Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
The most interesting one that I've
0:02
seen is is actually
0:05
a cure for a treatment for lazy
0:07
eyes. No, a lot of
0:09
kids are born with the two eyes that don't
0:11
have binocular vision. You know, they work kind
0:13
of independently of each other when they
0:16
should. So what this company is doing is
0:18
they created a game a game, they
0:20
gamified it. They worked actually with the gaming company,
0:22
you'll be solved, I think. And they said let's create
0:24
a game that exercises that forces
0:27
the eyes to align. So it's a video
0:29
game out, literally looks like a video
0:31
game, but they have proved
0:34
that it will actually fix your lazy
0:37
eye. So
0:49
thank you very much for joining us. My pleasure
0:52
and very honored to be here. so just to start
0:54
with, I wanted to ask you about something which is very
0:56
near and dear to my heart, which is sleep.
0:58
Vasanth. And I have had some debates about whether,
1:00
you know, there's all these companies that sell mattresses.
1:03
They'll give you a better night's sleep and all these sleep trackers
1:05
that are on your arm. And we've had some debates
1:07
about whether you could actually get like nine hours
1:09
of sleep in seven hours with these new pieces of technology.
1:12
I personally don't believe that's true.
1:15
What do you think? Are you bullish on this or bearish? I
1:17
don't know yet. I don't know yet because it's,
1:19
it's a little complicated. It's a matter
1:21
of the technology can only do so
1:23
much, right? The technology can help
1:25
you pick up what's your sleep patterns, right?
1:27
So it can help you. You wear all these sensors, you
1:29
sleep on the mattresses that are sensing how
1:32
long you're sleeping. What's the quality of your sleep.
1:34
Do you move around too much? Do you snore, et
1:36
cetera, et cetera. And it can
1:38
help you create a pattern
1:41
of your sleep habits. There
1:43
are, you can also look at evidence
1:45
and based on the evidence, you can create
1:48
apps and other kinds of software
1:50
that can tell you how to behave
1:53
so that you can make your sleep better. Right.
1:55
So if you're, if you're sleeping, You
1:57
know, really late after midnight and
1:59
your sensor picks that up. It
2:02
might give you a nudge that, Hey, you
2:04
might consider having an earlier bedtime, but
2:07
to really, for you to sleep the nine
2:09
at 10 hours, you have to follow that
2:11
advice. And. As
2:14
humans, we are notoriously
2:16
irrational, you're notoriously
2:18
international, especially when it comes to things
2:20
that are you know, to do with our help and
2:22
to do with our financial
2:25
wellbeing, things like that, things that are
2:27
really important. We rarely follow
2:29
the, the, the good advice on that. So
2:32
that to me is the bigger problem to solve,
2:35
you know, to have that behavior modification
2:38
that comes out of nudges. Now.
2:40
Today, there are a lot of There's a lot of
2:43
technology. There's a lot of software, a lot of mobile
2:45
apps that do a pretty good job
2:47
of understanding human behavior
2:50
and therefore trying to modify it. But
2:52
I haven't seen anything that's highly,
2:55
it's been highly successful at changing human
2:57
behavior. There really isn't a
2:59
great habit building app, which
3:01
is, which is what you would meet. Right.
3:03
So, which is what you would need that, that really understands
3:06
why you behave a certain way and
3:08
then give you that actionable insight
3:10
on what to do about it, to change that behavior.
3:13
Once we close that loop from
3:15
understanding our sleep patterns, to
3:18
getting the proper nudge, to actually acting
3:20
upon the nudge. It's getting better. Right.
3:23
I actually read a book by this by
3:25
this very prominent sleep scientist at Berkeley
3:27
and in Matthew Walker, he wrote this book called why we sleep.
3:30
So he said in it that most Americans
3:32
only get like six hours of, of decent
3:35
sleep at night. And it's not because they
3:37
don't get in bed for eight hours. It's because
3:39
they just blast their eyes with blue lights
3:41
because they drink a cup of coffee, like three hours beforehand.
3:44
So. It does make sense. It
3:46
does make sense that if you can actually have data that says
3:48
like, Hey, like you're in bed, which
3:50
we're not actually getting good sleep, then
3:52
that insight could at least in theory,
3:55
prompt better behavior. Yeah.
3:57
I've been thinking about the Berlin and FinTech,
4:00
right? If you have a great banking app,
4:02
which, which a lot of them are now, which
4:05
gives you your spending patterns, right. You
4:07
can look at it and with great data
4:09
visualization, you can look at it and go.
4:12
Dang. I didn't realize that I, I spent
4:14
so much of my money eating out. And
4:17
then you might do something about it yourself,
4:19
or if the app is even
4:21
better, or if you have another way of,
4:23
of reminding you in the moment
4:26
that the next time you're going out, it senses
4:28
that you're blowing out and tells you stop. You
4:30
know, you're, you're going to regret this. This is going to
4:32
blow your budget. That will drive real
4:34
behavior change. Right? So it's the same with helpful
4:37
behaviors. You know, if you have, if you have
4:39
the right insight and you have an aha
4:41
moment, you'll be that much more motivated
4:44
to change your behavior. I definitely
4:46
agree with that. I think that like billions
4:48
of dollars of startups have come out of like, not
4:50
necessarily brand new technology, but
4:52
UX solutions like Twitter is
4:54
nothing but a UX solution. Slack is nothing but
4:56
a UX solution. And I say nothing,
4:58
but as if that's a small thing, but it's a massive
5:01
thing. It completely changes like how human
5:03
beings interact with the world. Yeah,
5:05
no, I, I you're, you're talking about
5:07
a subject that I'm so passionate
5:09
about. I always see I'm a, I
5:12
usually say I'm a technologist, right.
5:14
But actually I'm not, I'm actually
5:17
much more passionate about the
5:19
use of technology. Right? How
5:21
do people use it? How, how do you
5:23
design it so that the impact
5:26
of technology is even greater that
5:28
the intended use actually drives.
