← Insights Unlocked
Insights Unlocked
Transforming healthcare UX with storytelling, data, and AI
00:28:29
Discover how healthcare leaders use storytelling and AI to improve UX, align stakeholders, and transform patient experiences.
Loading summary
Transcript34 lines
- [00:02]
Host
Welcome back to the Insights Unlocked podcast. In this episode, recorded live at the Human Insights Summit in October, we're diving into how innovative UX strategies are transforming healthcare. Join us as we chat with Amy Stillman from CDER and Logan Hartline from HCA Healthcare about leveraging AI, simplifying patient experiences, and using storytelling to drive meaningful change in a complex industry. It's an inspiring conversation you won't want to miss. Enjoy the show.
- [00:37]
Welcome to Insights Unlocked, an original podcast from User Testing, where we bring you candid conversations and stories with the thinkers, doers and builders behind some of the most successful digital products and experiences in the world, from concept to execution.
- [00:57]
Leah Hogan
What I'm going to do first is actually let you both introduce yourselves and we'll start with Amy, who's sitting next to me, and then I'll introduce myself.
- [01:04]
Logan Hartline
Sure.
- [01:05]
Amy Stillman
So I'm Amy Stillman. I'm the vice president of design at Cedar and so I own product design and user research. And CDER's mission is to help everyone be able to afford and pursue the care they need. So we're kind of at the intersection of finance and health care trying to help people pay medical bills. So I'm sure everyone's gotten a bad medical bill and if you haven't, then you've known someone that has so pretty gnarly problem that we're excited to work on a great user experience with a lot of room for improvement.
- [01:34]
Leah Hogan
Oh yeah, a lot of opportunities, Logan.
- [01:36]
Logan Hartline
Well, hey, good morning. I'm Logan Hartline. I'm the senior director at HCA Healthcare. HCA Healthcare operates in 22 states, hospitals, urgent care, surgery centers, freestanding emergency departments, and physician practices. So I lead our user experience center, which really enables product designers to focus on digital products for patients, consumers, nurses, doctors, hospital operators, HR professionals. If it helps our hospital group function, we are working on capabilities to enable those people to get their jobs done. We focus on accessibility design, systems research as well. So never a dull day.
- [02:20]
Leah Hogan
And our job, Yeah, I feel like for everybody's job theses, there's never a tall day. But just a little bit about me. I'm Leah Hogan and I'm a principal on our experience research strategy team at User Testing. And basically what that means is that I have this kind of hybrid role of working with our customers to be able to essentially execute on a lot of the complex research strategy that they have in place, whether they're setting up programs or scaling, and also working with our product team to help to evolve the product. So really looking forward to this conversation Today we're going to be talking about collaboration both across teams and internally across those silos. And then finally regulation. We all know that there are a ton of constraints that we have to work around in this space and so just want to embrace that and hopefully provide some ideas for strategies you might use. So let's get started. So I always love to start with success stories because people love to talk about when things went well. So I'd love to hear examples of a recent success story of a project where insights from either patients or providers really help to drive your decision making. And I guess I'll start with you, Amy, here.
- [03:45]
Amy Stillman
So we've been working on. So within billing, there's a lot of ways to pay your bill. Actually the best bill is not to have a bill, right? Like sometimes you get a bill and it's actually that the insurance wasn't properly added and so there's a denial. Other times we find users that have a high deductible plan and so they actually maybe need a payment plan to help them figure out how to pay over time. And then sometimes we do know that there's a large cohort of users that probably could apply for Medicaid and don't realize that they should, and they can. So earlier this year we worked on a tool and it was actually about Medicaid eligibility. And so we tried to, in our user database, identify users who we thought would be eligible for Medicaid. And one of the hardest parts about Medicaid is that the process is a horrible user experience. You have to go onto some state website and figure out where a form is, download the form, physically fill out the form, facts back in the form. It's really, really hard. And we knew that. But I think in our user research, what we learned is that before that even happens, there's this moment for the user in which they start to question like, is this even for me? Is the effort even worth it? And so the part of the problem that we decided we wanted to tackle first based on that insight was just getting people to a point of like, yes, like this is something that we think could be for you. And there is dollars and funds available. And so it helped. The insight helped us. You know, as experienced designers, we wanted to really go after like all the friction with the fax scene. And we did handle that too. But we thought as a starting point, the first starting point is just getting people to the table to know that this service is something that might be for them and do it in a really lightweight, user friendly thing the other thing we learned from the research was this is a scary conversation for a lot of people. A lot of people maybe had insurance, but then had job losses and they don't now. And so the emotional vulnerability of also being able to come to the table. We found this insight from one of our designers in research where the insight was it's not rude from a robot. And the idea is to have a vulnerable conversation with technology is sometimes easier than a financial aid counselor that's calling you and you have to go through this kind of heart wrenching story that might be very emotional. So getting people to the table was not only kind of providing them with that just first answer of that gut check, but also really leaning into the fact that it might actually be more comfortable for some folks to use kind of a digital screener and kind of a wizard. A clippy. A clippy, but not a clippy. We all love a clippy. We all come back to a clippy. But to do something like that versus tackling the whole problem, which later on we did that as well. We actually partnered with someone to actually help us execute the form part. So I don't know. So that was a success for us, knowing which part of the problem to tackle and focusing on really that human need of emotional vulnerability and understanding that there is a path for you that.
