[MUSIC PLAYING] KAREN DESALVO: Well,
hello and welcome. I am Dr. Karen
DeSalvo, and I lead the team that serves as the
backbone of health expertise across Google. We guide the company as it
develops products, makes discoveries, and provide
services that help improve the health of everyone. We focus on privacy
and equity by design, and we drive the highest
industry rigor in the work. I'm glad you can join us today
at our second annual Google Health event, The Check Up. Before we get into
the program though, I do want to take a
moment to acknowledge that there is a lot going
on in the world right now and many people are affected,
professionally and personally. Against the backdrop
of a global pandemic, we're witnessing a
geopolitical crisis with the invasion of Ukraine. My heart is with
everyone impacted. I'm proud to work
at Google, where we continue to show up and be
helpful in times of acute need and crisis. Our work to improve health
doesn't stop during a crisis. After all, people will be
diagnosed with cancer today and others will have a
heart attack tomorrow.
So thank you for taking
time to tune in so that we can show you what is
giving us hope for the future. Today, we're pleased
to share some of the amazing work happening
at Google Health, all towards achieving our vision of
billions of healthier people. I joined Google just
about two years ago, inspired by that very vision and
the opportunity to achieve it. I spent 20 years
practicing medicine. And in my work, I carry
the stories and lessons my patients taught me at the
bedside and in the clinic.
I learned early that
building trust is job one. As my patients
trusted me more, I could help them more on
their health journey. I also learned that my
patients had health drivers beyond medical care,
including their ZIP code and the color of their skin. I'm excited to be at
Google, where we understand the importance of putting people
first and not just building better health care but helping
people build healthier lives. COVID-19 stepped
onto the world stage less than 30 days
after I joined Google. At the time, most
people had never heard of flattening the curve,
social distancing, or mRNA vaccines. But they did know they could
turn to Google for help. In response, our whole
company pulled together to provide answers when
people needed them most. One of the many
ways we did this was to amplify the important
messages from public health authorities by donating Google
Search advertising offering expert assistance to
ensure their messages were effective as possible.
Through one of
these initiatives, the World Health
Organization was able to serve more than
2 billion public service announcements, lifting
up authoritative COVID-19 information. The reach still astounds me. Today, we find
ourselves entering a third year of the
pandemic and I'm inspired by the
resilience I've witnessed and I'm hopeful for 2022 as
we return to a new normal. For Google, the pandemic
strengthened our resolve that we could and should work
to help everyone everywhere live a healthier life. The pandemic underscored
that the right information at the right time
can save lives. We learned that Google
can be most helpful when we meet people where they are
through tools they already use Search and YouTube.
And finally, the
pandemic reminded us about the power of partnership. We want to be a trusted
partner for three key groups. First, we want to help people be
healthier in their daily lives with information and insights. Whether they are patients or
not, these are our consumers. Second, we also partner
with their caregivers, providing tools and technology
to offer the highest quality, most [INAUDIBLE] And
by caregivers, we mean all the professionals
and institutions who help you on your health journey. And lastly, since we
know that being healthy is impacted by where we
live, learn, work, and play, we also partner to support
healthier communities. Consumers, caregivers, and
communities, our three C's. Today, we're going to
take a closer look at how we're bringing Google's tools
and technology to support the three C's. First, let's start
with consumers. Billions of healthier lives
means helping everyone everywhere in their
personal health journey as they seek to discover,
understand, decide, and act. Today, you'll hear from my
colleagues, Dr. Garth Graham from YouTube and Hema
Budaraju from Search, about ways that we're
supporting that journey, whether that is
helping people view authoritative,
culturally-relevant health information on YouTube or
finding a local doctor who takes their insurance and
speaks their language on Search.
We'll also share how wearables
and other sensing devices offer personalized
insights that can motivate us to be healthier. For example, last year
we launched new features to sleep habits in the Nest Hub. And just over a year
ago, we welcomed Fitbit, a team that
trailblazed a new category for health and well-being. Coming up, I'll introduce
Fitbit founder, James Park, to share updates on new features
supporting heart health. The second of our three
C's is caregivers. We focus on supporting them
and applying Google's expertise in organizing information
through Cloud solutions and products, like
Care Studio, as well as through technologies, like AI. Last week, Care Studio
announced a partnership with MEDITECH, our
first EHR partner. With this exciting
progress, we hope to show how our tools
can give clinicians the information they need to
care for patients all in one place. Today, Dr. Charles
DeShazer from my team will lead a discussion
on the partnership. At Google I/O in 2021,
we shared research we're doing with Northwestern,
where we're applying AI to mammograms to help screen
patients for breast cancer.
This year, we partnered this
work with several trusts at the UK National
Health Service, exploring how I can help
radiologists prioritize higher-risk patients. The goal is to get the right
diagnosis as fast as possible by using AI to
support radiologists. And in so doing, we reduce
waiting times and cost. When I think about the
potential that can come out of this research, I'm
reminded of all of my patients who anxiously waited weeks
to get mammogram results. We're also partnering to study
how AI can assist pathologists in the diagnosis and
determination of treatment options for prostate
cancer and other cancers. For example, we're working with
the US Defense Innovation Unit to explore the use of
augmented reality microscopes that can bring AI directly
into the pathologist workflow. The goal is to improve
diagnostic accuracy and efficiency for
treatment of cancers. Today, you'll hear from my
colleague, Yossi Matias, from Research about some of the
ways we're using AI in health. We'll also hear from Ivan
Poupyrev of our Advanced Technologies and Projects
team who will give us a preview of early research
aimed at improving how people recover after surgery.