5:31
The expected output, right? So
5:33
the technology is almost, you
5:35
know, it's just it's just the infrastructure,
5:38
right? It's what you do with the technology. That's
5:40
more important and, and how you, how
5:42
you design it so that do the end user,
5:44
the technology should be invisible, right? So
5:46
it's not so much that technology for
5:48
technology's sake, but the it's
5:51
it's, as you said, it's the user experience
5:53
of that particular technology. So,
5:55
how do we solve this? How do we solve these problems? You said
5:57
that like you, you know, these apps will tell people
5:59
to go to sleep at this time, but they won't actually listen
6:02
to it. So what are some ways
6:04
that this can actually succeed in digital health? It's
6:06
the holy grail, right? How do you change human behavior?
6:08
It's it's not, it's not easy. You'll
6:11
see a little bits
6:13
of brilliance and the more you personalize
6:15
it, the more you understand individual drivers
6:18
of human behavior, the
6:20
better you can be at designing technology.
6:22
Right and designing the solution. So
6:25
in digital health, I'll give you a very
6:27
common problem is on adherence
6:29
to medication. You finally
6:32
went to a doctor, told them your symptoms
6:34
got a diagnosis. They wrote you a prescription.
6:37
You'll take it the first two days, and then slowly
6:39
your habit starts deteriorating.
6:42
And the number of people
6:44
who are what we call nonadherent or noncompliant
6:47
to their, to what the doctor
6:49
said about taking the medication about,
6:52
you know, lifestyle choices, et
6:54
cetera, is tremendous 50%
6:56
of the people in many areas in many fairly
6:59
significant diseases fall off
7:01
treatment. Or they'll start delaying treatment
7:04
and it's not always forgetfulness.
7:07
So non-adherence is not really a problem.
7:10
It's the symptom of something that's going
7:12
on in their lives. Right? So some
7:14
people might have beliefs about medicine
7:16
saying, eh, I don't think you
7:18
know, this, this has too many side effects. I'm sure
7:20
this is poison. I'm not going to put stuff in my body.
7:23
That's poisonous. I think my disease
7:25
is fine. It could be completely erroneous,
7:28
completely erroneous in that case.
7:30
It's a different solution than somebody who just
7:32
forgets. Right? So somebody like
7:35
me, I just forget to take medication on time
7:37
because I'm not disciplined. So for me,
7:39
a timely reminder would be a highly effective
7:42
solution in getting me to change
7:44
behavior. But for the person who has believed
7:46
that, you know, I, I don't really think
7:48
this, I should be taking this I'm okay. Even
7:50
though the doctor said, so. Then there
7:53
is you have to treat it differently that
7:55
the design should push perhaps proper
7:57
education for them, you know, and,
7:59
and maybe connecting them to people
8:01
like them or people that they trust
8:04
telling them more about the drug or whatever.
8:06
There could be some people who without telling
8:09
their doctor actually find the
8:11
copay too expensive. So
8:13
for them, if you send them a reminder, they're only going to
8:15
be more upset about it that, Hey, stopped reminding
8:17
me that I have to take this every day. I can't afford this.
8:20
Right. I can only afford to take it every
8:22
other day. In that case, perhaps
8:24
the solution is you,
8:27
you showcase some of the
8:29
financial programs that
8:31
companies offer government software, et cetera,
8:34
that they may not be aware. So the, the
8:36
point is that. It has to be that
8:38
much personalized. The intervention
8:40
has to be personalized. It has to be at
8:42
the right time. It has to be using the right
8:45
channels. Some people respond
8:47
really well to SMS messages, others
8:49
respond to their, their
8:52
child picking up the phone, calling them, right.
8:54
So somehow you have to really talk
8:57
fully create a very individualized
8:59
solution for them. Very difficult to
9:01
do, but we are heading towards that. You're
9:03
heading, we've already had it to where we already have
9:06
personalized medicine, right? People are doing,
9:08
finding out exactly what genetic makeup
9:10
you have and therefore you should be using this.
9:12
Therapy versus that. Similarly we
9:15
are getting better and better with the AI
9:17
and machine learning algorithms, but understanding
9:20
human behavior and even predicting
9:22
as I had as working on a project
9:24
where we were developing an algorithm, which
9:27
we. We're predicting, who's
9:29
going to drop off the treatment after the, you
9:31
know, out of the a hundred people,
9:34
which ones would drop off because
9:36
it's a machine learning algorithm and based on multiple
9:38
inputs, you could start
9:40
seeing patterns. If you have great data.
9:42
There are large amounts of it and smart
9:45
people looking for signals in that data,
9:48
you could perhaps pick up enough
9:50
patterns to figure out, ah, this person
9:52
they're respond better to this message
9:54
sent by this channel at this time.
9:57
And then they'll do something about it, right?
10:00
It sounds almost like you've built an algorithm
10:02
that tells you the people that are the most self-destructive.
10:07
Yeah, I guess. Yeah.
10:10
I think there's like some very obvious parallels. I mean, I think
10:13
it just extends beyond trying to change human behavior
10:15
to help other humans, like,
10:17
let me change your behavior so you can help yourself is
10:20
such a hard thing to do. Even, even in
10:22
if you look at like the financial markets, all the
10:24
apps that are being developed to help people trade really
10:27
the, they sort of incentivize self-destructive behavior.
10:29
I've talked to farmers about this a lot, but it's like, I
10:32
I'm not even certain that it's a UI. Whether a
10:34
UI fix will help people change their behavior.
10:37
I think it's like an incentive issue, which I think you mentioned
10:39
what would you say are sort of the best
10:41
incentives to that?
10:44
That actually change human behavior? This will incentivize
10:46
you to not be self destructive. Is there something
10:48
like that in the digital health space? So
10:51
that's what we look for. So in patients,
10:53
what we are finding that if you tell them, Hey
10:55
you will get better. Your blood pressure will
10:58
drop by this much. And your, you
11:00
know, this particular LDL will
11:02
drop, will drop by this much. If you stay on
11:04
treatment, that message doesn't
11:06
always resonate. What resonates
11:09
is something deeply personal? So,
11:12
you know, for somebody who
11:14
has a heart condition, let's say, and
11:16
they have a hard time keeping on medication, they
11:19
respond better to their own
11:21
metric and say, you will, you
11:23
have a higher chance of attending your
11:25
grandson's graduation. If
11:28
you take this medication versus not.