- [06:45]
Leah Hogan
I think is a really powerful story. So thank you for sharing that.
- [06:48]
Logan Hartline
Yeah, yeah, that's great. I'm curious, show of hands, anyone in the audience, how many of you have ever visited an emergency room or emergency department before? Okay. Hopefully you're all better now that you're here. So, you know, the emergency department is where a tremendous amount of patients become impatient. So it really is a front door to individuals that present with an issue such as a fracture, a laceration, something like that, or even one issue that then physicians and clinicians identify. There's actually something else going on that you may not be aware of. And so if you've ever been on the opposite side in an emergency department, it is a wild ballet of getting people where they need to go at the right time. So if you come in with chest pains, if you come in with a laceration or you come in with a burn, chest pains is going to get seen first, maybe the cut is going to get seen second in the burn, depending on the severity will get seen third. So there's a lot of coordination that has to happen to ensure patients are getting the right care at the right time with the right providers. And so, and one of our designers on our team was working on a project recently, essentially a master view, like a digital master view of the emergency department. So it looked like just a screen with boxes and numbers in these boxes that the charge nurse of the emergency department think of the charge nurse like the quarterback, making sure everyone is going where they need to go at the right time. They could look at that screen and know, we have 10 patients in behavioral health, we have 30, and fast track, we are overstaffed in trauma. So it helps give them the right insight to know who do I need to call to get help. And so we had all of these assumptions about, you know, hey, we need to use our design system and we have colors for certain things. You know, red is bad, green is good, blue, we're not sure, so on and so forth. Well, I happen to be shadowing one day in the emergency department, and I see in context what these charge nurses are looking at. It's a bay of screens. You know, it looks like, you know, some hacker in their basement. All of these screens. And I can see our product, and then I can see other products in concert alongside them. And I'm like, well, what's that? Oh, well, this is our electronic medical record that plays a role. Well, this is this other thing that plays a role. And so I started asking, what do those colors mean? Oh, well, green over here is bad, red is good. And that was something we couldn't solve for because we'd have to have, you know, a pretty significant update with that EMR vendor. And so just being there in context to see, you know, one of these screens is going to show green is good, the other is going to show green is bad, was like such an insight. To say, hey, maybe we need to bring some of these teams together just for something as benign as a color could mean the matter of not to be too dramatic like life and death for some patients if a charge nurse is not necessarily being able to distinguish what they mean. So that was a recent insight for us, Just kind of the power of shadowing, observing, asking questions, kind of bringing fruit to bear.
- [10:08]
Leah Hogan
That's really an awesome story and actually foreshadows a little bit something we're going to be talking about later, which is really around, like, other ways that you might think about listening to folks, but obviously shifting gears a little bit. AI is huge everywhere right now. And we were talking right before we started just about like, some of the other examples that we're. How we're seeing it play out in other places. And I'd love to hear your perspectives on where you're seeing AI in your spaces. And how does research play a role in helping you to implement or not? Because obviously sometimes we choose not to use it in some circumstances in your experiences. So here I'm going to start with you, Logan. Cool.
