#438: Does design matter more than tech in healthcare?

0:02Design is a mix of everything. Design is how it looks, how it works, how you feel welcoming that thing into your life. But it's also the foundations that went into its operation and its creation. Design thinking is essentially just how does this thing operate in the world in a way that people actually want to use and how do we apply that to business strategy and to economics and loads of other things. Human-centered design is a phrase that came out probably from 10 years ago for at least six years but what's the alternative what's the counter argument to human-centered design hey buddy delighted to be joined by ian wharton co-founder of aid health ian how are you doing sir very well yeah thanks for having me on pleasure we're big fans we've been listening for a long time so yeah really happy to be here appreciate you good to have you on because many different reasons aid health's doing some really cool stuff um you had your recent white paper which i want to talk to you about how to build trust uh in ai which uh that is not going to be any smaller of a topic for quite a while i think that's going to be a work in progress let's just say um you've launched mirror which is your patient facing scribe which is an interesting take on the whole scribe world um so yeah loads for us to talk about man um but yeah first of all a bit about yourself um whereabouts whereabouts are you speaking to us from where are you based i'm south oxford yeah so we we classic post-covid cliche we sold the london flat and bought a house in the countryside and we're very very happy here yeah so um oxford your countryside oh glorious do you have a dog we don't no but it was it coincided with a toddler yes oh perfect yeah yeah yeah that's why yeah we uh we did the same but we went out southwest so we we we went out towards sorry so same manoeuvre yeah those are the options for us yeah we wait like everyone yeah well literally literally come on so let's let's talk about you so i think this is really interesting because you obviously you found us you have a health tech company but you're from the more creative side of the world the artistic side of the world and we've had a few artists on here before and and creatives that have become founders and it's always fascinating to me the difference between art and science in terms of its root to founding a company and how it relates to entrepreneurship because science is so often the hard hypothesis and i need to know the problem first of all and then we're going to specifically design a solution to the problem and test it and this scientific method i guess you would call it whereas the creatives and the artists are very much like oh it's a cool piece of tech let's have a play with it for a bit and let's just play around and let's just do this and i'm being very i guess i'm negging that the artist here it's actually a lot a lot better than that however there is a lot more just experimenting from the artists that i've spoken to and the exploration of technologies and their uses to eventually arrive at something that then oh hold on a minute if it does this it could do this and it's just a really interesting difference I think between between art and science in that regard and in entrepreneurship but you're obviously from that side of the world so I'm interesting if first of all if that's your experience but yeah tell us the whole journey man where does it all start for you yeah so what you've explained is is true and there are traps within that actually so um very happy to talk about so so you're right principally I'm a designer I'm a classically trained designer through A-level and through university.

3:41But I've been doing this kind of professionally since I was 16. So, yeah, 24 years in some form. Both in startups, I've either founded or joined very early on, or working in design and innovation firms. All my clients in those firms were mostly kind of Fortune 500 companies, and we'd be designing the digital products and services for them. Quite often, millions of people would use those every day. So yeah, I've done that. Health tech was always on the cards for me from about 15 years ago. But to your point, you're right. There are very distinct differences here. So I'm not a clinician. We came at health tech from a different angle, which was one of frustration from lived experience.

4:27And necessity is almost always the underlying reason people start businesses. in some form that they witness some epic unmet need in the market or they feel slighted by something that wasn't a right solution for them so they go and try and fix it. Very, very common. But you're right in the differences. So from a design world, what we get is understand the market, understand the customer and the user, understand the reasons why people might stick with it or not, what are the barriers and burdens to choosing to adopt that service all that technology in your life and also how to communicate it so that's really critical so uh good i'm very fortunate with the people with the teachers i've had in and the mentors and colleagues i've worked with some very very gifted and capable people but if you get the right blend of all of that what you're taught is communicate the thing in the simplest possible terms so as many people as possible can understand it i it's probably fair to say in the academic world that isn't a core taught skill.

5:29But there is a difference there. However, you're right on the experimentation thing, but there is a trap which is kind of laid for us, a lot of us fall into, which is here's a piece of technology, let's just play with it and see what happens. And almost always that is an issue, which is you shouldn't just apply technology because it's within reach, because it has this idea of the silver bullet. And we've seen this over and over again. We saw it with the internet. We saw it with e-commerce. we saw it with mobile products we saw it now with ai and just it goes on and on and it will carry on doing that but that is the trap you just adopt something because it's within reach and that's what you need to be really really careful of do you know what came to mind there was the metaverse because i haven't facebook now just canned that or meta i've just i've just i've just canned it because it was like oh here's vr vr plus vr world vr world plus vr conference room and VR meetings and then they and then like oh no actually no one wants any of this was it 30 billion down the drain whatever it was so right so so what's the conversation around that is the conversation is should they have not tried probably not because it's they things should be experimented with and we should understand what are the limits of how we interact and want to interact and all the things that are buried within it but then there was a question of surely there was a better way to validate that idea than spending the tens of billions of dollars that they did spend.

6:51And again, what's underlining that? I don't know, hubris of a company that has unlimited resources, ego of a founder that's rarely made a, like from a business perspective, rarely made a wrong move, in a social perspective made disastrous moves. But from an economic perspective, his shareholders are the happiest shareholders in the world. So it's a weird thing, but that was an unmitigated disaster. And I think a lot of people saw the writing on the wall, which is, I don't want to be that disconnected from reality, from people sat next to me. I don't wear a fucking headset in my living room, sat next to my wife and my toddler.