The final category in
our three C's, community. We know that health is driven
by more than health care, so we have to support
healthy communities. We help people find resources
like healthy food, green space, and we're looking at the
impact of climate on health. Building tools for some of the
population isn't good enough. We cannot leave anyone behind. That requires us
to build tools that are helpful in any
community anywhere. You'll hear from
Googler Katherine Chou and from Garrett
Mehl our partner from the WHO, about new
open-source software initiative. Together, we aim
to help developers build interoperable and secure
applications for community health workers serving patients
in low- and middle-income countries.
We'll also give
you an early look at something else we're
working on, gender equity. We believe technology
can close gaps, including how women of color and those
from underserved communities access and experience
health care services, like maternal health. We'll hear from
Googler Dr. Ivor Horn, who will introduce an
equity-focused research initiative that
we're kicking off. She will also share a
conversation with Dr. Lasé Ajayi at Scripps about the
Fitbit-funded PowerMom study. So lots to cover as
we share our work to partner with consumers,
caregivers, and communities. So now, without further ado,
I'm delighted to welcome my colleague James
Park to get us started.
[MUSIC PLAYING] JAMES PARK: Hello, everyone. I'm James Park,
co-founder of Fitbit. I joined Google just a
little over a year ago, and it's great to be here today
to share more about how Fitbit continues to help people on
their health and wellness journeys. More than 15 years ago, we
created Fitbit with a vision to help make everyone
in the world healthier. Since then, we pioneered a new
category of wearable technology that makes health
more accessible, fun, and achievable. We do this through our
devices and services and by working with
health plans and employers to make health care more
accessible and affordable for everyone. We know that being healthy is
far more than just activity.
It's about holistic health
and how it all fits together. For example, tracking nutrition,
helping with a better night's sleep, or managing stress. Our Fitbit platform makes
managing overall health fun and motivating. Behind our wearables are many
technological innovations that continue to push the limits
of what wearable sensors can do to make the invisible visible. For instance, Fitbit was
one of the first to support continuous risk-based
heart rate tracking. We're also one of the first
to bring sleep tracking to wearable devices with a
multifaceted view that included measuring stages of sleep and
offering insights on sleep quality. We were the first to bring an
electrodermal activity or EDA sensor to wearable,
measuring hand perspiration to measure how your body
is responding to stress. We've also made metrics
like oxygen saturation, skin temperature, and breathing
rate more accessible, putting users
effortlessly in control of their health and wellness.
Among the areas where we
believe wearables can have the most impact is heart health. Heart rate is a foundational
metric for general health and well-being, and it's also
been a core offering for Fitbit users from the early days. PPG optical sensors are able to
measure the tiny fluctuations of blood that happen under
the skin with each heartbeat. We use these sensors
in our wearables to measure blood volume
changes to track heart rate and variability during a
workout or while sleeping. In recent years, with
an increased focus on supporting
chronic conditions, we tested and launched
our first FDA-cleared ECG, or electrocardiogram,
app that allows people to assess their
heart rhythm for signs of atrial fibrillation or
AFib, a heart condition that affects nearly 33.5
million people globally.
AFib can be difficult to detect
outside a hospital or clinic as episodes can be
sporadic and asymptomatic. Some studies suggest
that as many as 25% of people who have an
AFib-related stroke find out they have AFib only after
a stroke has occurred. Because it can be
so hard to detect, being able to more continuously
look for signs of AFib is essential, including at
night when the body is at rest. That's why, in 2020, we
launched the Fitbit Heart Study, a clinical study
directed by doctors at Harvard involving nearly half a million
Fitbit users to test their PPG AFib algorithm which looks
at heart rate data passively and can alert users to sign
of any irregular heart rhythm. The study results presented at
the most recent American Heart Association meeting showed the
algorithm accurately identified undiagnosed AFib
98% of the time.
We have submitted our algorithm
to the FDA for review. This is one of the
many ways we're continuing to make health
even more accessible to more people around the world. Next up, we'll hear
from Hema Budaraju on how Search is expanding your
access to the right information at the right time and
simplifying actions, like finding available doctors
and locking in appointments. [MUSIC PLAYING] HEMA BUDARAJU: Hi, everyone. I'm Hema, and I lead our
health work on Google Search. No matter where we live in
the world or how old or young we are, one thing holds
true for all of us, health is a universal need. Today, when we have
questions about our health, the internet is usually
where we start our journey. That's why, no matter what
you're searching for on Google, it's our mission to
give you relevant and authoritative information
exactly when you need it. COVID-19 has shown
how important it is to access timely
health information from trusted national and
local health authorities. Billions of people from
more than 200 countries and dozens of
languages have access to information about
COVID on Google. That's why we continue to
expand our Search experience, including things like
news, statistics, trends of the pandemic, and
information on how to access testing and
vaccines to protect yourself and your loved ones.
COVID testing remains a
critical need for all of us. On Google Search and
Maps, you can easily find details on tens of
thousands of testing sites worldwide. In addition, vaccines are one
of the most important tools to fight COVID. That's why we worked
with dozens of partners to highlight over
180,000 vaccination sites across the globe. If you're in the US,
a quick Google Search will show you where to
find different types of COVID vaccines
and boosters nearby. Or it can show locations that
have vaccines specifically for kids. We also launched a
feature in the US to show vaccine appointment
availability to make it easier to find a nearby location.
Beyond the immediate
challenges of COVID-19, we are working to
make health care more accessible to everyone. There can be a lot of
friction when you're making a doctor's appointment. In the US, the average wait time
for a primary care appointment can be 20 days or more. It shouldn't be this hard
to get access to care. We are really excited to
announce that, in the coming weeks, we'll make it
easier to find health care appointments by
showing available dates for local providers
on Google Search. Maybe, like many of us
during the pandemic, you've recently moved and
you're looking for a new doctor. Or maybe you've put off getting
your annual physical checkup or maybe your child
needs a same-day visit.
