Download - The Power of Sensors in health & healthcare
the power of
sensors in health & healthcare
Exploring Digital Health Trends Report #2: The Power of Sensors
In a series of reports we explore key digital health trends and related opportunities
for technology companies, healthcare providers and patients-consumers. We take
both an international and Flemish perspective, the latter based on interviews with
local stakeholders. In this report we focus on sensor-based applications.
A big thanks to Chris Van Hoof (Imec), Julien Penders (Bloom Technologies), Lars
Grieten (Uhasselt), Daniel Berckmans (KU Leuven) and Frederik Horemans (DSP Valley)
for valuable input. You can read our interviews with them on the blog at
www.micvlaanderen.be
Author: Frank Boermeester
About Microsoft Innovation Center Vlaanderen
Our mission is to stimulate ICT-related innovation and
entrepreneurship in Flanders, with a particular focus on healthcare.
We do so by supporting high-potential technology startups. MIC
Vlaanderen has two operational offices in the cities of Genk and
Kortrijk, Belgium.
www.micvlaanderen.be
The Promise
Not so fast!
Opportunities
ahead!
Conclusion
AGENDA
The Promise
Gartner Hype Cycle: A helpful framework
The Promise (or Peak of Inflated Expectations)
Dr. Algorithm is coming
“In the next 10 years, data
science and software will do
more for medicine than all the
biological sciences together.”
– Vinod Khosla, Khosla Ventures
http://techcrunch.com/2013/09/11/vinod-khosla-in-the-next-10-years-data-
science-will-do-more-for-medicine-than-all-biological-sciences-combined/
But sensors underpin the
promise of digital health
Data-driven health
applications require data.
Sensors make possible
automated, large-scale
monitoring & measurement
of health-related
parameters
And what a promise it is, in all health
domains!
DIAGNOSIS
MONITORING
TREATMENT
WELLNESS
Just imagine!
DIAGNOSIS Sensors systems will make diagnosis easier, faster,
cheaper and more accurate. They will also predict
health problems giving people time to intervene
Will X-Prize
contestants like
Scanadu (or one of
the other 9 finalists)
deliver Star Trek’s
Tricorder?
Just imagine!
Doctors will rely on sensors to monitor their
chronically ill patients, providing early warning for any
potential problems
Will chronically ill
patients benefit from
a 24/7 connection to
specialised medical
care, as being
pioneered by
Cardiology at
Ziekenhuis Oost
Limburg?
MONITORING
Just imagine!
Sensors will help improve medication compliance
and will be critical components in physical
augmentation
Will companies like
BrainControl help
people in a locked-in
state to ‘speak’ by
sensing their
thoughts?
TREATMENT
Just imagine!
Mobile & wearable health trackers will help people
overcome their lifestyle-based health risks
Will your personal
digital health coach
become as ubiquitous
as the mobile phone?
WELLNESS
By 2017 the number of wearable wireless health and
fitness devices will reach 169.5 million (90 mil fitness
devices + 80 mil health-focused devices). (ABI Research)
$47billion
44.4%
169million
Great Market Expectations
The sensor market in Consumer Healthcare is expected to
reach $47.40 Billion by 2020 growing at an estimated CAGR of
5.56% from 2014-2020. (MarketsandMarkets)
Approx. 3 million patients worldwide were using remote
wireless devices to display health-monitoring results in
2013. It is estimated that the number will grow at a
compound annual growth rate (CAGR) of 44.4 percent to
19.1 million in 2018. (Berg Insights)
Not so fast! (the Trough of Disillusionment)
User Disillusionment?
PwC reported that mobile health apps show immense user
drop-out: 67% of people who use an app with manual
data entry stop in the first 6 months. Remarkably, for
automated (sensor-based) apps the figure is even higher!
74%.
74% drop out
PwC report Emerging mHealth: paths for growth (2013/EIU analysis)
Tracker Efficacy Questions
“To date, no long-term, peer-reviewed study has
shown that people using activity trackers become
and remain more active.”
– Well/New York Times, March 10 2014
"The durability of the effect is still in question. We
don't have randomised [clinical] trials showing
improved outcomes or durability in influencing
behaviour.“ - Dr Eric Topol, author of The Creative Destruction of Medicine, director of
the Scripps Translational Science Institute (as quoted in Wired.co.uk, 23
January 2013)
Doctor Resistance?