7:24Like who? So, but again, you talk about iterations of innovation. Where does that get you down the line? So it gets you in something that like we're wearing, which are very similar glasses. And it's much more discreet that maybe does fit into our cultural norms. And maybe there is an exception of behavior that does become the norm over time. And we wouldn't have got there without the 30, 40 billion, whatever the hell it was with these giant things that take you out of the out of the world who knows but um strange time and i for one was pretty happy to see that pass quite quickly yeah same the the the the ut i put it into a heart of vision this week this week but like it feels like every single week i could have like a black mirror section of the newsletter with all the news stories that put us closer and closer to this utopia dystopia coin toss and it it genuinely that genuinely felt like one of them like hold on a moment if this does get mass adopted what do we then become what sort of species do we then become this escape to a different universe and oh yeah very very bizarre tell you what though one thing that's actually coming in the next two hours are my ray-ban meta glasses now there you go what's interesting is that i've taken very much the artistic route to going i think i just need the glasses in my hand to see what i can do with them in 30 days to put it into my life in a useful way or else they will be getting returned because i try to do the scientific method of what problem i am i actually trying to solve here and I could not answer that question now I think what I've actually done is now just found an excuse in order to still order them which is I'm going to experiment I'm going to see what I can do in order to make this useful in my life but I do fear that I'm not going to be able to do that and for the price tag of 400 quid, which yes, I will be using them in work, therefore they will be purchased for work, et cetera.

9:44I still don't think I'm going to be able to use them and I still think they're going to go back. Yeah, so what you just described probably still places you in the early adopter category, which is I'm not entirely sure. I'm excited enough to see if this has a role in my life or not, but also pretty willing to discard it if it doesn't. So that is early adopter behavior. The vast majority of people are not that. The vast majority of people are, I need a bit more proof. I need a bit more signal because me in making purchasing decisions as an individual is not just is that cool thing. It's related to timing, competition.

10:20What are the products are there? What else can I do with that 400 pounds? What's the opportunity cost of going through the motions of buying it and sending it back and all that nonsense? And so as an individual, that's probably a good place to be because you seem pretty detached from it as an idea and as a choice and a decision. Harder for institutions and organizations to be in that mode, which is a good reflection on health tech and healthcare writ large. But I'm interested to hear. Yeah. But again, the form factor there has played a big part in your decision. Design has played a big part in that decision.

10:55Not necessarily the underlying tech. you could argue it does in the sense it's been minimized enough to fit in something stylish but design has influenced you there not the technology explain that a bit more for me you've made a bet on the form factor of those ray bands right form factor what do you mean yeah so the design the look the size okay yeah weight all of those the material science that's gone into it i went for that you haven't made a better the headliner versions rather than the wayfarers or the others yeah because they're more there you go so what what you've done is closer to buying these one of these oliver people's glasses which is a decision process on what did these things say about me as from a design perspective do they say i appreciate design do they do they say do they communicate outwardly i i'm as a social animal that i take care of my aesthetics and my style and i'm i'm disciplined enough to be well dressed and all those things what you haven't really made a decision on is how capable that technology is and how it adds and how it which is again the brilliance of people like mark zuckerberg whatever you think of him i don't love him uh far from it in fact but the brilliance of his decision there is what do i do i partner with the best glasses brand in the world that carries a load of social value and i put my technology into that to validate whether or not this technology is really worth it super smart it's such a good point because even as i'm thinking now like i'm not sure that i would purchase meta glasses you would i'd purchase ray-ban meta glasses and actually the route in because i'm a because i was gonna say i'm a snowboarder i'm not a snowboarder i have snowboarded um i was actually i actually my routine was through oakley because my snowboard gear is oakley and so when when when i saw there was oakley meta glasses i was like oh interesting and then i went in through there and then i was like oh actually no i'd prefer these ones because they look more like my current glasses and like yeah it's really interesting so this is so this is this is such a cool insight i love i love these conversations because what is probably difficult for you to codify exactly is how you've arrived at that decision and that decision is based on design it's on sensibilities of design whether or not you've studied it or have the terminology doesn't really matter and this stuff matters so as a we are a design-led health technology company which is reasonably rare so we're a product and design-led health technology company what does that mean it means we do things a certain way we build our products a certain way we conduct ourselves a certain way that a non-design-led company might not because we believe that these things really matter and they matter in our sector because digital health has loads of power and promise the potential digital health is uh one to combat the the workforce ceiling that we have in health care and a million other things but people have to stick with these products yes and and the the unspoken reality of most digital health products and what we're really talking about here is mobile apps in in large part mobile apps the unspoken reality is that the vast majority of people just don't stick with them and if you depending on what evidence you look at but it's somewhere between three and a half to four percent of people still use a given thing after 30 days and you go okay well if we're talking about things like chronic disease well that's nowhere near good enough we're never going to get any outcome and i would argue one of the reasons why we haven't cracked that is because we just design hasn't been applied to health really until the last year or two when you talk about design there i think you you definitely know what what you mean by that but i think people listening will have varying definitions of design because some people think of graphic design so people think of like digital marketing design others think of product design others will think of the the biggest and widest you know version of design so there's design thinking there's design led this that and the other there's design system so the word the word's interesting so when you say you you know you design first design led company and and these terms what are you actually talking about there so in a sense it's kind of a mix of everything you've just said now the challenge is consultants over the past 30 40 years done a really good job of stripping everything down to component parts so they can sell it effectively to their clients so design thinking was a really good example of that design thinking is essentially just how how does this thing operate in the world in a way that people actually want to use and yeah and how do we apply that to business strategy and to economics and loads of other things but essentially it's like how do we solve a problem that actually works and it's there's other things like human-centered design is a phrase that came out very very popular probably from 10 years ago for at least six years and we like myself and my colleagues we always have the question of well but what's the alternative what's the counter argument to human-centered design what animal-centered exactly so like technology we'd be doing so design is a mix of everything design is how it looks how it works how you feel welcoming that thing into your life but it's also uh the the foundations that went into its operation and its creation so i'll give you an example we every time we build a pathway in aid so the chronic our aid is our chronic disease management platform and we support six conditions today we're just about to add three of the uh common hematological cancers so myeloma and two of the chronic leukemias and but every time we build a pathway we establish a patient panel and almost always it's through an advocacy group in this case it was blood cancer uk an amazing group of patients uh we've worked with over the course of the past kind of six months or so and what it means is we run hundreds of hours of workshops with patients to understand how they live their lives what what is the barrier if if you could wave a magic wand what would what happen if you could design a particular service to solve one ungodly challenge in your life what is it all those things and and then we put our initial thinking in front of them and get feedback then we build the prototype put that in front of them get feedback then we give them the real thing get feedback then and that just goes on and on that is design in in a in a very clear sense that is design nothing necessarily to do with the aesthetics the visuals of it but that being said you shouldn't limit or minimize those the how the thing looks is a massive influence on whether or not you keep that thing in your life or not whether or not you studied design and i can say that because some of the most ubiquitous companies sorry comes with those ubiquitous products rather apple really good example they they have been educating the world on what good design looks like since the 19 early 1990s and people just instinctively know and we get this all the time so we could be speaking with a potential customer in the nhs or an investor and they'll say wow yours just looks different and they don't they don't have necessarily or need any stronger vocabulary than that but they know intrinsically instinctively they know and users and patients are exactly the same so in your opinion then you sort of open by saying that design hasn't really you know found its way into the normal healthcare practice or the way that we do things in healthcare why is that is that something to do with a perception of roi first of all because what you just talked about there was hundreds of hours of patient panels via advocacy groups and in this new world of ai talking about hundreds of hours spent with human beings you're sort of making investors cringe at this point and going like hold on a minute can you not just blah and headcount is now perceived as not a flex it's like a it's like a oh you've got messy complicated humans with all their emotions and all sorts of stuff so there's design and the way that you're talking about design and the processes involved um i mean particularly this because you talked about the interface of human beings but it wouldn't necessarily need to be that but is it is it perceived as not not the best roi way of doing things so there's a couple of things overlapping there so there um there's a reason why i think there's some clear reasons why design hasn't hasn't made it into health as quickly as it did in e-commerce or or financial services or entertainment or news we're the last sector i think to get design a couple of reasons one is the um it is i think fair to say it's unattractive because of the risk and regulation so designers are not used to being in this world we're used to uh we're used so designers and i'm speaking mostly from like a design-led founder as opposed to hiring a design team as like a major health system in the u.s or something but design founders or or people who come from that world we're not used to academic measurement we're used to subjective measurement we're used to making a thing putting it out in the world and like kind of attaching some of our identity to it which is which is by the way the worst feeling because it's never it's all it never you never get it right all the time there are times where it just fails miserably and you go oh my god this is now an affront on my dna and my personality and it's really hard to detach from that so we're not used to the academic the rigor the science the evaluation the evidence i think that's one part i think um it's slow as a market if brian my co-founder and i wanted to build a startup just to make some cash we wouldn't we wouldn't have sorry to make some cash quickly we wouldn't have made we wouldn't have chosen health tech um and probably just come back to roi it's hard in any circumstance it's probably elevated in health it's hard to explain the value of design but but we can you can get there so i'll give you really stupid really simple example and this kind of overlaps design and kind of behavioral science so uh there's loads of studies done that if you give someone a course of antibiotics there's a there's a good proportion of people to never never fill a prescription right never never complete it they'll take if it's a seven-day course they'll do three days or they might do seven but kind of miss doses and it's yeah so not guilty yeah right like like 50 percent of the population statistically so half the time people aren't adhering to the prescribed care they've been given lots of different reasons now there's loads of studies shown simple behavioral intervention mostly governed by design which is if it's a seven-day course you go okay well take the white pills for the first six days and then take the red pills on the sixth on the seventh day but always end with the red pills now chemically there's no difference in the molecules in those pills, but there is a design process that's gone into that, which is, oh, I take the white ones, I take the red ones.