Our broader efforts are focused
on how to help you get care with less friction. We started by showing you which
doctors accept your insurance. And now, we are expanding
access to available appointments as the next step to help
you stay safe and healthy. We are just getting started
with a few early partners, and we are excited about making
your health journey easier. Now, I'd like to introduce to
my colleague Dr. Garth Graham to tell you about what we are
doing with high-quality health information on YouTube. [MUSIC PLAYING] GARTH GRAHAM: Hello, everyone. I'm so pleased to
be here with you. I've spent the majority
of my career focused on the social
determinants of health, across many different parts
of the health care industry, government and academia.
Now, you may be thinking,
that's great, Garth. But what does YouTube have
to do with health, and how is a video site
supposed to deliver better care to patients? It connects directly back to
the determinants of health and the things that
drive health outcomes. There are reams of
literature looking at the impact of
housing policy, food insecurity, and racial
equity on health outcomes. But there's another area that
we need to start focusing on, and that is information as
a determinant of health. As anyone who has worked
in public health knows, empowering communities
with education is essential to helping
people live healthier lives. This has been brought into even
clearer view over the past two years as the world has battled
a misinformation epidemic alongside the COVID pandemic. We need to consider where and
how people access information about their health, the quality
of that health information, how easy it is to understand,
and how engaging it is. Listen, the reality is the
majority of health care decisions are made outside
the doctor's office, in the everyday life
of our patients.
As physicians, we only
have limited opportunities to connect with people
within that exam room. And how can we better show
up at many more points within a patient's day,
bringing health information into their lives through the
spaces they already visit? YouTube is already a part
of a patient's journey. When they turn to us for answers
to their health questions or look for a community who
understands what they're experiencing or
find helpful videos to explain complicated medical
information more easily, health education can
come in many forms. And we've just seen
firsthand how powerful video can be used as a medium to reach
people with health messages. That's why the mission
for YouTube Health is to provide equitable access
to highly-authoritative health information that is
evidence-based and culturally-relevant
as well as engaging.
This approach also further
reinforces our ongoing efforts to tackle the very serious
issue of medical misinformation. We want to make sure that
people are finding information from evidence-based
authoritative sources and that information
is easy to understand and connects with them in a
personal and engaging way. We're doing this with a focus
on two key areas, information quality and information equity. First, let's talk about quality.
In the past year, we've
launched new features, kicked off new
partnerships, and worked with experts in the field
to help build foundations upon which we can greatly scale
our work in health in 2022. We continue to expand our
work with leading health organizations, like Mass General
Brigham, and public health leaders, like the American
Public Health Association, as well as with
clinicians and creators. Our goal is to further
increase the accessibility of high-quality health
content on our platform. Listen, these
organizations are creating engaging evidence-based
videos that really connect with people. In addition to
increasing the volume of highly-authoritative
content on YouTube, we're working increase the way
people find it and raise up its discoverability
to ensure people can find these sources easily
when they come to YouTube. To do that, we've added health
source information panels and videos to provide context
to help viewers identify videos from authoritative sources. And we've also introduced
health content shelves that highlight videos
from these sources when you search for specific
important health topics on YouTube. These context cues aim to help
people more easily navigate and evaluate credible
health information.
Starting just this week, users
in Brazil, India, and Japan will also see these new features
while searching on YouTube for top health conditions. As part of this effort, we
partnered with the World Health Organization and the
National Academy of Medicine on principles for evaluating
authoritative health sources who've also worked closely with
regional health authorities while doing this. We're excited these
features are rolling out to more people and plan to
reach additional communities across the globe
later this year. The second key I mentioned
that we're focused on is information equity. If we all agree, information
is a determinant of health, meaning that it drives
health care decisions and we must consider how that
information gets to people. Health leaders have
a responsibility to keep pace with the changes
in where and how people find information. We must meet people where
they are and connect with them in ways that are not
only scientifically accurate but also
culturally-relevant and engaging.
Through video, we can provide
free and equitable access to the best and
brightest thinking in science and medicine. And we can make it
engaging, interesting, and compelling to watch. We can become part
of people's lives and reach them in the
moments that matter. More and more leaders are
taking up that challenge. I'd like to share a couple
of examples with you guys. The first is the New
England Journal of Medicine. As many of you already know, the
New England Journal of Medicine is not only recognized
for publishing practice-changing
medical research, it's also become a cornerstone
in providing equitable access to information for
many years by allowing online access to
readers in the world's least developed countries. New England Journal of Medicine
has just taken a new step forward in providing access
to the best minds in medicine by launching a new
channel on YouTube. They're creating short
summaries of key findings from their latest
articles, along with interviews and animations
to provide the best and newest medical research
from the journal and make it available
at no additional cost to audiences around the world. YouTube Health also
has experts making the most of digital video to
support public health education efforts.
The American Academy
of Pediatrics has used YouTube creatively
to reach younger audiences. One particular project, they
created short-form videos called "YouTube
Shorts" to connect with young women, girls,
and gender-expansive people about key topics
related to their health, such as menstruation
and consent. The AAP has launched a video
series to empower parents with the information they
need to make decisions about their child's
immunizations. The videos cover a commonly
asked question but also address how disinformation and
misinformation about vaccines can spread online
and help parents know how to find
accurate information. Lastly, the American
College of Physicians has also tackled
vaccine misinformation but from the perspective
of the physician. Their physician-to-physician
conversation series teaches doctors the effective
communication strategies to build vaccine confidence and
address patient concerns that are rooted in misinformation. All of these are great examples
of how medical experts are reimagining how and where
health information is shared to increase access
to information for everyone. It's really an exciting
time to be working in health communication. We have more tools available to
help people than ever before. The scale and reach of
platforms like YouTube can radically increase
equity of access to high-quality
health information by breaking down barriers
between the ivory towers of academia and the
everyday people who want to understand how to
take better care of themselves and their families.