53%
24%
PwC report Emerging mHealth: paths for growth (2013/EIU analysis)
53% of younger doctors (who one would think are more
open to change) worry that mHealth will make patients too
independent. (This compares to 42% for all age groups)
And 24% of younger doctors actively discourage patients
from using mHealth applications. (13% all age groups)
Economist Intelligence Unit research (in PwC Report:
Emerging mHealth)
“Don’t come to us telling us you can upload [data] into our electronic
medical record. We don’t necessarily want it there… Our physicians don’t
want it all there. They really don’t need to know how much exercise each
of their patients is getting on a daily basis; they just don’t have time to
process all of that.” - Christine Folck, Lead Innovation Designer, Kaiser Permanente (quoted in
mobihealthnews article July 9, 2013)
What are you measuring?
Ebola airport screening using thermometers. Photo: Melissa
Maraj/U.S. Customs and Border Protection
Non-invasive,
physical sensors are
inherently limited in
a health context.
There is only so
much that
temperature,
movement, light and
electric current can
tell you about a
person’s health.
The inherent limitations of physical sensors
What are you measuring? From sensor data to health interpretation is an arduous route
Most fitness trackers
and apps rely on a
single sensor: an
accelerometer which
measures force along
3 axes { } An algorithm tries to make
sense of this data, but cycling
doesn’t generate the right
sort of data
“You’re being lazy,
time to get moving!”
And thus makes the
wrong interpretation
about energy use and
physical activity
Real-world behaviour
generates data, much
of which is ‘noise’
Prediction is difficult
In a clinical setting, making complex interpretations and
predictions on the basis of one or two parameters is a
dangerous game.
“Blood pressure alone can’t predict your
health. In our clinical context we deal with
patients who often have multiple diseases
but we only focus on the cardiac issue, and
that limits our ability to predict the general
condition of the patient.”
– Lars Grieten, Mobile Health Unit/Dept of
Cardiology, University Hasselt
Quoted from interview, published at http://www.micvlaanderen.be/word-geinspireerd/door-interviews/
Prediction is difficult
Analytical approaches that are based on ‘snapshot’ measurements
and comparisons against population norms are inadequate. To
improve predictive power one has to model behaviour at the
level of an individual
“Population norms are not that relevant for the interpretation
of individual data; the population average is a purely
theoretical concept that in practice no individual will
comply with. For example, there is little value in comparing
the heart rate of a football player with the team average; you
have to measure variation over time in the individual player
and make conclusions on that basis.”
- Prof Daniel Berckmans, Head M3-BIORES, KU Leuven, research
group specialised in biological response monitoring and prediction
A new analytical approach is required
Quoted from interview, published at http://www.micvlaanderen.be/word-geinspireerd/door-interviews/
Security & Privacy Fears
Hackers have demonstrated vulnerabilities in pacemakers, implantable
defibrillators, and insulin pumps both by attacking wireless connections and
by manipulating sensors with electromagnetic interference. Former Vice
President Dick Cheney had his pacemaker’s wireless capabilities deactivated…
Changing the System is Hard
“You have a research environment
that produces papers, a business
environment that produces
expectations, and a healthcare
environment that creates healthcare.
But so far they have not met. This
will happen, but how long it will take
I’m not sure.”
Widespread adoption of sensor-based technology, especially for remote
monitoring, is hobbled by reimbursement gaps, regulatory issues,
privacy & security fears, poor interoperability and a lack of clinical trials
- Steinar Pederse, CEO Tromso Telemedicine Consult (in PwC
paper: Emerging mHealth: Partners for Growth, 2013)
Opportunities Ahead! (or Scope of Enlightenment)
Smaller, More Accurate Sensors
“The key enabling technologies have
been miniaturization and low-power
sensors. Sensors have got smaller
and smaller and need very little
electric power, and that makes
possible wearable sensors, ingestible
sensors, etc.”