22:23If I haven't taken the red ones, I haven't finished it. And you see the uptick. You see the uptick and completion and adherence. Right. So really straightforward, really simple. But then, okay, so bring that back to what we've built. Where does design play a role here? Okay, so adoption. I mentioned the terrible stats on retention. Our retention rates in aid are about 20 times the sector average. So after three months, about 70 % of people are still using our platform more or less every day. And in fact, we've had people use aid more or less every day for over two years. And some of these people have conditions like heart failure and are quite dependent on it.

22:58So how do we get there? Well, the method of interaction, we use conversational AI, which we can talk about, and we made that bet five years ago. That's definitely part of it. But there are small, tiny elements of design that I think add up to this. I'll give you one example. our brand, the visual presentation of aid in both product and brand is white and blue. And it's a very deliberate choice. It's very deliberate, not white and green or white and purple or white and red. Blue reflects health. Every healthcare indication signal that most people have, not just in this country, is cleanliness from white and blue from calm.

23:36And those colors present physical reactions in most people. red is kind of angry or fast moving green ecological you get the idea so we chose blue because it's a clinical product and patients need to feel like this is an extension of the healthcare system if it was green it wouldn't so there are thousands of these decisions we've made over the past five years but they all add up to our goal people stick with it for long enough to take a therapeutic effect. I definitely think the blue, I mean, the blue thing is funny. It's almost like parodied in our side of the world in like comms and media and all that sorts of stuff just because there's sheer volume of companies that use it.

24:19But there's clear reason for it. I definitely think it's more on that, you know, it looks like an extension of the NHS if you use the same hue as the NHS logo, more so than any color theory around calm and things like that and anger and all that sort of stuff. I know that there's many believers in color theory and I'm neither here nor there to be honest but yeah it's definitely that people just feel comforted by the fact that it's actually advice I gave I can remember like on the accelerators that I used to run like digital.london I can remember all those years ago just saying to someone that they should actually just copy like the font and the colors like pretty much exactly the same and I don't know if that was like legal advice at the time like that was okay but like for a small startup I was like look just make it completely nothing you feel like it's probably an NHS thing probably not the uh most ethical advice but no it's and these things these all these things matter so there's a good argument to be made that like you said most of the market in health is all kind of blue and white okay well what's a good pivot for that a good pivot is to be differentiated so what typically is a company from a design perspective what you what you want to typically avoid is fatigue you don't want to just feel like you're part of the ferment and like everyone else in order to succeed as a company from a brand perspective you typically need differentiation as one of the pillars so you could argue okay well we are going to be red against the blue just so we stand out and there is value in that economically commercially like visually but you have to weigh these things up with the downsides so you mentioned i saw you just you interviewed the co-founder of neco so neco does not need to be blue and white.