11:30
Right. That resonates. That's a big
11:32
driver for them because that's really personal.
11:34
So again, everybody's motivation and
11:36
drivers are personal. That's why it's so
11:39
hard to solve. So not everybody
11:41
thinks about how I'll do better,
11:43
because look, my LDL will get better.
11:46
They have different metrics that they measure. So
11:48
people who have mobility issues. Every
11:51
time we talk to them about, you know, certain
11:53
validated biomarkers
11:56
that, oh, this particular level of this biomarker
11:59
in your blood is reducing and your RH
12:01
factors reduce. It didn't mean anything
12:03
to them. We told them, Hey, do you remember that?
12:05
You used to be able to walk the dog one
12:08
block. Now you walk at five blocks and
12:10
they go, oh yeah, that makes sense. Yeah. My medicine
12:12
is working. So that drives.
12:15
Motivation is very
12:17
different and it's very personal and
12:20
that's why it's so hard to design for. But,
12:22
but if you can personalize that motivation for
12:25
someone like me, I don't,
12:28
I don't respond not deliberately, but I don't
12:30
respond well to Hey, you'll feel
12:32
better. I respond better
12:34
to, Hey, you, you need to be better
12:36
to, to be better for your,
12:38
to be able to help your kids. Right.
12:41
So it, it, to me, it's a little
12:43
bit more external, so it really depends
12:45
on your personality type. And what drives
12:48
you? I think this kind
12:50
of goes back to the UX thing that
12:52
I was mentioning earlier as well, in the sense that
12:54
like Facebook has absolutely
12:56
dominated adherence. Like I think the
12:59
average person spends like two hours a day or something insane
13:01
like that on social media, Facebook, Twitter,
13:04
Instagram, whatever. And they employ
13:06
a lot of, kind of maybe dark shady tactics
13:08
to do this, but a big one that they use is emotion.
13:11
That's why they show you stuff that gets you angry
13:13
or gets you excited or joyous.
13:15
And it, yeah, it kind of sounds
13:17
like you guys are. Maybe employing some of the similar,
13:19
similar techniques, but for good, rather than
13:22
for lining your wallet. Yeah.
13:25
I mean absolutely. When you design for
13:27
behavior change you have to connect
13:29
with the emotion. Right? Most of human behavior
13:32
is driven by us feeling
13:34
good about ourselves doing something.
13:36
If it doesn't make you feel good, you're not gonna do
13:38
it. Right. Most people don't stick to exercise
13:40
regimens because it's so darn
13:42
painful the first few weeks it's so uncomfortable,
13:45
right? If you get past that threshold,
13:48
you start feeling good. You've got that endorphin
13:50
rush in and things like that. Apparently
13:52
it becomes addictive. I said, apparently
13:55
because I haven't been, I
13:58
haven't treated that, that threshold.
14:00
It's always painful for me. So people will do
14:02
things that make them feel good. So
14:05
for Facebook and that instance. The
14:07
number of likes, they were getting the
14:09
number of people responding or commenting
14:11
to them about things that they posted
14:14
is a huge driver. You feel good? Oh,
14:16
I got noticed, right. Similarly
14:18
able to express your anger about
14:21
other things and your frustration, et cetera,
14:23
by, you know, commenting randomly on
14:25
things and, you know, making angry faces.
14:28
It also makes you feel good. You know, you got
14:30
to let out all your angst. So,
14:33
so they, everything has to tap into behavior.
14:35
That's why, I mean, you have to have been doing
14:37
the designing helpful
14:40
products and helpful technology. You
14:42
really have to be with really very
14:44
empathetic to the patient. It's
14:47
it's it's consumer technology is actually
14:50
easier because most consumers
14:53
will, as you said, you know, 90%
14:55
of behavior will be the same. Most people feel
14:57
really good about being able to get noticed
14:59
and in when you're a patient
15:02
you're dealing with so much else,
15:04
right. You're, you're dealing with the
15:06
disease burden. In addition
15:08
to being irrational human beings. So
15:11
it's that much harder to have empathy
15:13
for the disease burden and what it's doing to the
15:15
life, to their life. And
15:17
also to kind of figure
15:20
it out that they're not just a patient
15:22
and most of them don't want to be reminded of
15:25
their sickness. They want to be reminded
15:27
of their health. That's. All
15:30
of that has to factor in, into designing a good
15:32
digital health solution that,
15:35
that that can actually drive behavior change
15:37
and have an impact on, on
15:39
their health outcomes. So are there any
15:41
kind of like takeaway home tips for this,
15:43
like, forget building something digital, but just what
15:45
you've learned about habit forming. Like if
15:47
you wanted to stick to some exercise program
15:49
or follow some good diet, like what techniques would
15:51
you use to self motivate that like I could
15:53
employ a restaurant could employer, our listeners can employ.
15:56
It's, it's fairly well known
15:58
now because of there's. So many people have been working
16:00
on it. You have to
16:02
divide, you
16:04
have to set a goal, an
16:06
achievable goal. So there's a target
16:09
goal that you have to help the person set.
16:11
Right? And then you have
16:13
to divide it up into small achievable
16:16
steps and, you
16:19
know, help them. Get to their
16:21
final goal by, by these incremental
16:23
steps so that they feel successful at
16:25
every stage. That's number
16:28
one, you have to constantly remind
16:30
them of why they are doing that. So
16:32
what is that inner motivation? So
16:34
if their motivation is to attend
16:37
their daughter's reading and therefore
16:39
they need to feel better by then by next
16:41
year, keep reminding them of that so
16:44
that they stay motivated. You have to
16:46
do it in such a
16:48
way that they feel supported.
