رونوشت 🔗
Chris Nelson: Before we get into it today, a quick note. We're two men talking about a women's health platform. So I want to make sure that upfront I was going to say that we're not going to pretend to speak for women's lived experience of reproductive health, fertility, pregnancy, menopause, or the healthcare system in general. Our lens today is narrower and I think still very important about what it takes to build safety, evidence, regulatory discipline, and trust around digital health products that are used. enormous scale. And because Tau Sif, our guest today, is both a GP and a regulatory leader, he's unusually well placed to talk about that. But for a deeper dive on FemTech, check out the podcast FemTech Focus and the Pink Pineapple podcast or search on Substack for very deep dives on many of the topics we'll touch on just briefly today. Over the show. Welcome to Prove It, the show where digital health leaders stop hand waving and explain how ambitious claims survive contact with evidence peer reviewers, regulators, and the real world. I'm Paul Wicks and today's guest is Dr. Tassif Murali. Tassif is vice president of regulatory science at Flow Health, where he leads regulatory strategy across a fast moving women's health platform used by tens of millions of people worldwide. Before that, he was VP of medical safety and regulatory affairs at Ada Health, where he helped build an always on compliance culture, overseeing MDR and ISO certifications across the product portfolio. And he's also a GP. a visiting lecturer on ethics and AI and healthcare, and according to his own LinkedIn profile, an optimistic optimizer. Towsief, welcome to Proof It. Thanks so much, Paul. I think I found my vibes man. ⁓ Brilliant. So let's start with the human bit. I'm really interested in, you know, the Towsief ⁓ origin story, not villain origin story, obviously superhero origin story. ⁓ Tell us a little about your background and how you came into medicine in the first place. Sure. Okay. So I grew up in Birmingham and the accent left me sort of a long time ago, but ⁓ once from Birmingham, always from Birmingham. But in fact, story begins before Birmingham. So my parents were displaced Ugandans from the Exodus in 72 when General Idi Amin sort of gave Asians 90 days specifically to leave the country for a variety of reasons. And ⁓ as a result of that, my parents sort of with suitcases in hands were ⁓ expelled from the country and found themselves think my dad was in an RAF camp ⁓ somewhere near Ipswich and my mom found herself in an RAF camp near Saffron Walden in Essex. And that was their beginning of the of the displacement journey. And I think that sense of displacement, that sense of dislocation, that sense of ⁓ what drives people ⁓ towards embracing risk, but making something of risk as it were, ⁓ this idea of this optimistic optimizer, I think, origin story starts very much from what's hardwired into my DNA as a result of that story. In fact, my son recently surprised me and a lot of people around me by talking about that very honestly in a TEDx talk that I thought twice about sharing. I don't often share about my family publicly, but I did on this occasion because it resonated so deeply with me and the rest of the family. And I know that there's a lot of people in the diaspora to whom it spoke very directly. So I trained as a medic. And that was much to the, I'm going to say disappointment of my mum actually, which make him as a bit of a surprise. It does. That doesn't go with the cliches of mums or mums from certain cultures for sure. No, indeed. Indeed. Because it's often the, know, the, the stereotype is that you're given sort of three sort of cards, right? Engineering, doctor or lawyer. yeah, no, this, was, my mum was very keen on me not becoming a doctor. The reason being that she was a diabetes research nurse. And so she'd seen the life of doctors ⁓ in their sort of unfiltered reality ⁓ quite clearly and didn't necessarily want me to be exposed to that life in her eyes of sort of, know, sleepless nights, relentlessly being on call. And so she did everything that she could to dissuade me from becoming a doctor. So she sets up more and more elaborate ways of putting me off. And that was usually in form of ⁓ work experience. that at places that she thought would sort of really put me off the idea. So I remember doing work experience at like a homeless shelter. I remember doing work experience at a drug rehab clinic and I fell more and more in love with medicine and what it was all about and how it would allow me to serve people in a way that I'd never seen before. ⁓ So I mum rapidly realised that her plan was backfiring and stepped away and allowed me to scratch this itch. And my dad, I think influenced very much by the Ugandan experience. did want me to do medicine, but very much for the reason of, well, not if you find yourself sort of relocated in the world somewhere, but when you find yourself relocated, you can land on that set of skills and practice again wherever you are. So that was the background to it. I studied medicine at Cambridge and in my third year, as you often can there as part of the tripos course, you can sort of do a little bit of pick and mix actually and take something random off the shelf and have a go at it. And I did ethics and law for a year along with mechanisms and disease. And that I think so to see that I only sort of germinated ⁓ much further down the line. I set up a research project though, which I know may interest you and some of your listeners, ⁓ which was on hyperglycemia and awareness syndrome. So this is this paradoxical situation where people who suffer from diabetes and uncontrolled diabetes or poorly controlled diabetes, they experience episodes of hypoglycemia where they've over injected on the insulin. And they find themselves physiologically in a state where they're sweating, they can't make sense of things, but they're unable to register the fact
Load full transcript
