Can AI Improve Clinician Well-Being?

Why This Episode Matters

Healthcare organizations are investing heavily in new technologies, yet many implementations unintentionally add complexity to clinicians’ daily work. This episode explores a different question: what if we deliberately evaluate tools for their ability to reduce friction and support clinician well-being?

Dr. Chris Dale and Dr. Ryan Dix discuss the development and evaluation of MedPearl, a clinical decision support tool built to streamline referrals and support frontline clinicians. Their conversation highlights why system design, not individual resilience, is often the most powerful lever for improving workforce well-being.

Key Ideas Explored

• Micro-frictions in clinical workflows accumulate into meaningful cognitive and emotional burden
• Organizational interventions often outperform individual resilience strategies
• MedPearl was designed to capture and operationalize “tribal knowledge” in referral workflows
• Technology adoption spreads socially through trusted peer networks
• Measuring well-being impact requires using existing data thoughtfully
• The future of innovation must include workforce impact, not just efficiency metrics

Takeaways for Quality Leaders

• Treat clinician well-being as a system property, not an individual responsibility
• Look for “sticky note problems” that signal hidden workflow friction
• Use existing organizational data sources before launching new surveys
• Expect heterogeneous impact. Not every tool benefits every group equally
• Pair product design thinking with traditional improvement methods
• Monitor indirect indicators of well-being, not just annual survey scores
• Recognize that meaningful improvement will come from many small changes, not one solution

Continue the Conversation

Connect with Dr. Ryan Dix through the Wellbeing Trust website to learn more about Providence’s workforce well-being initiatives.
Follow Dr. Chris Dale on X (Twitter) or LinkedIn or visit Arborgenie.com to explore his work in AI and clinical data.

 This episode is especially useful for quality leaders, CMOs, CMIOs, operational leaders evaluating new clinical technologies, and anyone interested in the intersection between AI, data, quality improvement, and clinician wellbeing.

If you found this conversation valuable, consider rating, commenting, or sharing with a colleague.

Resources & Frameworks Referenced

• MedPearl clinical decision support tool
• Quadruple Aim framework
• PDSA cycles
• Epic EHR integration concepts
• Physician well-being survey methodologies
• Physicians Foundation research funding

Leading Quality is a podcast for healthcare leaders committed to improving systems, culture, and outcomes.

If you found this episode valuable, follow the show, rate and review the podcast, or share it with a colleague working to improve care.

Connect with Jason Meadows on LinkedIn for more insights on healthcare quality and leadership.

Help us build this podcast  community from the ground up: share your top insight from this episode and where you’re seeing it in your own work. I read every response and will share what we’re learning over time in future episodes and other ways.

New episodes published every other Thursday at 7AM Eastern Time.

Credits:

Host, Writer, and Executive Producer
Jason Meadows, MD

Produced by
Thrive Healthcare Improvement

Edited by
Milan Milosavljevic

Transcript

Opening And Three Forces

SPEAKER_00 0:00

If you're gonna do nothing else, like focus on a great work environment for people, everything else will take care of itself. So as we think forward, I would predict that practice patterns are going to radically change over the next two to five years at different paces at different organizations, but well-being undoubtedly is gonna be part of the conversation because of what we've all been through together. So that makes me actually quite excited.

SPEAKER_02 0:19

Welcome to Leading Quality, the podcast spotlighting the people moving healthcare forward from the front lines to the C street. I'm your host, Jason Meadows. Today's conversation sits at the intersection of three forces that are important in shaping healthcare right now technology, clinician well-being, and the often invisible friction built into our daily workflows. My guests, Dr. Chris Dale and Dr. Ryan Dix, both of Providence, bring complimentary lenses to this challenge. Chris is a pulmonary and critical care physician, former Chief Quality Officer and Chief Medical Officer, and now a leader working at the intersection of data, AI, and care delivery. Ryan is a psychologist and senior leader focused on workforce mental health and well-being across a large health system. Together, they've been studying something deceptively simple but deeply important. What happens when an What happens when we design technology not just to improve efficiency, but to remove the small cumulative burdens that clinicians experience every day? In this episode, we explore MedPurl, a clinical decision support tool designed to reduce referral friction and the broader question it raises Should clinician well-being be treated as a core design requirement when we build and deploy new technology? This is a grounded, thoughtful conversation about systems measurement and what it really takes to make care better for both patients and the people delivering it.

unknown 1:56

Dr.

SPEAKER_02 1:56

Chris

Meet Chris Dale And Ryan Dix

SPEAKER_02 1:57

Dale, Dr. Ryan Dix, welcome to the show. Thanks for having us. Yeah, thanks for uh thanks for inviting us on. Well, I've been we had we had one false start in the scheduling process. I'm looking forward to getting uh this conversation under our belts. And uh I did want to start maybe with just a 10,000 foot view of each of your backgrounds. Chris, maybe we'll we'll start with you. I know you've had kind of an interesting path from ICU physician and then a variety of leadership roles, CQO being one of them, but not the current one. Maybe you can just give us an overview of your career.

SPEAKER_00 2:29

Yeah, sure. Uh thanks, Jason. It's great to be with you. And Ryan, I always love hanging out with you. Um yeah, I mean, I'm a simple country pulmonologist, basically. Like I do pulmonary and critical care medicine here in the Seattle area, uh, work at Swedish, which is part of the bigger Providence um family. Um, and then on the non-clinical side, uh, I don't know. I've had a meandering journey. Um, I started out as a flight surgeon uh in the US Navy uh with the Marine Corps, and that got me interested in um kind of aviation safety or the implementation of medicine. I think that uh aviation safety is kind of a gateway drug for a number of us into quality. And that led me to health services research, which led me to quality. And then I was a chief quality officer here at Swedish for a period of time and the CMO um for a period of time. And then I kind of took a bit of a turn and I've been um trying to use data to improve quality. And that's what I've been really focused on for the last, I don't know, four or five years. I've been the CMO of two different health data startups, and now I run uh my own kind of with my wife actually, um, an AI company focused on small practices called Arborgenie. Um, and then also work at Providence doing um a variety of different health services research and other uh quality improvement sorts of projects.