25:56They are producing a completely different product, different area of the market, different consumer set, different mindset behind health. What they've done, absolutely perfect for what they're building. So all of these things matter. But I guess the umbrella to everything we've just discussed is these things need consideration as a health business. And they're typically not there. And they're done as an afterthought. Or if the business manages to get big enough, they employ one of the one like a very very capable brand design company and then they get it get it kind of in line but it can stymie growth initially if these things aren't considered 100 and it's that attention to detail i think that matters um you're right at the early stages are because there's so many people that say oh there's nothing that just name it whatever like it's fine just name it whatever get it off the ground get a few customers sort all this stuff out later and it's just so the opposite the like i called our newsletter health tech pigeon knowing that most because i've had newsletters previously knowing that 50 percent of people sign up with gmail and hotmail addresses that are personal email addresses so i know that they are receiving this in their personal email if they're receiving it in their personal email and by the way best time sites start you know send a newsletter back then was 6 p.m on a Sunday it's remained that way even though that was you know six seven years ago but um if I'm sending a newsletter to someone's personal email at 6 p.m on a Sunday this thing needs to be read by them while they're in the mode of I'm at home and this is a homely this is an interesting thing for my life it's not a work thing so immediately i'm like well the tone needs to be this the tone needs to be humorous and amusing this needs to be the oh we're just you know we're we're at work on monday like where we've got to behave but like let's just let's just say some stupid stuff about the news and like like the banter basically and so all of that went into the the name health tech pigeon and And then the original like cartoon bird that I picked was just some bird off Canva before I got Fred, our creative director, to actually do the proper branding for it.

28:09But I just picked like the most expressionless, like hilarious image of a cartoon pigeon that I could possibly find. And it's probably nobody that remembers that even now. But it was just that eyes looking in the wrong direction and like all this sort of stuff. And it was just like, yeah, here's this stupid carrier pigeon that's brought the news to you on a Sunday and it's written like as if it's a voice note to a friend and and but but all of that was details and if if I'd not applied that level of detail I don't think I'd have had 200 subscribers on day one which is what I had and then from there it built right and it built and it became this personality underneath those original foundational decisions that I'd made it then became what it was and then when Fred got his hands on it he was like oh let's do let's do some like you know looney tunes color palette and let's lean into that side of things and like if we're going down the cartoon route it can be there so it just it was so enabling to make those decisions early on that i just i can understand people when they say name it whatever and you're gonna you're gonna rebrand the name like by the activity that you do whatever and it's like well but why not take advantage of what you can take advantage of right at the start when you have complete carte blanche to do whatever you want like surely you want to just optimize all your advantages at the beginning right like yeah just my version of what you've just said i guess these things matter and what you like these things matter what you're illustrating there is good instincts which is differentiation this needs to be unique when it hits people's inbox it needs to arrest their attention and that's why you don't want fatigue if it was called health tech newsletter sunday's health tech newsletter i don't know i don't know if you the success you'd have had.

29:51So these things are crucial, especially crucial in health technology, in the healthcare world writ large. And I think like to come back to where this began, there are more and more people who have design sensibilities or design capabilities or more companies with those things coming into health. I think that's a great thing. And we might get to where we need to be because of that. It's not zero sum. That doesn't undermine the rigor of science. doesn't undermine the rigor or need of evidence all those things are as important if not increasingly so but it's the unison of these things that's i think that's where we need to be it's really interesting and i do think people are coming around to this and actually when i when i actually think now like on a serious note when i look at somx and i look at the some of the strategic projects that come in through somx now i'm not joking like we're almost in double figures of people that have like wanted naming the company naming the product to be a specific work stream and you know we've identified 12 different things that contribute to the naming of a company and actually we have all those conversations about those 12 different elements and all those feed into you know multiple so it is it we can turn the art into science in a way and actually like it becomes incredibly important um before we disappear off into this which i could easily do for for another three hours um let's talk about aid so you've you've talked about it a little bit building pathways um and blah blah blah so let's just do that properly so aid then how how did the idea for aid come about talk to me about the the founding of it the founding principles of what you want to do with the company and you know a few examples maybe end to end of you applying everything that we've talked about into something patient facing aid health the company aid is our flagship products so started the business five years ago in january yeah so we've been we've been going a a while i guess or still relatively young as a as a health technology company but the founding reason was uh personal experience so i i had asthma as a child it went away when i was 16 it came back in my mid-30s, didn't really understand why.

32:06And I was given a peak flow meter by a GP, which is the okay measure of lung function. It's what we've got. It's not brilliant, but it's what we've got. And I said, I'm really happy to do this in combination with my new inhaler. And I said, well, how do I get the results back to you? And he just went, that's a great question. Let's come back in two weeks. If you write them down, I'll take them off you. i went okay sure now i like even then that didn't make sense and now it seems insanity because yeah we've got more sophisticated systems now but then even then it seemed ridiculous so i yeah i basically just put all that into a spreadsheet i made i made the dumb version of aids which was an excel spreadsheet of my adherence to my medicine uh my peak flow readings and what i saw them was stabilized and then go up over time as you'd expect but i was also tracking diet sleep lifestyle and I took all of that back to the GP and I said here's this thing maybe you can make some sense of it and because of that we identified a trigger and I came off the therapy a month later and I've been fine since so that was and you could see the look on the GP's face you could just tell A, who is this lunatic and B you could see the potential if this was done more often so that was the instigation originally aid was purely a tool to improve adherence to medicine because that is a giant low-hanging fruit in healthcare that has not been solved.

33:31And regardless of therapy, demographic, disease, severity, 50 % of people aren't taking their prescription as advised. And what we get from that is complication and additional healthcare resource use and all sorts of things. I just think this is worth highlighting a little bit. Adherence. It's one of the only things that I've heard, That and fax machines is like the only things I've heard about since I, my first day at medical school to now, and it just being a problem that seems impossible to solve. Is that because there's like from, from the work you've done on it, is that, is that because there's so many contributing things to it?

