Ironically, patient care coordination itself is often uncoordinated. It's spread across multiple providers and administrators, causing frustration for care teams and patients.
Lumeon founder and CEO Robbie Hughes has developed an approach that makes the care journey more reliable rather than settling for the current craziness, where patients are expected to be satisfied as long as they get some time with a doctor.
One of the really interesting elements of this interview was the discussion of whether we should be willing to trade off some efficacy for reliability. For example, if we had a treatment path that was 80 percent as good but could be followed by 100 percent of the patients, would that be better than striving for the perfect solution that only 1 percent can attain?
From a population health perspective, the 80/100 calculus is probably correct, but most of us want to be in that 1 percent. I've certainly had that experience myself of navigating the system with help from well-placed insiders plus my own knowledge of how things work.
Robbie is an avid reader who recommends The Experience Machine, Chip Wars, and Ultra-Processed People.
Host David E. Williams is president of healthcare strategy consulting firm Health Business Group. Produced by Dafna Williams.
Ironically, patient care coordination itself is often uncoordinated. It's spread across multiple providers and administrators, causing frustration for care teams and patients.
Lumeon founder and CEO Robbie Hughes has developed an approach that makes the care journey more reliable rather than settling for the current craziness, where patients are expected to be satisfied as long as they get some time with a doctor.
One of the really interesting elements of this interview was the discussion of whether we should be willing to trade off some efficacy for reliability. For example, if we had a treatment path that was 80 percent as good but could be followed by 100 percent of the patients, would that be better than striving for the perfect solution that only 1 percent can attain?
From a population health perspective, the 80/100 calculus is probably correct, but most of us want to be in that 1 percent. I've certainly had that experience myself of navigating the system with help from well-placed insiders plus my own knowledge of how things work.
Robbie is an avid reader who recommends The Experience Machine, Chip Wars, and Ultra-Processed People.
Host David E. Williams is president of healthcare strategy consulting firm Health Business Group. Produced by Dafna Williams.
0:00:01 - David Williams
Care coordination is actually remarkably uncoordinated, scattered across multiple providers and administrative staffers. It's frustrating for care teams and patients. Can technology and automation streamline and unify the workflow? Today's guest says the answer is yes. Hi everyone, I'm David Williams, president of strategy consulting firm Health Business Group and host of the Health Biz Podcast, a weekly show where I interview top healthcare leaders about their lives and careers. My guest today is Robbie Hughes, founder and CEO of Lumeon, which brings predictability and consistency to care delivery by orchestrating teams and automating tasks. If you like the show, please subscribe and leave a review. Robbie Hughes, welcome to the Health Biz Podcast.
0:00:54 - Robbie Hughes
Hey, david, thanks for having me.
0:00:56 - David Williams
It's my pleasure Listen. We're going to talk about Lumeon, but I want to wind the clock back and talk about your childhood and what that was like, and you know any childhood influences that have stuck with you throughout your career oh so I, I had a very blessed childhood.
0:01:11 - Robbie Hughes
I, my dad was a concert pianist, my mom was a medic, a doctor. Um, we traveled a lot. I grew up in in the uk, I went to school in in france. I also went to school in germ, germany. I went to a French school. Like I, I had the most amazing childhood. Um, I think the the thing that stuck with me the most was when I was 11.
We moved to Germany and I didn't speak a word of German. Yeah, um, I had a choice between going to an English school which was full, an American school which would have set me back a year relative to the English education. So, so that was kind of out. We're going to this French school. And so I went to this French school and when I was there, I discovered pretty quickly that it was a very, very difficult school to get into, for very good reasons, and that they were all incredibly intelligent people and they all worked incredibly hard. And so I, I went in speaking minimal French, no German, and I had half my classes in French, half in German, and let's just say the experience was character building.
0:02:13 - David Williams
Yeah, yeah, yeah.
0:02:15 - Robbie Hughes
It, um. It taught me the power of persistence and I you know. I'll fully confess it was miserable at the time. But having kids now, I really want them to go through that kind of experience because it teaches you what's possible and what you can do when you put your mind to it, um.
But at the same time I don't want to just have to go through any of that just I wouldn't wish it on my worst enemy, but that as a as a kind of a formative device for me, yeah, um, it was really. It's one of those things where I still have. Even at at the age of 44 now, I still have very vivid memories of those years.
