Episode notes
Entrepreneurs have a problem. They love comparing what they do to waging a war. In the management literature, it's almost entirely written for peacetime CEOs. On the other hand, if you're running out of cash — if you're like Apple when Steve Jobs first took over and they had three weeks of cash left — that's when the wartime techniques come into play. It's not called rapid growth. It's called blitzscaling. Even our beloved health tech messiah, Eric Topol, wrote a whole book on the creative destruction of medicine. What's with all this aggression? Like, bro, you're building a widget that might be acquired by a large enterprise healthcare incumbent one day, not fighting off the Persians. This is Sparta!
That's Tom Whicher, CEO of DrDoctor, who describe themselves as a patient engagement platform — a term which, well, I don't know what it means. Basically, patients use DrDoctor to book their appointments and receive some care. But the whole point is to reduce DNA rates. Did Not Attend. DNA rates — when patients don't turn up for their appointments — really piss off managers, because they can measure it, and because you're wasting capacity in an often already overburdened healthcare system. So if your startup can fix this, that's a lot of ROI.
Tom started DrDoctor ten years ago. They've had 20 million patients use their platform, recently raising a £10 million Series B. He's always done things a bit differently. Thanks to Isabel Paret from Fight or Flight for making this introduction.
In this conversation
- Why DrDoctor bootstrapped for its first seven years while its 2012 peers — Babylon, Huma — raised big and blitzscaled: Whicher's case that in healthcare you keep the equity, go slow, and prove the value before you scale.
- The counterintuitive playbook for cutting no-shows: solve patient engagement first, get 60–70% of a hospital's patients onto text, then layer higher-value products on top of that engaged base.
- Real numbers from the field — a global NHS DNA rate of 12% brought to 3–5% in DrDoctor hospitals, a 60% cut in face-to-face oncology follow-ups in Nottingham, and 91% in Liverpool dermatology using patient-initiated models.
- Carrot beats stick: why the Netherlands' fines leave a stubborn 5% of patients missing appointments, and how the Freakonomics daycare study explains it — treat people like adults and they treat access like a scarce resource.
- A founder's framework for building in health: get on the train and go see users in person, be weakly opinionated on the journey but clear on the outcome, and "fake the automation" — DrDoctor's first product was literally a developer named Steve replying to texts by hand.
Transcript
So you started with a very specific problem in a specific healthcare system — patients booking appointments within the NHS. Can you talk me through the reasoning behind that, and in hindsight whether "let's solve a specific problem really well" was the right approach?
With hindsight, a lot of things have changed, but I've always believed in focus. One of the interesting things about DrDoctor is we bootstrapped for the first seven years. That was really intentional. I looked at other health tech businesses that raised a lot of money and weren't necessarily getting a correlated return with the equity they'd put into their organisations. So we decided to stay really tightly coupled to a problem we knew had value.
We actually shifted slightly from booking and looked much more at patient engagement as the first part of our product set. Patient engagement wasn't really a thing people talked about in 2012. Now, in the UK, it's written into national policy that every organisation has to have some sort of patient engagement platform.
The reason we focused on engagement was really purposeful. We set out to solve the booking problem, but what we discovered when we talked to patients was: the issue isn't so much that I can't book an appointment. It's that I don't know why I'm turning up, I'm not sure I value this, and no one's telling me where I am in the queue. So we built a lot of our early technology on text messaging. We still send about 20 million text messages a month to patients across the UK.
That engagement step was vital to being able to build higher-value products. That was always our product strategy: solve patient engagement, get 60–70% of patients in an organisation using our technology, and then leverage that engaged user base to develop and deploy higher-value products — booking, data collection, remote monitoring, and ultimately moving to a data-driven, value-based healthcare system. But if you don't get that first step right — the onboarding, the channel shift, taking people used to the phone and post and moving them online — none of the other value comes through. The focus was intentional because it gave us a base on which we could scale everything else.
In creating a patient engagement platform, are there any quirks of human psychology you must have noted? Any strange ways people engage better or worse based on little quirks of how you design the product or pitch things?
There's a lot of that stuff. We do a huge amount of work on behavioural insights — everything from the way you word a message to a patient. Different people react differently to different wording, so ideally you customise the message based on someone's age or social demographic group. You can use nudge. There's a famous bit of work the Behavioural Insights Team did, funded by the UK government, that found including the cost of a missed appointment in a text message reduces the non-attendance rate significantly — I think it's around 20%. Adding nudge stuff like that is massive.