that they're experiencing those symptoms. And so behind the scenes, there's a physiological adaptation that's taking place that makes them less and less able to recognize the horrible situation that they find themselves in. And so I set up a research project with a professor in Birmingham. As a result of that, finished my clinical medicine in Birmingham. And I decided that pediatrics is the way for me. ⁓ until I came to a crossroads where I realized that if I do carry on down this career trajectory, then I'm probably going to end up seeing more of other people's children than my own children. And I need to have a real step back and think about, what is, what is important to me and how do I strike that life balance? So I retrained as a GP and have been a GP for the last, I think almost 15 years. And it was towards the back end of that journey where this love for medtech grew and grew and grew. Yeah. So, I mean, you know, it's incredible really in the ways that if you're not in medicine, I think people think a doctor is only one thing, don't they? And then when you're on the other side of it, you realise there's actually hundreds and hundreds of different types of specialties. you know, the sub branch that I used to do research with was in motor neurone disease and neurodegenerative diseases. you know, although the people there might be ⁓ neurologists, they would be super specialised neurologists. And you'd sort of get to this point where if you asked them a question about a rash or cancer or something, they'd forgotten all of that information. You know, it was in, it was an ancient history. And, ⁓ you know, I would find that in areas like oncology, seemed super, super deep. Like, you know, I only deal with left-handed pinky toenail cancer and I've forgotten everything about everything else. So, ⁓ yeah, but then the GP has to do a bit of everything. Right. So I suppose you've got your pediatrics fix. I'm sure you had your moments of ethics and law. ⁓ Yeah. Was that giving you sort of more of that breadth of experience and exposure? Yeah, very much so. Very much so. it's only sort of looking back that I realised that I was probably beginning to set up one of these T-shaped careers where I was happy to specialise in a couple of areas in particular, but I was also very comfortable. ⁓ And I think that came from this. think GPs are naturally very curious, very nosy individuals who like to get plugged into people's lives. And where I practised, it was very much, know, Willow Tree Family Doctors. It's a family medicine set up where I could follow people in their journey from... ⁓ not just cradle to grave, actually it was literally that and even beyond that. ⁓ So I remember sort of seeing pregnant women helping them on their journey. And then, you know, a little human being would pop out and I'd carry on helping them through their journey into the world. And they wouldn't necessarily need to share bits and pieces of their life with me for me to appreciate
how they plugged into their own family context, their community context. Something that even probably the world's greatest EHR as it stands right now can't really deliver. that nuance of knowing somebody in their very, very specific context. And in the early days, those EHRs were for good or for bad, like relatively hackable by clinicians. So I could hack into our EHR to make it work better for me and my colleagues around me. So could I set up automations and macros that would allow me to spend more time engaging with my patient rather than with the screen that was next to me? ⁓ And that set off a little bit of a chain reaction in that I got quite fascinated by what was possible with regards to making a clinician's life ⁓ more efficient, more effective so that I could deliver medicine more effectively to my patients. ⁓ And it got to a stage where we started inviting ⁓ med tech companies, both hardware and software to the practice and created like a little hub of innovation basically in Northwest London where Innovative companies would come and share what they're up to and sometimes iterate products with us in real life. ⁓ And a couple of examples that come to mind, one is Fibris. It's a company that's led by Elena Nadenova. ⁓ It's doing wonderfully well at the moment actually that ⁓ I think really kickstarted its product journey during COVID, where all of a sudden we found ourselves having to deliver medicine remotely, including virtual home visits. ⁓ Now, I miss those days where you would visit people, spend a lunchtime running away from dogs. was usually the situation I found myself in. But if I did have it, actually make it into someone's house and avoid the dogs. You could really appreciate much more about what someone was bringing to the table by seeing them in their own sort of setting. But that rug was pulled from underneath us during COVID. And so we found ourselves having to get... information, clinical information in a different way. you know, temperature measurements, blood pressure measurements, heart rates, all then delivered through these technologically steroid boosted equivalents of what they used to be, you know, so, and that's where Fibros sort of come to the fore. Another company was AcuRx that revolutionized communication between doctors and their patients. And I think they've ⁓ yeah, that's who I get my messages from for my, you know, a middle-aged man health check or the COVID vaccine or whatever it is. There you have it. There you have it. Like a deceptively simple concept, but their footprint in the NHS is remarkable. ⁓ And I remember a session particularly well actually where I was in clinic with a patient in front of me, ⁓ a developer next to me, a product manager alongside me. And I was able to see what it feels like to ⁓ create and refine a product in real time. and that was a, that was a scratch. That was an itch that I couldn't ignore scratching. That was, ⁓ that was, that was, was, know, that, that, that really was