There's always more
work to be done, and we hope to make a
difference through our focus on information
quality and equity. Of course, the whole world
still needs better access to health care. Yossi Matias is here
to talk about how we're using technology to expand
access and the important role your smartphone will play
in things like screening and affordable care. [MUSIC PLAYING] YOSSI MATIAS: Hello. I'm Yossi. And I lead Health AI team
along with other teams who are advancing AI and using AI
for the benefit of society, including conversational AI and
tackling the climate crisis.
Our teams focus on input-driven
research, innovation, and [INAUDIBLE]. So here's what we know. Health care is not equally
accessible across the globe, whether due to cost,
proximity to experts, schedule flexibility, or
a host of other factors. So when we think of
making health care more accessible to everyone, people
like Santi come to mind. Santi lives with
diabetes in rural India. She's been diagnosed with
diabetic retinopathy, a disease that can cause
blindness if not treated early. Diabetic retinopathy
is asymptomatic in its early stages,
making screening critical for diabetic patients. Unfortunately, challenges
in access to health care means getting a diagnosis
could take months, if she is diagnosed at all. Santi is not alone.
Diabetes affects
400 million people across the globe, 70
million in India alone. With the help of new
technology, an expanded access to health care diagnosis of
diabetic retinopathy and other diseases could be
vastly improved. In fact, it's already happening. Santi could get same-day
screening and treatment because of these advancements
instead of waiting for months. Over the past few years,
we've shared progress on our work in this space. Last year, we shared updates on
our Automated Retinal Disease Assessment, project
in Thailand and India also referred to as "ARDA." ARDA uses AI to help
detect diabetic retinopathy in patients like Santi. This is done by developing
machine-learning model to look for signs of
diabetic retinopathy in photographs of
the back of the eye. For patients, it's
a simple process of sitting in
front of a tabletop camera that takes the picture.
The clinician then asks
ARDA for an interpretation and uses the
information to decide whether to refer the patient
to an eye specialist. We're so excited to share that
this project has been tested with the Thailand
national screening program in a prospective study,
which found that ARDA can be deployed safely and in real
time across multiple regions and achieves comparable
accuracy to eye specialists. Across the globe,
ARDA has screened more than 75,000 people,
the majority of whom, including Santi, live in remote
and rural parts of south India with 10 times lower
access to eye doctors than in urban areas.
This innovation
marks great progress. Our work with ARDA has
reinforced the need for regular eye
screening, especially for those who live
with diabetes. But it's also made
clear that we need to reach people who are
unable to get to a clinic, so we've been reimagining
the entire patient journey and investigating
what's possible at home. Our engineers and
scientists are envisioning a future where people with
the help of their doctors could detect, diagnose health
conditions from their own home. Today, our research uses
large tabletop cameras to make high-resolution photos
of the outside of the eye under uniform lighting. We found that a machine-learning
model could sometimes detect when a patient has high
blood sugar levels, high blood cholesterol levels, or
diabetic retinopathy on the inside of the eye,
just from these photos.
This research has
recently been published. We're in the early stages with
our academic and non-academic partners, such as
EyePACS and CGMH in hope to build on this promising
proof-of-concept research. This includes
further exploration to see if smartphone
images could also be used to detect these and
other potential health issues. With over four billion
smartphone users globally, there is an enormous
opportunity to increase the accessibility of screening
and affordability of care. A device that was once used
solely to make phone calls or sending texts has transformed
to a fully-featured computer with many sensors and functions
that have the potential to unlock accessibility. One way we are doing
this is with DermAssist, our AI-powered skin
condition search app that helps people
better understand their skin concerns by answering
a few questions and taking three quick photos. Importantly, this technology
comes with the responsibility to build a product that
works for everyone, people of all skin tones, skin
types, ages, genders, and more. We have been working hard to
build DermAssist with health equity as our priority
and we're excited to get DermAssist in the hands
of product testers later this year. As with DermAssist,
growing from solid base is something we're
already working through on other projects.
Last year, my
colleague Shwetak Patel introduced some of
Google's work to extend the usefulness of
everyday sensors to support daily
health and well-being. At the time, he shared a
new feature in Google feed that uses your smartphone
camera to measure your heart and respiratory rate
to track fitness goals. I'm delighted to share
that these features, which were first available
on Google Pixel, are now available on more
than 100 Android device models as well as iOS. Outside of the
camera, a smartphone has many features and sensors
that may give insights into health or wellness. Another capability where testing
is whether a smartphone's microphone can correctly record
the sounds that a person's heart makes when placed over the
chest, just like a stethoscope would. Listening to the heart and
lungs with a stethoscope, auscultation is a critical
part of the physical exam during a doctor's visit.
Healthy hearts give a
characteristic rhythm, lub-dub lub-dub sound. [HEART BEATING] But hearts with valve
diseases, like aortic stenosis or other conditions, can
make different sounds, such as whooshing or rubbing. [WHOOSHING SOUND] Heart valve disease can
sometimes go undetected. A study from the
Journal of Cardiology found that 40% of
patients are not diagnosed with aortic
stenosis, or AS, until they start
to show symptoms. A different study from
the journal Circulation found that by the
time AS patients show symptoms half are
already too sick, so sick that they will die
from the disease within two to three years. Screening for AS typically
requires specialized equipment, like a stethoscope
or an ultrasound and an in-person assessment. We're currently in the early
stages of a clinical study where we're testing whether
a smartphone application that uses the phone's
built-in microphones can record a participant's
heart sounds when placed over the chest. We look forward to sharing
more on that in the future. Remember something, ARDA helped
ensure her diabetic retinopathy was detected early enough
to possibly save her vision.
For future patients, we
hope to bring Santi's access to care closer to home,
with no commute needed. Google Health strives to build
more affordable, convenient, and accessible care to
people around the world. We can't wait to share all
of these groundbreaking developments in the
technology with you. Next, we'll show you what
Google's Advanced Technology and Projects team is working
on with UCSF researchers. Together, they're using AI and
affordable consumer electronics to monitor knee
replacement recovery and create a future where
surgery recovery can be continually and
accurately measured outside of medical labs. [MUSIC PLAYING] JACK SIMS: My name is Jack Sims. I was a high school
art teacher, moved to the desert, built a studio.
