A strong argument can be made that the health care field is one of the most information-intensive sectors in the U.S. economy and avoidance of the rapid advances in information technology is no longer an option. Consequently, the study of health care information technology and systems has become central to health care delivery effectiveness. This course covers the modern application of information technology that is critical to supporting the vision and operational knowledge of the health care leaders in managing the health care organization. Heath care decision-makers have to meet head-on the dynamic challenges of health care delivery quality, cost, access, and regulatory control. Additionally, this course integrates the Healthcare Information System as integral to the Quality Assurance Tracking Programs including measurement of systems inputs, processes, and outputs with special emphasis on systems outcomes research and organizational accountability to its various stakeholders, not the least of which are government regulators.
Module 3: Wearables
Loading...
From the course by Icahn School of Medicine at Mount Sinai
Health Care IT: Challenges and Opportunities
107 ratings
From the lesson
Module 3
Meet the Instructors
Bruce J Darrow, MD, PhDSENIOR VICE PRESIDENT MSHS IT | ASSOCIATE PROFESSOR of Cardiology & Population Health Science and Policy
Cardiology
Nicholas G Genes, MD, PhDASSOCIATE PROFESSOR | Emergency Medicine ASSOCIATE PROFESSOR | Genetics and Genomic Sciences
Emergency Medicine
[MUSIC]
Hello, in this lecture we'll be talking about wearable devices and
their potential for impacting patient health.
We'll start with the potential of wearables and some of the risk.
We'll delve into some technical considerations.
We'll look at some unique challenges about the data collected.
And then we'll look ahead to the future, and see what the future has in store.
So in terms of potential, as I mentioned in the health lecture.
There's enormous interest in detecting disease before
they become clinically apparent.
If a patient has a chronic illness and
we're worried about them getting hospitalized.
A wearable can monitor that patient and help assess risk, and
allow earlier intervention.
You can imagine a dash board, much like this one.
This is from a company that was called Pulse.io,
that I think has gone out of business, but I found their slide very compelling.
They attempted to kind of collect data from a bunch of different devices.
From smart pill bottles from fitness trackers, from Wi-Fi scales and
develop a composite picture that could be then given to subscriber.
Like a family member, a loved one or a doctor or a healthcare team.
And someone would been get a snapshot of the patients well being, and
it's kind of Orwellian look in to a patients life.
But you can get all kinds of information such as sleep quality, and how often
the refrigerator door is being open, how often the pill bottles are being open.
When someone is coming and going, how many steps they're taking, and
things like blood sugar and weight and other scales.
So I can understand the appeal of wearables, and kind of giving a snapshot.
What's not clear though is when numbers don't add up or
don't trend to the direction that we want to see, how do we intervene?
That's still very much unclear.
Wearables, we can imagine them as just a consumer version of what we have in
healthcare.
We do telemetry monitoring in the hospital,
we track all kinds of vital signs over time.
We can view wearables as an offshoot of that field or
we can view it as an entirely new field,
where we're collecting data that doctors never used to collect on a regular basis.
And we don't really know the value of monitoring everyone at all times for
things like heart rate and step counts and so forth.
And it totally, it seems that this data has value, but
the truth is the large trial aren't there yet to help us make sense of some of this.
I'll give you an example.
This is a tweet from Eric Topol where he talks about
a patient who had an Apple watch that was tracking heart rate.
And it found the heart rate was going up and down and all over the place.
And occasionally, the patient would feel some dizziness and
seemed to correlate with the variable heart rates.
And so he thought he had sick sinus syndrome which is a real entity.
It's actually pretty rare and unusual, but he went to his doctor.
They confirmed the diagnosis, the patient got a pacemaker.
Kind of sounds like a victory.
It sounds like the Apple watch unearthed a condition that
could have led to harm in the patient had not had it checked out.
I would argue the opposite, that if the patient was feeling dizzy and
checking their own pulse,
they would have been able to achieve everything that the Apple watch did.
And there are times when you work out wearing an Apple watch and
the heart rate detection goes off kilter.
Turns out that it’s very reliant on a steady blood flow to your hand.
So if you're lifting weights, you disrupt the blood flow slightly and
the watch can figure out your actual pulse.