– Julien Penders, Biomedical Engineer,
ex-Imec, co-founder Bloom Technologies
Miniaturization and ultra low power sensor systems are enabling
new formats and better integration with the human body, which
also improves accuracy
Murata's World's Smallest Chip Ferrite Bead (on
the far right) (Photo: Business Wire)
Quoted from interview, published at http://www.micvlaanderen.be/word-geinspireerd/door-interviews/
CASE MC10
www.mc10inc.com
Stretchable skin patch sensor
MC10 developed a flexible, stretchable patch
(Biostamp) with sensors for monitoring
temperature, movement and heart rate. MC10 is
seeking FDA approval and recently announced a
partnership with pharmaceutical company UCB to
explore applications for neurological diseases
(e.g. to better measure response to therapies).
Ultimately, pharmaceutical companies could
bundle such technologies with drugs or therapies
to enable personalised therapies.
Almost invisible. A sensor that is less obtrusive than a band-aid
CASE Imec and Dutch affiliate Holst Centre
www.imec.be www.holstcentre.com
Nanoelectronics research institute
Holst Centre & Imec, with
Shinki Electric Industries, have
developed a flexible, low-
power health patch that
accurately measures physical
activity through real-time
ECG, tissue contact
impedance and accelerometer
data. Furthermore, data is
processed locally and then
transmitted via Bluetooth
Smart.
Taking accuracy to the next level
CASE Proteus Digital Health
www.proteusdigitalhealth.com
Developer of an ingestible sensor
Proteus developed an ingestible sensor
for medication adherence. The
accompanying patch tracks the
ingestible sensor and detects heart rate
and activity. The ingestible sensor
secured FDA clearance in 2012. Proteus
is one of the best funded digital health
companies (approx. $400 million).
Medical grade ingestible sensor & patch
“The information we measure is verifiably accurate and not just consumer-grade
or a toy.” - Proteus CEO Andrew Thompson quoted in MobiHealthNews
CASE Google X
www.google.com
Google X life sciences division
Google X is working on a system for
detecting diseases like cancer early. It
involves ingesting “painted”
nanoparticles that target specific
biomarkers. If they find such
biomarkers they send out signals that
can be picked up by a device such as a
wristband. In another project, Google
is testing a smart contact lens
designed to measure glucose levels in
tears. Early prototypes generate a
reading once per second.
Nanoparticles as sensors
“We’ve done a lot, to be quite
humble about it. Enough to give us
great confidence that this is all likely
to work.” - Andrew Conrad, Google X,
talking about the nanoparticle project in
an interview with Steven Levy, Oct 28,
2014, Medium/BackChannel
CASE Biocartis
www.biocartis.com
Automated, real-time molecular diagnostics systems
Biocartis develops molecular diagnostics
systems that integrate (and partly automate)
multiple clinical tests in a single device. The
systems can conduct diagnostic processes,
from sampling to results in very short time
frames (40-150 minutes) and with very little
hands-on preparation time. Ultimately this
promises faster testing, more accurate
testing (less prone to human error) and more
testing in primary care settings.
Biosensors emerge: automating the lab
Combining Multiple Sensors
“If you take an individual sensor I
don’t see a lot of change there. What
is changing is the ability to integrate
the perspective from multiple sensors
and come to new conclusions because
of that integration, that's what
changing.”
– John Oliver, Senior System Architect,
Intel (quoted from Curiosity.com video)
Capturing and analysing data from multiple sensors improves
predictive power and enables the measurement of more
complex parameters.
http://www.discovery.com/tv-shows/curiosity/topics/j-oliver-how-is-sensor-
technology-evolving.htm
CASE Microsoft Band
www.microsoft.com/microsoft-band/
Microsoft’s smartwatch with multiple health-related sensors
Most fitness trackers have a single sensor.
Microsoft Band includes an accelerometer,
an optical heart rate sensor, GPS, an
ambient light sensor, a skin temperature
sensor, an ultraviolet light sensor, a galvanic
skin sensor, and a capacitive sensor. It also
has productivity functionality like
messaging & notifications. And it connects
to the broader HealthVault ecosystem.