SPEAKER_02 3:32

So yeah, I don't know. That's that's me in a nutshell. That's great. And Ryan, how about for you?

SPEAKER_01 3:38

Yeah, so I'm a psychologist by background. Uh have uh on the clinical side mostly spent my career in some form of hospital or primary care clinic to uh providing integrated primary care. Uh so had the the pleasure of working alongside lots of physicians and uh APPs and a number of other care providers. I transitioned into um leadership roles lots of years ago. Um and really since then uh really kind of focused on one overseeing clinics and a number of different programs. Um but over the last couple of years uh had the opportunity to partner with a number of uh other parts of our organization to oversee all the mental health and wellness initiatives uh for Providence. And that's been a part of my career uh really for the better part of the last 10 or 15 years, uh at least in some form or fashion, focus on the well-being of all care providers and anyone associated with uh with the clinical team. And so uh focused on that currently and uh really enjoying uh being able to focus on the mental health and well-being of our workforce.

SPEAKER_02 4:47

Great. And and then so if I'm understanding correctly, it is the the workforce itself, so not as much about mental health um initiatives for patient care specifically, but for the the people providing care.

SPEAKER_01 5:00

Yeah, so I'm on a team that has um three pillars essentially. And so the pillar that I'm responsible for is the focused on employees or caregivers. And so uh our team does focus on patient initiatives, but that's not my role specifically. So so my work is focused on uh caregivers specifically.

SPEAKER_02 5:19

Yeah, and so with these kind of two different in two very different backgrounds, I'm curious how you first came together to do work that we'll be talking about today, and and how do the these two lenses kind of complement each other.

SPEAKER_00 5:33

Yeah, it's like how do you start a band? Uh actually uh Ryan put a little poster on a uh on a lamp sign that's like drummer needed. And I was like, Yeah, I play the drums. And that that's how we got together. No, it actually started. Uh let me I can maybe talk a little bit about the about the product that brought us together. And then, you know, maybe that kind of dovetails into the origin, sorry for Ryan and I um collaborating. But,

MedPearl Origin And Referral Friction

SPEAKER_00 5:53

you know, both Ryan and I work for Providence, which is a big healthcare system here on the west coast of the US, you know, 51 hospitals, and any and number of uh clinics. I don't even know how many uh physicians, APPs, nurses we employ, but it's a lot, it's big. One of my uh colleagues, a lady named Evie Cunningham, um, who now works at Cadence, um, had the idea. She's an OBGYN doctor, uh, practicing clinically and doing uh leadership work. She had the idea that it would be very helpful if when patients came to their specialist that the tests that they needed prior to referral to the specialists were kind of taken care of, and that the kind of the work that could be done in primary care was done in primary care. Um, and that the patient's journey would be aided by uh making the furrils kind of right-sized and at the appropriate time. And with that background, she, I mean, to her immense credit, with a gentleman, uh Adrian Giannis, uh, spent their nights and weekends developing a tool that ended up being called MedPearl, like Medicine and then Pearl because it was a little bit of wisdom. It was supposed to be Pearls of Wisdom, I think was the origin story for the name. And then they implemented that in, well, first in Teams, like we are a Microsoft partner. And so we did a lot of Microsoft-based, or we do a lot of Microsoft-based, um, I don't know how you say it, like IS or back office um sorts of things. So they built a prototype in Teams and um debuted it with a small number of doctors, like a couple of uh, you know, two dozen or so doctors and APPs, um, and got really high remarks or high grades in terms of the kind of the PDSA there was just based on improvement in the in their uh in the self-perceived quality of their referrals was the the primary uh outcome asia there. And Evi and colleagues had some other ones. And then that thing done well, like I have a saying, like small things done well earn the right to do big things, like that small proof of concept then led to the system wanting to scale it within the Providence Clinical Network or PCN, kind of the physician APP medical practice arm of Providence. So then Evi and Adrian uh hired up some additional developers and built a tool working with uh content providers within the Providence family. So they'd try and find uh like you know, Ryan's and behavior health. So they try and find a friendly psychologist or psychiatrist to maybe let's say write a depression algorithm. And that would include both the diagnosis of the condition, um, kind of like yellow flags, red things, red flags, like things to think about as you're um seeing the patient in clinic, really focused on primary care and urgent care in the ED as well. And then uh suggested tests. So if there are any kind of like imaging studies or lab studies that you'd want to do prior to referral, um, and then uh and then that would tee up then the kind of next step in the patient's journey. And uh through the wizardry of Adrian's team, they were able to integrate in with Epic, and also through the our kind of our Epic team, um, able to integrate into Epic. So it would pull contextual data about the patient from studies that they already had performed. So imagine if you saw information about the re the thing that you were looking up, kind of like an up-to-date or an open evidence sort of monograph in context with what we already knew about the patient. And again, like the whole premise here was to remove friction and help uh ease patients' way in terms of their care journey. So that was uh that was the tool that was developed. Evi invited me to join um kind of from a wearing my kind of like health services research hat or my technology assessment hat to understand the impacts of the tool in terms of practice. Um this is something that's been, you know, I don't think widely discussed in our industry, but you know, there's a lot of stuff around like AI or technologies in terms of like, hey, we got to measure it, like we got to do stuff, we gotta measure, we gotta measure it. And there are precious few studies on the kind of the real world implementations of things. Um, that coupled with just the tremendous financial pressures that organizations are on. I think people who innovate within the context of organizations have an obligation to serve the organizational mission. And part of that's being you know fiscally responsible. It's you know that saying in Catholic healthcare, you know, mark no margin, no mission. It that applies, I think, to innovation and that our innovations need to show that they produce their intended effects. Otherwise, you just kind of go off in a billion different um directions. And so um Evie and I partnered then to do an analysis of MedParl's intervention. And one of the main things that we kind of picked up, or she picked up at the beginning, Evie's the main driver. I'm just like some you know random kid in the backseat. But one of the things that uh Evie picked up from the beginning was just this kind of like spark of feedback or this intuition that med Pearl ease people's way. There's so much happening in AI, like things are moving so, so quickly. And you can have this very kind of utopian view of things or this very dystopian view of the thing of the future. And it's hard to know necessarily for you know specific uh for the future what's actually going to happen. And so I think it's really for people who are interested in the space incumbent upon us to study stuff as things progress. And so the kind of the challenge that we laid down before us was to um to study the well-being implications of uh this MedParl tool. Yeah, and so that's kind of what led to our partnership. We um got uh extra mural funding through the Physicians Foundation. Um, there's a nonprofit in the States um called the Physicians Foundation, but it's uh it's interested in aspects of uh physicianhood and physician practice, uh including well-being. And they put out uh an RFP specifically focused on kind of technology and AI and well-being, and then um uh through the grant writing folks at Providence who responded to it. And then um, I was just like kind of shaken my professional tree within PROV to find co-investigators who are working on the same thing. And I'd kind of heard of Ryan, but Ryan, I'd never met you before before we started kind of collaborating on this. But there's a gentleman, Mark Rosenberg, uh, who's done well-being work in in the state of Oregon within Providence, and it was a mentor of mine from Residency. And Mark knew Ryan introduced us, and that's how we um kind of jumped jumped into the uh more formal assessment of the well-being uh aspects of this uh technology intervention. So I don't know. Ryan, I feel like I'm blathering on. What what do you got to add? Subtract, clarify.