My midlife crisis,
sports cars, is this. First sort of documented
damage to my knee was in middle school, and
I developed the disease of the growing end of the bone. So that coupled with just
being very active in sports, my left knee started acting up. By my early 60s, it was becoming
a problem where some days it was hard to walk around
without limping on it. When I went to the
orthopedic surgeon and they talked about
knee replacements, I was actually
thrilled that there was a possible way of
fixing it once and for all and alleviate it. STEFANO BINI: There are
approximately one million knee replacement surgery
procedures performed in this country every year. It's predicted
that 50% of adults will have advanced knee
arthritis requiring knee arthroplasty at some
point in their lifetime and that that number of total
joint replacements performed in this country will grow
to almost 3 and 1/2 million over the next decade. While total knee
replacement is considered one of the most successful
procedures done, there are some challenges
that we still need to solve. One is that somewhere between
10% and 20% of patients still have problems
after surgery, but we have a hard time
measuring those outcomes.
We have subjective
measures, how patients feel. And then we have traditionally
seen patients in the clinic, but that's twice, sometimes
three times after surgery over the course of a year. What we've never had
before is the ability to track how people do over
time on a continuous basis. JACK SIMS: My
recovery was longer than other people's recovery. I didn't really feel like
I had a partner in that. RICHARD SOUZA: We
really want to be studying those people who are at
risk for having a poor outcome. The UCSF Human Performance
Center can evaluate patients. We can reserve the resources
like the 3D Gait Biomechanics lab for those individuals
that most need it. However, access to
that is quite limited. Generally, only people who
enroll in research studies get to experience 3D
biomechanics testing.
So our standard equipment
in our laboratory is motion capture data. We put markers on specific
areas on people's body. And then as they walk
within the laboratory, the cameras can
track those markers. We create a 3D model,
which tells us exactly where the joint locations are. Measuring these key variables
in the laboratory is great. However, it's not enough. What we really need
to know is what people are doing in the real world. STEFANO BINI:
We're not the first looking at the opportunity
to use sensors to recreate a gait-lab-like experience. We can start tracking
how people do but not have them come back
into an academic environment or a lab environment to do so.
I actually had the
opportunity to attend TED and Ivan was showcasing
how the Google Jacquard sensor could be
used to interface with their environment. I went to Ivan to mention that
I've been doing this research and I've been very disappointed
with the results of using commercially-available wearable
sensors to track patients in the gait lab. And he said, there's
a reason for that. You're simply collecting
the wrong data, but we can help you with that. IVAN POUPYREV: My
first reaction was, we can do so much more with more
than artificial intelligence techniques and
machine-learning techniques. Solving problems,
such as the one we're working on with
Dr. Bini, requires us to have a broad
variety of perspectives and broad variety of expertise. And I'm leading an
innovation group at the Advanced Technology and
Projects division in Google, also known as "ATAP." Our team is highly
interdisciplinary, we have hardware engineers, we
have signal processing experts, we have machine
learning experts, and we have
interaction designers.
And by working together, we
can approach this problem holistically. The modern methods of
artificial intelligence, particularly in real time and
on-device machine intelligence which allow us to run very
complex machine learning models which understand
human emotions, run very inexpensive and
very easily accessible computing hardware,
such as the one we're using in a Google Tag. Tag is a small
computing module, which is designed to go anywhere
we want to put computing in. NICHOLAS GILLIAN: So the goal
of this early R&D project is to see if we can replicate
the result of the GitLab using the Google Tags. If we can do this,
this means that we can replicate the same
type of analysis in the lab outside in the real world. So one of the big technical
challenges here is, how can we replicate the output
of the UCSF GitLab analysis system with a Google Tag? So to achieve that, we use
our custom neural network architecture. This allows us to take
our raw time series data and actually be able to estimate
directly from this in blue what the knee angular
velocity could be.
RICHARD SOUZA: What the AI
team at ATAP was able to do is match our 3D
biomechanics data to within one degree throughout
the entire gait profile. I didn't really think that
was going to be possible. NICHOLAS GILLIAN: One of the
key things that we've learned is that the single
sensor alone we are able to accurately
replicate the output for some key variables,
like knee angular velocity. IVAN POUPYREV: If
we are successful, these techniques will
have massive impact on how people recover
after complex operations, particularly with
orthopedic environment. STEFANO BINI: This idea of
collecting data real time 24/7 during recovery phase and
using that information to improve the
recovery process, we're on the cusp of a
revolution about how we think about patient care. [MUSIC PLAYING] [MUSIC PLAYING] CHARLES DESHAZER: Hi. My name is Dr. Charles DeShazer. I'm an internist by
training and the director of clinical products at Google.
Patients rely on their doctors,
nurses, and other health care professionals to make
well-informed decisions about their care. But oftentimes, the
information caregivers need isn't easy to access. It may be buried deep
in a patient record or spread across IT systems
that don't speak to each other. And that's why
doctors often don't know your full
medical history or may be unaware of the care
you're receiving elsewhere. This is why you may need
to repeat your history when you see a new doctor
even in the same system. So fragmented data impacts
everyone, the patients, the nurses and doctors
delivering care, and more. Improving on the flow of data
across health care systems, referred to as
"interoperability," is key to ensuring
optimal patient experience and outcomes. Last week, we announced a
new partnership with MEDITECH to take a step forward
in interoperability. This partnership aims to combine
our data harmonization, search, and summarization capabilities
from Google Health's Care Studio product suite
and integrate them into the MEDITECH Expanse
electronic medical record system.