- [10:56]
Logan Hartline
So we are looking at AI across a continuum of use cases. I would say the biggest ones are within our clinical space. So trying to leverage AI, not to take jobs over, but to leverage AI as a means to make processes more efficient, to make things more expeditious. And so one, one project that comes to mind is a project around labor and delivery. So my wife and I, we have four kids. So I'm not an expert in that, but I've watched it happen four times. And so if any of you who are mothers or know women who have had children, it really varies from pregnancy, labor, delivery, from child to child to child. So you can't always assume either the water breaks or we give Pitocin and like, we're going to have a baby in 10 minutes. It could be 10 hours, it can be a few days. And so from a staffing standpoint, this is really hard for nurse leaders to know. When do I need to find and get nurses here when mom is going to be delivering this baby? And so we were working with a clinical team around, based on our clinical data repository, figuring out with AI, could we predict, based on some markers, when moms were going to start kind of getting over that threshold, kind of getting over, you know, a lot of them say, you know, I forget what the actual term, but it's like the pit of despair for them. And once they get past the pit of despair, baby is getting ready to come. So could AI start to predict rather than having Amy sit on, you know, her phone waiting for mom to deliver, like, hey, Amy, you could actually be at home or working on clinical education, you know, upskilling to bring you in when we know that this is going to happen. And so there's always, for us, a human as a part of that process. We never always fully take what AI says, and as gospel, there's always a human in the loop. And so that's, that's a recent project that comes to mind where we were both marrying kind of, you know, the experience of a mother delivering, you know, a child and, you know, shadowing and talking to nurses with AI and figuring out what is that right kind of product to put into production. So that's when it comes to mind.
- [13:20]
Leah Hogan
Yeah, there's a lot of data there to work with.
- [13:22]
Amy Stillman
It's a lot of data. And we actually are looking at it in a data example similarly. So we think a lot about what is AI good at. And large language models are especially interesting. And we see our focus, we see AI as a tool and the call center is really our focus. And so in the call center, there's a lot of emotional labor that goes along with also just the busyness of, of tons of calls and volume. And oftentimes a call center employee is incentivized on like the number of calls and how fast they get through calls in a day. And those are the metrics, right, that managers are looking at. Those metrics aren't great because if you think about like you might have someone that spent a little bit more time with someone, or maybe you had someone that gave a really quick answer and answered it, but they got off the phone and it's not a great tool. And also then the call center employee also is incentivized to end the call and go to the next one. So they also oftentimes aren't writing up great notes or actually like taking the time. We did desk sides and I watch people picking the generic reason for like why the call happened every time. So the data out of it that we get from the call center on reasons why folks call in isn't great, along with then the burden of the actual call center employee feeling rushed to deliver volume. So one thing that we looked at with call center is actually looking at the call volume reasons. So can we actually take that part out of the call center person's job? And can we with AI, listen to what they said and provide a summary at the end to say, here are the things that this person called in about. We think these are the actual tags and the reasons why and kind of provide that for the call center employee. So that should be kind of like a nice thing that they may not have done or would have been rushed to do. The other thing we can do is AI is great at understanding tone. So we can start to understand, did someone answer this in a way that was really productive and respectful and professional and on brand, or were they really curt and short and leaving the user with a bad experience? So we think those are really interesting analytics to start to switch the paradigm from call centers just being about volume, a volume gain into more of a quality game like this is actually a touch point for your customers. And so how do we measure that touch point with. With more analytics that can help us there? I think the other thing that we're doing with the qualitative research is to understand on the patient side what their tolerances are as well. So we did a bunch of qualitative research to figure out as a patient. We kind of also don't want to put you in some loop where you're talking to the AI to the point of frustration. And so we did some frustration studies on how long and what types of things we felt like users would have tolerances towards. And so it turns out three minutes, no longer than three minutes at three minutes. Kind of like when you call the airline, you're like, operator, operator, operator. And for some people that might be different and that might change over time. But for right now, we found that people often like to engage if the, if the LLM felt like it was smart and it was understanding them and it was down a point of resolution. But at some point we would need to cut over and you might need a real human at that point too. So the, the research helped, the qualitative research helped us inform that. And I think also then our decision making on kind of where we would use AI and what it's best at as a tool were kind of our decision factors.
- [16:49]
Leah Hogan
Yeah, that's a really like, fascinating story because I'm thinking both, there's a lot of the, like, operational analytics. But that emotion piece again, is a, is a really significant thing because first call resolution, I know, is something that is usually very high on the list of analytics in focus. And if you can look at the motion, first call resolution. Yeah, that would be an interesting thing.