16:52
Right. So for people who are, who need
16:54
that coaching and that constant support
16:56
coaching really works. So whether it's a human
16:59
or it's a, or it's a chat bot doing
17:01
the coaching, it just depends on the person. That's
17:03
really powerful too. You have to design
17:05
a coaching program, a coach that's constantly
17:08
encouraging and. You have to tell them if you
17:10
fall off the track, it's okay. It's
17:12
okay. You'll get back. Right. So that they don't
17:14
feel like complete failures at any point. So
17:17
there's a bunch of things, but the first and foremost
17:19
is to help them understand what
17:22
things are getting in the way of them not achieving
17:24
their goals. You know, so that the collection
17:26
of data, we are using sensors,
17:29
et cetera, and visualizing it for
17:31
them. This is what you actually
17:33
do is a huge motivator.
17:35
You know, they themselves go, my God was I really
17:37
spending. Four hours
17:39
a day of sitting on the couch, watching TV.
17:42
I can't believe it. Right. So I didn't think
17:44
it was that long. And so once you start
17:46
looking at that you, you, you can get
17:48
a better sense of the design. So apple
17:51
has done a pretty decent job of driving
17:53
healthful behaviors with the apple watch
17:55
and, you know, closing the rings and all that. This
17:57
apply a lot of those principles. The data visualization
18:00
is really simple and, and.
18:02
You know, immediate, so you have to keep it simple.
18:04
Don't make it complicated. Don't make it seem hard
18:07
to then just be kind of positive
18:09
about it. Great example.
18:11
I, at first I, I
18:14
found the stand-up every hour notifications
18:16
on my apple watch to be like the most annoying thing
18:18
in the world. But once I closed
18:20
those rings and got those really fun animations
18:23
a few times now, like I'm
18:25
chasing that high every single day, right?
18:28
Yeah, exactly. So
18:31
speaking of apple, I think they're making,
18:33
I've heard, they're making a lot of plays in the, in the health
18:35
space. So can you kind of expand on what
18:37
exactly their role in this? Yeah,
18:41
so it was interesting deal a
18:44
few years ago. Let's say 10 years give
18:46
or take, right. It's been 10 years now. I can't
18:48
believe it. Well, the, the
18:50
use of technology in health
18:52
care or rather sick care, you
18:54
know, we call it healthcare, but really we
18:57
it's sick care, the use of technology
18:59
and healthcare. Can we, sorry, can you,
19:01
can you expand on that? Like what do you mean by sick care versus
19:03
healthcare? So, so when
19:05
I think about the healthcare ecosystem,
19:08
I'm thinking of doctors, I'm thinking of
19:10
pharmaceutical companies, et cetera, right?
19:12
So these are not talking.
19:16
Actually, if you think about, are they taking care of sickness?
19:18
They're not care of health as
19:20
such what we call. Health
19:22
care actually takes care of sick people.
19:26
Right? And the other piece
19:28
where you're promoting healthy habits and prevention
19:30
of disease is doesn't
19:32
fall under the, probably more
19:34
like I've heard that's more effective. Preventative care is better
19:37
for you than, than like reactive
19:39
of course. But think about it. How many times would
19:42
it, would your doctor, would you go see your doctor
19:44
and get prescribed? I'll go to the gym five times
19:46
and be monitored, right? It's
19:48
it's we do take care of, we
19:50
do. You know, we've got our language
19:53
mixed up here, I think. But apple
19:55
was so the use of technology helped the
19:57
healthcare ecosystem. Let's say all the
19:59
players in the healthcare space, doctors,
20:02
pharmaceutical companies, nurses
20:04
insurance companies, not
20:06
early adopters of technology, not
20:09
early adopters, they're way behind
20:11
the way behind it, the adoption of, of
20:13
cool technologies. So.
20:17
But about 10 years ago, all that changed, right?
20:20
Everybody decided that, Hey, you
20:22
know what, we can't, we can no longer
20:24
just be prescribing drugs. We, it
20:26
has to be more holistic because we're spending
20:28
a lot of money and it's costing our healthcare
20:30
system, a lot of money because people
20:32
are ending up in hospital and having
20:34
all these challenges. So we have to do more. We have
20:36
to do more preventive preventive care, as
20:39
you said, and we have to also make sure
20:41
that once a person has been diagnosed with
20:43
a disease, ensure the best
20:45
possible outcomes, because not only
20:47
is it better for the patient, it's actually better
20:49
for the entire healthcare ecosystem.
20:52
It's cheaper, right? For overall,
20:54
it's cheaper to keep people out of hospital
20:56
and maintained and, and, and on a healthier
20:59
path then B that have them admitted
21:01
in the hospital. So when that
21:03
happened The tech companies
21:05
started going, Hey, I can do stuff. I know technology.
21:08
I can do a lot of cool stuff here. So
21:10
they started getting into the healthcare space and
21:12
then pharma companies, I worked for a big pharma.
21:15
We started going, oh, I can do technology
21:17
and I can, you know, so, so
21:19
all the, all the technologies wanted
21:21
to become healthcare people and the healthcare
21:24
people, pharma, et cetera, wanted to suddenly
21:26
become technologists. So apple
21:28
was one of the first ones because they had
21:30
hardware and they had software and they have
21:32
a fan following of their products
21:35
there. So they've a beautifully
21:37
poised to make a difference.
21:39
So a couple of years ago, Tim cook, I don't know
21:41
if you guys remember. He even said
21:43
I think it was 2019 is talking about the apple
21:45
watch and other services coming up, that
21:48
if he looks to the future, he
21:51
will he predicts that apple
21:53
will be known for making the biggest difference in healthcare.
21:56
Which at that time was like, wait, what? And
22:00
then you started looking at the investments that
22:02
we're making and what they're doing
22:05
is actually, it's actually pretty phenomenal
22:07
because they have the reach, right? So
22:09
to be able to influence lots of lots
22:12
of human lives, they have the fan
22:14
following the habit captured patient.
22:16
Patient population, right? So
22:18
they do not have to reach out to two
22:20
people and force them to adopt anything. People
22:22
already use their products so much.
22:25
Right. So what they do is what
22:27
they've done in the, in, at recently
22:29
is They've added more and
22:32
more health related features
22:34
into their products. So the apple watch
22:37
it has it has all kinds of sensors in
22:39
there that can predict even heart conditions
22:42
and pick up and diagnose heart conditions.
