In a wide ranging conversation, Sleep Junkies talks to Colin Lawlor, CEO of Sleepscore Labs, about sleep, digital health, artificial intelligence and more.
Colin Lawlor is concerned.
On an early morning trans-Atlantic phone call, the SleepScore Labs CEO ponders, in his buoyant Irish brogue, how recent high-profile failures of several consumer sleep-tech firms may have rattled the nerves of those considering investing in the future of sleep technology.
Colin Lawlor: “It’s kind of a funny story, you know. I’ve been up to Sand Hill Road, and we’ve knocked on the doors of some of the biggest VCs [venture capital firms] in the US. And I’ve said, listen, I want to talk to you about sleep….but the first thing is, I do not want any money.
And they look at me like I’ve got ten heads. They say: ‘you don’t want any money?’
I say no, I don’t want any money. Here’s what I want to do;
… companies like you have bank rolled sleep technologies like Hello by Sense, Zeo, Provent and Theravent and between those three companies, have collectively lost somewhere in the region of £160MM to $170MM.
And we are very concerned that because of the recent high profile failures that investors may stop supporting sleep technologies.
And we think that would be a travesty because there are many sleep issues that need solutions and we want to encourage and support those. So we’re saying to you, the VC, if you want help to assess something that’s in the early stage, we can help you.
We have the scientists, we have the experts, we have the team and we have the technology to measure it. And let’s work with you and support you. We don’t even want a cut of the deal necessarily. We just want to increase your chances of being successful in bringing new solutions to the market.
We were not looking to control all of this. We’re actually hoping to stimulate, but we want to stimulate serious players doing real things that we really want.”
Partners in sleep
At first this trio may seem unlikely bedfellows, but the company’s structure and playbook actually makes a lot sense.
ResMed is the largest sleep medical devices company in the world, and as a result of connecting all of those devices in recent years, they now have over 1.5bn nights of sleep and breathing data.
Dr Oz brings not only huge marketing reach, but he also has a record of being a successful serial entrepreneur in the health and wellness space, including the online platform Sharecare.
Pegasus Capital is a less obvious choice of partner for a sleep technology firm but as Colin explains;
CL: “The interesting thing about Pegasus, a lot of people asked me ‘why is ResMed partnering with a private equity firm?’ And it’s true; it’s not an obvious thing for ResMed to do. I mean, it has very substantial cash resources so it doesn’t need a co investor.
And so it’s partnering with Pegasus because if you think about it, ResMed is the global leader in funding sleep research. It funds the Harvard Chair of Sleep Medicine, the Farrel Institute at UCSD, it funds research all over the world and the best universities with the best academics. So globally it is the biggest investor in sleep research anywhere.
But, Pegasus has built this amazing network of people who intersect with sleep but who aren’t sleep experts. So, for example, Pegasus introduced us to the Harvard professor for Circadian Rhythm, which we’d never really focused on in Resmed, we tended to focus more on pure sleep and breathing disorders related research and there’s a whole group of people working on research programs in other areas, for example, ADHD with kids, for example, fatigue programs and for example, astronauts and so on. So what actually happened was they [Pegasus] brought a whole bunch of additional related expertise to us.”
Over 10 years in the making
Even though SleepScore Labs was only launched in 2017, the origin of their first product, SleepScore Max, goes back more than ten years. The technology was borne out of a medical research project. Much later found its way into a consumer sleep tracker called the ResMed S+. (You can read our full S+ review here)
CL: “It all started with the formation in Dublin of a technology company which was called BiancaMed and those guys set up a company which was spun out of University College Dublin and they were really interested in using a variety of different types of sensors to measure sleep.
They actually started out by using ECG to measure sleep and sleep disorders [but] they quickly realized that the challenge of needing to place lots of electrodes on people’s bodies to measure stuff, wasn’t going to make the technology easy to adopt or use in the long term.
And so they set out way back then over a decade ago to find better technologies to measure or to use or to sense [sleep]. And so they found a technology which they initially liked and then began to develop, which was based on radio frequency sensing.
A number of partnerships were entered into and we went through a series of stages which brought us to the point that ResMed acquired the business in 2011.