And if you look at that trend over time,
you would may be diagnose yourself with sick sinus syndrome.
Apple watches not marketed as a sick sinus syndrome device detector.
It's marketed as a fitness and health tool.
So if you happen to use it to detect a disease, that's great.
But you have to ask yourself,
how many times is the Apple watch leaving to unnecessary workups or
scaring patients that actually are quiet healthy?
What we would call a false positive with expensive or
even risky interventions that are triggered by a readings that you get on
a consumer device that was never really even intended to detect disease.
So in addition to false positives that lead to sometimes unnecessary work ups,
expensive or even risky medical interventions.
There is also such a thing as lead time bias where detecting
a condition early seems to give an opportunity to intervene.
However, it turns out that no interventions are effective, and
just knowing about the condition earlier.
Well that might have some benefit on its own, but
doesn't actually improve mortality or improve outcomes.
However, we are pretty confident that by preventing exacerbations of chronic
disease, we will ultimately improve care, reduce cost, and improve outcomes.
We know that wearables like apps, they can improve insight to disease.
They can monitor conditions, and they can facilitate communication and
early intervention.
Let's go through some examples of wearables, and
see the broad array of devices that are available today.
The most popular are undoubtedly the fitness wearables, step counters.
And now some are equipped with heart rate monitors, and
also even GPS positioning devices.
So that you can actually really track how far you run or bicycle.
This in the form of the Fitbit, the Jawbone,
the old Nike FuelBand and the new Apple Watch.
But beyond just movement detection, you can actually drill down and
get some fine grain movement in terms of, for instance, Parkinson's.
You can actually assess the effectiveness of your Parkinson's
medication by how much a watch or a wristband is assessing tremor.
And there's actually epilepsy bands now that will detect whether you have
a seizure.
And they can alert care coordinators or
family members that you had a seizure, just based on the movement of the band.
There's also posture detectors,
like the Lumo Lift that will vibrate, if you're slouching.
They can be programmed to fire every few minutes, if you're not sitting upright.
And with related concepts, there's a lot of wearables that will
measure your respiratory rate, your running style.
And make sure that your cardiovascularly doing
as good as you can, in terms of your fitness.
And we use in the hospital pulse ox, pulse oxygenation devices
that can detect the level of oxygen saturation in your bloodstream.
Those have been converted to wearables and can bin information to your phone.
But in a related concept, there's also indoor air quality monitors.
We have outdoor monitors that ran by the government and
the National Weather Service, but indoor air quality.
I mean, we tend to spend most of our time indoors, so it's interesting
to kind of see if our pulse slot can be affected by local air quality and
our devices like this Zoat TCOA that can do that fast.
In addition to monitoring heart rate on a watch or
wristband, you can actually monitor your
heart rate in a more detail fashion with your life core and the cardiac.
This is an app and a companion device that slides on your phone or slides over
your finger tips and lets you record a one lead EKG, an electrocardiogram.
So, you can actually see the rhythm that your heart is beating in and the access.
You can't diagnose a heart attack with this, but
it is FDA approved to diagnose dangerous dysrhythmias that can be apparent.
There are blood pressure casp,
that instead of the homemade single manometer that you has a old fashion dial.
Instead, you can record the data directly to your phone and transmit it.
You can trend it overtime.
There are even temperature sensors to round out the vital signs.
A temperature patch that you slide on your infant would kind of
give you a continuous read out of the baby's temperature.
Some nervous parents are really excited about this.
However, all the pediatricians I've talked to think that this is technology that's
looking for a problem.
I've never seen a pediatricians that have only we detected this child's fever
sooner.
We would been able to have a better outcome.
For diabetics that are reluctant to stick themselves in the finger repeatedly or
have an infusion pump monitor in their blood sugar,
there is a smart contact lens that's been under development for
some time where you actually wear a lens that's monitoring the blood
sugar in your tears or just in the fluid that bathes your eye.
And this could provide a continuous method of collecting measuring blood sugar or
measuring sugar that can help guide diabetes management and
then there are devices like the Wi-Fi scale to track your weight.
Allow the devices to put the Wi-Fi broadcast in link.