34:13Is it that it's just so hard from, from a human behavior perspective? Like what, what is actually the problem? Why is it adherence so difficult as a problem to solve? so shifting human behavior full stop is hard but uh you would argue the first studies in this were done in hiv therapy right and the clinicians were like i just i don't understand this is saving their lives how are people not adhering to the therapy and the kind of the the the research mostly speaking was born from that point on so and people on with cancer on cancer therapies also become non-adherent some people with life or death heart failure become non-adherent so the question is what the hell is going on yeah and as you've rightly pointed out it's different for everyone in every different scenario right so there are people who don't take their medicine because they just forget okay that's one category of people there are people who don't take their medicine because of fear of social stigma right they don't want to take their their inhaler out in public so they just don't take it because it makes them look poorly or sickly or weak or whatever there are there are there are reasons you run out of your medicine then there's lots of weird and wonderful behavioral reasons so um all of which we capture as a platform so the adherence was is our foundation is just helping people remember to take the medicine help them understand what it does and why they've been prescribed it but also capture the reasons why they might become non-compliant for instance we in the in the initiatives we've run in the nhs all all kind of live real world deployments of our product where everyone studies if you take an asthma patient for example 25 % of patients tell aid that they stop taking their inhaler when they feel well so classic uh behavior with preventive inhalers and asthma but it's also what leads to exacerbations and why we have such a terrible rate of of um asthma attack deaths in the in the uk and in europe because of behaviors like that so it's multifactorial and it's really complicated the reminder is one tiny bit but there's loads of social behavioral reasons why it's just hard to form a behavior.

36:19I don't think it's insurmountable. We've shown in our deployments that we can get adherence to kind of low 70%, 72, 73%. Now, in a lot of medicines, that's therapeutically effective. In others, it needs to be 80-ish percent, low 80s. But it's the right direction. That's one intervention with one product. so adherence costs the nhs as near as makes no difference a billion pounds every year due to exacerbations or complications or medicine waste there's a whole host of other amazing papers on it um but it's a it's it's one of the biggest problems in healthcare that most people don't understand um and it needs that that before anything else needs solving so you're capturing everything that you possibly can in terms of the contributors to that and so what else are you capturing there so yes adherence was the was the original thought behind aid we just help people with their medicines but then it branched out very very quickly with some of the amazing clinicians we brought on board who are still with us today um to become more comprehensive chronic disease management so uh with a particular focus on comorbidity so we like spotting we like picking things that haven't been solved that are the the the longest hardest problems to try and solve yeah and i don't know if we'd have done the same thing if we're starting now with all of this knowledge again it's the beauty of ignorance right in these things so we said comorbidity is underserved one in three adults globally have two or more conditions disease care care today is still very disease specific no patient tells you they want that they want holistic care um so we said okay what we can do is build a platform single platform where you can manage multiple conditions No more point solutions.

38:04You can't, as a patient, coming back to product thinking and design thinking, you can't have five apps for your five conditions that don't talk to each other, some of which might send something to a clinician, some might not. It's a nightmare. And no patient wants that. Again, obvious determination of social behavior and product behavior. So we launched with asthma and types of diabetes. We then added hypertension, COPD, chronic kidney disease, obesity, and these blood cancers are coming. So aid in a really simple term, it's a patient mobile app that helps them do one thing, which is improve their capability to self-manage, to take more control of their health.

38:40That is the only scalable model of healthcare is to help put the patient at the center of how their care is delivered. So Don Kemper, who founded a company called Healthwise, championed this. Samir Gray in the NHS championed this, which is the greatest underused resource in healthcare is the patient. That's also our belief. So give the patient the tools and knowledge they need to take control of their health. We also have a clinician platform where those insights go, all the insights on adherence and behavior and all the monitoring and tracking and everything else that you can do on our platform goes to them.

39:13And then the key thing is the interaction method. So five years ago, when we were thinking about this problem, just very quickly identified what you identified, which is this is really, really individual. and no two people have the same presentation, progression, therapies, lifestyle. You can't design an interface that suits all of those scenarios or at least you couldn't up until five or so years ago. And the bet we made is that the way you do that at both an individual and a population level is through conversational AI. So aid in simplest terms is having short daily conversations with patients to help them take their medicine, capture reasons why they become non-adherent help them track their health things like biomedical stuff blood pressure peak flow glucose but also things like patient confidence mental well-being some psychosocial stuff and also educate them so we have a we have a library about 100 pieces of content across all of our conditions now all in kind of bite-sized format which all narrated to improve access and the accumulation of those things we hope is what's going to drive patient activation patient engagement um who buys it and where does it fit into the the the experience of a patient in their clinical pathway as of the most of our customers are in primary care right now um so it's we sell mostly to primary care networks yeah uh we have relationships with icbs but mostly the contract falls at pcn level and not gp level either so they can get the population benefits of this stuff because we generate a load of population level insights that being said blood cancers obviously is a secondary care pathway so in that regard it's it's almost certainly trust in hospitals but where does it fit in so think of aid as all the interactions that should be happening between clinic visits that resources don't allow so things like medicine adherence so you are a let's say you're a type two diabetic and you or let's say pre-diabetic and i don't know maybe obesity whatever it might be um you will see your your clinician once a year maybe you get some blood tests done get an annual review if you're an asthmatic patient um aid is everything in between all the educational stuff all the information all the monitoring um and it's and it's typically used every day but it depends on the context so if you're if you're a type 2 diabetic aid will have a conversation with you once a year about the importance of of retinal screening and how that's different from just going to spec savers and getting your eye eyesight tested but that only happens once a year whereas things that happen every day every the week so it's a real mix of conversations so the conversational capture is is really interesting and important here i imagine and this sort of plays into the white paper that you guys have done about building trust in AI with patients and I think it's interesting because I said this on this podcast before but in the in inverted commas you know good old days of Wobot I used Wobot and it was very conversational which was innovative at the time and um it propped up my mental fitness incredibly well um when I was starting my you know entrepreneurship journey and and challenging time as I'm sure you well know and it did that incredibly well and it did it it captured a lot more data from me this is anecdotal n equals one but it captured a lot of data from me because it was so conversational and far more I can definitely say far more than if it was up to me to go onto the app and then figure like then I've got a document and like almost like in a proms type way of of like metrics and what's your mood out of 10 and like all all that sort of stuff I found it very good at for extracting things out of me and then it would offer me teaching on do you want to talk about catastrophization and what that is you're like yeah okay it's like yeah this is the notion that just because one thing's happened the next 10 things down the line are not going to happen that's not a fact and actually you need to start separating fact from fiction here and here's a way to do that and here's an exercise so it was very good at sort of linking the two things as well and i found n equals one i i personified it and actually it I felt like I had a relationship with this, with this little robot and, and it did its job incredibly well.