So let's dig in. I'm here with Aashima
Gupta, Director of Healthcare Solutions
at Google Cloud; Mike Cordeiro, Director of
Interoperability and Strategy at MEDITECH; and Hien Brown,
a licensed RN for 13 years and now a Clinical
Implementation Specialist at Google Health. Let's start by grounding in the
importance of interoperability in terms of patient outcomes. Hien, can you talk through
some of the challenges you've faced in delivering
optimal patient care when data is fragmented? HIEN BROWN: Sure. Previously, I worked at
a pediatric hospital that didn't have an
emergency department, but we were connected to
an adult hospital that did. So kids that came in
through the ED and got admitted to our
unit, we couldn't access any of the data
that was from the ED because they were
on a separate EHR.
And it was a
frustrating experience for the providers
and also the patients who were asked the same
questions upon admission that they were in the
ED, just down the hall. And in the ICU, I took care
of a lot of chronically ill patients, and I
remember one young boy in particular that came in. He was admitted on night shift. His mom was at the bedside. And he was chronically ill,
so he'd been in and out of the hospital a lot. And his mom had a
binder at the bedside. And she opened to
a specific page, and she wanted to let us
know that her son was taking a certain medication because
her son's doctors had told her to always make
sure every care team knew about this med.
And it was important because it
was a steroid, which you never want to stop taking abruptly. And the moment gave
me pause, because I felt really conflicted. As a nurse, I thought
this woman is incredible. She has all the
information that I need. But as a mom, I
just thought this is the middle of the night. She's exhausted. Her son's in the ICU. But she's so accustomed to
this lack of communication among health care organizations
that she took it upon herself to make this binder and bring
it with her to every facility that she goes to and to be
here in the middle of the night just to ensure her son
has continuity of care. And I just couldn't
help but think that's not her burden to carry. These are just a
few examples of how the lack of interoperability
impacts our patients. But unfortunately,
it's far too common. CHARLES DESHAZER:
Thank you, Hien. That's a great example,
and that is a burden that a mother in the middle of a
crisis should not have to bear. Our Care Studio
product is designed to provide a unified
view of patient records, making them more accessible
and useful for clinicians.
Mike, what do you view
as the top challenges in interoperability
and what are your hopes for this new partnership
with Google Health? MIKE CORDEIRO: Sure. So clinical information
is often scattered across many care settings,
stored in various forms, structured, semi-structured,
unstructured. It's difficult for providers
to review all the data formats in a quick and effective
manner, especially when coming up from multiple
systems and multiple locations. Clinicians need the
information most relevant to the needs for their
patient at their fingertips so they can provide
timely, targeted care. So leveraging search
summarization tools within Expanse will pull you
information that provider needs into a clean summary
without having to hunt for it and building on. And it builds on native
streamline workflows already available in Expanse, so
I feel that the tools are complementary to each other. CHARLES DESHAZER:
Thank you, Mike. Something that's
discussed frequently is how to prepare for
increases in data loads. And many hospitals keep old
applications running just to preserve the data. Aashima, can you speak to some
of the trends, especially how tech like Cloud AI,
are helping manage this influx of data in digital
transformation trends underway? AASHIMA GUPTA:
Thank you, Charles.
We have come a long
way as an industry with digital transformation. And with moving to
Cloud and tools like AI, we are able to extract
deep inside action from health system data. For example, Cloud provides
the ecosystem in two big ways. Number one, it's providing
the secure encrypted data foundation to support
data interoperability to support the use cases
that Hien mentioned. And the second is lowering
the barrier for AI accessibility for
technologies like Cloud NLP to get meaning out of
that unstructured data. CHARLES DESHAZER:
Thank you, Aashima. Mike, can you speak to the
creation of MEDITECH Expanse and why it was important to
have a native Cloud solution? MIKE CORDEIRO: Sure.
So when MEDITECH
designed Expanse, our goals was to create a
modern mobile health care platform with an open ecosystem
that embraced greater data liquidity. And it also wanted to streamline
data access from anywhere over the web and capitalize
on Cloud native technologies and tools like those
provided by Google Health. And finally, a Cloud
solution simplifies the IT infrastructure for
health care organizations. It provides much greater ease of
scalability, improved security, and is especially
important in today's rapidly-changing
technology landscape. CHARLES DESHAZER: Absolutely. I want to bring this
back to patient care. For the Care Studio
team, success will be when clinicians have
a holistic and complete view of the patient record
with ease at a glance.
Hien, looking ahead, how
will interoperability impact the everyday
workflow of clinicians? HIEN BROWN: I think it's going
to have a tremendous impact, saving us time, looking
for information while we're at the bedside, reducing
costs by avoiding repeat labs or procedures. In pediatrics, that's especially
important because we're always trying to minimize
the number of needle sticks that a patient needs. And taking it a step
further, I think health care providers
having a fully-formed view of their patients history
makes health care more equitable across the board. It takes the burden off the
patient and their families and helps ensure everyone
gets the care that they need, independent of their ability
to recall all the details of their medical history.
They've got enough
to worry about. CHARLES DESHAZER: Absolutely. Thank you, Hien,
and thank you all. Thank you, all of our panelists,
for joining the discussion. Next, learn how our partnership
with the World Health Organization and
Android developers is helping health care workers
worldwide access and share information empowering them to
make more informed patient care decisions. Thank you. [MUSIC PLAYING] KATHERINE CHOU: Hi. I'm Katherine Chou, Director
of Research & Innovations at Google. Mobile devices are essential to
delivering data-driven health care, particularly in low-
and middle-income countries, or LMICs for short. Today, I'd like to share with
you our work with the WHO and mobile developers
to build an SDK, or a software development
kit, that makes it easier to develop data-driven apps
for community health workers.