So we then launched a product called S+ by ResMed. [Read our detailed review of the S+ here] That product was launched in 2014 and the product aimed to be – and still is – the most accurate way to measure sleep outside of the sleep lab using the radio frequency technology.
And so as ResMed’s person -I was the Vice President for Global Consumer Strategy, my role was to figure out how to and essentially approach and the wider opportunity so that we could figure out how to make a wider difference and essentially establish for ResMed an opportunity to learn about how to be successful in the consumer sleep wellness area.”
CL: “And so what happened was the S-plus went to market and it was quite extraordinary. The product sold – enough to collect a lot of data, but it certainly didn’t sell as many as we would’ve liked.[But] what it actually also did though was it made a lot of other people aware of the fact that there was a company called ResMed.
Today, ResMed is a $14bn market cap company. They treat more than 13,000,000 patients every night somewhere in the world with one of their technologies. And it’s surprising how many people have never heard of the company.
As a result, some of the best companies in the world started knocking on our door because they were all looking at the sleep space at the same time.
So as we began those conversations, what we found was that there were a number of those companies we felt we could really work with and do something really amazing together. And that’s kind of how this all came about. One of those was Dr Oz and the other was Pegasus Capital.”
Introducing SleepScore Max
So having established the background of the company, I asked Colin about their first product release, a standalone sleep sensor and tracker based on the same technology used in the ResMed S+.
So what exactly is the SleepScore Max work and how does it work?
CL: The Max itself consists of a hardware device that pairs with a downloadable app. The hardware is a frequency sensor and it scans the body 16 times per second and it sees every movement down to just 1/10th of a millimetre. Which means that it’s so sensitive, that we can see the pulse for example.
But we don’t use the pulse because we don’t see it consistently all night long – in fact we are working on this R&D-wise as we speak.
What we do see and use is the respiratory pattern. So from a measurement point of view, it is equivalent to the gold standard in respiratory monitoring and because the device is essentially detecting all movement in your upper body we’re able to see all of the things that an actigraph would normally see [ie when] your wrist moves or your leg kicks or your central torso moves.
So it’s a highly, highly sensitive bio-motion signal collection device and it sits beside your bed and you don’t need to touch it. Basically you put it within arm’s reach, point it towards your torso and then you forget about it.
In addition to the radio frequency sensor it also has a temperature and light sensor built in because we know how room environment are important factors in sleep quality. All of that data is then streamed to a cell phone or a smart device, a tablet which collects data and insights.
Our algorithms convert the data every morning into a sleep score out of a hundred . The higher the score, the better your sleep is. Our sleep score profile is built on data collected from the Ohayon sleep study in the early 2000s which is the largest meta study of true sleep in general consumers. We provide the score based on the best available clinical reviews [but this] is the just beginning of what we do with Max.”
Turning your sleep data into something useful
Anyone who has used a sleep tracker will tell you of their initial fascination of being able to see their sleep data for the first time. But after a while, the novelty can wear off.
That’s because with the majority of consumer sleep-tech, you’re not given any actionable insights to work with. Hence the data on its own is essentially useless to most people.
SleepScore Labs are attempting to address this criticism, as Colin Lawlor explains:
CL: So [SleepScore Max] is a whole advice engine that we have built based on a review of over 600 clinical papers and we’ve taken evidence in terms of advice and direction based on all of the major sleep issues that we see in the data. We understand your sleep data and then deliver personalized, customized advice on how to improve your sleep.
If you’ve got an underlying medical issue, for example, well then you really need to speak to a doctor and for that reason we built in a doctor’s report.
So we collect your objective sleep data and we also have you complete all of the key clinical screening questionnaires. And we put that information together into a report that you can take to your doctor if you need to see a doctor.
But equally if you’re snoring and it’s bothering your partner, we will triage that snoring issue and try and help you to find a solution that will work for you. Or if you’re tossing and turning a lot and your mattress is seven years old, we’ll take a look at what mattress you’re sleeping on, and we’ll make a recommendation as to considering changing your mattress.
Point is that whatever it takes to help improve your sleep based on scientific data, we will connect you to the solution.
Consumer sleep tech: accuracy vs mass adoption
One of the biggest challenges for sleep technology companies is dealing with the ‘expectation gap’ between clinical-grade and consumer sleep monitoring.