SPEAKER_01 11:33

I think you captured it. I'm I'm still wondering which uh which part of the band I'm in, though, if you're if you're the drummer, but what would you be? Do you think you'd be a I could see as a I don't know. Gosh, I hope not for everyone else's sake. But um, yeah, I don't know. No, I I think you captured it well. I mean, I I think that's um at least as far as the origin story.

SPEAKER_02 11:50

It's it's Ryan, I'm curious. So uh your work, it's it sounds like is is primarily focused on clinician well-being. And so this is uh you know a beautiful intersection with with the work that you're doing. When someone when they come to you and say, hey, we'd like to assess this tool that we've created called MedPearl from the perspective of physician mental health from burnout, uh, how do you start to conceptualize what that that whole thing looks like?

Measuring Well-Being Without More Surveys

SPEAKER_01 12:16

Connecting with Chris was really first around um what types of assessment around well-being would we want to look at? And there's lots of ways that you can assess well-being, you know, whether it's a survey, whether it's um you know, impacts on lots of other metrics that that exist. And so uh one of the things that we do within our organization is uh is a survey that's provided annually for our physicians and APPs. And so identifying uh not only the that tool, but how within that tool uh would we want to assess the impacts of of really any intervention. And so MedParl allowed us one to to identify what that what the intervention was going to be, but then identifying whatever type of outcome you're gonna really track as part of that. Within that, there's lots of questions in terms of uh well, we know about the use of technology in the in the betterment of well-being. And uh it is a space that's evolving really quickly, but is a space that a lot of organizations have invested a lot of time and and money into. And there are certainly some areas where it shows uh significant impacts, and there's other areas that it's maybe still a little bit up for a question, either in terms of its uh its impact more broadly speaking, or is it the same for all groups? And what we know from other studies is that uh some interventions work really well with some groups and some work less well with other groups. And so trying to identify, and so those were some of the additional questions that when you look at the broader well-being literature, I was able to bring into the conversation and and have Chris and others perspectives on as we think about the med Perl tool, there's of course, you know, does it impact well-being, you know, before and after? But then are there opportunities to further refine that the utilization of that tool? And one of the things that I see as well is when when you're thinking about an intervention, really deciding that there's of all of the things that are available, which ones make sense for which groups. And I think that's the some of the additional work that's being done in the well-being space is is identifying maybe this can't be everyone needs this one specific tool, but other specific groups within that. And so those are some of the additional parts of the med-pearl conversation that that we were able to bring in.

SPEAKER_02 14:43

It sounds like the rather than maybe the way that well-being and and burnout for healthcare providers is is framed a lot of the time, which is in terms of things like reducing workload or building resilience at an individual level. I mean, this tool was conceived to kind of reduce microfrictions that that exist within within your system. And maybe I missed it, but I wonder if you can fill me in, Chris, on kind of where you said there was a small pilot that started and then they wanted to spread it. Where did it start? Kind of why did it, do you know why it started there? And then how did the the spread look from there?