0:02:47 - David Williams
That's interesting I was going to ask you. You kind of answered my follow-up question within your response, which was you know, when you look back at the childhood, it was fantastic. Did you realize that at that time, or is it like how old did you have to be to have that reflection?
0:03:00 - Robbie Hughes
I don't know. I don't think as a child or kind of growing up, you fully appreciate anything, Actually. I don't even know it's limited to children, frankly.
0:03:07 - David Williams
I think as an adult.
0:03:08 - Robbie Hughes
you don't appreciate the value of an experience until you've been through it and learned the lessons from it. But I do think that every I think there's a lot of value in reflection. I think there is a huge amount of value in taking stock and trying to learn lessons and always being curious around what you do. But I, it is definitely, I'm definitely the product of my upbringing and my childhood, and the choices I make today are. You can kind of see, you can trace it all the way. Yeah, yeah, yeah, it's kind of interesting.
0:03:38 - David Williams
Do you play any musical instruments?
0:03:40 - Robbie Hughes
Do you know what? I have played many musical instruments I played piano, the clarinet, trombone. I used to sing for a long time. Turns out, having a concert pianist for a dad is not a conducive environment to learning the piano, however, that's my guess yeah, he's got the gift and I never had it, but I'm happy to say both my kids love music and they both play so that's good.
0:04:04 - David Williams
That's good. Well, it skipped a generation without being like completely destroyed in the middle. So I guess I you know. So I think I saw you studied aeronautics and was that because it was difficult and so you were used to the hard things after, uh, after high school or what was the what led to that?
0:04:15 - Robbie Hughes
um, I wanted to pick something interesting and I was actually going to go and study electrical engineering and, um, I just couldn't. It didn't really interest me and, frankly, you know, you're, you're a teenager, you're, you know 1920 and learning the science of how to make things go fast, that's cool, yeah, that's a really cool thing to do, and so, um, it also happened to be very difficult, which again, was probably one of my many mistakes I've made in life, but it was, it was cool, it was interesting, um, it was a great university, a great experience and, again, like I said, it got to, I got to go off and study abroad for a while, um as well, which is very formative.
0:04:57 - David Williams
That sounds good. What was abroad to you?
0:05:01 - Robbie Hughes
So this was back to France. So I studied in Toulouse, where Airbus was based. Got it, so I went to a school there and again made a great many friends, some of which I'm still in touch with today.
0:05:12 - David Williams
Very nice. Okay, so after school I saw a couple of things Cambridge Consultants, dynamic Systems what were those about? Why'd you go there?
0:05:22 - Robbie Hughes
So Cambridge Consultants was an engineering design consultancy based owned by Arthur D Little, and so what's fascinating about that company is if a customer of theirs or a company has a hard design problem that they want to try to address, they'll send it to CCL Cambridge Consultants and those are the guys that figure it out consultants and those are the guys that figure it out. So I worked on everything from, like, a new kind of shower head to the first video phone, to um inhalers, um, and it was just amazing like getting to solve problems, industrial design problems, uh, putting these things together and then um again, because I was, because I was young, I got put on lots of different projects and so I got a lot of exposure to different things, but really smart people solving fascinating problems, and I would have loved to have stayed, but it was based in Cambridge, which is where, again, I spent a lot of my childhood and I didn't really want to stay there, so moved to London, so, before you go there, so moved to, moved to london and so before you go there.
0:06:28 - David Williams
So I have to ask you so if you're working on all these kind of weird contraptions and stuff during the day, did you have any strange dreams at night?
0:06:35 - Robbie Hughes
um, not that I can recall, david, but if any spring to mind I'll let you know to me I went.
0:06:41 - David Williams
Sometimes, when I'm working on a bunch of intense things, some kind of a crazy thing may come into my head overnight, but maybe not for you, or maybe it was so dumb that yeah.
0:06:51 - Robbie Hughes
No, it's kind of funny. It's interesting what different skills you develop. So one of the things I was always very good at was computers. I grew up in a time where I had the first home computers, had a Mac, learned how to program it, and I'd always realized that being able to use computers gave me an advantage, and never was that more pronounced than at CCL.
And because I got really because I was good with computers, I got really good at CAD and 3D modeling, which then meant I got put on all these projects as a very cheap resource to do, to do a lot of the a lot of the design work.