something that I couldn't ignore because on the best of days in clinic, I could hopefully make a difference to the life of the patient in front of me. But that realization that I could actually do something similar, if not more on a scale that was unfathomable prior to that was something that I really sort of decided to lean into. ⁓ And it was during the pandemic then that I took that jump into the relative unknown and joined Ada Health ⁓ as their director of medical safety. And that's where our paths crossed actually. ⁓ so listening to the podcast, we'll have heard an interview with Daniel Nathrath, who's the founder and CEO of Ada, but just to say quickly, so they are ⁓ a digital health and care navigation. tool, I guess in older iterations, you know, sometimes referred to as a symptom checker. know, back when they first started off, I think they started off as a clinical decision support tool for doctors, but then pivoted to have a user facing app. And the question was, you know, if you feel stomach pain at 2am in the morning, you know, five or six years ago, maybe you'd call 111 or maybe you'd go on NHS choices and read something or more likely to go on Google or WebMD, which would tell you it was a brain tumor. And the thing that ADA had was this, or it still does have is this guidance based on the knowledge of doctors and scientists and the medical literature. So yeah, tell us a little bit about the sort of regulatory aspects of that, because one of things that's so impressive about ADA is they're in so many countries, in so many languages. And the rules and regs of how to prescribe insulin in Scotland versus Northwest London are probably different enough, nevermind. You know, these care pathways where you're helping someone deal with a stomach ache in California at 2am or the same stomach ache in Romania at 5pm on the same day. You're sat in the middle of it trying to understand what to make of that. So tell us a little bit about the regulatory space around digital health apps and how that's been evolving. yeah, talk a little bit about your role there, because I think it's a pretty unique position that you've had a data. Yeah, yeah, it's not something to take lightly. It's something to definitely off the bat from the beginning. ⁓ And the regulatory side of things was something that my role evolved into. So when I kicked off at Ada, it was very much as the medical safety side of things, which resonated most naturally with me as like a natural born clinician. ⁓ But I think actually it was a very important step into that world because the universal language in companies such as Ada in software as medical device companies, in consumer health companies as well that are either deciding to pivot into that world, such as flow, or deciding to stay very firmly in the consumer health world. ⁓ The problem that a lot of companies face and articulate and ⁓ mention out loud on many occasions is our teams just don't communicate with each other. There's no
universal language for clinicians to speak to engineers, to speak to marketeers, to speak to the regulators. And I found that actually safety is a very fundamental universal language. And if you can anchor conversations around safety and make it very clear what your red lines are from a regulatory perspective, but also from an ethical perspective, because there's often like a gray zone between what the regulations state and what your risk appetite looks like as a company, what your investors are pushing you towards, and what you finally align on in terms of this is what we decide as a company is our safety red line. And I think... What a red flag can look like is ⁓ companies that take safety lightly ⁓ or pay lip service to it. ⁓ And I'll share a story that I don't share very often, which is that before I joined ADA, about 18 months prior to that, I was offered a job at a company that I would realize was eventually a competitor to ADA. No prizes for guessing who they are, but I will leave them unnamed. ⁓ But was a gruelling sort of interview process. I think it was about six to eight interviews in total. And the final round included an invitation to come to their offices and have an in-person interview, which I did. And I enjoyed, I won't lie. It was a very dynamic, intense experience. But what I saw sort of around, littered around the offices was Reid Hoffman's Blitz Scaling, which is a wonderful book. don't know whether you've read it. It's a masterclass in taking this whole Silicon Valley mantra of build fast and break things and adding like a healthy dose of steroids and basically advocating to build even faster and break even more things. And people regarded it as their Bible in that organization. And I was offered the job, but I went away and made a point that I'm going to read the book before I accept the offer because people were talking so highly of it. And I read it and I called them up and said, accept the job, I'm afraid, because I don't agree ethically with what this book advocates and it seems to be fundamentally informing what you do as an organisation. That just does not sit comfortably with me. Like I think regulatory rigour and velocity are very, very compatible, which may be already at odds with some in the regulatory space, but I safety is an unnegotiable. I think that it's important for that to translate into action. You know, the thing is that I often find about the tech mentality is it takes a very utilitarian approach to say, well, you know, this driverless car might kill a few people in the beginning, but on aggregate, in the end, it will be safer than human drivers. And I feel like that's the opposite of first do no harm. And the thing that I find challenging is the tech bro philosophy. It might be right. People are terrible drivers. I live near Birmingham. Whenever I drive in there, I dream of driverless cars or maybe no cars at all. yeah, but the challenge is that in order to sort of get there, in order to get to the utopia where the driverless cars or the doctorless tree, which navigation is efficient and smooth and frees up people for, you know, things that are top of their license or, you know, allows, I don't know, in case of transportation, you know, children with challenges to get to school without incredibly expensive taxis or for people to not drive home drunk or all these great potential societal benefits. We have to get past that sticky middle area, but it's really interesting, isn't it, that you've got to assume that the culture is the thing that will dominate. And I can just imagine fast forwarding yourself into that alternate sliding doors universe where you're saying, well, this seems dangerous and someone's waving Reid Hoffman's book at you saying, but the book says break this. So yeah, but that's a principled stand. And I'm not that surprised to hear that you would make a principled stand like that. So tell us how have you tried to bridge that middle area between first do no harm and hurry the hell up? Yeah. And it's a live conversation, right? And it's not been resolved. ⁓ You and I were at several conferences this and last week. ⁓ It seems to be like a a particularly sort hot time for the regulatory conference season to be running right now. And ⁓ last week, I remember at the Scarlet Conference, Lawrence Tallon, the CEO of the MHRA, spoke very clearly