- [17:19]
Amy Stillman
It's a different. Yeah. And so enabling to, you know, us as a product team, like, we want to know what people are really frustrated about and what they're calling in for, but also for our partners, our health care systems, helping them manage their employees and be more empathetic to, to what's happening in someone's day as well, I think is like an opportunity.
- [17:37]
Leah Hogan
Yeah, yeah. I did want to just talk about an example that you've had recently of getting people to like, like show and tell is a great storytelling strategy.
- [17:49]
Amy Stillman
Right.
- [17:49]
Leah Hogan
And one of the things that especially qualitative research enables you to do, but sometimes quantitative research does this too is like show and tell. Really effective stories that get people to come along with you. So do you have any recent examples of that working for you? And I think, Logan, this was a question for you.
- [18:10]
Logan Hartline
Yeah, I'll be interested what Amy has to say as well. So at HCA Healthcare, when I first started probably almost 11 years ago, we had 1500 websites that we were managing independently from one another. So, so that would be like you going to target barton springs.com or target targeting south congress.com for those in Austin rather than just Target and then kind of focusing and filtering in on that location. And so we've gotten much more mature and much more streamlined over the time that I've been there. And so we always have questions from our business stakeholders around behavior. You know, we're not seeing individuals from an analytic standpoint kind of hitting our targets, like, help us understand what's going on. And I'd like to sit here and say that it's as kind of open ended as that there's always oftentimes an assumption around what's broken. So there's the question with this assumption on you kind of barnacled on with it. And so we recently ran a study around how patients are navigating to find care. So that kind of seemed like we're all people, like we all need, you know, physician services at some point in our life. And so they had some assumptions around what people were doing, what they were searching and why. And from working in this space for almost 11 years, I was already thinking, like, okay, I don't think their assumptions are right, but let's go ahead and kind of show the power of like the people's voice, like the patient's voice, like, let that be brought to bear. And so of course, like, patients were not searching against some of the criteria that our business stakeholders thought they were patients were interested in specialty. Like, my foot hurts, I want a podiatrist, or I want an orthopedic doctor. I want to understand, is that doctor going to cover my care from an insurance standpoint and location? Like those were some, some three, like three elements that we saw overwhelmingly across a few of our markets that patients were looking at. And we could have stopped there and just said, hey, you were wrong, we were right. Like, this was the data. I have learned, you know, in almost 14 years of marriage just to be quiet sometimes and let certain, you know, data speak for itself. And so we pulled together, you know, a really simple highlight reel letting individuals in some of these markets kind of show and say what wasn't working with the existing experience. And of course that's not to vilify or demean our business stakeholders. They are very smart. We all have bias and hypotheses. I think what it comes down to is helping tell those stories, like at the right time, especially when you have stakeholders that are ready to hear them because there was some concern. Like we could hear the tone in their Email like, you know, I'm getting questions from my, you know, senior leadership, like, we got to figure out what's going on. And so that was a really ripe opportunity to kind of let the patient's voice, the customer's voice kind of lead. And then we were there to say, hey, we've got some recommendations. You know, what do you think about these recommendations? So I think anytime you can allow the customer, the person to speak for you, and you're there as kind of playing copilot, I think that is a tremendous strategy rather than always having to be right. What are your thoughts on that?
- [21:32]
Amy Stillman
Similarly, I think the hardest thing in healthcare oftentimes is change management and that you have the really good idea. But how do you get everyone else aligned to like, that is the good idea. We're gonna focus on today and put the IT resources towards. And so we like that too. So as much as we use user testing platform to create videos for our own knowledge to know, like, hey, we're refining the product. It's also just a great way to paint the problem space. And so one thing that we've been doing on our team, Jayas on my team, she's wonderful, is creating these highlight videos more around the problem around and creating urgency. You know, one of the biggest things with change management that I think folks struggle with is alignment to the problem across stakeholders and then creating the urgency of like, now we should act. And there's nothing better than hearing directly from the patient specifically on this problem and something that can be attached in the email. So. So we actually, we had a client and he actually brought this idea to us and that he was like, could you package that up in a way that it can just be an attachment? And he even had like a certain number, like, don't make it more than like 2.5 gigabytes. Because I need it to show up, not as something that then you have to. I want it to be like the highlight clip that all the person has to do is click on. That'll pop up and it will show them. And so when I go then to present this feature that we want to update and use with you all my conversation is really easy because the patients have already explained why this is such a pain point. So we've really loved that video storytelling as a mechanism to create the urgency for change management.