SPEAKER_00 15:17

Yeah,

Innovation In Big Systems

SPEAKER_00 15:18

well, that's a good uh a good question. Um, and you know, Ryan can talk about it, but you know, he introduced me to the the notion of uh pickle jar and pickle juice in terms of like, you know, how do the paper cuts of daily practice impact individual physicians or APPs or nurses, or you can just extend it generically to healthcare workers. Um and so you're right, that a large, a large part of it is the brine, the juice that we're all uh emerged in. I think a whole thing, like you'd have like a whole podcast, there probably are out there, on like innovation, specifically within big systems. Um, you know, I think uh, you know, that Paul Batalden quote of like every system is perfectly designed to get the results that it gets, like that applies to the big systems, which you know, Providence is around for 150 years and has um, you know, very established ways of doing things. And so sometimes to push against that and to do things in different ways can be challenging, um, particularly in environments where there's a lot of financial pressure pressure, because the organization really kind of sometimes can have a tendency to take a kind of a cost mentality and like, okay, what's the cost of this thing or that thing? And so to uh to answer your question about like, well, where did MedPro get started? Again, this comes back to Ebbie Cunningham. And, you know, I think that a lot of us who do innovation within bigger systems, it's uh, you know, politics is described as the art of the possible. And and I think that same label applies to innovation in big systems sometimes. It's like it's the art of the possible. If you're kind of mid-level person or a person with a good idea, it's like, what can you get done? You know, what's within like what are the tools that you have available to you? Constraints can be very helpful from an innovation perspective. If you say, like, hey, the rules of the road of this specific thing, I'm just gonna talk generically, are we're not gonna touch Epic, you know, we're not gonna play in the EMR and ask for like a new big module, we're not gonna contract with like a new external vendor for a brand new service, X, Y, or Z, and going through the internal vetting, contracting, you know, IS process, because that'd be a you know, a year or two anyway. We're going to say, uh, we're given the tools that are available to us right this second that we can kind of get our hands on, um, what are the things that we can do in terms of improving and then whatever the thing is that you're wanting to improve? So in this case, it was looking at the the, I think Jason you said it well, the kind of the micro frictions of the practice environment with specifically um knowing that a growth, and you know, I just say this as like someone who's been practicing pulmonary critical care for I don't know, over 15 years now or whatever, is you accumulate knowledge about how to get things done in your in your system. And I think a little bit of the why that was starting, and and Evie actually has pictures of it too, would be the doc who's been around for a while or the APP who's been around for a while, she take pictures of their computer, their workstation, and they'd have like sticky notes about like, hey, if you're gonna refer to cardiology uh prior to that, you're like for AFib, like get an echo prior to referring or something like that. I don't know. I'm not a PCP, so that's probably a stupid example. But they had these sticky notes. And so Effie's like, well, there's all this tribal knowledge in there in everyone's sticky notes about like the way that patients actually flow through the system. And her insight was like, well, if we could capture that knowledge and we could say, like, hey, you know, dear OB, like what do you want to see, or dear GIN doctor, what do you want to see prior to the patient arriving at your door? It would not only be better for the receiving doctor, it'd be better for the sending doctor, and most importantly, it'd be better for the patient because the patient, like sometimes there's like an extra step, is like, you know, you come and see me as a pulmonary doc, and I'll just be like, well, I need a CT scan and PFTs. It's like, you know, it's like a one neuron reflex for a pulmonary doc. It's like PFTs, PFTs or CT. You know, it's very easy for us. And so it's like, uh, if if the person shows up with all that stuff kind of already done, then it just eases people's way. And so it was really the art of the possible. So Evie designed this like literally nights and weekends, and she built it, like I said before, in uh as a Teams app so that people could kind of select through it. And then she, you know, using the power of persuasion, can convinced uh, you know, a dozen of her friends or uh two dozen of her friends to try it in in kind of the southwest Washington area around Olympia, like our state capital, just south of Seattle. And she was thoughtful about it and did kind of before-after surveys of uh of people's self-perceived kind of I don't know, quality of life. And I I can't recall all the different variables that she assessed in the in those surveys. But uh at the end of the day, she got this little bit of data that said, hey, you know, there's there's probably something here. It was enough of a signal then to go to uh the leaders of the of the Providence Clinical Network Network of the PCN and kind of get the the funding or the internal authorization to build it up um bigger. So that's kind of how it started, is just somebody's good idea, somebody using the constraints in a local environment to focus their thinking. Again, I think constraints, like you can view them as a negative thing, but they can also be viewed as a very positive thing because they can help help focus innovation. Uh and then uh and then you know, small things done well earn the right to do big things. Like if someone who has dealt with healthcare data for you know 15 or 20 years now, I tell you, just like the start of things is just like count things. Like if you can count things, like it seems stupid, but just like tally stuff. And if you can tally things in one column versus another column, like you you probably have a data set that like tells a story reasonably well. So don't overcomplicate it. And then that that can be maybe the the ember, the fires that you blow on to then earn the right to do something bigger.

SPEAKER_02 20:30

Yeah, and and you know, Ryan, I I wonder if this, so I can I can very much imagine, you know, as a clinician myself, kind of thinking about someone referring to me, I really need something done, it's not done, and that's this is happening over and over and over again. And this is you know a real kind of maybe cognitive and emotional drag. Was this already kind of part of the work? Was this already on your radar that that consults were a big point of of difficulty for clinicians?

SPEAKER_01 20:58

I think the the broader theme that it taps into is what are a lot of those system uh components or environmental factors? And Chris uh provided that metaphor as he was responding it. And so there's all of these things about systems that have significant impacts, sometimes small little little bits at a time, uh, but sometimes that's you know, they're larger pieces that we know is really where uh well-being work should focus. Um there's a lot we can do, of course, in terms of teaching individuals more resiliency, but the literature really directs us towards focusing on organizational or environmental interventions more so. And so uh consults, you know, being one part of those, but there's so many of those things, so many of those sticky notes on various people's computers about all of the little parts of systems that sometimes we uh adapt or kind of learn ways to adjust to it. But The reality is, is if we really want to impact well-being in the most substantive way, we have to focus on adjusting those pieces and not expecting Chris or any other physician to just continuously adapt. Of course they can, but that's really goes against what we know to be most effective when we impact people's well-being is what are those pieces that that yes you can and have adapted to, but boy, if we could address that and we if we could modify the system in a way that removed that from you, the amount of impact that can have on their well-being is markedly more uh significant than if we offer you know any types of personal resilience stuff. Again, it's not to discount those parts, but we know that the vast majority should focus on that. So again, consoles is is an example of that, but there's so many of those things that are built into systems, particularly big systems, that can have a really significant impact on well-being. And so it's something that we are uh constantly focused on and in terms of that well-being work.

SPEAKER_00 23:03

One one thing, Jason, I've wondered about that just in terms of quality. I don't know if you have a comment on this one or the other, but like Ryan, you're talking there about like I think it was sometimes like the paper cuts, like just the difficulties of going through life and like the extra clicks. Um I wonder about it from the patient perspective too, is you know, all these clicks that it takes to get care and all the, you know, just kind of the the steps and the hassle that that all these things are are going through. I mean, it seems like it's tremendously deleterious to the patient aspect of well-being, in addition to the you know, the healthcare worker aspect of well-being. So I like, I think that insight that Ryan has is just like, you know, focus on people's journeys and making the journey easier if that applies more broadly.

SPEAKER_02 23:43

I I totally agree. And I think it's it's important, even at a basic level, that you're highlighting essentially that redesigning the work, like the nature of the work, how the work flows, how the dots connect, you know, is is such a big part of this, and not just insisting that people do things at an individual level to be more, you know, quote unquote be more resilient. Before we get into the the details of you know what you've measured and how you've you've measured it, I'm curious kind of how the you know, I guess how the sausage is made or what what is going into the the MedPurl uh product itself.