I was, frankly, terrible at the maths and the hardest stuff that was proper engineers are doing, but I could do the CAD pretty well and, and again, that that was kind of instructive and interesting because and interesting because a lot of that thinking then got turned into what I subsequently did afterwards, which I think you were getting to Exactly the dynamic systems thing. Basically, the idea there was that I'd observed that there were companies and, frankly, industries that had sort of computerized and there are companies and industries that are digitized, and what I mean by that is there were industries where the product itself was a defined, specific, knowable thing, and in the creation of these products products there is a strong incentive to reduce cost, reduce variability, produce to a higher standard, potentially a higher volume, because that all drops to the bottom line in terms of improved margin and those industries have digitized really really well.
Other industries healthcare one, the legal profession being another were more time and materials kind of industries where you are essentially paying for time, and so what they've done is they put in computers, but the computers were effectively replicating paper processes, so you got all of the disadvantages of paper plus the disadvantages of a computer that would crash half the time and none of the advantages that you'd seen in in that digitization, uh, in those digitization processes. And so I I spent a bit of time doing that and kind of working with a few clients, and I landed then from that into healthcare, seeing through a client that I worked with, that they have really computerized their processes and they hadn't thought about how to drive efficiency, quality et cetera at the kind of level that I was seeing elsewhere. And so that's what was the genesis of the formation of Lumeon. It was essentially born from the problem of this clinic that had multiple different sites. They had a product that they were delivering for a fixed price to their clients which was interesting and different, but because they didn't have the operational tooling to make sure that that service they were delivering was predictable and free of variation, they they were suffering from a margin and um and an efficiency problem.
And so I went in and I looked at what they were doing and I said, well, really, what you're, what you've got is you've got very high quality clinical decisioning, but you've got a gap in terms of how that's then realized from an execution perspective. And so, unless you can couple the clinical decisioning and the operational execution together, you're going to continue to have this variation and that's going to be perceived by your patients and your customers as a quality problem or a patient experience problem, et cetera. And so that was kind of the founding thesis and you know, all this time later it remains an unsolved big problem. Got it All?
0:10:50 - David Williams
right. So, Lumeon, so first of all, you decided you're going to found a company, which is a big decision. I would say, yeah, maybe I assume you realized that at a time, but you know why did you need? To go and start a company. It's like there wasn't another job, so you'd just make your own job.
0:11:13 - Robbie Hughes
I will confess to being as arrogant and naive as to have more or less exactly that thought. I mean, I it. It didn't seem very complicated at the time. I kind of set my mind on saying, right, this is something I want to do and that someone will pay me some money for solving it and that I can see other people who have the same problem, so this seems like something I should do.
I mean, that was kind of the sum total of of it. Now I will say that's an incredibly stupid thing to have done and an incredibly stupid way to think about it, like in retrospect. And again, this is a really interesting difference, I think, between british entrepreneurs and american entrepreneurs. British entrepreneurs go after the problem. They see something interesting, they say right, I can solve that, I've noodled that out. Whereas american entrepreneurs again, this is a horrible generalization, but stick with it for the moment are much more disciplined in understanding the market, understanding what the market needs in terms of proposition, how that can be segmented and defined into a category and so on.
And that kind of science of marketing is much more well established in the US than it is in Europe, or I should say in England, and I think it's just because it's a bigger market. If you're not across those things and doing them well, you're not going to do very well. Yeah, and I think if you look at British companies that scale internationally, they have to take on a lot of that science of marketing in a way that you just don't need to in many cases, in the UK at least, depending on the market you're going after, etc. So yeah, it was as simple as this is an interesting problem. I think I can solve it. I'm going to have a go at it and I gave myself a year to see if I could do that and I did, and I got some customers and we kind of went from there.
0:13:01 - David Williams
Yeah, so it sounds like you had at least some concept of a customer that might need it. So who was the first customer or what was the first kind of approach that you did? Did you have a customer in mind that had this need? Did you seen it?
0:13:12 - Robbie Hughes
Yeah, so it, like I said, it had been sort of presented on a plate to me by these clinics and by speaking to the practice manager at those clinics I then met a bunch of other practices and they had the same issue and it's kind of interesting.