This might not be the first thing to come to mind.
However, it turns out that it makes it easy to plot progress overtime.
If you are beaming your weight to your personal computer,
you can plot it on a spreadsheet.
You can view it on an app on your phone and you even have the capability of
putting it on your social media pages, your Facebook page.
If you want social support, if you want people to comment and
encourage you to continue to lose weight, that's certainly one way to go.
There are other wearables that claim to track mental state.
The device on your left is called a spire and it's clipped onto your shirt,
and it can possibly measure how you're breathing, and
they claim that you'd be able to sense to whether you're stress and you're
taking a lot of rapid shallow breathes versus relax, and calm deep breaths.
And it will actually bugs you with the things you're stressed and
actually prompt you to take some deep breathes, and refocus.
I haven't tried it.
I'm not sure it works as advertised, but compelling.
There are other devices too that slide on over your head, like a cap and
can monitor your brainwaves.
This is important for epileptics, but also has mood tracking capabilities as well.
And of course, the same brainwave detectors can monitor
how much you spend in restful quality sleep.
Let's talk a little bit about some technical considerations
when we look at wearables and especially
we're confident that these questions have been answered for medical grade devices.
Not yet confident that the consumer devices out there
are doing a satisfactory job of this, but friends to start with,
we should be careful and define accuracy and precision well.
Accuracy is how close to gold standard or true value a readout gives you.
Precision is about reproducibility and whether running the same device,
the same sensing over and over is giving you similar results.
They're two important concepts.
Obviously, we want a device that does both well, but they are separate concepts.
Also when we talk about devices, there's a lot of signal and noise and
signal to noise ratio.
A lot of filtering that has to happen as we think about detecting a heart rate or
electromyography, a muscle signal.
You are passing through a lot of skin and passing through a sensor, and
then feeding that to a phone or a personal computer.
That signal has to be filtered.
A lot of noise has to be filtered out and
the signal has to be digitized for transmission.
Every step along this process is an opportunity to introduce error and
to make the signal farther from the source of truth.
And then finally, there's aliasing and sampling when you can actually,
if you're not sampling a signal at the right interval, you can actually miss
a trend altogether and might have an inaccurate heart rate measurement or
any number of errors can creep in if you're not sampling properly.
I don't want to delve too much into the details, but these are things that occupy
us a lot on the engineering side of consumer wearables.
But there's also challenges in just aggregating this data,
analyzing it and sharing it.
The folks behind the jawbone lift.
They had a really cool device that people wear to track their movement and
wear it to sleep, and there was an earthquake in the Simi Valley area.
Actually, I'm sorry, the Napa Valley area a couple of years ago.
And the engineers at Jawbone, they noticed it because they
saw a lot of people waking up in the middle of the night.
In fact, they could kind of make their own little seismograph by seeing
how many people woke up in different areas and plotting it over time.
And so they were able to kind of say, the epicenter of the earthquake
was in Napa Valley, because that's where the most people woke up and
it took them the longest to go back to sleep.
But who owns this data actually, because researchers were really thrilled
to kind of see this blog post from the job owned people.
However, the job owned people were not willing to share their data.
They held onto it, because it's according to the terms of use,
it’s their data and they wanted to do with it what they wanted to.
So opportunities for research and delving into this and
studying the accuracy, the precision, the sampling, the error.
These are lost on regular investigators.
Instead, we have to rely on the companies that stand to make the most money.
They're the ones that are controlling the flow of information and the study.
So the data scientist which I think has an enormous role here in interpreting
the data that's collected from patients They are sometimes at the biggest
disadvantage because the companies are reluctant to release this information.
We've seen time and again that sometimes wearables are used for medical purposes.
But when we try to study the reproducibility of this data we're
stymied, because the companies simply don't want to release the data.
Ultimately, if we want a situation where patients are acting responsibly on
the data that's collected, if we want a situation where care coordinators and
doctors know to interpret the data and
what steps to take when they see a trend heading in the wrong direction.
We're going to need to open up this data to researchers and
data scientists who are going to be able to properly assess this.
And yet we're not there yet.
Also, there's the question of how
to share this data effectively with the right people.