How MedPearl Works In Epic

SPEAKER_02 24:16

So how does it use, because now it's integrated with Epic, you said it's it's been deployed and it's been spread further. It's using real-time Epic data, it's using kind of semantic AI, maybe something akin to a large language model. Is that is that how it's working? How is it actually kind of taking in information and then making it useful to people?

SPEAKER_00 24:35

Yeah, there's a bit of AI in terms of the search functionality of it. Um uh, but the but the the core of it is actually, you know, relatively, although it's been historically very labor intensive, it's it's relatively simple kind of conceptually, in terms of um there's a uh a large body of monographs or just kind of collections of writings on different topics within within MedPerl. And so uh again, like the primary customer um was conceptualized as primary care providers or urgent care providers kind of bleeding into uh the ED um space. Nice to say uh too is that I'm like as kind of time has evolved, Provence is now spinning or has spun out MedPurl to a private company. Um and so a lot of these RD works that Prov did to get it off the ground, like they're they they've been transferred outside of our organization at this um at this point in time. Um and so uh, but historically, um each one of those monographs then was was hand labeled for the different uh fire resources or the different kind of uh cookie crumbs or the connectors into the the EMR for the various kind of semantic objects that were necessary to populate the monograph. So, for example, if um like let's say you have a uh one, I don't know, uh on GERD or question of uppergy, I believe, or something like that, you you might link that to hemoglobin. Um and then uh the the we'd pulled all recent hemoglobin values or the five most recent hemoglobin values so that the the provider in the context of of the window, which was an iframe within Epic, would see their data in terms of the patient's data in terms of um, hey, here's the you know, most current hemoglobin isn't is 9.6, the previous one was you know 10.2, the previous one before that was 11.5 or something like that. So um, I don't know. I'm choosing an inpatient example uh because that's where I work, but like similar sorts of things. Like if you're training, I know TSHs over time or something like that, or A1Cs over time, or other parameters you might be interested in for any one of those things. And so there was, you know, a collection of like hundreds or thousands different sorts of monographs that there were uh organizationally, there was kind of a senior person who was kind of in charge of the whole body of knowledge and then different topic specialists, kind of like quarterbacks. Um, and then they'd kind of farm out. And so, like I would review some of the ones on pulmonary, like, hey, look at this one on you know obstructive lung disease diagnosis or COPD management. And then the work there was concordance with guidelines and um, you know, any, and then they each had kind of like ticklers to come back to at a certain point in time to make sure that the the evidence stayed um up to date. But yeah, that's how the whole that's the the process that Evie and Jessica Slitcher and and others built to keep the kind of the knowledge graph of up-to-date up or of uh the knowledge graph of Med Perl up to date. Up to date being a different product from multiple.

SPEAKER_02 27:24

Right, right, right. Exactly. Um yeah, no uh want to make that that distinction very clear. And so in in your collaboration and and this this work became focused on, you know, how is this product, which already exists, which is already spread, I guess at this point, to all different types of consultations, like primary, so primary care consulting and urgent care consulting, every type of specialist at a couple of hospitals. Is that kind of where we are when when you guys start to work together?

SPEAKER_00 27:53

Yeah. So

Adoption Patterns And Product Design Lessons

SPEAKER_00 27:54

uh Evy and I started talking, you know, a couple of years ago about it, and I knew about it. She was um kind of to show me uh some demos of it like very early on in the team sort of space. And I was the CMO of this uh kind of affiliated health data startup. So we were kind of crocking out over the data together and how to pull data and how to put data in context and what things I I've I've done some work on the more ML side of AI, and so you're talking about predict predictions and prediction engines and scores. Uh, there's some odd like if you think about the things you can pull from uh from Epic, there's some automatic scores that you can calculate. So that's part of it as well. I guess I didn't mention that. But anyway, but at that point in time it was still in Teams, and then she um got it implemented in Epic. Providence has three different, they're kind of like operated in parallel, but three physical different builds of Epic, just because Providence is so big. I don't know, I guess you can't have Epics over a certain size or something like that. I don't really know. But there's one in Alaska, there's a one in Washington, Montana, and there's one in Oregon, California, um, or the three different builds. And so she got it put into each of the three different builds and was going through a um kind of an internal change management sort of process, um, where she and the team were basically doing the dog and pony show, like, you know, talking to anybody at any lunch meeting to tell people about it. And we had internal uh usage data. And I mean, actually, from a QI perspective, it was fan fascinating um that we uh we produced some data where we were graphing the spread of med Pearl, and it was like uh, I don't know, it was like John Snow in the well or whatever. It was like you could see a hotspot emerge in a practice, and then it would, it would spread to the the usage would spread to the other people in the practice, and then it would seed like the practice down the road, and then that one would go or whatever. So you saw this like very geographic, very person-centered, like, oh, how do how does behavior change? Well, it changes when you kind of bump into your friend at lunch or whatever, and they're like, Hey, have you heard about this? And then you're like, nah, like how how's it going for you? And then uh through that conversation, then you begin to adopt it and stuff like that. And and then we began to get profiles of like super users, like you know, these people that were in med Pearl, you know, every patient every day, versus some more kind of like recreational users who would like touch it occasionally for maybe certain things, but not, you know, not much. And then like people who basically never uh never used it, you know, never user sort of people. And they began to merge these like fingerprints. And the story I kind of told myself about them is that like the highest users are people who are relatively new to practice. You know, they're they're kind of they they tend to be sooner out of training, which kind of makes sense to me a little bit. Like you have this technology who's designed to aid people in their way in terms of like easing their way practicing. And by the time you get to be a little bit more established, you probably have certain ways, at least in your mind, about doing activities A, B, and C. And so it's like there's less utility in like someone coming along and telling you what you already know. So it's it tended to be people who are uh who are uh newer to the organization. It tended to be a little bit more, this was interesting, I wasn't expecting it. It tended to be uh a little bit more APP versus physician. And that was biased just the way our organization is organized. Urgent cares are kind of disproportionately staffed by APPs versus physicians. And so there's some hot spotting about APPs in urgent care. And that, like if you're gonna have like a single phenotype of the most common user, um, that that would be it. Um, residents, uh, similarly, uh, there are a number of residencies that prov's got, I'm not, I don't know the total number, but half a dozen to a dozen different uh family practice and internal medicine residencies, uh, in addition to other ones. But uh, those would be ones where it would kind of take off and like you'd see like basically the whole cohort using it. It was like this is how they practiced, um, sort of things. And so anyway, so that's kind of how it began to spread, which I thought from like, you know, how does how do you how do you understand user journeys? I think there are a couple of aspects of it that are really interesting to me as a quality improver is like one, just being able to map it out, um, because we were able to display the data just just graphically. We did a like this on the data nerd side, but we, you know, we have a bunch of stuff in Snowflake and we pull it out and we like uh geocoded things for locations and we created these maps uh that were really kind of cool and interactive. And you could see like changes in use over time uh based on uh geography, which I thought was just kind of interesting for us as we're trying to tell the story about like and trying to trying to understand too. Um so part of that's on the like quality improvement data nerd side, and then part which is interrelated to quality improvement, but it's like how do you build good products? Um, and I think sometimes like the I've done a fair amount of work over the years just in in designing products, and that skill is maybe underappreciated, I'd say, in traditional quality improvement, which tends to be much, I mean, it comes from you know deming and industrial processes and things like that, not from a product design like UI UX um sort of perspective. And I think that bleeding it over and saying, oh, like how do I design an intervention where my users love to, you know, be in the thing and they love to do stuff, and it's like they get some kind of dopamine hit out of doing it, like that's a really cool design principle to have in terms of quality improvement. Um, and so anyway, so we combined all those sorts of things, and that's how we kind of like told ourselves a story about who used MedPro and its growth.