44:00And, you know, for me that, that was a great example of, of a very early version of me building trust with this thing and leaning on it. And I, I've even heard Keith Grimes talk about something similar where he just said, you know, a time, and he said this publicly, but you know, a time where he needed it, it turned up at 9am every day when it said it would turn up at 9am every day and it did and and that that just in itself was helpful and it's a really interesting time that we can lean on this and academically decide what the best elements are to develop trust with an individual and a patient and so do you want to talk to that a little bit because you've you've clearly done a lot of work in this area the term inflection point is used a lot in entrepreneurial worlds it still has meaning this is one for for so many reasons so we our tech has been in the hands of patients for four years i think we've got a reasonable handle on how ai fits into someone's life or not um and so yeah so at the end of last year we wrote this white paper called building patient trust in ai and essentially what we did we looked at our own programs in the nhs kind of real world deployments what are our findings we looked at peer review studies two in particular for this particular white paper and kind of behavioral and economic models around trust and novel technology and the long and short of it is that there is this there is this paradox of trust with with patients and an ai and the what that essentially means is attitudes are not fixed anytime there's a novel technology put on a sector or a market or a or a populace or whatever it might be there there is this assumption well this is it's a blanket that just sits on here and it can do its thing and operate within that world and we'll figure it out and it and it's just not that straightforward there are nuances and ai with patients we be we build patient-facing tech right so um as much as we can talk about clinical efficiency from those models what i'm interested in is patient adoption and what we what this paper talks about is those attitudes are pretty malleable in fact any any individual patient can have completely contradictory reviews about AI depending on the setting they're in.

46:13So we group them in the papers, two things, two sides of the same coin regarding this paradox. One is willingness to confide in AI, which is essentially how patterns of patient disclosure are different in certain environments and why that matters and what levers we can pull to change those things. And the second side of the coin is, well at what point does confidence dip and trust fractures with ai that's very interesting and it's and in that instance it's moments of pressure so any any kind of high stakes uh interaction so if we take the willingness to confide things so what underpins that i think it's i think it's relatively reasonable to say that care is quite largely dependent on a patient's willingness to disclose information about them their symptoms how they've been feeling and yet disclosure is at best inconsistent and it works just kind of downright inaccurate so the literature illustrates this really well 70 of patients will admit to withholding information that is otherwise medically relevant and there's loads of reasons why that might be it could be embarrassment because it's symptom driven or relates to their lifestyle or it might be they disagree with the diagnosis or the therapy or they they don't want to ask what might be a silly question whatever the reason they all are grouped under one thing which is patients don't disclose information through fear of judgment whether or not it happens right this is well documented i don't want to say that thing in case the doctor in front of me thinks i'm less than i than i than i am that withholding information directly undermines care quality so there's this amazing study done by the university of southern california i think it was actually one of the first controlled studies measuring human interaction with virtual systems in a clinical setting.

48:05So they took about 200 patients, a little over 200 patients, and they used a virtual human, so a 3D human on the screen to conduct semi-structured interviews, semi-structured health interviews. The manipulation in this study wasn't the system itself, but it was whether or not the human believed it was acting autonomously or if it was tele-operated by a human. But either way, their interaction was with this virtual human. The findings were really remarkable. It was a really great study. So what the researchers showed is that when participants believed they were interacting with an autonomous system, they had lower fear of self-disclosure.

48:46So they were more willing to share stuff voluntarily. They had reduced impression management, which means they were they were less likely to amend their words to come off in a better light and importantly it was comparable in usability to to a human so like really great study so what does that tell us well what it tells us is that that detachment is a is a is a tool is a bonus in ai in that setting so uh the fact that there was no judgment there was this this pre-built neutrality to the system meant that patients were willing to share things that otherwise they might not, which means their care could be enhanced in ways that it might not.

49:27So that's great. That is trust building, right? We've seen similar things I mentioned with aid, right? We've had patients tell aid willingly that they don't have an asthma action plan. Half of all patients do that. I think it's fair to say many patients wouldn't say that in front of their clinician for whatever reason. So that's one side. The other side is a different study, which is where trust breaks down, which was done by University of Arizona. And it was a couple of thousand people. It was a large national survey, kind of qualitative to look at attitudes of when they might choose AI versus a human.

50:00And this is really interesting. The finding there was people fear the coldness of AI more than the potential for human error. So most patients acknowledge the AI would be more accurate, but they worried the AI would not understand them. and therefore I want a human. And the circumstances here are really important, which is moments of diagnosis, prognostic conversations, anything really high stakes. And it's in that mode that AI should really have no role. And we would argue the same thing. But crucially, that second study did it across demographics and ethnicities, and that's not consistent either.

50:44So black participants in the study were much less likely to engage with the AI than kind of white middle-aged. And now the reasons for this and what's interesting in the study is they say it's not due to the technology and how the technology is built. It's due to how the healthcare system treated those patients before they interacted with AI. So any of those perceptions and those values from the traditional healthcare system were then translated, imparted onto their interaction with AI. And that's how they evaluated it. So there are these fingerprints that are translated to interaction with AI.