I looked at practice management software and I looked at EMRs and the practice management software was it was all kind of it was always specialty specific, so you could get a piece of practice management software to do one thing for one clinic, for one specialty.
But actually if you pull these things apart they're all functionally quite similar. So the first thing I did was like OK, I can create a general purpose tool to solve this problem. And then the second thing was when I was looking at the EMRs, well, they were all kind of more general purpose but they had really been designed from the idea of documenting and billing what needed to happen. They weren't really designed from a process efficiency point of view. So they were taking two different approaches on the same problem and I realized that if you could build a tool that could kind of sit above it and coordinate the things that need to happen around the right context and the right clinical information, then you could have this sort of go-to-market approach that would enhance the EMRs without competing with them and drive a huge amount of efficiency for those people that needed it. So that was kind of how we thought about it, Got it.
0:14:43 - David Williams
So I see language on your website like proactively and continuously coordinating patients and the care team. Break that down. What does that actually mean so?
0:14:53 - Robbie Hughes
it's kind of funny like there's so many weasel words when you look at how things are presented on the web, it's very difficult. The way I tell this to people is if you think about the flight path of a plane, what Lumeon's doing is it's calculating the flight path of a patient based critically on a subset of the data. It doesn't need all of the data to do that. It can do that reliably. So for any given patient it can work it out time and time again. It's not guessing. It knows deterministically what to do and it can do that instantly. And then, based on that calculation, it can then inform the care team what they need to do, communicate that out either into the EMR or out to the patient and then ultimately, if the thing can be automated, it will then automate it. And the point of all of that is it introduces this horribly boring concept called reliability, because we know then predictably for a given patient in a given presentation and a state, this is what's going to happen and we can make sure that happens every time. And the hard part about that is doing it a in real time, b with a subset of data, because none of the data no, emr has all the data and then the final bit is personalizing it to the patient, because you're not interested in factory medicine. You've got to make sure that every patient gets the right thing for them.
So that's very precisely what it's doing and it works incredibly well. It's almost embarrassing how well it works in terms of the efficiencies it can bring, but it it just speaks to the size of the opportunity um around, how much um kind of fragmentation and discontinuity there is in care delivery, and I mean we've all experienced it. We've all had had someone in our lives who's maybe been discharged too early, or or they they've had their visit moved, or you know the wrong thing happened, or God forbid you know the wrong surgery took place, or something like that. These are really basic communication and process failures that frankly don't need to happen, but they're a symptom of the fact that we don't think about care delivery as a continuous process, but a series of small, disjointed things that aren't connected. So that's what we try to resolve.
0:17:08 - David Williams
The airplane example is interesting because you hear in health care sometimes people will point to airlines from a safety standpoint and they'll say you know, go through a checklist and that kind of thing.
A you know, go through a checklist and that kind of thing. But what? What you also see that we have just described here is that the plane typically, you know, takes off and makes it pretty smoothly along its path and then it lands and a lot of that is done, you know, with an autopilot and it's done without having all information at all times and it still goes pretty smoothly. And, you know, sometimes there's a little turbulence and so on, but you expect you're going to get to your destination, whereas in you know, on I guess the equivalent would be if you think about your care journey, you know it's pretty likely that there's no pilot or co-pilot or navigator around and if you're lucky there might be a flight attendant that you could flag down from time to time. But if you think about it as a journey that you're really supposed to get on and have a soft and smooth landing, I think that's great.
0:18:02 - Robbie Hughes
Yeah, and I think you know, when people think about checklists in particular, they tend to focus around the problem that, okay, you can use checklists in airplanes, because it's always the same thing.
0:18:14 - David Williams
Yeah.
0:18:15 - Robbie Hughes
And therefore, that's easy, you're always checking for the same things, you know, that works fine. And the argument in healthcare would be well, that's easy, you're always checking for the same things, you know that works fine. And the argument in healthcare would be well, every patient's different. And my counter to that would be well, yes, every patient is different, but the way in which we look at them and the way in which we diagnose, manage them should not be different. For two identical patients with an identical presentation, we should get the same outcome, but the fact that we don't is a symptom of failure to understand the need for consistency of decisioning that should result in personalization of care. The job is not to do the same thing for every patient. The job is to have consistent decisions that result in that personalization.
0:18:57 - David Williams
Okay, and so on the personalization side, what, what, what would be an example of personalization?