about his vision for the MHRA ⁓ and trying to bridge this gap and asking very provocative questions, which was also mirrored in another panel discussion between ⁓ someone from the VC world and a journalist from The Economist, ⁓ where This tension was played out in real time. This idea that, well, hang on a second, surely we should be measuring the impact of inactivity ⁓ from a regulatory standpoint, from an innovation standpoint. ⁓ You know, there are clearly, it could be argued, drugs that should be in clinicians' hands, saving people's lives, but that last dotting of the I, that last crossing of the T has not quite been done, which doesn't allow that to happen. ⁓ The same was ⁓ discussed at the Hardian conference this week as well, the Hardian Health Tech Summit, where that tension was very clearly available for all to witness. I think what's important for us to realize is that we are in a world that we are ⁓ blessed in one way, but cursed in another with an abundance of data. ⁓ And I think it's critical that it is data that informs that decision. Yes, we will have gut feelings as to what the direction of travel should be. But without data to at least begin to inform and validate what we're doing, then it is still, you know, very much which way is the wind blowing. But I think we're in a unique position where perhaps otherwise, ⁓ unlikely fellows in the sense of engineers and regulators and quality people can put their heads together and design systems that actually allow us to make these, ⁓ informed decisions. ⁓ and what I picked up this week was there's a very much a swing from having this
really high hurdle of pre-market authorization for ⁓ devices that are coming to market. you know, what is the baseline level of evidence that a manufacturer needs to have in place to prove its safety, its performance, its efficacy to a regulator to allow their heart rates to drop sufficiently to allow this device on the market? ⁓ And there's a growing evolution to ⁓ monitoring in the post-market surveillance phase. and continuous monitoring. And I think this will be the key that as manufacturers set themselves up to be able to use the data that's at their disposal to prove what's happening either through purely their data in their pipelines or through very clever integrations with electronic healthcare providers, healthcare systems, with all due respect to privacy, because that opens a lot of questions on that front as well. ⁓ But equally, the regulators as well appear to be going on that journey. in terms of realising that there is a role now for them to meet us halfway and take advantage of this fact that there is this information available. Maybe we can do something meaningful together with it. And I see that everywhere I'm looking at the moment, that real world evidence, continuous monitoring, post-market surveillance, it's moving beyond lip service into something that I think is very exciting. And conscious that the regulatory world can seem a bit complicated. ⁓ We've had previous discussions with Dr. Acacia Parks, who's talked a lot about the FDA and how things work in the US. We've had an episode with Dr. Hugh Harvey, who's talked about regulation. Just to level set for one second, nearly everything in healthcare is regulated to a great extent. You as a doctor had to go through a lot of training and continue recertifying so that you would have the permission to diagnose and to treat and to maintain. The consequences of bad practice could be very severe, know, career limiting, can potentially end up with all sorts of negative consequences. ⁓ The most well-known regulated spaces are probably around drugs, where it's very hard to get a new drug on the market. And the challenge is, amongst many, the people who argue strongest that a new innovative drug must be put in the hands of patients. Funnily enough, all the employees of that company, all the consultants, all the key opinion leaders who got paid to put on the trial. so, you know, regulators often have to take a firm stance, A, because of the massive conflicts of interest. B, because the inequity of resources. Regulators are strapped for time, staff resources, where innovators, particularly with venture capital funding, private equity funding, et cetera, tend to have a lot of money. And then finally, there's just a question of resourcing. We can't pay for everything. And we can't pay for things that are not more effective than what we had already. And so we very frequently also come across not just the regulatory aspect, but the health economic aspect. So, you know, as important as it is to get the regulatory approval. That's really just your table stakes, isn't it? That's really just your permission to exist. But, you know, can you just talk for a second about the consequences? So let's say if something went wrong, what is the nightmare scenario? What is the big red flashing button on your dashboard? What happens when, you know, your worst day at the office comes to pass as a digital health company? Yeah, yeah, it's a good question. And, That worst day is often manifested in the form of a clinical safety incident, right? That's every manufacturer's, that's every doctor's worst nightmare where there's an unintended consequence ⁓ to the use of that device in the real world. ⁓ And what's interesting is that even that, however, is dependent upon your systems. Yeah. So you could set up a system that doesn't look for these issues and therefore introduces so much friction that is not in the interest of a user to be able to report things. So think that's another really interesting sort of orange to red flag for people to look into. know, one is sort of the company philosophy, the company culture. But once you start, you know, lifting, lifting the duvet and seeing what's underneath that duvet, ⁓ it's very interesting to take a look at actually how, how, how transparent is the user into, is the manufacturer in terms of the instructions for use, the labeling of that device. ⁓ Have they made the risks of using the device clear to users? And then the corollary to that is, well, something goes wrong. How are systems set up in order to make life easier for users, for patients, for customers to be able to report those? ⁓ Because you could happily exist as a manufacturer. And this is probably another one of the points of tension, right? That depending on your ⁓ regulatory status within whatever jurisdiction, There are