Community health workers often
form the bedrock of health care in LMICs, where about
half of the population lacks access to vital
health care resources. These frontline workers who
are embedded in the communities are the cornerstone
to connecting patients with essential care. They take on jobs ranging from
maternal and child health care to malaria prevention
to helping families access health information. Over the past few years,
smartphones and mobile apps have become
invaluable care tools for community health workers. They support community outreach
and home health services, including triaging patients by
accessing immunization and care records, conducting
health screenings, and distributing medications.
Today, community health
workers have these services at their fingertips,
helping them deliver accurate and
timely evidence-based care. These efforts aim to improve
health outcomes at a population level across these countries. With these services come
a new set of challenges. Often, community
health workers will have multiple apps
or even phones to manage the same patient
across different disease or outreach programs. Incompatible data and
lack of interoperability leads to situations where
decisions for patients are made with
incomplete information. We're applying technology to
help accelerate a paradigm shift, where it's easier to
develop health care apps that are seamlessly interoperable. By open sourcing
and supporting tools that ensure the exchange
of high-quality data, developers can focus resources
on designing applications to meet the needs of
their local communities without having to worry about
managing clinical content and data requirements
for evidence-based care.
To advance this mission,
we're collaborating with the WHO and an
ecosystem of tech providers to create an open-source SDK. This SDK will help developers
build secure mobile solutions using a patient-centered global
data standard called "FHIR," also known as "Fast Healthcare
Interoperability Resources." This, in turn, makes it
easier for health workers to deliver data-driven care. With over 3 billion active
Android devices globally, there's an opportunity to
support a huge number of health workers and their patients.
Some health workers often
need to deliver care in areas with
unreliable connectivity. We've designed the SDK to allow
Android apps to run offline, with no connectivity, by storing
and processing data locally. The FHIR SDK is
part of our mission to democratize access
to high-quality care through mobile devices. Let me introduce a key partner
in this work, Dr. Garrett Mehl, Head of Digital Health, an
innovation unit at the World Health Organization. Thank you, Garrett,
for joining us. Can you share a little
more about the WHO's vision and how developers
are using the SDK? GARRETT MEHL: Hi. Thanks so much for having me. I'm happy to share what
we're doing at WHO to make things easier for developers. We want to bring evidence-based
care to health care workers everywhere. First, we're helping our member
states with care guidelines. These guidelines consist of
life-saving interventions. Our approach applies to
those already investing in digital transformation. Guidelines are designed to
capture rigorously tested and adapted clinical and
public health recommendations. Our goal is to
provide an easier way to adopt high-quality,
evidence-based care guidelines into digital systems.
Now, translation, we know,
has been a challenge. We need an accurate translation
of narrative guideline recommendations into
executable digital systems. And it's difficult. It's both
resource-intensive and prone to mistakes. We have focused on updating
guidelines for the digital era. We call these digitalized
guidelines "SMART" because they're standards-based,
machine-readable, adaptive, requirements-based,
and testable. So we've embraced open
standards like FHIR to support interoperability
across patient-centered health care applications and
CQL for consistency in calculating decision
support and indicators. It isn't yet easy
to build software that can make use of
digitized SMART guidelines. So to solve this, we've
engaged the community and collaborated with Google
on developing the FHIR software development kit. We think this will make it
easier to bootstrap health care applications with
SMART guidelines, offline capabilities, and
mobile-first functionality.
Take, for example,
WHO's Em Care project. This is a project that's
leveraging the FHIR SDK to develop a reference
software platform. It's focused on health care for
children in emergency settings. So when we combine SMART
guidelines and the SDK, we enable technologists
to accurately integrate evidence-based content and
allow that interoperability. This is an open source and
open standards approach that will ensure every country's
digital transformation has access and can leverage
the latest care recommendations to deliver
essential interventions to all. KATHERINE CHOU:
Thank you, Garrett, for the insights into
where we're headed. In addition to
partnering with the WHO on accelerating the development
of next generation mobile-first tools for data-driven
health care, we'd like to share some examples
of other innovative tech collaborators who are currently
doing amazing work using the FHIR SDK to help community
health workers better connect with their patients.
One example is ONA, a
team working in Liberia building a FHIR-native health
app for frontline workers doing health and humanitarian
relief efforts. IPRD Solutions is a health
care technology group that we partner with to
enable health innovations. This partnership helped over
2,000 community health workers allocate malaria prevention
nets to more than 700,000 people in Nigeria. And finally,
Lattice Innovations, a health tech solutions
company based in India, is joining the State of Tamil
Nadu's population Health Registry Consortium. They're building a
community-led health screening and primary care
mobile application that will help fulfill
the state's vision to provide data-driven
health care for all. We look forward to engaging with
these and other regional app developers in our mission
to empower community health workers and enable equitable
access to high-quality care. We've spoken today about
the remarkable partnerships and the incredible
progress we've been able to make together.
Next, you'll hear how we're
working with Northwestern Medicine to put the power
of performing ultrasounds in the hands of midwives to
address maternal mortality rates. [MUSIC PLAYING] IVOR HORN: Hi. I'm Dr. Ivor Horn, and I lead
Google's Health Equity efforts. You've just heard about
some amazing innovations in health care. But even with all
these advancements, women around the globe
still die during childbirth and in high numbers. 300,000 maternal deaths and
2.5 million perinatal deaths occur each year, with
an astounding 94% happening in lower- and
middle-income countries. Uneven access to care can
lead to unnecessary harm to birthing parents
and their babies. Early ultrasound screening is
a safe and effective method to plan for pregnancy care
and identify issues early.