22:44
Before the patient themselves is aware that I might
22:47
have an issue. Right? So
22:49
one of the, so there they're
22:51
started partnering with a lot of pharma companies. They
22:53
did a lot of health study and, you know, apple
22:55
health study with millions of people at
22:57
bear. They're collecting data through the sensors,
23:00
in their watch and through the
23:02
demographic data for the, from the use of their phone
23:05
and based on all that. They
23:07
can, they're coming up with algorithms that
23:09
are pretty good at predicting certain diseases,
23:12
they're diseases or disorders
23:15
or conditions that can be picked up. Depression
23:17
can be picked up. As I said, heart-related
23:19
a lot of things can be the sensors can sense.
23:22
They're starting to actually have go
23:24
into the space of medically regulated
23:27
devices, right? So you're, you have
23:29
to get an FDA clearance. If you're going to say, Hey,
23:31
I can predict heart failure. But you have to have,
23:33
you have to get the evidence and show it to the
23:35
FDA. So that's happening. So I think
23:37
that they're doing a fairly good job of this. The,
23:40
the, where
23:43
I think they can do better is.
23:45
Working in the regulated space, because at,
23:48
at its heart, apple is a, is a consumer
23:50
company. And they'll have
23:52
to work with the FDA and with with the
23:54
regulating bodies to make, to
23:56
figure out how do you regulate a consumer
23:59
device, which has so many medical
24:01
features in it, because almost everything gets
24:03
regulated as a medical device. And then
24:05
it's really difficult to update
24:07
it and to test it and to produce all the right
24:10
evidence, et cetera, et cetera. I
24:12
wanted to talk about that because Yeah, sort of, sort
24:14
of, one of the hardest things about building in healthcare is because the
24:16
biggest bottleneck being the regulation,
24:18
you have to go through all these sort of watchdogs
24:21
and the FDA you were mentioning. And
24:23
so for us and I always talking about, well, when is
24:25
the quote, unquote two
24:27
guys in a, or people in a garage
24:30
going to build the next great software company.
24:32
Is that even possible in the healthcare space?
24:35
How are you going to go up against the sort of behemoths
24:37
that are no other way around all these regulations?
24:39
It's possible. If you partner with the right people,
24:42
that's the business model. I've seen a lot of the
24:44
startups take. They, they
24:46
have a great idea that really innovative.
24:48
And I work with a lot of startups, right? Because
24:50
the real innovation comes out of
24:53
this kind of passion that people have. So,
24:55
what they'll do is, and there's a lot of investment
24:57
money right now flowing into digital health, a lot
25:00
of investors and kind of, sometimes
25:02
it cockamamie things. You go really, they got,
25:05
they got funded for that idea, right?
25:07
So you will always get seed
25:09
money if your idea's good. And if you can show value.
25:12
Right. So the question is now,
25:15
how far can you take it as two
25:17
kids in a garage? You have to start
25:19
getting in and talking to people and leveraging
25:21
the external partnerships. So
25:23
that's, that's a very common business model. So
25:25
I rebook with a lot of startups. They
25:27
have cool ideas. We know
25:29
the FDA, we know what it takes to do clinical
25:31
research and to produce the evidence
25:33
that's required by the FDA. And
25:37
it's a win-win. So, so
25:39
what happens is these partnerships
25:42
made in heaven, but they're not always successful
25:44
because, you know, if you're talking to a large
25:46
company, there's a lot of bureaucracy. There's a lot
25:48
of things like that. And, and smaller
25:50
companies, just the startups
25:53
with two kids in a garage have a very different mindset.
25:56
So it has to be done carefully,
25:58
but that's one Bible will do it. The
26:00
other way is if you start showing incremental
26:03
value, you know, so if you start
26:05
showing incremental value, you will get the investment.
26:09
So I have a lot of examples of companies
26:11
that are, can you give us, can
26:13
you give us some something you'd like to highlight? Yeah.
26:16
So there's, there's there's a few different companies.
26:18
There's one that we worked with in the heart failures.
26:21
Young peoples decided that let's
26:23
start collecting data of
26:25
all kinds, temperature blood
26:27
pressure, heart rate, heart rate, variability,
26:30
and weight, weight, gain
26:33
shortness of breath, things like that, whatever
26:35
we can collect and see if we can figure,
26:38
feed it, to do an algorithm and figure out if it's
26:40
heart failure. Right. So figure it out and,
26:42
and, and be able to predict
26:44
heart failure before. It becomes
26:46
really apparent. So there's
26:48
a company called bioform. This has a lot
26:50
of good AI algorithms. They
26:54
just have a great credibility
26:56
because they partnered first with the right academic
26:58
institutions to, to went
27:00
to the really credible institutions
27:03
and Did studies with them
27:06
first they're smaller studies that perhaps don't
27:08
require as much of an investment.
27:10
And then they started talking to other bigger
27:12
companies and other bigger investors saying,
27:14
Hey, we do a larger study
27:16
than perhaps this this particular
27:18
algorithm that we're using might get widespread
27:21
adoption. And, and
27:23
the journey continues from there. So there's
27:25
a, quite a few companies like that.
27:28
It's a hard field only because
27:30
of there's so much uncertainty around how,
27:33
what is the exact business model? How do you make money?
27:36
You know, who pays for this? Did that's the harder
27:39
problem right now? You know, people have a lot
27:41
of ideas. What is unclear and what
27:43
you don't have control over is
27:45
who pays for this stuff. So
27:47
would a patient. Pay
27:49
for something. Patients are so used to their
27:52
insurance companies, paying for everything, and
27:54
they're paying just a small copay. That's
27:56
psychologically you, if you tell
27:58
them, oh, this will help with your heart failure.
28:00
The first thing they ask is, wait, is this covered
28:03
by my insurance? And
28:05
insurance companies are harder to convince.