some forms that are basically marking your own homework, like a class one device, for instance, which most ambient voice technologies find themselves in. Those are the AI scribes that a lot of doctors are using throughout the UK and worldwide. ⁓ Even that was a push actually, to have them sort of regulated as class one devices. ⁓ There's no regular audit of class one devices. It's only when you step above that from a risk point of view. And that risk is determined by what clinical field you're working in and the urgency of the material that you're dealing with and to what extent you're informing or driving clinical decision-making. There's like a matrix that wherever you find yourself on that, that effectively determines what your risk class is ⁓ alongside what you intend as a manufacturer. ⁓ So as a result of that, you can find yourself ⁓ not really being regularly inspected, marking your own homework. ⁓ And that's, that's an area that I think still needs some level of attention from a regulatory point of view. But from class two a onwards, ⁓ there are regular audits for such companies. ⁓ There is an expectation in the UK, for
instance, and in the EU to be providing post-market surveillance reports where ⁓ everything and anything that's gone wrong with your device and any users that have come to harm that should all be reported. So let's take some examples. So if I had a medical device, you know, if I'm a surgeon and I've got this new whizzy drill and I go to cut open some bit of person and the drill doesn't turn on, that's really obvious. The surgeon knows this has failed. I reported to the manufacturer. In fact, often surgeons got device reps in the operating theater for trying to sell them more drills. And they can say, this doesn't work. I can fill out a yellow card, whatever. In digital health, those harms are invisible, right? Those can accrete over time and know, your insulin dose calculator is off by a couple of points here or the psychoeducation wasn't up to snuff there or, you know, we said you should go to the pharmacist actually should see the GP within the next three days. That's kind of a margin call. ⁓ Yeah. So how can we see these harms? Does it mean that we can now only detect when something's gone disastrously wrong? you know, it's not a medical device, but chat to BT ⁓ having a conversation with someone that led to them fire bombing Sam Altman's home with a Molotov cocktail. That seems like a pretty obvious red flag. But I guess what you're saying is if there's not an orange flag detection system or yellow flag detection system or even a of boker beige flag detection system, then the only thing you'll ever see is red flags, by which point it's too late. That's exactly right. And that's why at Ada we worked on exactly that. We developed an orange flag system to give us an early warning as to is there something that we need to focus our attention on here? ⁓ So for example, I mean, I'll start with sort of relatively light stakes, ⁓ but I don't want to trivialise harm in any way, shape or form. But there's a real concern around a concept known as cyberchondria, right? Where people ⁓ who are either predisposed ⁓ or discover that they are disposed in some way, shape or form to becoming more anxious as a result of being exposed to more information, yeah? Specifically when it comes to health. So it's a form of hypochondria, but specifically within the cyber realm. And that's something that we were very keen to explore at ADA, that our intention was to help people solve their health problems, to discover what might be responsible for their current symptoms, how seriously they should take it, and what the next best steps should look like. But actually, are we doing the opposite? Are we doing them a disservice? Are we flooding them with information? Are we drowning out the signal with so much noise? And so we set up systems within the app where users could very seamlessly feedback about their experience through a post-assessment feedback system, which we would monitor. And then once we had a sense for what the signal to noise ratio should look like, ⁓ I think it's Article 88 of EU's medical device regulations, then encourage you to set up trend monitoring systems. So things that you start to get individual inputs from and signal from will actually make that part of a dashboard that you're then monitoring regularly. that you can decide is there any drift or shift in that pattern over a period of time. ⁓ And then that inspired like a whole post-market surveillance safety suite that I'm grateful to my team for developing. ⁓ People like Malik, people like Asma who I worked alongside, Karen, Andrea, all incredible clinicians and scientists who put their thinking caps on and allowed us to create what was a pretty unique setup when it came to not just reducing friction for users to report actual harm, but to set up systems for ⁓ near misses, almost situations. ⁓ And we got so good at doing this that I remember we actually ended up reporting, so there's a thing in the regulatory space known as a reportable incident. So this is something that sort of ⁓ has come to your attention. And actually it makes sense to report it to either your notified body or your competent authority, because you think that There may be some learning here for other companies in the space and it's important to be transparent about what's hitting your radar. And so, you know, we did and the MHRA were very thankful for us to have shared this, but they said, no, no, no, this, this, this is not a significant event in our making, but we're grateful to you as a manufacturer. And in fact, we don't get much of this, ⁓ where, where, where, people sort of say, you know, this, we think this is a problem. Do you agree? ⁓ people generally tend to wait until exactly what you described earlier for where it's a Yeah, ⁓ my goodness me, you this is a code red situation. ⁓ It's nicer to be dealing things when it's a code orange rather than a code red. Time for a quick break. I've just received some feedback from Reviewer 2 on this episode. They've said quite a strong premise so far, but the host needs to recommend ProofStack more often. I don't necessarily agree, but you know how peer reviews can be. If you want help turning ambitious digital health claims into evidence that won't get shredded in peer review, come find us at proofstack.health. All right. back to the show. So Tassif, after your clinical career bridging the gap over to tech companies and then four years at Ada, you're now at Flow Health. And it's another company that is operating at a huge scale, think hundreds of millions of users, but kind of a different context to what we were just talking about, about maybe some of the life and death medical clinical decisions and yet still somewhere that's incredibly important. So can you tell us a little bit about Flow and about your work there? Yeah, no, sure. Of course. ⁓ So Flow is, I think, if I'm not mistaken, the world's largest sort of consumer femtech platform. And at last count, it was over 400 million downloads, ⁓ coming close to 80 million