Then there's understanding what a patient values in an appointment. Are they coming to be told something? Do they feel like they're being checked up on? The time of day makes a really big difference. If you're a senior citizen with free travel, there's no point booking that person an appointment outside their free travel times — an early appointment before the free bus pass kicks in is a total waste of time. Getting that individual-level knowledge and applying it to a platform at scale is one of the reasons we've had such broad success.
Have you got any more on this? One thing you talked about reminds me of the "start with why" idea. It's interesting how the phrasing — what the appointment will cost the health service, or what benefit the patient will actually get — is an interesting route. It reminds me of the Apple ads back in the day: Zoom were advertising "this is a two-gigabyte MP3 player which can play this MP3 in WAV format," and then Apple just put up an ad saying "a thousand songs in your pocket," and immediately you get it. Do you have anything more on the nudge psychology of getting patients to turn up?
“None of us want to feel like we're a sausage in a sausage factory. The more agency you can give a patient, the more likely they'll engage — and have a great outcome.”
Tom
I've always believed really strongly that if you empower patients so they're equal partners in their decision-making, you get better outcomes. None of us want to feel like we're a sausage in a sausage factory. If you want a patient to have a great recovery, or to adhere to a treatment routine, you need to include them from day one, right at the very beginning.
The fact that most healthcare systems are still very paternalistic in the way they approach that first interaction is a massive challenge. If the first thing you get from your health provider is a letter telling you to turn up at a certain day and time, you're already on the back foot — you've had agency taken away from you. The more agency you can give a patient, the sooner, the more likely they'll engage later on and have a great outcome. So we're really thoughtful about how you make people feel like they have choice, like they're being done with and not to. A lot of that is wording, but also the back end of the appointment journey — monitoring patients remotely, whether in low-tech ways or using devices. You want that patient to want to engage with the remote monitoring tool in a way that's equitable. If they feel like they've been treated like an adult at the beginning, they're much more likely to treat a patient-initiated follow-up model, for example, with respect — and not abuse it. Because, hey, guess what: you've treated me like an adult, so I'm going to view this as a resource I'll treat with some scarcity.
The whole thing is about treating people like grown-ups and being really clear where the value is. It goes for clinicians too. As a clinician, if you sit in an outpatient clinic and ten patients turn up, a lot of them could probably have been done asynchronously. Why see ten patients for 20 minutes each when you could see two or three for half an hour or longer — if you use technology to make sure you get the high-value patients in first? If you can manage the risk remotely and keep people informed on both sides about how someone's adhering to their treatment, you end up with much higher-value in-person interactions. It's all linked. It's a holistic problem — you can't expect one part of the system to work without getting it right from day one.
Interesting. I guess the counter-force is that when you're building a big organisation, or thinking at a systems level, there's something about not individualising stuff too much — keeping the variance down. If you can keep everything to a ten-minute face-to-face slot where the patient turns up at this time, speaks to reception, comes in, and the doctor has five minutes history, five minutes exam, and then "see you later" — on a system level, that seems to work better, reducing that variance. But maybe that's the bit you've spotted where, on the personal-connection level, it doesn't perform so well, because people feel like a sausage in a sausage factory.
Yeah, exactly. The cool thing about technology like ours is that standardisation is helpful — you don't want everybody with a hernia to have a different route through the hospital, that would be a nightmare. But what you can do with technology is understand the intricacies of a particular patient's needs and slightly bespoke the frequency of their interactions.
I personally would much rather have the majority of my interactions online and by text, and just come in to see my doctor. I'm going through an allergy pathway at the moment, and honestly every single interaction that isn't remote frustrates me — I'm like, okay, I've got to take half a day off work, this is rubbish, we could do this better. Whereas I know my dad, who has an autoimmune skin condition, values those interactions with his clinicians very much. He'll happily travel into London to go to Guy's Hospital to see someone. He doesn't mind. We have different needs, and the great thing about technology is it lets you standardise a pathway but personalise the way different patients interact with it based on their preferences. You raise the quality of the interaction for everybody while actually reducing the net load on clinical practice — because a whole bunch of people are going to go do it asynchronously. I'm cool with that.
So you can spend longer with the dads who want a longer chat.
Absolutely.