- [23:08]
Logan Hartline
I think there is power if you don't have the access to a great capability like user testing. Just asking co workers like, hey, like, I heard that you went to the hospital this past week or hey, do you have a bill? Like, I'd love to hear what your experience is like, because there is power in story versus data. And that's not to say data is bad, but I think us being able to tell narratives to individuals that might have become abstracted from those stories really helps bring them to the bedside in a. In a meaningful way.
- [23:40]
Leah Hogan
Yeah, absolutely. Totally, totally agree with you on that.
- [23:43]
Logan Hartline
And it.
- [23:43]
Leah Hogan
And I think, you know, some of that urgency you can amplify a little bit and say, here's an example. And this costs us $20 million. Doesn't hurt to be able to tell that story either.
- [23:57]
Amy Stillman
Yeah, that and. Or the population size. And, like, this is like, a set of 10 users, but actually represents a similar problem that we think that, like, 40% of your actual user is, like, qualify for Medicaid today. Yes. Kind of thing. Yeah.
- [24:10]
Logan Hartline
Yes.
- [24:10]
Leah Hogan
Yeah. So storytelling, very important. So shifting gears a little bit. So what advice would you have to share with folks to start their journey down this path or continue, you know, if they've started to build a foundation?
- [24:25]
Logan Hartline
I would say, I think it was Michelangelo, you know, the famous Renaissance artist. On his deathbed, he's quoted as saying, I think Ankora emparo, which loosely translates, and yet I'm still learning. So, you know, this Renaissance master, as he is dying, is still kind of having this idea that, like, I'm still learning. And so, you know, for us, we are very curious. As an organization, I'm a very curious person, so I'm always asking questions. And so I would encourage anyone that's starting on this journey, start small. Like, it doesn't have to be a huge, you know, a mixed modal heuristic study. You know, three months. Like, it can be something very small that you're curious about. Like, I wonder why. Fill in the blank and then just start. Like, I think sometimes UX practitioners will get in their way if it's not exactly following, you know, the Nielsen Norman group model. Like, any feedback is better than no feedback at all. And so for us, we have tried to democratize research with our UX designers. So rather than have, you know, we have UX researchers that will focus on really significant studies, and they help coach our UX designers to make sure that we're not leading the witness or leading the bias. You know, that we want users to tell us, like, to make sure we're asking the right questions. You know, we. We do a lot of education. So I will try to help set aside time for my team on a calendar year just for them to be trained. So we brought Nielsen Norman Group in this past year to kind of do like leveling courses for everyone, individuals that may have never run a user test and others that might have been a little bit more experienced. And it's amazing. You get to learn just by being in an educational session. Oh, wow. That was something I never thought about doing. And so if you don't have the resources to do that, like they have free resources on their website. So I would say learn in your spare time and in your free time, but I would just say start, like, just start learning how to figure out the right way to ask good questions to be able to get some of those insights.
- [26:36]
Amy Stillman
Yeah, I think that's great. And I think that's so much of the storytelling piece. I think the other thing that I think we got to remember about health care and you have to think about not your boss, but your boss's boss's boss and what they're looking at. So much of it in healthcare is a numbers game. It's metrics, it's numbers. And I would encourage you to figure out what those numbers are, what do they care about, and figure out how you can draw the through line to the compelling story. But I do find sometimes it's hard if you only have the compelling story. It's hard to express the value to stakeholders that are. Design is a fuzzy thing. And so the more you can get crisp with a number or a metric or even suggesting a number that could be a proxy to that objective that they have, but a hard measurement. So I think you have to have the quant because health care speaks quant. Health care companies also run on really, really, really slim margins. So oftentimes any savings or any user acquisition lifetime value that you can prove is worth something. And then I think our superpower is the storytelling. And that's like get people, bring the patients to the front. That's our superpower that I think we ran to the table.
- [27:50]
Logan Hartline
Yeah. Cool.
- [27:50]
Leah Hogan
That's a great point to end this conversation and I thank you.
- [27:54]
Host
Both Amy and Logan want to keep the conversation going. You can find the show notes@usertesting.com podcast if you haven't already, don't forget to follow us on Apple Podcasts, Spotify, Overcast or Google Play so you never miss an episode. And if you enjoyed today's show, please share it with a friend or leave us a rating and review on Apple Podcast. And until next time, this is Insights Unlocked, an original podcast from User Testing.