SPEAKER_02 32:47

Yeah, I mean, a lot to dig into there. It sounds like to use an example, essentially, how do you build, you know, how do you build Google search or how do you build something that that kind of has this inherent uh appeal to it? It kind of reduces friction in an obvious way. Is that kind of what you're getting at?

SPEAKER_00 33:03

Yeah, yeah, yeah. I mean, I think that's a very good example. I mean, I think two things about it. Like one, like Google, like Google search is not like it doesn't come with a manual. Like none of us ever had to go to some sort of like web-based training to use Google search. So it is possible to design great products that are widely used without a without a lot of friction in the uptake process. And then the second thing is like if you look at the like particularly the in the in the earlier days of Google, just like its exponential growth in terms of like, well, I mean each of us could probably think back individually, like and maybe we can't recall because it's so ubiquitous now. But like, how do we hear about first hear about that thing called Google? And like, you know, there was a bunch of stuff before, like, you know, Lycos and web crawlers and Yahoo, and like Yahoo was like the the was was the bees' knees, right? Like Yahoo was gonna kill everyone back in the day. And then Google is just like upstart with like this like stupid looking homepage that had nothing on it except for like the word Google. Like, what was that? Like Yahoo had all this stuff, and like now, like Yahoo, I don't even know. Like, I think it's around. Uh, and so like I think that's interesting to think about is like how can you get to something where the organic growth of it is just so compelling that it tells you that you're on to I I think it's like kind of the dominant solution is like the thing that has better outcomes for lower lower friction cost, like you know, economics a dominant thing is like better outcomes, lower cost. And you can think of that same thing sociologically or in terms of product development too.

SPEAKER_02 34:22

So then

Study Hypotheses And Early Findings

SPEAKER_02 34:23

as this is spreading, um the two of you know find each other through this work and you get grant funding. And Ryan, I'm I'm curious what kind of specific measures you started to to put together so that you could be able to measure whether this tool was uh was making an impact.

SPEAKER_01 34:40

Well, I uh as I mentioned a little while ago, we uh use an organizational uh assessment that allows us to um assess well-being and so forth, uh, amongst a number of other things. And so those metrics had already been selected organizationally. And so from our perspective, it was really identifying which ones we felt like could be most useful in terms of asking the questions that we were looking to ask in terms of the uh use of MedPurl and then really being able to extrapolate beyond that. And so I think as Kristen pointed out earlier, uh just kind of to trying to find what was already available rather than uh trying to add yet another survey or something like that. We know that there's a lot of exhaustion when we think about asking people additional pieces of information. And so that was the other part that that was important to us was really trying to identify is there is there a data set that already exists in some kind of way that we might be able to tap into that would allow us to answer our questions.

SPEAKER_02 35:39

Yeah, yeah, there's there's a there's always an inherent risk in decreasing well-being by asking people too much how their well-being is going. So I'm curious if you're at the stage now of having some results that you can share maybe on specific dimensions of of clinician well-being and maybe seeing areas where it's helping more or or less. Uh are you kind of at that stage already?