28:07
Doctors are harder to convince is
28:10
this recommended by my doctor? That's the other
28:12
question? Right? So those
28:14
are some other aspects that will prevent
28:17
people from building
28:19
a successful business and that have to be sorted
28:21
out. That whole thing is just being sorted out. Would
28:24
this be easier in a country like Canada,
28:26
where they have government provided healthcare?
28:30
Well as long, but then you have to
28:32
cover the government, right? The government is still the
28:34
payer. So you have to, who is the payer?
28:36
You have to convince them. So
28:38
it either has to be the patient it's actually
28:40
easier. And whether their self pay. So
28:44
where people in countries where people are so much
28:46
used to just paying out of pocket for medicines,
28:48
they'll go. All right. One other thing I have to pay for,
28:50
for being healthy. So they're
28:52
there. The psychology is a
28:54
little better to overcome. But in
28:56
countries where you're used to somebody else.
28:59
Paying a majority of your costs, either your employer
29:02
or your insurance company or the
29:04
government, it becomes much harder
29:06
because now you have other people to convince you have
29:08
got the whole ecosystem to convince that, Hey,
29:10
that, so this is, this is one of the challenges
29:13
in digital therapeutics. So
29:15
It's a relatively new area where we
29:17
are not just saying in digital
29:19
health where we are not just talking about solutions
29:22
that you give in combination
29:24
with the medicine. You're actually making
29:26
claims that the software is
29:28
the medicine. Right? So it's
29:32
yes. And imagine the evidentiary
29:34
burden there, because now you have to show
29:36
that the piece of software, somebody
29:38
using the piece of software compared
29:41
to somebody who's getting the standard of care.
29:43
Right or not using software, but getting the normal,
29:46
whatever treatment they get is will
29:48
do better because it, software is actually curing
29:50
whatever disease. So exactly. Yeah.
29:52
Yeah. So this is for like mental health things
29:55
or it's also for physical elements. It could
29:57
be for physical elements, especially for physical
29:59
ailments as well. Right. Because mental health
30:01
is great. Right? Mental health is the obvious one.
30:04
Because meant a lot of mental, mental
30:06
health related conditions can be treated by
30:08
counseling. Now whether the counselor
30:10
is human versus AI,
30:13
if your AI is really, really good and you're
30:15
using the same principles of cognitive
30:17
behavioral therapy, et cetera, the standard
30:19
psychologic psychology principles
30:22
that that that our counselor would use. You
30:25
know, it, it will be effective. So that's
30:27
one area. The other area of physical
30:29
is a lot of things can be solved
30:31
with physical exercise. So, so things
30:33
that are more actually physically
30:36
structurally you know, effecting your structure
30:38
you can provide the same kind
30:40
of. Physical therapy to an
30:42
app and you could through through digital
30:44
means than you would through having them go
30:46
to a physical therapist. The,
30:49
the most interesting one that I've
30:51
seen is is actually
30:54
a cure for a treatment for lazy
30:56
eyes. No, a lot of
30:59
kids are born with the two eyes that don't
31:01
have binocular vision. You know, they work kind
31:03
of independently of each other when they
31:05
should. So what this company is doing is
31:07
they created a game a game, they
31:09
gamified it. They worked actually with the gaming company,
31:11
you'll be solved, I think. And they said let's create
31:13
a game that exercises that forces
31:16
the eyes to align. So it's a video
31:18
game out, literally looks like a video
31:20
game, but they have proved
31:24
that it will actually fix your lazy
31:26
eye. So there's a lot of things like that.
31:28
Similarly, ADHD, you know, with
31:30
ADHD, there's a, there's a, there's
31:32
a company Akili interactive. That's
31:34
done clinical research to show
31:36
that there. Quote unquote
31:38
game, but you know, it isn't really a game
31:41
it's, it's basically software as
31:43
a therapy. We'll train those
31:45
areas of the brain that can help people focus
31:47
better. No. So, so
31:50
it's like with neurofeedback and
31:52
those mechanisms it's based on solid science,
31:55
but the manifestation is instead of eating
31:57
chemical, now you're making those same changes
32:00
in the brain, but using software. So
32:02
yeah I think going back to this idea of having a really
32:05
high evidence burden that the FDA
32:07
requires, or a lot of these sort of companies require.
32:10
Is that a good or bad thing? I almost feel like having
32:12
the bar be set so high would disincentivize
32:15
people from even attempting to
32:17
solve real problems. I like the idea of solving
32:19
like ADHD or lazy eye, but you know, like the really
32:21
big sort of inherent problems that
32:23
exist within large, large societies.
32:26
Yeah. Is evidence burden too high, but
32:28
you're risking people's health too. Exactly.
32:31
Exactly. I think the evidence burden is high
32:33
because the cost is right. If you,
32:35
if you don't have, if you can't
32:37
prove safety and efficacy to
32:40
to a high standard, then
32:42
you might cause harm and in
32:44
healthcare, that's what you have to be aware of.
32:46
Right. So if you're, if you're. If you have a bug
32:48
in software and in a mobile
32:51
phone and it drops a call, no big deal
32:53
right here, it could kill people
32:55
or, or, you know, that's at its
32:57
at its worst case. Then you start to
32:59
start, have to worry about, Hey, what's the evidence.
33:01
So yes. We,
33:04
we always, as pharma, we,
33:06
as a pharmaceutical working for a pharmaceutical
33:08
company, I always feel like, yes, it takes
33:10
forever and sometimes frustrating, but I can
33:12
totally understand it. I think that there needs
33:15
to be a standard of evidence, truly burden.
33:17
If you're claiming it depends on what you're claiming.
33:23
You're claiming that your, that
33:26
your piece of software is curing
33:28
disease. You have to,
33:30
you have to be pretty sure that it is. I
33:33
think that's a, well, I think that's a good answer. Like
33:36
since I've been developing software for some time,
33:38
I think most other software engineers would agree with me
33:40
that. Once you read enough
33:43
software, you realize how crazy it
33:45
is that we're depending on other software developers
33:47
work for like driving a car
33:49
or like flying an airplane. Like that's some code
33:51
somebody wrote, they probably Googled answers
33:54
and copy and paste it from stack overflow.