So Tom, I did warn you that I just ask adversarial questions, because I think it's interesting. First criticism: when I see a company like yours solving this problem, the cohort of patients who frequently do not attend — who, in quotes, "waste" these slots — are sometimes characterised as chaotic patients with chaotic lifestyles. I wonder whether a technological innovation would actually work on that group. On people like ourselves — young professionals — it probably works great. But for those groups, are you actually hitting the chords that'll help them, or is it missing the point there?
I'm going to split that into two, if that's okay. How frequently have you sat in outpatients — has your clinic run late?
I did four months of general practice, and I think every single day, every single clinic, was late.
“If you gently disrupt a healthcare system from the inside out, so it doesn't even know it's changing, it doesn't put up antibodies — it actually gets behind the change.”
Tom
They all run late. And the reason is you have this variability. Some days your DNA rate could be 20%, some days it can be zero. Because you don't know which you're going to get, you have to book to an average — and if you book to an average, most of the time you'll have too many patients and run late. So if we can reduce non-attendance — and we've also got some really cool AI which predicts the people likely not to come, I'll talk about that — you can plan and book clinics correctly from day one so they're much more likely to run to time. You get a better experience.
Combine that with the AI tools identifying the specific cohorts who aren't going to come, and you can put targeted interventions around them. You're quite right that different people don't come for different reasons. Actually, the highest DNA rate is in young professionals — people who are working, who think "oh man, I forgot," or "I can't take the day off, I've got an important meeting." They're a massive issue, and they respond really well to a tech-based solution.
Then you have a different group who are harder to access, and we've found SMS works really well for them. If you're transient, move around a lot, live in a shared house, you probably don't open your post — or your post doesn't reach you before you've moved. Shelter did a piece of work around homelessness: how do you reach someone who's homeless with healthcare needs? Text messaging is the best way. They might not have any credit, but they can receive a message. So we try to make our platform work for all of those people.
When we roll the AI into that, we can give you a clinic list and tell you: these are the patients who aren't going to come. You look at that list and decide what to do. If the person at the top is someone with diabetes who hasn't been in for a long time, badly controlled, you can get a community diabetes nurse to reach out and give them a much more bespoke experience — which ultimately saves money.
That works pretty well for airlines. It'd seem to me that if big airlines are doing this, why do we need to do something different in healthcare? Is there a specific reason it doesn't work as well?
One of the big things is it's very rare you'll book a flight a year out, whereas if you book a follow-up appointment, it's highly likely it's six, eight, twelve months in the future. When that appointment gets booked, you don't know what you'll be doing that day — you might not even remember you've got it in six months. Whereas a flight, you're going on holiday, you book it three months in advance, maybe six if you're really organised — if you're me, two weeks before. So the booking horizon makes a massive difference.
When someone's been booked a year out, they just don't know what they're doing. And even if they want to change that day, it can be really hard — particularly if a hospital is fully booking all its clinics. If you've got a clinical need to follow someone up at six months, and two weeks before they ring up and say "I can't make it, I've been invited to a wedding," and the next available capacity is three or four months out, you actually have a clinical safety problem. So the way we book appointments has a massive effect on non-attendance.
The other thing: some people are scared of flying, but lots of people are quite nervous of coming in for their appointment. DNAs aren't just about today. You often see endoscopy has super high DNA rates, because people wake up and go, "I don't want bum-cam today, I'm not up for this, this is not what I'm about." That goes back to our behavioural point — demonstrating that by coming in they'll ultimately have a better outcome, and making them feel secure and safe and like they have some agency.
The last bit is that the DNA rate on flights is low — single-digit percentages. Whereas in the NHS we have a global average DNA rate of 12%. I'm really proud to say that in a DrDoctor hospital we're between 3% and 5%. But when you're at a 12% DNA rate, you can't just overbook to solve it — otherwise you end up with the late-running clinics we described earlier.
In general, have you found the carrot or the stick works better for motivating people? It comes back to theories of loss aversion — if you tell someone "if you don't do this you'll lose a thousand pounds" versus "if you do this you'll win a thousand pounds," they'll be much more motivated by losing the money. So which have you found works better, carrot or stick?
Definitely carrot. Definitely carrot. The stick thing — in the Netherlands they charge people, you get a fine if you miss your appointment. Every now and again it comes around in the papers here: "we should charge everybody for missed appointments." They have a stubborn 5% DNA rate. It doesn't go below that in the Netherlands. We can get a hospital in the UK to a 5% DNA rate really easily, just using simple tech — and it doesn't go any lower than that, because the stick stops working. Whereas — and I'm going to sound like a stuck record — I go back to agency. Making people feel like a valued and equal partner from day one means they're more likely to come, more likely to have a positive interaction with you as a clinician when they turn up, and more likely to have a positive outcome. So carrots all the way.