0:19:03 - Robbie Hughes
Really simple one. Um, a patient's coming in for surgery. We know cause. We've built an order, set that when a patient comes in we order this battery of tests, 20 tests, whatever it might be. Um, that's the thing that is done, kind of as normal. And again, kind of from an industrial perspective, it makes sense. We're going to do the same thing consistently for every patient.
But maybe the patient's a primary care patient in our network, maybe they've had all these tests before, maybe the surgery they're coming in for doesn't need all those tests. So maybe we can look at the results we got on hand, see that the patient's low risk, see we don't need any other tests from that, from that procedure. Therefore we don't need to order those tests. And that's an example of very simple personalization.
But the cool bit comes from the fact that if you don't need to order those tests or maybe the rest of what you could do, you could do over the phone or you could do it remotely you don't need the visit in the first place. So suddenly you look at these things and they start compounding. You're saying, well, if that doesn't need to happen, then this doesn't need to happen, and so on and so on and so on, and you do all that and you're releasing huge amounts of capacity, all because your previous process was centered around the fact that you had this anchor thing that needed to take place, and there you, therefore you designed around it, but if you remove that constraint, it frees up all the other things you could do. So that's a simple example, but I mean, there's millions of them.
0:20:32 - David Williams
Okay, yeah, well, not everyone would think that's so simple, but it's simple in principle. So you know, one thing I see with care coordination sometimes is a focus on automation, you know, because there's such staff shortages and so on. Theoretically it should achieve some of what you just described there, right? If it's automated, then the nurse doesn't have to do this and they can move on to something else and there's more capacity, et cetera. In practice, what I have often seen is that the automation without personalization actually might create more work and creates problems. So I'll give you an example.
So I know someone that was discharged. They had a hip replacement, and then they would get a series of notices from the physician's office about this medication or that or whatever, but it wasn't really personalized to them, and so they got something. That was something that I don't even know. Maybe it's true for 80% of people, let's just say and they would get it. And then you know the spouse is saying the doctor changed your prescription, and then the patient is all upset about it and that if they actually do that, they're actually going to have damage and be back in the hospital. Never mind, they need to get on the phone, and so how do you get? Can you practically get beyond that what you're doing?
0:21:46 - Robbie Hughes
get? Can you practically get beyond that and what you're doing? Yeah, absolutely, and I would say this is one of the single biggest points of why. But, people, you're presumably familiar with the term pathways. Um, pathways are a vitally important tool in delivering consistent, high quality care. But if poorly applied in exactly the way you've just described, what it results in is this standardization and then people falling off the pathway, and then you get this rework issue. In fact, there's a lot of this. Stuff works really well in either a pilot or a clinical trial, but you put it into yeah, into the, into the wild and the whole thing falls apart because, guess what, people are different and processes are unreliable and things don't go to plan. And da, da, da. The whole point about what we've built is that, because of the fact that we can do this in real time and because of the fact that we can do this with only a limited amount of data, it means that at any, if the data comes in saying that something else needs to happen, we can use that instantly to redo or to redirect that patient into the right activity. And once you've got that personalization, then the thing starts to deliver the benefits. Kind of an interesting digression.
But in the context of AI, like every company under the sun right now, we're spending a lot of time looking at AI and generative AI and all the various different use cases. But actually it's harder than you might think. So I'll give you an example. We one of the things we're looking at is ambient AI and one of the use cases for that is around rounding. So, for example, it makes all the sense in the world that you might have a facility to record the round and have the round then documented in the system, so you don't have to do that with a human. That makes all the sense in the world.
Problem is that let's kind of work that through in pieces. So the first part is let's record the round. Okay, assuming it's 20 issues, we can do that Now. Let's transcribe it Now. I don't know transcription Now, looking at a fairly long transcription, is going to take a few minutes, so it can't be done in the round, it has to be done after the round. So now what's happened is we've said, okay, before we were documenting in the round, but now we're going to not document, but someone's gonna have to do that after the round, but they're gonna do that with out being in the round, so if there's any ambiguity they don't get an opportunity to correct it. So we've introduced an extra work step.