SPEAKER_00 36:00

I guess the short answer is probably no. Like um, so we're working with a PhD researcher who's been just absolutely wonderful, uh, Amara Mohamed, and she's uh leading in the project. And we have kind of we have a couple of different research questions about one is the association between uh well-being and product use, uh, which we came into this with a kind of a an association that people who had had potentially greater indices of well-being, maybe greater self-efficacy would be more likely to try to use the product. And then the second is that the our hypothesis was that the product is associated with uh improvement and well-being over time or greater rates of improvement and well-being over time. So those are the things that we're kind of like testing through the data right now. And we're at the stage where we have the data, we have it together, and we're analyzing it. So I guess I don't know, your the audience will have to stay tuned for the like the great reveal, but we're in the like in the throes of analyzing the data at this point in time. But it's been really uh like a fun and interesting journey. And again, like I'm deeply appreciative of the Physicians Foundation for the funding that makes it um possible. And I do think that you know, I've done a lot of unfunded research work over the years, and like as Ryan was saying, it's a bit kind of like catch-us, catch-can in terms of the data. And I think that that's like a good health services researcher is like someone who's like they kind of understand the data structure of their organization and and like the stories that can be told from it, and then they kind of take advantage of things that are going to happen anyway. And then uh me and a colleague Michael Melbahl used to joke that it's like you our job is to put a p-value on it, is like this like natural experiments are going on all around us every day. And if we look with the right lens or we nudge things in the right way, we can we'll be lucky enough to put a p-value on it. And so, really, that's basically what this this project is designed to do. And I'm I mean, just as excited as you are in terms of like seeing the end results of it. We have results on like the uh med pro, like the non-well-being uh aspects of medParl. Like there's a number of just like fascinating things about it. Like one, you know, it's a it's associated with a decreased referral. We were talking about referrals before, like that was one of the primary parts of it or whatever, but it's associated with deep the use of med perils associated with decreased referral rates out of urgent care. And the story that we are, you know, told ourselves there is that it it was it kind of steered people away from having to like knee-jerk refer if they understand, oh, I can take, I can order these tests, and maybe you know, the person can follow up with their primary care provider, or the story can just kind of end there for their care journey and they don't they don't need a referral. Conversely, in primary care, it's actually associated with a very slightly but statistically significant higher rate of uh of referrals. And I don't know, maybe the storyteller shows there's a somewhat the opposite is that it helps kind of refine people's thinking and then say, oh, you know, these people should be referred. And then when people are referred, there are a number of interesting things. Like one, the organization was interested in kind of the the business impacts of it. And the use of med Pearl is associated um with improved surgical conversion rates and improved contribution margin on people who had uh conditions where med Pearl was used versus those uh where the same conditions where med prol wasn't used. And the story we kind of tell ourselves there is that you know, these might be patients for whom they're they're better kind of prepped for whatever might come um downstream. And the organization might be able to get to the and the patient might be able to get to the end result a little bit more efficiently. Um so there's just a bunch of very you know rich data, you know, like slightly improved RVU, site uh decreased uh after hours charting time by an hour a month. Like there's just a bunch of like very interesting practice data about the about the tool.

SPEAKER_02 39:25

Wow. Any any hunches on which we don't have the the data kind of fully assessed yet? Any hunches as to whether this is going to be you know seen as a you know overall positive impact on on primary care and and uh urgent care physician well-being? Or is it just too early to say?

SPEAKER_00 39:44

Yeah, I mean, I think I wouldn't be surprised if like the use of my perils associated with well-being overall, like like enhanced well-being overall, but but maybe not like a tremendous amount. And you know, I think it that that wouldn't surprise me just because the it's very complicated. You know, there's so many factors that go into it, uh, you know, the the model of the environmental thing versus the personal things um in each one of those. And this is just like very one very small aspect of the environment of the environment of care is you know, how do you uh how do you smooth the referral process or how do you smooth the caregiving process uh for patients in front of you? You know, that's just like one very, very small slice of the pie. So probably illogical to think that it would have like a tremendous outsize effect on on well-being overall. But it but I think it, I mean, what I think about those things is like uh it takes all of us. Like it, you know, it's it's not gonna be one thing, it's gonna be a bunch of things. And so like we gotta like our our for people who care about healthcare delivery and care about patients and care about the you know, healthcare workers is you know, how what are the things, the changes that we can make in terms of the whole care production pipeline to decrease all sorts of different frictions? And then how how can we know those things? And that like putting a p-value out of the scientific process, like that's what I take maybe most away from quality improvement overall, is just like having hypothesis and measuring the outcome of it. Like that process there, that's the thing that really ends up winning the war is you know, individual like kind of hypothesis we have from time to time may or may not be right. Like that's the whole point of having them. But the overall approach is one that really tremendously has legs.

SPEAKER_02 41:10

Having had this experience, and I think so often, you know, if we're talking about a new tool or a new care pathway or something, then like so often we're talking about increasing some kind of efficiency, right? Decreasing a wait time or increasing some kind of you know, bed turnover rate or whatever it might be. But this is, you know, using the data you already have, this is very focused in on uh on clinician well-being as a result of this tool.

Making Well-Being A Design Requirement

SPEAKER_02 41:35

Should clinician well-being be kind of treated as a formal design requirement by healthcare institutions before deploying new technology? Is this should this be or or is this something that is kind of routinely a lens through which your organization or or organizations that you've seen generally um view new technology?

SPEAKER_01 41:54

Yes, it should be required. And just to elaborate on that a little bit, I mean it the data, at least from my perspective, is pretty clear of the impact of the well-being of our physicians, APPs, nurses, medical assistants, all that has impacts on every every data point you could possibly care about. Whether you're a clinician, whether you're a healthcare administrator, whether you're a patient, whether you're a patient family member. And so it it of course we care about the person experiencing you know well-being or or absence of well-being. But when you look at all of those downstream impacts, to me the the picture is very clear that that should be an absolute requirement of um really any new technology that we that we look to implement. I think the questions then become which technologies for which groups? And that's where I think organizations hopefully are I feel like the conversations have progressed a little bit from where I started, whereas initially well-being I think was seen as a nice to have, or like as a not at not at Providence specifically, but just kind of more naturally, part of the conversation was more it wasn't really recognized for the level of importance. And I think you know, for better or worse, one of the things that the pandemic kind of shined a light on was a lot of the importance of well-being on the healthcare workforce. And so I feel like the conversations have progressed a little bit past, although they still happen in certain places, um, the need for it, but it's now kind of what what's most efficacious, and beyond that, which groups would most likely benefit from you know, X tool or Y tool. And I think that's where the literature is hopefully moving towards is being able to allow us to more specifically articulate those groups that would most benefit from something like mentor or another piece of technology.