The stubborn 5% DNA rate despite charging people in the Netherlands — I think this links to a study mentioned in the original Freakonomics book. A psychology study — I think it's Israeli nurseries. Parents are working, they end up picking their kids up late, so the nurseries start charging them — and what happens instead is now parents don't even feel guilty about being late, because they're paying. They're like, "why not?" The late rates actually get worse. It's interesting how psychology and human behaviour just contradict themselves in so many ways. I guess you've just got to run the experiments to find out what will work.
You've got to run the experiments. And in health — yeah, exactly, there's a give and take. I've even noticed it with the school fines — you get fined if you take your kid out of school these days, and suddenly you hear people talking about "okay, there's a balance here." The school fine says this, but the extra cost of the flight says that — and it becomes an economic equation rather than "hey, my kids should be in school."
People don't always do what you expect.
People don't always do what you expect them to do, so being able to run rapid iterations and A/B tests across different clinical settings is really important. Somebody treats a two-week-wait oncology clinic differently to how they treat a physio appointment — that's just a fact. People have a different fear factor going into it, so you have to personalise and customise the experience around all of that. One of the cool things about the DrDoctor platform is that because we have such a high volume of interactions — tens of millions a month — and the system is super configurable and granular, we're constantly running small experiments at a specialty level, a hospital level, even a locality level, working out what we can change to reduce friction for the patient, make it easier for them to access the information they need, and ultimately save some money and time.
You should get some people from the behavioural unit, or Daniel Kahneman or someone like that, to run some interesting experiments — you have some really cool data.
We really should. I talk to the behavioural insights guys a lot, actually — we've got this amazing resource in the platform that we could leverage in all sorts of ways. There's never enough time in the day. That's the problem.
All right, Tom, I think this conversation has become too nice, so let me add some more criticism. I've got this skeptical take — and it's not just mine, I think a lot of people have it. Specifically, in a socialised healthcare system, that friction in booking appointments and follow-ups exists for a reason: to temper or control demand in some way. The very clear example is that in primary care, in GP, the only way most people in the UK can get an appointment is calling at 8am and sitting in a queue for maybe 30 minutes, eventually getting through. So my skeptical take is that perhaps the reason innovations like DrDoctor don't already exist everywhere isn't because we don't have the technology or the people with amazing ideas — it's because the system, or "the man," or whatever nebulous bad actor, doesn't want it to exist. What do you make of that?
The more time you spend working in any complex environment, particularly in health, the more you realise there is no big conspiracy — because no one has enough control over it to manage that sort of thing. If anybody could pull off a stunt like that, it'd be very different. So I don't think there's a big conspiracy. I do think, particularly at the front door, people are trying to make sure they manage demand. That is true.
What's really interesting is there's latent demand out there for healthcare services — people who aren't being seen who should be. And the great thing about a socialised system, where the payer and provider are one, is that there are incentives to invest in prevention. That person who can't get a GP appointment, if they leave it too long, ends up turning up at A&E as an emergency admission — and that costs a lot more than it would have to just see them early and solve the problem at the front door. So the system we work under is incentivised to make sure people get access right at the start, because it costs less ultimately.
Now, the reality is those incentives aren't lined up on the ground. The GP doesn't mind if the patient goes to A&E, because they're not paying for it currently. But as we've moved to integrated care systems, which do have sight of the patient across the locality, I think we'll see more of that. We're going to see commissioners thinking about how to put incentives in place so everybody gets access earlier — a big shift towards a value-based care model, which I think is exciting. The NHS has an opportunity to lead the world in delivering true value-based care.
Twenty years later, value-based care still isn't a reality — but every single conference up and down the Western world is talking about VBC. I think the NHS could grapple that and show the world what good looks like, because we have an integrated payer-provider.
“In Nottingham, we moved to a monitored patient-initiated model in oncology services, and we saw a 60% reduction in face-to-face.”
Tom
That's one thing. The other thing is, when I speak to clinicians about freeing up booking, they actually get more worried about follow-ups. When it comes to patient-initiated follow-up booking models, clinicians say, "if I allow patients to book, they're going to be in every single week talking to me." We've seen the opposite.