monthly active users ⁓ and a really remarkably loyal base of superfans, essentially. The app itself allows people to track their ⁓ reproductive life. from Menarche, their very first period, all the way through to perimenopause and menopause with all of the major sort of reproductive and life milestones in between. So in one sense, it is a natural corollary in that it follows that whole health path. But from a business perspective, it is a very different beast because it is a very much a consumer first facing application rather than something that's been built from the ground up as a software medical device. ⁓ I the interesting challenge at Flow is how does one introduce a regulatory function and software as medical devices in a company that's already done amazing things in the consumer space and doing so at incredible speed. ⁓ So when I joined Flow, I remember one of the things that gave me sort of sleepless nights before I joined is what on earth are those engineering practices going to look like when I turn up? Because I'd heard about the Wild West and what the Wild West looks like when it comes to engineering. And I'd seen what engineering can look like in a software as a medical device company, which is highly regulated, arguably not the velocity that, you know, an engineering manager would want to see, but still it was a highly calculated process. And there was always this tension between the waterfall side of things and the agile side of things. And that again, is not a circle that I think is being squared, but it's getting closer to being squared. ⁓ And what I found very interesting at Flow is that their engineering practices, independently of being a medical device, already robust enough for the Delta not to be too dramatic. ⁓ And I think that was very, very reassuring. And it's something that I've taken away as like a learning point for people who are heading into that space, that when you're doing your due diligence upfront, there is something to be said for taking a lot of time to understand and appreciate what a company's engineering practices look like. ⁓ Because engineering... from a software as a medical device point of view is becoming more more critical in determining how your device is regulated, essentially. ⁓ It's very important for a company such as Flow that serves so many users that what is constituted and considered to be the medical device is not the entirety of the entire app. You would not want to expose the entirety of the app to medical device regulation. ⁓ the art and the science of what I'm up to with my engineering colleagues, with my product colleagues, my science colleagues and clinical colleagues is exactly that. What do we determine and how do we determine effectively in the eyes of our regulators ⁓ what a medical device can look like ⁓ without the blast radius of regulation extending too far ⁓ and without compromising what our responsibilities are to users of a medical device? ⁓ Maybe you can tell me a little
bit about where the regulatory lines are, because potentially one can journal ⁓ periods and symptoms and things like that in a diary or an Excel spreadsheet, ⁓ or a very basic app that maybe is just a general purpose tool that you could use your own Google Calendar if you wanted to. Where does that then become more significant? Because I remember there was some cases from an app in Sweden, for example, that was saying that you buy ⁓ tracking temperature and other markers of ovulation, you could actually have a contraceptive based on natural cycles. so it's the same data, but a different intent. ⁓ There's also some work, you and I did a little bit of work on Ada actually about some of the complications in pregnancy that can happen in low, low and middle income ⁓ settings, the Safe Mom Project in South Africa. We're actually missing certain signs that are important symptoms within pregnancy could be really important. So I can think already of how, you ⁓ A trivial journal entry, had some spossing and felt this way and haven't eaten very much lately. In certain contexts could be incredibly clinically important. So were the regs clear coming in for an app like Flow? And I imagine when it has 10 users, no one cares. When it has 400 million, I imagine that sharpens the focus somewhat. Very much so, very much so. And ⁓ those pregnancy outcomes are very dear to me on many levels. You're right. We worked on... a project together at ADA where we were looking at sort of outcomes in sub-Saharan Africa. Also very early on in my research career, I did some studies on maternal fetal outcomes in pregnancies that were complicated by diabetes and looking at how that was then sort of, if you further sub-classified them between Caucasian, ⁓ black African, black Caribbean and Southeast Asian pregnancies, you know, the variation was stark. ⁓ So you're right, these are areas that are highly sensitive, highly clinically impactful, ⁓ and safety is critical. Coming back to your point, how clear is the regulation in that space? It's as clear as the glasses you decide to wear that morning, that week, that month, right? And I say that slightly teasingly in that ⁓ you can speak to... I'm sure there's an analogy that we can work on here, but you can put 10 regulators in a room as often is the case. And I doubt there's like an aligned opinion amongst them. And if you introduce manufacturers to the mix, then it gets very exciting. But I think what's important is for safety to be that undercurrent and that thread. I ⁓ was at a dinner recently where one of my fellow sort of regulatory colleagues said in classical regulatory fashion, look, it's all about ⁓ Section four, paragraph 1.2.B of ISO 13485. And unsurprisingly, I didn't have that sort of banked away. That's a