SPEAKER_00 43:53

Yeah. I d I do wonder a a bit about that though, Jason, because um like you posited before uh Um, like the you know, the off-sided survey fatigue is like, oh yeah, like you know, the we don't like people don't really want to be out like surveyed all the time about various things. And and and I've always kind of wondered the kind of the counter of that, which is like Ryan, if what you're saying is true, that we need to, and and I like a hundred million percent agree with this, is that like if you think about the quadruple aim, like the people are the fourth part of the quadruple aim. Like there is no organization without the people, like it's the if you're gonna do nothing else, like focus on a great work environment for people, everything else will take care of itself. Like, I like if you just wanted to have that as an organizing principle, like that would that is a recipe for a successful organization. So if that's true, then like other than kind of like asking or or or measuring, like in some ways, and maybe maybe I'm answering my own question, is like, how do we get at that part of it? Because I feel like if we just kind of like you know, cut ourselves off the knees and we're like, oh, well, people don't want to be asked about well-being, like, therefore, we can only ask, you know, once a year on a structured sort of thing. And then the counterveiling point though is like, well, all of our interventions need to have a well-being component to their assessment. It's like, well, what's the instrument then? Like, how do you like, you know, like I'm a simple person, like just tell me how we're gonna measure it, you know? And so it's like, I just kind of wonder about that. Is like, do we use indirect measurements like you know, uh pajama times one that is like sometimes cited for EMR things, or like, you know, how do we get it well-being like outside of these like very formal, it seems like almost too formal, too rigid, like annual surveys. Like, what are the who's doing it the best?

SPEAKER_01 45:24

Yeah, I think we need to look at additional ways to identify the impacts of whatever interventions we apply. And uh, you know, surveys are serve an important role for all for a number of different reasons. But I think to your point, uh one, we know that those that are most burnt out are like least likely to actually fill out a survey. Um and most likely, if they do fill it out, they're they're pretty far kind of down the path of burnout, and so there may be limited interventions that are available for that particular person. And so I think really focusing on what are those other ways that we can ascertain that information that uh not only informs what interventions we're going to apply, but what was the ROI on those interventions in terms of the well-being for that individual. And again, I think that the literature provides a number of different outcome measures that we can look at that give us that information. That again, large health care organizations uh oftentimes have that data and collect that data just as part of their normal daily functioning. And so I think those that's another area where organizations are moving their some of their well-being workers being able to identify what are some of those other markers that we can utilize to supplement or or perhaps um more accurately inform some of the work that we're doing.

SPEAKER_02 46:47

Yeah, I guess some of my own curiosities about how you would, you know, like stagger surveys throughout the year or have some kind of like, you know, in-person interview components to assessing people or sampling small samples of people that uh uh you know around the organization to see kind of how well-being is evolving over time. But uh I'd be I'd be so curious to know kind of what the future looks like when you're once once you're there, maybe Ryan, you can let me know um in terms of the uh in terms of how you uh you ultimately decide to measure that. I I am conscious of time since uh Ryan, I know yours is is uh limited. We're running

Future Practice Changes And Closing

SPEAKER_02 47:23

towards the end of your uh your time window here. But I I did want to just uh get your take, fast forwarding a few years, if this work has has really achieved you know kind of its its highest aspirations. You know, how how are things different maybe for for primary care and and urgent care clinicians at Providence?

SPEAKER_01 47:43

Hopefully it it has at least some positive impact. Again, I I don't know that it by itself would be the kind of one solution for all challenges, but I I would hope that it would be part of that solution that really brings them to a place that restores the joy in medicine and rejoice, you know, restores that joy in practice. And I think those are the parts that any type of well-being work, that's what we hope for, is that you know, even if in a small way it provides some positive impact. And so I think with this uh with this work I would say the same.

SPEAKER_02 48:18

Yep. If there was a place that you could lower friction, so uh an a next area of focus after the uh the consult workflow, is there a uh a next place you would choose to start lowering friction after this work is uh is complete?

SPEAKER_00 48:34

Yeah, I mean I I like I totally agree with what Ryan said, just in terms of uh you know, the the this shines a light on the importance of well-being, but I'd go even further and say that like you know, it's what's that saying like uh uh sometimes attributed to Berwick is like the future is here, it's just not evenly distributed. Like the best care in the world is somewhere, it's just not everywhere. And that's what I think the next the like the issues of the next you know two to five years is gonna be. Like information technologies are changing at such an incredibly fast pace. Um, I'm very thankful in retrospect for the fact that the pandemic happened, specifically with the focus that I think it had or the enlightenment that it had in terms of the importance of well-being. Like the well-being conversation, I think, from my limited perspectives and ICU doc specifically, like changed materially by virtue of going through this pandemic together. And the importance of like Ryan's activities and well-being, I don't, we're not gonna erase it. We're not gonna go back on that. And so as we think forward, like like I would predict that the practice patterns are going to radically change over the next, you know, two to five years, different places, different paces at different organizations, but well-being undoubtedly is gonna be part of the conversation because of the work of folks like Ryan and because of the fact of what we've all been through together. So that makes me actually very quite excited.

SPEAKER_02 49:44

That's a great place to round out the conversation um respecting uh time limits that we have today. And and uh I think there's so much more we could dive into here, honestly. I uh I always find I bite off more than I can chew with uh with these topics. But for listeners who would like to uh follow each of you and connect with you or or uh follow your work more closely, um, what's the best way to do it? Maybe Ryan, I'll start with you.

SPEAKER_01 50:05

I think the best way would be through the Wellbeing Trust website. Uh you're able to see my information there as well as a lot of the other great work that the Wellbeing Trust uh does for Providence.

SPEAKER_00 50:15

Yeah, and for me, you can uh find me at SnowCallMe on uh X or Twitter and then uh Arborgenie.com.

SPEAKER_02 50:22

Very good. Well, thanks so much again. Really appreciate your your time for the conversation. Um grateful for the rigor that you're bringing to this work, and I'm excited to see where MedPro goes and how you can uh improve clinician well-being. Uh, thanks so much, guys. Thanks, Jason. Thanks so much for listening to today's episode of Leading Quality. If you enjoyed the show, please take a moment to like, subscribe, and share it with someone who might find it useful. You can find all our episodes at leadingquality.budsprout.com or in your favorite podcast app. The show was written and hosted by me, Jason Meadows, edited by Milan Milostavievich, and produced by Thrive Healthcare Improvement. See you next time.