51:15You can't divorce the patient's prior experience from their experience with AI. It was an awesome study. So where does this all come out? Which is neutrality is a lever we can pull to drive patient trust. And in some instances, neutrality and detachment is really valuable. And in others, it undermines care and patient relationship and all the stuff that we wouldn't want to interfere with. So these are the types of discussions that we need to start having. Yeah, and it is that nuance, isn't it? because every clinical interaction is slightly different for many reasons the purpose of it the information you're trying to either impart or um extract or like there's there's so many different types of clinical interaction and this is what i found when i looked at it i did this keynote in australia about this and about how all the literature points to a you know a positive correlation between the level of empathy and the amount of disclosure so when you have a really empathic clinician patient relationship which by the way improves and strengthens over time so it's not as if you can just turn up and be super empathic and get all the information then disclosure improves the quality of history taking improves but that research was done pre the boom of LLMs and actually you're right how do you make the most use out of maximizing empathy and also this element of the anonymity being a superpower in some elements and I think there is no single answer there's no single type of conversation or consultation that you could engineer that does the best of both it will be a very good detailed analysis and actually standing on the shoulders of giants most probably and building on all of this information from now for the next 30 40 50 years as to how best to do this what do we reserve human to human empathy for and you might say well if information is going from the clinician to the patient and that information is anything but positive then we might say that that is an example where we reserve that for human-to-human interaction but then we might say that in these specific clinical areas and in this specific you know severity of potential case and the information is going from the patient to a clinician and the clinician's job is to extract well perhaps that's a very good use for AI where it's low severity high embarrassment and we take like maybe like you know i'm i'm just making things up here and and but but i guess what i called for in that in that keynote was perhaps we've entered a world where that level of analysis and scrutiny becomes incredibly important in order to make use of the resources that we have in healthcare because we we have we can copy and paste the ai we can't copy and paste the human beings and create infinite numbers of them we can't do that so actually is it not the responsibility of us as the health tech community to actually put the time the effort the resource the money into figuring this out is this not the area of research that becomes the most important so that we can put the foot on the gas with ai safely knowing it's better than the standard if we can find those areas where it's being done by humans with varying levels of empathy so there will be people at the lower end of the scale and perhaps we can knock some of them off for a better AI experience so that we can reserve those areas that we instinctively know and we can define with data but clearly instinctively we know that should be reserved for human to human I think that that that as a means of distributing resource I think is so important yeah that's where we have to get and we have a firm belief and and by my perspective that the best ai company health tech companies that adopt or use or build ai share it which is it's not this isn't replacing clinicians this this is taking burden off them giving them insight that otherwise they wouldn't have had extending their reach all the other great things certainly what we believe the question is like but it comes down to decisions on the company level so i'll give an example we don't diagnose and we don't suggest changes to prescription yeah clear we've carved a clear regular like we're a class one medical device at some point probably the other year would class two probably we don't necessarily need to be but we'll do it anyway good but we don't diagnose and we don't suggest any therapy changes or prescription changes really deliberate move deliberate choice by us so we know where ai where we can we can we could accommodate and build trust from how we interact with patients that being said the second study from arizona showing that the high stakes moments, the diagnosis, the discussions about therapy or finding a pathway.

56:31I don't think that's out of bounds of AI doing that either. I just think it's really rare and it's really hard. I know the work the guys at Limbic are doing. I know you had Ross on the podcast. Yeah, Ross, yeah. But they're unique, right? So they do diagnose. They do give treatment, right? That demands a level of rigor and evidence that I have not seen anyone else even remotely get close to. So they're doing it. They'll get there. I just don't see other people. I'm sure everyone looks at them and goes, who want to diagnose and provide therapy. But they're so unique. And it's supposed to be so fucking hard for them to do, probably.

57:13I hear you, Matt. I think about that conversation a lot. I still think about that conversation because I - It's very, very rare. I agree how unique it is. I don't know anyone else doing it. Just the level of detail, like as he was explaining it, of everything that they do to set up that clinical AI layer. And you're right, you know, you saying, by the way, that, oh, look, we don't need to do class two, we'll probably just do it anyway, as somewhat of a throwaway comment. I actually think that's a really important thing to just pull out here of that's what good regulatory practice looks like. That's what building brand looks like as a rigorous company that can do clinical activity safely in that we don't fear getting class two regulated.

58:03By the way, like I'm absolutely fine with it. And in fact, we're doing it. And actually, I think that's partly the answer to what you said previously as well, in that I think this is going to be more of a community movement from within the community than it is going to be imparted down through rules or regulations or laws or anything. I think it more is the case that as a community, we decide what the best level of this is. And then that information getting to the, you know, venture capital layer or the enablers, you know, not just VCs, but there are many other enablers of companies growth. and them actually then starting to ask these questions because they perhaps fear what would happen if they were to back a company that did not have this sort of rigor or did not get in line or was not uh you know a vocal player in you know what the new world looks like i think that that's partly the value of media it's partly the value of you know you look at i don't know even like climate change movements or back in the day like anti-littering movements or you know stopping smoking movements like these are these are all movements that get championed by people that have power in the community and they're boosted by media and then it becomes this this force of its own that ends up having all of this kind of um effect i guess more more widely and and that's that's partly where i see our role as well it's partly where i see our role as the wider version of Somex that owns Health Tech Pigeon and this podcast and Biotech Dodo and all these the ways that we're reaching more than just our clients and and and stuff I consider that like quite a privilege to be honest it's why I've taken like a lot more heavy-handed view in Health Tech Pigeon the newsletter over the last six months as well because I just think with the amount of people that reaches and the amount of people that read it's basically like the biggest publication in Health Tech like I can't imagine that other people are getting the same amount of views as we're on that newsletter but like it's i like it's a it's a really important thing that actually to have the ability to influence at that kind of level and to start a conversation like we're having now that people then listen to and this finds these messages find their way out i just think yeah it's going to be more of a community movement i think than it is us lobbying and hoping for any kind of broader change down that line but i just wanted to pull that out because just saying like yeah we'll do class two just because whatever i actually that that's that essence is what i think we need to respect as an entire industry because that's what's going to force other people to want to do it if that makes sense yeah it's i i agree it's it's hard but it's also rewarding and it's so that we're the challenges with health tech startups is we probably reflect them quite well we're a six-person team we've done everything we have with six people so we're a multi-product company we're in the u.s as of april uh we've been doing okay scale-wise in in the nhs in england we've done okay in evidence but the six of us and two part-time you can still do these things it's just hard and it's like it's just but it is hard and it and it's costly like going through iso 2701 it's not i wouldn't i wouldn't go as far as saying it's a fun exercise but it's a it's a necessary exercise like we we got a clinical safety case as soon as we could pre i think even for our pilot we had a clinical safety case and again wasn't needed but it's what we felt was right these are entrepreneurial decisions that that are just demanded of being in this field yeah um but yeah if if you want real clinical intervention limbic is a brilliant blueprint and a lighthouse for for what needs to be done completely agree um it's in two parts if people want to go back and listen to that part one and part two you'll find in the uh in the list of episodes with ross um just quickly before we go and i've got five minutes left mirror patient facing scribe three interesting words put together there patient facing scribe what's the thinking behind that again born out of born out of frustration uh my again, we don't sit around and go, hey, where's the market going to go?