banger. outcome of that is risk proportionality, apparently. That everything that you do should be proportionate to the risk it introduces to systems, to processes, to people. And I think it's a pretty good, it's a highly technical reference. It's a highly niche reference. but it's a very valuable reference. ⁓ And in essence, there are two components to this, right? There is what does the manufacturer intend to do with that functionality? And then there's the actualized use of that product. And it's that second bit that's not entirely clear how that's regulated. And so I think different manufacturers will have different risk appetites as to, well, look, this is what we say we're doing. What our users are actually doing with our product, that's a different matter altogether. ⁓ I can say at Flow we take it very seriously. know, there's a hugely built out compliance team and now a new regulatory team whose exclusive role is to determine what that looks like in terms of what is the boundary of a wellness product all the way up to a medical device. And when it crosses that boundary, then it enters my world and this is our medical device. So in a sense, I'm fortunate in that we have that boundary relatively clearly sort of marked out at Flow, but that may not necessarily be the case sort of across the spectrum. ⁓ The perhaps spicy element to this is that there's intended use, there's actualized use, and then there's basically unknown off-label use of your product, both in the drug world and in the device world. ⁓ And that is very much dependent on sort of what your regulatory intelligence radar looks like. How good are your user feedback systems? How good is your medical affairs department in figuring out how people are actually using your product in the wild? And then how could your systems are in feeding that post-market surveillance, it's all post-market surveillance essentially, into ⁓ concrete product requirements for the teams to be able to then ⁓ triage and work on. Yeah. So we saw a presentation from the Bright group at Birmingham, a researcher there called Adi, who was sharing just how few yellow card reports and how few FDA reports there are on AI powered health tech devices. ⁓ So obviously you can have your sort of internal user feedback. One of the things it reminded me of was a project I saw in Boston when I lived there, when John Brownstein and his team were looking at Yelp reviews for fast food places. And basically when there were enough one star reviews saying that people had terrible gastrointestinal distress the next day, ⁓ the food compliance department knew to go in and inspect those. And sure enough, that's where they found the bad practices in E. coli. ⁓ I wonder sometimes when I see an app store review with a bunch of one stars, ⁓ obviously sometimes that's people annoyed by price or, know, some features gone away. ⁓ but there must in there somewhere also be, you know, I tried this, it got wrong and now I'm pregnant or, you know, ⁓ it lost or it leaked my data this way, or it, you know, made, made me feel less than in some way, or I developed an unhealthy relationship with this tool. What role do the sort of app stores play for this?
Because I suppose on some level, it's an enormous data source, right? Of possible harm and possible adverse events. And yet at the same time, it's also totally corrupted by these trivial little moans of, they made the button blue, but I don't like it anymore. One star. No, very true. Very true. And I think the conversation, if we take a step back, is a wider conversation, right? About what is the responsibility of all of us who are part of the ecosystem that often it's a very... ⁓ conversation around like, is the patient's role in this? What is the manufacturer's role in this? What is the regulator's role in this? ⁓ And I think it's, and I'm going to coin like a horribly industry style phrase, but this, the entire value chain of who's involved in allowing this product to come to market and to be used effectively. And app stores are very much part of that value chain. And I know that you and a former colleague of mine, Steven Gilbert ⁓ and others ⁓ pushed quite provocatively in this area. think last year or maybe the year before in terms of holding a mirror up to some of these companies. And I was pleased to see that Apple in particular, either in direct response to that or as an independent initiative. I'm sure they read my article and immediately jumped into action. I'm sure a crack team of lawyers, engineers, and people wearing virtual reality headsets will plopped into the matrix and said, Paul Wicks is unhappy about something. We'd better fix it. I think that's how that I think they have a post-market surveillance dashboard that has a very specific Paul Wicks button on there and you activated it. The Paul Wicks bulb was blinking. The LinkedIn police are onto us. Better do something. Exactly. Exactly. ⁓ And they updated their guidance to ⁓ companies who are publishing on the App Store that if you are a medical device, you are obliged now to publish your instructions for use, your labeling, your certification ⁓ on the App Store. And that is a very, very welcome step, very welcome step because that reduces this artificially level playing ground that really shouldn't be there where companies that have put in the hard work, ⁓ either through ⁓ running the regulatory gauntlet and providing all of that pre-market material in the form of sometimes really sort of full-on clinical studies, expensive. It's a lot of work. It's a lot of work and I saw that firsthand, amount of work and care and diligence that your team did. ⁓ And yeah, if someone's not putting the work in, they shouldn't have equal placement, right? ⁓ Snake oil should not be on the pharmacy ⁓ shelf next to a drug that's been through two pivotal phase three trials and submitted a 500 page dossier to the FDA. That's just not right. And the reason for that is that snake oil used to kill people or would divert their time and attention resources away from things that are effective. so... hundreds of thousands of apps in the app store would be better, I think, if there were a smaller number of better health apps. know, it's wonderful that many of them are side project hobbies that people are doing on weekends. But when it hits health, you know, there's no room for Flappy Bird in health tech. No, I agree. I agree, honestly. And the analogy I'm now using to