33:56
And we're like, depending on that for our
33:58
lives. But the reason we can is because they have
34:00
these. Yeah. I, I think, I think regulation
34:03
is important. The more, the more you like the
34:05
more code you write, the more you realize, like we should have more.
34:08
So let me ask you this. See me. If you were starting a digital
34:10
healthcare company today, is America the
34:12
best place to start it? Would you go somewhere
34:14
where, like, if you personally wanted to create a company
34:17
that made money help people, is America
34:19
the best place to start it? I don't see why not,
34:22
you know, we are. We
34:24
are one of the most innovative
34:26
nations in the world. We are risk
34:28
takers. So I,
34:32
I actually think that we can drive.
34:34
So evidentiary burden is important, but
34:36
the other thing that's important is
34:38
that you have to right size, the evidentiary
34:40
burden don't make me produce.
34:44
You know, mountains of evidence that
34:46
something works. If all I'm doing is reminding
34:48
a patient because see, what's the worst that can happen.
34:50
That I send them a reminder of the wrong time.
34:52
Right. Big deal. Right. So,
34:54
so in those you have to right size
34:57
the risk with evidentiary burden. If
34:59
I'm creating like something that would
35:01
affect their heart or their lungs
35:03
or something like that. Great. Hold
35:05
me to a high standard. But if
35:08
all I'm doing is very slight
35:10
modifications in their habits, then
35:13
you have to right-size that regulation. So
35:16
the good news is the FDA
35:18
is actually working very hard
35:20
and understanding how to regulate
35:22
digital software that
35:24
is closely tied to diseases
35:27
in the right way. So that's why I think
35:29
it is the right place to start a digital
35:31
health company because you'll get the
35:33
investors are here. The risk takers are
35:35
here. People are willing to put their money
35:37
where their mouth is. So you'll get a lot of. You'll
35:39
get a lot of encouragement. You'll
35:41
be had credible
35:44
testing test beds. So
35:46
all the most of the big
35:48
healthcare systems and universities
35:50
and the, and the culture of research
35:53
and, and really high quality research
35:55
all exist. So I
35:58
actually think this is a wonderful
36:00
place. I think digital health solution
36:02
developed in the U S as a higher
36:04
credibility than in a country where
36:06
the regulations were relaxed because
36:09
you know that you've met a higher bar. So you
36:11
have a higher chance of making an actual impact
36:13
on a patient's life. That's fair.
36:16
Totally fair. So,
36:18
one question I did want to ask is with
36:21
regard to digital health particularly apple
36:23
they have all these incredible sensors
36:25
and all this incredible technology, but not everyone can
36:27
afford an apple product. So as digital
36:30
healthcare, just for the rich. Sometimes I do
36:32
sometimes feel that we are solving first world
36:34
problems. But apple is, is
36:36
not the best example if you're talking
36:38
about affordability agreed,
36:41
but there are a lot of solutions that are really,
36:43
really inexpensive and
36:46
they are actually meant
36:48
to be deployed in the
36:50
underserved areas of the world. So
36:53
I was. Blown away by some
36:55
of the really great work that
36:57
a small digital tool
37:00
can do in tuberculosis, for instance.
37:02
So there are all these remote villages let's
37:04
say in Africa or in Asia
37:07
where it's very difficult first. There's not
37:09
enough hospitals. It's not enough doctors. So
37:11
what happens is you've got somebody going on a
37:13
mobile clinic trying to diagnose people with tuberculosis,
37:16
and then there's. The added issue
37:19
that first you have to diagnose them. And then the
37:21
treatment regimen is fairly intense. They
37:23
have to be on all kinds of drugs for six
37:25
months or so. And you have to monitor
37:28
whether they're making progress or not very,
37:30
very difficult in an underserved area,
37:33
but under-resourced area. There
37:36
is a digital, there are several digital
37:39
solutions where the,
37:42
they run out AI algorithms
37:45
and machine learning algorithms to
37:47
to kind of pick up signals of tuber
37:49
closest. We are just how the patient
37:51
cops. So they can pick up a phone
37:54
and a doctor in a central
37:56
place anywhere in the world can listen to their
37:58
cough and the AI will read
38:00
it and say, tuberculosis, this cough
38:03
is due to tuberculosis, versus that it
38:05
can also pick up whether or
38:07
not the treatment is working. If the cough is getting
38:09
better, just because it gets the
38:11
digital signature of a tubercular cough
38:14
versus other types of cough that
38:17
is life changing for these areas.
38:19
So. There are, I don't know
38:21
if apple is the best example for four,
38:24
but digital health is actually should.
38:26
The maximum impact is it'll have is
38:28
in these underserved areas. This technology
38:31
is relatively cheap, right?
38:33
Humans, doctors. And
38:35
human resources are expensive, right?
38:37
Specialist resources are expensive. Technology
38:40
solutions are the great equalizer,
38:42
right? Everybody in, in, in a lot of
38:44
countries, data is become very, very
38:46
inexpensive. Everybody has a cell phone.
38:49
So using the cell phone as a
38:51
means of it, doesn't have to be an apple phone,
38:54
but using a cellular phone as a means
38:56
of delivering healthcare through different
38:58
technologies is going to be really
39:00
impactful. If the, if the willingness is there
39:03
and it is, I mean, there's a lot of. Lot of
39:05
initiatives in Africa and other
39:07
underserved areas that is doing that.
39:10
We have deployed a lot of digital health solutions. I
39:13
think so technology has this sort of great pulling
39:16
power that over time it tends
39:18
to begin, you know, it goes into like some sort of
39:20
new industry, it builds some new innovations
39:22
and then before, you know, it they've taken over the industry.
39:24
We're seeing that with Hollywood right now. There's any number
39:26
of examples. And so if we look at sort
39:28
of the digital health space right now, the big tech companies
39:31
like apple and the Googles of the world, maybe
39:33
in terms of like the total GDP being built. In
39:35
the space, they take up 5% and
39:37
the remaining 95% is being done by
39:40
I dunno if this proportion is correct, but the remaining 95%
39:42
is being done by the existing sort of big pharma
39:44
companies, 30 years from now. What
39:46
does that proportion look like? Where does
39:48
tech stand up against the big pharma
39:51
companies? I don't think
39:53
that it's tech versus big
39:56
pharma. Really. I think
39:58
it's tech versus. A physician.