I'll give you a couple of statistical examples. In Nottingham, we moved to a monitored patient-initiated model in oncology services, and we saw a 60% reduction in face-to-face —
Sorry — so that means that instead of the patient automatically getting booked for a follow-up in six months, they're completely left to their own devices to book it themselves using some kind of link. Is that right?
This one was monitored, so it was slightly different. If they got out of range on a PROM, they were prompted to book an appointment — but at any point in the journey they always had the option to free-book into a link as well. That's one example. Another is dermatology services in Liverpool, where we saw a 91% reduction in face-to-face follow-up appointments by initiating this model. It works really well in chronic conditions.
The challenge everyone raises is, "are people just going to abuse this?" But the opposite is true. People treat it like a scarce resource and only use it when they need to — which takes us back to: empower people, give them agency, treat them like adults, and you get a better outcome all through the journey.
That's super interesting, because I'd have guessed the complete opposite.
Yeah.
Tom, let me ask you a really difficult, unpopular question. You're clearly a good person, and you've got this whole thing about building everything patient-focused — making it better for the real people dealing with healthcare at a difficult time in their lives, treating them like adults. Totally agree that good things happen when you do — who'd have guessed, right? But are there any things you've come across, in your healthcare or company-building experience, where the more traditional, paternalistic approach — "this is what you need to do, you need to be here at this time, you've got no choice" — actually works better?
That's a great question. I don't think anyone's ever asked me that before. I don't think I've ever seen it work at a system level. There are individuals in the system who respond better to that — who want to be told where to go, to just turn up, and are happy with that. They exist, for sure. I've met them, and I'm sure you have too. They don't want the complexity, they don't want to be offered the choice — the sort of people who, when you go out for dinner, just want you to choose for them.
But if you scale it up to even a clinic level? No, I don't think so. This is better for everybody — we just have to remember that different people have different preferences. When we design our products and platforms, we design with everybody in mind. For example, I use the word "hybrid healthcare" rather than "digital-first healthcare," because whenever we design something we make sure there's a phone option, a post option, a face-to-face option — because for some people that's better. You have to think about the whole problem, not just the digital part, otherwise you exclude people. So at a system level, I think no. At the individual level, you've got to design for everybody and their needs.
I remember seeing a patient — fresh out of med school — and there was a decision around a few drugs they could start for a long-term condition they had. I gave them all the pros and cons, and they just looked at me with this WTF face: "which one do I take?" They didn't like that. It didn't inspire confidence that I was giving so many options. "You're the doctor."
So it's interesting — you're right.
Have there been any habits, or ways you approach problems, or ways you think, that have helped you get to where you are today? Any secret sauce?
An open-minded approach to the problem space is key. I still think in healthcare, people building products don't spend enough time with the users. It amazes me how many health tech businesses are built by product managers sat in offices in cities. I used to have this thing — I was going to try to collect every single mini train station in the country, because I believe if you get on the train and go and see a customer, you're going to learn something. So my number one is: get out there and visit people in person. You can't do enough of it in health — it's endemic that people don't.
Number two: be really clear on the outcome, but weakly opinionated on the journey. As you go down, things change. We thought we were going to build an app for outpatient appointments; we built a text-message-based engagement platform initially. We solved the problem, but went about it in a totally different way. Never optimise too soon.
The MVP thing is really interesting in healthcare, because the bar is quite high. But you should still — I love the Airbnb example of the founders going and taking photos of the units themselves. We did a lot of that sort of stuff. We literally faked the automation in the background by just replying to messages, to see if patients would engage. Initially we were literally just Steve in the background — one of our developers — replying to messages.
So it's a long journey — it has its ups and downs, it's a rollercoaster.
I think it's why we do it. We're addicted to it. In healthcare you have to have a long-term horizon, and for me the mission has always been that I am so motivated by the thought that we're actually making a difference to patients every day. That allows me to motor through the painful bits — and there are lots of painful bits, where you want to scream and throw your laptop out the window because the system is being so useless. But long-term, we're making a difference. This is going to help the person on the street. It'll help my parents, it'll help my sister. If we can improve the NHS, that gets me out of bed, and it gets me through the challenges of working with a slower, enterprise-type organisation like the NHS is.
I hope you enjoyed that episode. You can find all my links by going to bigpicturemedicine.co.uk, and if you've been enjoying the podcast, please consider leaving a review. All right — that's it, thanks for listening.