Right now, every
separate device that you're collecting tends to go to a separate portal.
So your Fitbit, you have to login.
If you want to see your Fitbit data, you log into Fitbit.
If you want to see your Apple Watch data, you've got to use the Apple apps.
And there's no common portal or interface.
And certainly, the electronic health records that we use, like Epic and
Cerner, while some of them are taking steps to handle this data and
receive it, that's certainly not universal.
And a lot of physicians are actually unwilling to receive this data because
they don't want to be liable for acting on it, especially if worrisome data comes in
in the middle of the night or while someone is on vacation.
And also EHR Integration tends to be very expensive and
some of the proprietary algorithms and visualization tools get lost.
So a lot of challenges in how to best interpret this data.
Still, I want to highlight at least one success story and
take a look toward the future.
There was a case just recently in New Jersey where a patient,
40 something year old man, previously very healthy, no medical problems,
he had a seizure and he was brought to the emergency department.
And he was found to be in a dysrhythmia called atrial fibrillation.
And this was all news to him.
And it turns out a lot of doctors and a lot of emergency departments,
if the patient is young and healthy and can state when their atrial fibrillation
started, if it's within 48 hours, they'll get cardioverted.
They'll get shocked with paddles, and they can be sent home on their way for
further restratification.
But if the patient doesn't know the onset of the interpolation,
if they don't remember when the palpitations first started,
then they might have to be anticoagulated.
And they might have to be started on medications to slow their rate down.
And this is sometimes risky, and sometimes unpopular with patients.
In this case, the patient did not know the onset of their atrial fibrillation,
they did not recall feeling palpitations at any time.
However, the patient had a Fitbit, and so
the Fitbit was interrogated by the doctors in the emergency department.
And they were able to say that the rapid
heart rate started at the same time as the seizure.
In fact, triggered the seizure and at no time in the weeks preceding was there
a rapid heart rate detected, at least not one that wasn't associated with exercise.
So the doctors were reasonably confident
that this was brand new onset afib, and the patient was cardioverted.
And this is being hailed as the first time, at least that's being written
up in the peer review literature, the first time that a wearable fitness tracker
is being used to influence emergency medical decisions.
That's great, but it's also a little concerning because certainly,
no one at Fitbit is building these fitness trackers with the intent of serving
doctors in this particular circumstance.
As the doctors who wrote up this case say,
they used their discretion in interpreting the data.
They know that it's not a medical grade device, but
it seemed like it fit the story.
And so that's why they went ahead and cardioverted the patient.
They shocked the patient, and the patient had a good outcome in this case.
So in today's practice, we're seeing patients,
they are wearing these devices, tracking their fitness,
tracking other aspects of their health, tracking vital signs.
And they are even in some cases sharing their data with loved ones,
with care coordinators and with providers.
Still too often this data is siloed, it remains stuck in different apps or
portals or devices.
It is not being shared in a universally recognized way and it remains isolated
from the data scientists and researchers that can make the most use of it.
A lot of providers are reluctant to sign on to the additional responsibilities and
some of the liabilities that would come, especially since we don't have a lot of
demonstrated evidence that these devices are improving outcomes or lowering costs.
But we are seeing some forward thinking healthcare centers like ours and
others that are taking the initiative with pilot programs to start to
collect this data and make sense of it.
In the future, we do think that consumers, and I use that word in quotes,
patients and consumers are going to start tracking more and more of their health.
In part to motivate fitness, but if it's aggregated and collected in a broad scale,
we might start to see new trends, and make new sense of data.
We might be able to detect disease earlier than has been detected before,
before clinical symptoms are manifesting themselves.
And we hope that this
kind of data really leads to new therapies and new cures down the road.
Patients will be able to use their wearables to not just detect their
condition and optimize their health, but actually prevent deterioration and
prevent exacerbations of care, hoping that patients live longer,
happier lives, and while spending less time in the hospital.
Thank you.
[MUSIC]
Explore our Catalog
Join for free and get personalized recommendations, updates and offers.
Coursera provides universal access to the world’s best education,
partnering with top universities and organizations to offer courses online.
© 2018 Coursera Inc. All rights reserved.