1:02:14And where's the technology going to go? This was born out of real world. My father, who's 80, has Alzheimer's. He fell and broke his hip. He was in the John Radcliffe Hospital for two weeks. Worst two weeks of most of our lives. And he's out, he's okay as best he can. But myself and my mother, who also lives in Oxford, we lived together. We were by his bedside for most of the days in those two weeks, but not all the time. My sister flew over, she was there. my biggest fear during that time was when a clinician came to the foot of his bed said a load of words in clinical speak that even I struggled with so I think I've got a reasonable grasp on this stuff and I class myself as an activated patient and someone who on a good day is reasonably okay with language and intelligence and this stuff and I even I was struggling and like anyone who knows if anyone's got anyone with Alzheimer's in the family know the second you take them out their environment things get demonstrably worse uh quite often short-lived There can be echoes of that even when they get back in their own setting.

1:03:13Anyway, my biggest fear was my dad being told something. We weren't there. And he'd very politely, as someone who was a pretty capable lawyer, just going like, okay, thanks. Yep, great. And then it would be gone. And we'd never know. So Mira was born out of that. And then we dived in again like we would from a product perspective. All the evidence shows up to 80 % of information given by clinicians is forgotten the second the patient leaves the room. and and it's loads of studies that illustrate those numbers and what that leads to is poor adherence poor care management um it disproportionately affects people who are old or who are dyslexic or who have any form of dementia and it leads to all sorts of downstream problems and immediate problems so we just thought well this is ridiculous problem to have these days so we built mirror um which is as you say it's the as far as we understood we launched in October last year.

1:04:09It was the first of its kind medical scribe for patients. It listens quietly during a consultation. It provides a summary in plain English, not clinical speak, patient-facing language. You can interrogate the transcription. You can say, what did Dr. Smith say about metformin and why I'm taking it? It will look over the transcript and give you a generated response. You can share your summary with family, friends, caregivers, so you're not having to explain the same thing over and over again. The response to that has been monumental i'm not surprised and the speed of it compared to aids which is annoying because it took us years to build a it took us six months to build this product but we we've all become power users i can't go into a consultation of any kind without using it and uh and it's it's completely transformative and we've had some patients again like the feedback we get on patient use of this is um very motivating i i think that's fascinating man because i can't go into a meeting now without turning granola on so even even in person i i open my phone and i open granola just to make sure and this isn't an ad for granola sadly they are not sponsoring product sadly nobody is sponsoring the podcast but we're open if anyone wants to um i need to stop people turning people down but anyway but the point is it's changed it's changed the way i do meetings because i know it exists i can actually just be more present and listen the notes i take are very different the notes are level two level three thinking not documenting what people are saying my notes are now you know in a kickoff meeting i had earlier today my notes are messaging could be around this or actually let's add these two things together and go down this route maybe for this content strategy not like so and so said blah like so so actually it's changing the way that i do things and i can see that in this you know knowing that it exists you can try and you can try and listen you can ask the questions knowing that all of the answers are going to be documented you're not trying to like it's it's just interesting what it does to the two-way relationship and i see this happening now with the ai layer in between the health care data and the person we're seeing it with what fitbit and google have just won the race of which is the ai layer between your wearable data and you to contextualize it and all that sorts of stuff so they released that yesterday and all the others are going to do the same.

1:06:32So it's really interesting now, the AI layer that can take loads of data that's horrible and complex and messy, not least a clinician just sprouting jargon at you at the side of the bedside. It will contextualize things for you and give you that stuff. I think, honestly, I think this is where I think the legacy of AI will be in this, like, hold on a minute, everyone now understands everything and the whole thing just runs a lot better. I really think it's the one. I'm not surprised it's been such a good response, man, because I think it's one of the most impactful things that we can do for patients, genuinely.

1:07:07Because you're at the mercy of a good or a bad communicator clinician currently. Or a good or a bad day, a good or bad minute. Yeah, and so everything we've had from patients is empowering. I kind of forgot it was there the second I said it running. And I don't have to repeat myself over and over to friends about this damn thing I've got now. But most importantly, where does this go? so right now mirror is passive right it's just there it listens yeah summary where does this go in the future it won't be passive it'll pipe up and say hold on that's actually contradictory to the other consultation you had and actually maybe that medicine has a contraindication to this thing or you should definitely mention this bit before you go in that's what we're building now that's where that goes so and again like patient trust every time you open the app and then every session you're in with the app and you you go into one of your summaries we tell the patient we say ai can make mistakes even the best designed ai can make mistakes if you're unsure don't guess speak to your clinician like we just we we tell the patient these things are imperfect so then okay well why well then why make it well because there is already a baseline of risk which is people not remembering what they're told in person and that risk is pronounced and and verifiable and in the literature not to mention just the patient but the clinical risk 80 of severe clinical errors are in information transfer and a large portion of that is in patient handover so this is where good design good technology can take these things that are that everyone appreciates the second you say or it just helps you remember what was said in the consultation everyone goes oh god yeah like i've never i've yeah god and anytime we tell us about this product they go i'm going to give this to my mom or i know someone who would benefit from this or oh that would have been great when or when we're going through all our maternity god if only it's instant and universal and it's like why hasn't this been done before and it's like well the tech wasn't quite there the adoption wasn't quite there timing is always a part of innovation but this is where health is going like a unified patient-owned record of health that that is there for them that's where this stuff goes love it man um i better let you go but before i do thank you for this this has been awesome um if people want to learn more about what you're up to at aid what is the best way for them to find you or learn more about the company just aid.health that's that's where we are we're on we're on uh It couldn't be more simple, really.

1:09:43So aid.health, A-I-D-E.health. We're on LinkedIn. We do have some social media stuff, but it's mostly LinkedIn, and we're there. And white papers, both available for free on our website. You can get them. We're not charging for those things. And, yeah, and just thanks for having us. We're doing this because we love it, and it's because it's important. And the promise of this technology in health is real. It's just hard, you know, and we need communities and discussions to get through it, I think. awesome appreciate you man thanks for coming on thanks man