Polysomnography (PSG) – the ‘gold standard’ of sleep testing. But it’s very expensive and requires an overnight stay in a sleep lab, with dozens of sensors taped to your head and body. Hence, although there are many people who would like to know more about about their sleep, they are prohibited by the cost and inconvenience of PSG.
Consumer sleep tech, on the other hand is affordable, and offers a potential way for millions to learn more about their sleep and wider health. The trade-off, of course is that consumer hardware lacks the technical sophistication of clinical hardware, which has led to much criticism about the efficacy and of sleep monitoring devices designed for the home.
So I asked Colin about how he felt about criticisms from sleep clinicians and professionals who insist that consumer sleep tech has little value in measuring sleep;
CL: “Well the answer is that it depends. The first company to launch a wide-scale consumer sleep tracker was Zeo. And you’ll remember that was an EEG (electroencephalogram) monitor to wear on your head.[EEG] may very well be useful data and indeed it is the best way to measure in particular REM sleep. There is no doubt about it.
But people don’t want to wear those sensors on their heads every night. And in fact they don’t wear them; only a very tiny percentage of people stay wearing them because they are uncomfortable, they fall off and people just don’t like wearing stuff on their heads, particularly technology that’s transmitting data from their heads to someplace else. So people just don’t like it.
From our point of view, our answer is that there are always very perfect technologies which are necessary and useful, particularly in the lab , for diagnosing, you know, a complex sleep disorder.
But for the mass population, we want to help them to improve their sleep health. So what we need are extremely accurate technologies that help us to make a difference to most people – and those technologies need to be sensitive enough to identify people that need medical intervention.
And at that point, perhaps you can have somebody go into a lab and wear EEG and all the other sensors because they need to, but for the rest of the time, it’s not necessary and people won’t do it any way and it’s highly, highly expensive.
Our vision for this is, you know, we want to put good sleep health in the hands of millions, so we don’t want to put barriers in the way to make that happen.”
Sleep data, artificial intelligence and machine learning
One of the big trends we’ve seen in sleep technology is a move away from sensor-led measurements towards insights gleaned from big data and AI. I asked Colin whether he thought there might be a point in the future where artificial intelligence and machine learning might outpace traditional methods of measuring sleep, based on multiple sensors.
CL: “You know what? I spoke on this at a conference recently. I’m so sick of hearing about artificial intelligence because if you feed rubbish data into artificial intelligence, you will get rubbish intelligence.
So the bottom line is [that] artificial intelligence, machine learning that’s trained on rubbish data is essentially gonna give rubbish results and it’s been proven time and time again.
So just to clarify about our technology, we have two critical measures which we use in our algorithms, one is movement just like actigraphy, which has been the clinical standard for many, many years outside the lab.
And the second one is the respiratory pattern. And we believe that those two are the minimum necessary. And when we test them, we test them against thousands of nights of parallel PSG data. So not tens and twenties but thousands. And those thousands of nights are on multiple subjects; age, gender, BMI, ethnic differences, people with disorders, people without.
If you don’t have a massive data set of high-quality PSG data that you have your algorithms and your sensors trained and tested against, you are going to make a guess and you’re going to over-train your algorithm on one study in one lab, which is typically what I see happening a lot.
Our mission is to improve sleep health and to advance the science in the process.. and a lot of studies have been done with 20 subjects and it’s just not good enough.
So my point is that I think that there is no doubt that artificial intelligence could be useful but not by itself, it actually needs real intelligence. Otherwise you’re going to have this algorithm which is looking at rubbish data and it’s then going to deliver a rubbish result.”
Looking to the future
A big part of the SleepScore Labs mission is to nurture commercial and research partnerships that help to advance sleep science and to provide research-backed product and app recommendations to help consumers improve sleep quality with the best technology available.
CL: “So we have been making our platform and technology available to researchers and to companies so that we can help them to truly test their products or services in the real world with real people with sleep issues. And we’re obviously going to grow that and over the next number of years you should expect quite a lot from us and probably every quarter or so we’ve scheduled one or other major product or service, significant service enhancements.
The technology we have is available to be licensed to serious players and we actually have already licensed for some very significant global companies and they just haven’t launched their products yet. But we’re making the technology available because we want the standard to be higher, we think that’s in everybody’s interest, including our own.