Multi-functional, multi-sensor, and platform based
CASE Sense
www.hello.is
A sleep analysis system that also collects environmental data
Relying on multiple sensors, the Sense
System tracks sleep behaviour (through a
clip-on movement sensor) and monitors
the bedroom environment (through a
nightstand device that senses noise, light,
temperature, humidity, and particles in air),
to take sleep analysis to a next level. Also it
replaces your classic bedside alarm, with
the extra that it wakes you up at the right
point in your sleep cycle. Sense raised a
record $2.4 million on Kickstarter.
Embracing contextual data and functionality
CASE Samsung Simband
www.samsung.com
Wristband sensor module
Samsung (with Imec in Belgium) developed
what is touted as the world’s most advanced
wrist-based sensor module, an open
reference design for 3rd party developers.
The band is equipped with multiple sensors
(optical, electrical, physical) that together
enable new biometric measurements such as
blood oxygen levels (and possibly other
complex parameters such as blood pressure).
The world’s most advanced wrist-based sensor module?
Interoperability & Platforms
OPEN WINS! The Tech Boom is a story of platforms: the internet, the
world wide web, mobile APIs/SDKs, app stores, open data, ... Digital
Health is finally getting the message.
“Data has traditionally resided in silos
belonging to specific applications delivered
primarily by device vendors themselves. New
cloud platforms capable of collecting data from
a range of vendor devices and sharing it
securely with a range of related parties
including patients, healthcare providers, and
payers will drive adoption and bring more
connected devices to market,”
- Jonathan Collins, principal analyst ABI Research.
Interoperability & Platforms
The “Connected Elite”:
Best selling sensors connect
to as many apps as possible
CASE Angel Sensor
www.angelsensor.com
Open health & fitness sensor
Angel is a Kickstarter funded project to
develop an open sensor for health and
fitness. The band has sensors to measure
heart rate, skin temperature, blood oxygen
and physical activity. Contrary to most
trackers which come with a proprietary app,
Angel plans to open its communication
protocols, API/SDK and sensor data streams.
The goal is to attract developers and thus
create a wide range of apps and application
domains
A sensor as platform for innovation
CASE Hexoskin
www.hexoskin.com
Biometric shirt
Hexoskin developed a biometric
(washable) shirt for tracking heart
rate (with ECG precision),
breathing, activity & sleep tracking.
The product is an Open Platform
for developers.
Biometric clothing as Open Data platform
“Hexoskin is an Open Data device.. We have an open web API that you can use to
build 3rd party apps, because we’re not going to build all the health apps
ourselves.” - Pierre-Alexander Fournier, Hexoskin, quoted in Montreal Tech Watch
CASE Samsung Digital Health Initiative
www.samsung.com
Open reference design wristband + open data platform
Samsung’s Simband is an “Open Reference
Design” so that others can build their own
devices and applications without having to
reinvent the wheel. Complimenting the
Simband, Samsung has also developed a cloud-
based and vendor-agnostic platform (S.A.M.I.)
with open API that enables data exchange with
any device/resource. Furthermore, the platform
is equipped with a set of tools for running
analytics on data.
Building an open ecosystem
CASE Human API and Validic
www.humanapi.co www.validic.com
API aggregators
Human API and Validic both aggregate APIs from multiple data sources
(sensing devices mainly) so that app developers can easily integrate multiple
data sources via a single authentication and API (as opposed to making direct
connections to every device in the market).
Letting apps connect to multiple devices with a single API
CASE Intel
www.intel.com
Hardware components
Intel Edison is a low-cost ($50), low-power, stamp-sized microcomputer for
wearable devices and Internet of Things.
Microcomputer platform play
"If you want to build something, doing
your own board is expensive and time
consuming. We've produced a fully
power-managed operating system for
this chip, and you can literally build
something and go to market in
months” - Mike Bell, head of new devices
at Intel, quoted in CNET Sept 14, 2014.
CASE Qualcomm Digital Health Initiative
www.qualcomm.com
Healthcare data platforms
Qualcomm launched two integrated
cloud-based health information
platforms. The 2net platform is an open,
non-exclusive and interoperable platform
that connects with multiple devices and
applications. It enables users and
healthcare professionals to collect, store
and share biometric data securely. On
the front-end, HealthyCircles is a suite of
services for sharing data with care circles,
logging medication, setting alerts, etc.