Now let's take it further. We've gone into this transcription let's assume that it's okay. Now we're going to take that and probabilistically turn that into actions, orders, activities et cetera, off the backend. Well, that's not going to be a hundred percent reliable either, because we don't know that the transcription is reliable necessarily. So someone's gonna have to check that too. So we've gone from a place where we had a process that wasn't maybe perfect but it was actually pretty efficient, to now creating new processes that are less efficient and much potentially more error prone. So you've got a low probability, high impact event error prone. So you've got a low probability, high impact event, whereas before you had a fairly reliable process. So that's not a great way to think about automation.
And again, unfortunately, a lot of what we see is that kind of activity you try to solve one thing and then you cause another problem later and dialing this all the way back.
The reason this stuff works or it doesn't comes down to trust. If you can trust that the thing is going to work perfectly 100% of the time, then you don't need that safety netting, those error correction or validation steps and the thing will deliver the benefit. But as soon as you introduce the possibility that it might not do exactly what you expect, that's when trust falls over, that's when you introduce more work, that's when you get rework and again your business case falls apart. So long a complicated way of saying it. The example you gave is very typical. There's a lot of there's a lot of reasons why that kind of stuff happens and will continue to happen. The approach we've taken is very specifically designed to prevent that and it doesn't rely on kind of these probabilistic models that will say we're mostly certain that's the right thing. What we're interested in doing is saying, 100% of the time for a patient like this, this is what you do and automate it. That creates trust, it creates reliability, it creates efficiency. That's it.
0:26:27 - David Williams
So there's all sorts of differences in how health systems in different parts of the world especially, let's say, us versus everywhere else, let's say just, let's say, us versus UK Very different, right? There's a very different ratio of specialists to primary care, the funding is different, the cost level is different, the culture is different, the facilities are different. The funding is different, the cost level is different, the culture is different, the facilities are different. And yet there are quite a few companies the ones I'm familiar with, more sort of US based that also work in the UK, and you seem to work in the US and the UK as well. So how do you explain if it's the case that the systems are very different and yet you've got a value proposition that works in both markets?
0:27:03 - Robbie Hughes
you've got a value proposition that works in both markets. So I would suggest that there is more variation in practice, in process not necessarily clinical practice, but in process within different US health systems than there is necessarily between the US generically and the UK. So if you take a fee-for-service place versus an ACO, for example, like the way that those two are run, assuming some level of maturity will be completely different and the UK is different Culturally. You've named a bunch of different reasons or a bunch of different things that you'd see it is very different, but the core problems are the same. Patients can still get discharged without their labs being read. Patients can still get cancelled on at the last minute in surgery because of a workflow screw-up. These things happen. Humans are humans and the problems we're interested in addressing are fundamentally human and communication issues. They are humans and the problems we're interested in addressing are fundamentally human and communication issues. They are not ones necessarily related to a funding model, although you know that does create interesting kind of prior auth problems, which is a topic for another day.
0:28:18 - David Williams
Yeah Well, we'll cover that with some other guests, should they be so inclined. Good Well, robbie, I want to ask you a last question, which is about any books that you've read. I ask every guest if they've read any good books lately. I don't always ask them what language it's been in, but have you read anything lately that you would recommend, or anything you would recommend to stay away from?
0:28:42 - Robbie Hughes
Oh, so I read quite a lot. I spent a lot of time on planes. Um the last four books I've been reading have been the experience machine um by andy clark. Uh, guns, germs and steel yeah um fascinating book. Chip wars, ultra processed people, um I mean, I also read a bunch of spy stuff because it it entertains me, I read. I I'd recommend all those first four.
0:29:13 - David Williams
I'd certainly recommend um they're fantastic books great, okay, well, I won't push you on anything that you would, uh, that you would not recommend, but sometimes people warn us off of those too. But but no, that's all sound good, all right, well, robbie Hughes, founder and CEO of.
0:29:30 - Robbie Hughes
Lumeon, thank you for joining me today on the Health Biz Podcast. Thank you for having me, david, it's been a pleasure.
0:29:33 - David Williams
You've been listening to the Health Biz Podcast with me, david Williams, president of Health Business Group. I conduct in-depth interviews with leaders in healthcare, business and policy. If you like what you hear, go ahead and subscribe on your favorite service. While you're at it, go ahead and subscribe on your second and third favorite services as well. There's more good stuff to come and you won't want to miss an episode. If your organization is seeking strategy consulting services in healthcare, check out our website, healthbusinessgroupcom.