the folks that are working with these things. The way that you ask a question would be how we program these languages or how we interact with these systems. So, if I ask a particular question one way, I may get a different response or a slightly different response if I asked it a different way.

So, we have to train ourselves to ask questions that are as clear as possible because we kind of know what the response should be, you know? And the best way to do it is if, you know, if you talked to Alexa and you go, hey, Alexa, blah, blah, right? And it goes, well, I don't understand what you just said.

Right? And so you find yourself Alexa, and then you enunciate every little phrase, or you say in a different way, and then it gives you a response, right? You kind of go, oh, all right, that's what I wanted. The next day... So, you've programmed yourself to ask the question in a specific way. And I think that's where we're going to see some trade-off coming off. We're going to be smarter about the way that we ask these questions because we pretty much know what the response is going to be looking like.

Kieran Gilmurray: I might explain why I got that laundry deodorant ordered when I was trying to get Alexa to turn on the lights. But that could be, that could be a very dodgy Irish accent as well. So, we don't know what it is.

Joe, absolute pleasure, sir, talking to you. And thank you so much for taking the time to share all that wisdom. I never realized the number of uses for conversational AI. That was a pleasure! Although I have to warn you, at one stage when you were talking about whiteboarding and following the positions, and you leant into paediatrics and new babies, I thought, how far do you go for your job, Joe, to monitor everything that's happening? I'm delighted you have such dedication to your profession.

I think as an IT doctor, I think you're right up there at the top of the medical profession. But thank you so much indeed for taking the time and coming to talk to us today!

Joe Nieto: Mm-hmm. Thank you so much!