Building the ecosystem
Making Business Sense
Viable business models are possible by solving real, short-term
user needs (as opposed to long-term unproven health benefits) and
by working with clinicians and providers to integrate (validated)
technology in clinical processes. Remote Monitoring is a key
growth opportunity.
“I think we’re entering a
phase of application pull,
instead of the earlier
technology push phase.”
- Prof. Daniel Berckmans, M3-
BIORES, K.U. Leuven
CASE Bloom Technologies
www.bloom-life.com
Wearable sensor for measuring contractions
To drive user adoption, Bloom targets an
audience with a very specific short-term
need: expectant mothers who want to
know whether they are having
contractions. Also, the company is seeking
FDA approval and will rely on Samsung’s
open data platform (SAMI).
Addressing short-term need to drive adoption
“The product we are developing will be able to measure contractions - that will drive
adoption - but it will also be able to measure a range of other relevant parameters such as
activity, stress, sleep and movement of the baby. We are using those five parameters to
model the user's behaviour and subsequently push personalised messages.”
- Julien Penders, co-founder, Bloom Technologies (Quoted in interview MIC Vlaanderen)
CASE Qompium
www.qompium.com
Smartphone app to detect irregular heart rhythms
Qompium developed a
smartphone app for detecting
irregular heart rhythms. In lab
conditions (using a high end
Android phone) the app has a
93% reliability factor.
No fear of the regulator
“The app definitely tries to make a diagnosis and therefore will be classified as a
medical device. That is exactly our ambition; we want to differentiate ourselves
from the thousands of unregulated apps out there.” - Lars Grieten, Qompium,
quoted from interview at MIC Vlaanderen
CASE Live!y
www.mylively.com
Elderly care monitoring solution
Live!y developed a remote care solution targeting consumers (monthly
subscription fee) using a simple and well-designed set of sensors and a
wristband interface (for sending reminders and emergency response + a
pedometer).
Ingenuous remote care solution targeting consumers
CASE GrandCare Systems
www.grandcare.com
Elderly care monitoring solution
GrandCare integrated multiple (mainly off
the shelf) technologies to develop a
complete remote care solution. The
system includes a touchscreen providing
communication tools, instructions,
reminders and medication prompts. It
also connects with wireless health devices
(blood pressure, weight, pulse, glucose,
temperature) and motion sensors (e.g.
alert if no motion or wondering motion)
Integrating today’s technologies to solve a problem today
CASE Healthsense
www.healthsense.com
Elderly care monitoring solution
Healthsense developed a remote
monitoring solution for senior
living communities. Multiple
wireless sensors are used to
monitor a resident’s activity, and
activity trend data (patterns over
time) is then analysed to spot
potential health problems.
Healthsense has a strong market
presence with B2B business model.
Working with senior living communities to improve business
Key Lessons (Ride the Wave of Enlightenment)
Ride the Wave of Enlightenment
1. Harness the main technological advances: miniaturization, low-
power sensors, more accurate sensors, better integration with
human body (e.g., wristband, patch, textile, ingestible),
emergence of biosensors)
2. Sensor-based technologies are rapidly turning into reusable,
interoperable components and platforms. Exploit that trend.
3. Model individual behaviour and use data from multiple sources
to maximise predictive power and thereby make possible
actionable user feedback that is ‘surprising’ and relevant.
Ride the Wave of Enlightenment
4. Embrace the regulator. Sensor technologies have advanced
sufficiently to compete on accuracy. Resist the ‘gadget’ label.
5. To drive adoption, look for ‘low-hanging fruit’ opportunities
that address real, pressing needs among consumers and
patients. And/or combine health functionality with functionality
that is already ubiquitous (e.g. wristwatches, bedside alarm
clocks).
6. Work with the process owners and payers to develop solutions
that improve healthcare, streamline processes and reduce costs.
Thanks!
Author: Frank Boermeester
About Microsoft Innovation Center Vlaanderen
Our mission is to stimulate ICT-related innovation and
entrepreneurship in Flanders, with a particular focus on healthcare.
We do so by supporting high-potential technology startups. MIC
Vlaanderen has two operational offices in the cities of Genk and
Kortrijk, Belgium.
www.micvlaanderen.be