The WHO recommends an
ultrasound assessment before 24 weeks of pregnancy. This helps to assess the
health of the birthing parent and baby, plan the
pregnancy journey, and monitor for early
risks and complications. Ultrasound technology is
becoming more portable and more affordable. Yet up to half of all birthing
parents in developing countries are not screened at
all while pregnant. Among the barriers to access
is a shortage of expertise in reading ultrasounds. Our goal is to simplify barriers
to acquiring and interpreting ultrasounds. How do we do that? By supporting more people like
midwives and community health workers to provide care. Imagine anyone anywhere being
able to perform accurate ultrasounds with just
a small handheld probe and a smartphone. Throughout the past year, we've
been researching the use of AI to assist with maternal
health diagnostics. We will soon publish
foundational open-access research that
validates the use of AI to help providers conduct
ultrasounds and perform assessments. Our hope is that this
work and work to come will increase
access to ultrasound and improve experiences
for birthing parents and their babies.
That's why we're working closely
with Northwestern Medicine on further developing
and testing these models to be more generalizable across
different levels of experience and different technologies. We launched this
partnership with the goal of cutting global maternal
mortality rates in half. The goal, we aim to accelerate
research and development in two key areas, identifying
high-risk patients during pregnancy more
quickly and reliably and building tools to allow a
broader range of professionals to confidently provide
access to these assessments around the world. Together, this research
can potentially drive automated and accurate
AI evaluations of maternal and fetal health risk.
It can also address barriers
to scaling programs, enabling timely care in the
right settings. To scale this technology
and reach the many patients and providers globally,
partnerships will be critical. We'll be working with
health care organizations and local champions to
develop novel technologies and get the technology into
the right models of care to reach patients. We're thrilled to
work with Northwestern Medicine on these early
stages of development. Together, we're making
continuous improvements toward increasing access to
care for everyone everywhere. Part of that is
helping researchers with their work on
maternal health. Dr. Lasé Ajayi is the director
of Diversity Initiatives and Clinical Researcher in
Digital Medicine at Scripps Research, and you'll
hear from her next.
She's also an awardee of the
Fitbit Health Equity Research Initiative, a program aimed at
supporting researchers working to address health disparities. She's the principal
investigator of PowerMom, a smartphone app-based
research study that enables pregnant
parents to share data about their
pregnancies with scientists. This story is an
important example of Google's efforts to make
health care more equitable. Dr. Ajayi is focusing
on how PowerMom can help underrepresented
birthing parents who are more likely to die
during and immediately after childbirth
in the US compared to their white counterparts. I am so excited for you to hear
more about her incredible work. [MUSIC PLAYING] TOLUWALASE AJAYI:
When I was pregnant with my first daughter,
I had terrible fibroids.
They caused a lot of pain. As a Black woman,
I actually avoid going to the emergency
room because of the things that I face. I try to avoid it. In this particular
instance, it was so bad that I didn't have a choice. [MUSIC PLAYING] I arrived by Uber by myself
in excruciating pain. Nobody came to help
me off her wheelchair. I was treated brusquely
by the intake nurse. When I was getting
my ultrasound, the technician was so rough and
brutal I was actually crying. And it wasn't until
the person who was coming to check
insurance realized that I was a physician
within the health care system that the demeanor between the
staff, the physician, everybody changed. [MUSIC PLAYING] The problem facing Black
and Hispanic people, it falls into two
really big categories.
It's the lack of access
and then how we access the medical system in general. Black women are more
than three times likely to die in childbirth or
immediately after childbirth than their white counterparts. Black babies, again
more than two times more likely to die than
their white counterparts. We're more likely to have
gestational diabetes, preeclampsia, and die
from those complications. But why is that? It's about Black
and Hispanic women not feeling heard or seen
with the medical system. It's important to
me because it's health care and it's
something that we should all be able to access, no
matter what you look like. PowerMom FIRST is a study
from the PowerMom platform. It stands for "Fighting
Inequity & Racism using Supportive Technology." We work with Fitbit
to actually look at structural and
systemic racism throughout the
pregnancy journey. [MUSIC PLAYING] Do you mind if I
touch your belly? PATIENT: Absolutely. TOLUWALASE AJAYI: The
goal of this study is that every person should
have a healthy pregnancy and a healthy baby.
The data that we'll be
able to collect with Fitbit is 24/7 heart rate, heart rate
variability, oxygen saturation, body weight, sleep, sleep
stage, and activity. PATIENT: It looks like I've
been averaging about seven hours a night. TOLUWALASE AJAYI: That's
actually pretty good. And then we'll look at
outcomes of your pregnancy. Were you someone who
went into pre-term labor? Did you have high
blood pressure, hypertension,
increased weight gain, whatever during your pregnancy? That allows us to collect
this objective data with the subjective information
that they're providing to us, and it correlates
with the questions that we asked them
throughout the study. We asked them about
everyday discrimination. We ask about vigilance. We ask about how
they're perceived in their everyday lives. So are you able to find
time to exercise at all? PATIENT: I've been
doing better with that. TOLUWALASE AJAYI: What I
appreciate about Fitbit and this initiative
is it allows me to give these devices to people
who may not have access to it.
It allows people who don't
traditionally participate in research to give their
voice, share their information so that we can be more informed. And by having a more diverse
informative database, you're actually able to break
down health disparities. If you don't have
representation in your cohorts, you can't have a
representative treatment. PowerMom FIRST and PowerMom
is about everyday people working with researchers to find
an answer, to find solutions. We can't fight systemic
racism without the people who are experiencing it,
without their input, without their feedback,
without their information.
And we, as doctors,
as researchers, want to work with
this population. We want to make it easier. We want to hear you, and
we want to make it better. [MUSIC PLAYING] KAREN DESALVO: I want to thank
Dr. Ajayi for sharing more about her personal story and
her incredibly important work advancing equitable
health, research, and more equitable health. Today, you've heard
about our approach to partnering with
consumers, their caregivers, and communities towards
a vision of billions of healthier people. Together, we're building safe,
effective, and trusted products that help everyone everywhere
live a healthier life. To learn more about our work,
please visit health.google.
Thank you for joining
us and be well. [MUSIC PLAYING] .
Free Coupon for Discounts on Pharmacy Medications