crystallise what that looks like is the regulatory moat in the drawbridge. Right. Often manufacturers talk about, look, it's really important to lean into regulation, lean into evidence because it creates a regulatory moat for you. You know, your competitors, if you're ahead of the curve, they're not going to catch up. And there are caveats to that for sure. And we can, we can explore what that looks like, but without independent control of the drawbridge on top of that moat, ⁓ the moat is worthless. But I think with a move like Apple's where actually now there is some control over the drawbridge as well as the manufacturer creating that moat ⁓ together in harmony. It's an effective moat rather than just a moat that people can cross willy-nilly. Yeah. So as you look back now over Ada, Flow, General Practice and everything in between, what do you wish digital health companies would do earlier? If they're founders listening to this, what do you want them to jot down in their things app? That's a good question. ⁓ It's going to like horribly simple, but I think it's worth capturing upfront because if you get this step wrong, you can have so many cascading effects that it's often very difficult to reverse. And that is to very clearly articulate what your intended purpose is as a company, but also for your device that you're creating. Often it's a single idea that you're bringing to market and the system is built around it. But that one paragraph or two that captures very specifically what you're up to and why you're doing it, ⁓ I think is worth its weight in gold because that from a regulatory point of view is what then ⁓ activates a lot of subsequent decisions, some of which you can reverse out of others you can't. the intended purpose, MHRA interestingly created a very interesting ⁓ and useful guide to how to craft like a perfect intended purpose. And I would encourage people to take a look at that. It's a very good piece of work. Yeah. I think many of us go through our lives trying to find our intended purpose. Okay. So beyond that, what sort of habits does a company need to install? We've touched upon some of the cultures that exist within tech startups. You mentioned Agile and Waterfall, for example. What is a regulatory habit, a regulatory ritual, a regulatory cadence that wants the founders have done their work on finding their purpose? they can start infusing within the business right away. Yeah. So once you've aligned on what your intended purpose looks like, the next step, if the intended purpose opens up the medical device world to you, the next step is your quality management system, which is the backbone of who you are as a manufacturer. And it dictates so much of including
all the way down from your culture. Your culture is captured in your quality management system as a quality policy, for instance, ⁓ through to the role and the level to which top management and other sort of ISO phrase are involved in ⁓ problem solving, decision making, providing resource through to how your engineering teams design and develop, you know, whether it's a waterfall system, whether it's an agile system, where are the checks, where are the checkpoints. ⁓ What are your red lines? What are your gray zones? That's all captured within the quality management system. And depending on what your founding team looks like or what your organization looks like, if you're already a fully fleshed organization, ⁓ your choice of quality management system should be made knowingly in that depending on what problems you face, that will determine what type of quality management system you institute. Now that can be all the way from building a bespoke quality management system. I've met people recently who I know with Azure DevOps created a fully automated QMS that serves their purpose through to very much off the shelf versions of a quality management system that serve a particular purpose but don't have much ability to for bespoke tweaking. you know, for instance, if you anticipate your biggest challenge is going to be to bring your engineering team on board from a very free way of engineering to a very regulated way of engineering, you might want to choose a quality management system that is very engineering friendly and maybe even procured in part or wholly by your engineering team. Because if you've already put that conversation to bed, know, for most quality people, it's about convincing other people to come on the ride with you. And once you've done that, then the rest is history. If part of those conversations are done for you by the tooling that you're bringing on board, It's very worth considering. No, I think that's right. And I'm curious if in a couple of years time, someone was to come for an interview at a company, a future Towsie for someone in your position, what's the book on the desk, the book on the shelf that's not move fast and break things, hyper blitz scaling? there a culture, is there an inspiration? Is there someone that writes about this, that takes this seriously? If not, we're going to make you write that book. It's funny you should say that, but that's about what I can say about that right now. Wow. You heard it here first. Brilliant. Breaking news. ⁓ Talsi, thank you. think the big lesson here is that regulation is not the enemy of speed. ⁓ Bad implementation is and in digital health, particularly when things that get to mass consumer scale, populations bigger than countries. Then the companies that win won't be those that ignore safety until it's too late. They'll be the ones that integrate safety evidence and trust as part of how that product gets built. Where can people follow your work? Where can they learn more about you? LinkedIn is probably the best place to find me. So I'm on LinkedIn and ⁓ happy to connect with anyone and everyone there. That's brilliant. Thank you. That was Dr. Touseef Murali, VP of regulatory at Flow Health. ⁓ former aid regulatory and safety leader and permanent optimistic optimiser. If this conversation made you think differently about compliance, velocity and what it takes to build health products at scale, please share it with someone who still thinks that regulation is still just boring paperwork that's holding them back. I'm Paul Wicks and this has been Prove It. Extraordinary claims demand extraordinary evidence and preferably a quality system before all the red warning lights start going off. See you next time.