40:01
Hmm. So there is the
40:04
tech will produce algorithms,
40:06
very good algorithms. It's already
40:08
producing algorithm that can diagnose disease
40:11
that can produce treatment
40:13
regimens and also
40:15
treatment plans, and then provide
40:18
the continuous care continuum
40:20
to keep a patient healthy as possible. So
40:23
there is that. A lot
40:25
of discussion going on in the space of is
40:27
it is an AI bot
40:29
going to replace a physician.
40:32
So that's, that's been debated
40:34
forever. And I don't think, I think it's a matter
40:37
of how you view it right now,
40:39
at least in the short term. I
40:41
think I think of tech as
40:43
enhancing the position's powers. Right.
40:45
So giving them additional information,
40:47
additional knowledge, making their life easier
40:50
eventually ever evolves
40:52
to get the same kind
40:54
of you know, features,
40:57
et cetera, that, that experience
40:59
and common sense
41:01
and empathy that a human has
41:04
that a human physician has. I
41:06
don't know, but that's way into
41:08
the future in the short term, I don't, I just
41:10
see tech as being highly helpful in,
41:13
in removing some of the inefficiencies
41:15
that are inherent in healthcare right now in optimizing
41:18
healthcare. That's the first thing that
41:20
it should do and is doing now right.
41:22
Slowly. It didn't. That is pretty slow. DEC
41:24
versus pharma currently, the
41:26
way we use the pharma's core businesses
41:28
creating is making medicines
41:31
right now. Our drug
41:34
discovery process is very
41:36
slow and very labor
41:38
intensive and time incentive and
41:40
resource intensive. That
41:42
is helping a lot there too. It's helping us
41:44
identify targets to
41:46
and what kind of molecules to create. It's making
41:49
our manufacturing, operations faster
41:51
R and D operations faster or clinical
41:53
trials faster. Will it ever replace
41:55
it where it can, you can create a model
41:58
of of of a compound and then create
42:00
a medicine out of all. That, to me sounds
42:02
very futuristic right now. And
42:05
so futuristic and I mean,
42:07
I'm sure it'll happen, but it's so futuristic that I
42:09
feel like we shouldn't be worrying about all that right now.
42:11
Let's use tech for good and optimizing
42:13
and helping get better
42:16
patient outcomes today and
42:18
tomorrow we'll take care of it. So
42:21
let's say that one of our listeners
42:23
is software engineer and they hear
42:25
this and they say rather than making
42:27
my own social media app, I want to
42:30
make something good for my legacy and for the
42:32
human race. And I want to work
42:35
in health and digital health. What
42:37
advice do you give to them? Who should they link up with?
42:39
What should they start reading about? How do they do this?
42:41
So get educated is
42:44
the first thing. Right? And then there
42:47
isn't anything. I
42:49
think, keep that motivation alive if
42:51
they're doing it for the right reasons, right.
42:54
They're doing it truly because they feel like
42:56
I need to leave a legacy and leave something
42:58
behind. They're already
43:00
way ahead. Like they're already way ahead
43:03
and they're already poised for success. So
43:05
all they have to do is now kind of use
43:07
this motivation and use this drive.
43:10
To, to get educated
43:12
and into what are the real patient
43:14
problems to solve, right? What are those things
43:17
that really make an impact and start
43:19
small, choose one thing and usually
43:21
choose something that's really personal to you. If,
43:24
if, if you live in there's a lot of
43:26
digital health solutions that came out of somebody
43:28
watching their grandfather struggling
43:30
through a condition, right? So choose
43:32
something personal. If you can choose that and want
43:34
to work in that space, that would be fantastic. I
43:37
do you are, you have a very high chance of success.
43:39
I think working in pharma is a great way
43:41
to go. I think working in
43:43
health tech, meaning the,
43:45
the health or life sciences
43:48
side of big, technology's a great
43:50
way to go and
43:52
help and working in
43:55
the intersection of these two is
43:57
the ideal way. You know, where you're looking
43:59
to see where is big
44:01
tech partnered with. With pharma
44:04
and biotech to make really cool
44:07
solutions right. And working
44:09
in that space is, is always a good
44:11
thing. I don't know. I haven't really given it too
44:14
much thought about. Digital
44:16
health. I just don't see it as very
44:18
different than any other career. Right.
44:21
Other than the motivation part, it's just the use
44:23
of tech in a space that
44:25
in a new space it's, it's the use case
44:27
is different. So get really if, if, if
44:30
it's a software engineer get really good
44:32
at software, it's really
44:34
good at the user experience and
44:36
then go for it all comes back to
44:38
user experience. I, you know,
44:40
I love this and I'll always
44:42
start and end on that. Well, I think that's a great note
44:45
to wrap things up here. So to start our
44:47
final question that we wanted to ask,
44:49
what is a great piece of technology
44:51
that you want to highlight that you think
44:54
is one of the most interesting, fascinating,
44:56
best piece of technology you've encountered
44:58
in recent memory or even of all time? It could
45:00
be software, hardware, and whatever, whatever
45:03
floats your boat. Okay. If it's
45:05
software, I would hands
45:07
down, hands down, say the Google
45:09
search engine. And it's probably
45:11
an answer you've heard before. And that's because
45:13
I think it was truly transformational
45:15
for everybody. Right. For everybody, everybody
45:18
Googles, everything. Right. And
45:20
it has become part of your vocabulary. So I
45:22
think that, I think attracts, I
45:25
think it's by far the number
45:27
one piece of impactful software
45:30
that everybody uses in their daily lives.
45:33
That's our episode for this week. Thank you so much for listening.
45:36
Make sure to subscribe to us and rate us on Apple
45:38
podcasts. We would really appreciate
45:40
the support. You can also follow me on
45:42
Twitter at F Z from Cupertino
45:45
and Busan. The ad next facade.
45:47
See you guys next week.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More