Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
What we were able to do once with Ardy and
0:02
Seqster is we didn't need to go out to all those
0:04
patients' healthcare providers . We already knew who the patients
0:06
were . We could go directly to the patients and ask them if
0:08
they'd like to contribute their data to research , and
0:10
if they consented to that , we were able to gather
0:13
up their medical records so you're
0:15
able to do a study that otherwise just would not
0:17
have been very possible . That's the big
0:19
evolution we've seen . I call it
0:21
patient , you know , patient agency over their medical record
0:23
. Their ability to do so will transform
0:26
the way we execute all
0:29
sorts of research , from pre-registrational
0:31
through post-approval , and we're just starting
0:33
to see some of that .
0:35
Welcome back to the Chilcast , a healthcare
0:37
podcast from Chilmark Research , helping
0:40
healthcare leaders make the best decisions
0:42
for the populations they serve . Welcome
0:50
back to the Chilcast . I'm the managing
0:52
partner of Chilmark Research and your host
0:54
, John
1:01
Moore III . Chilmark Research is a healthcare technology industry analyst firm founded
1:03
in 2007 to provide objective , expert research and guidance
1:05
on those emerging data-driven tools and services with
1:07
the potential to fundamentally improve the experience
1:09
of care for all . If you are new
1:12
to the show or a return listener and you appreciate
1:14
our programming , don't forget to subscribe on your
1:16
favorite podcast service and leave us a review . If
1:19
you have a topic you'd like to hear us discuss on the show
1:21
or questions about how to participate , shoot
1:24
us a message at podcasts at chilmarkresearch . com
1:27
. For this episode
1:29
, we're going back to standard programming after
1:31
our initial six-part miniseries for the Health
1:33
Impact Project . Be sure to check
1:35
those episodes out , because we had some fantastic
1:38
discussions around defining healthcare
1:40
value and what the value
1:42
is of health IT In
1:44
the fall of 2023, . We
1:46
released our first real-world data real-world
1:49
evidence market trends report . I
1:51
will soon be following that up with the accompanying
1:53
buyer's guide report that goes into more detail
1:55
about the differentiators between solution developers
1:58
, how customers can think about defining
2:00
their specific use cases , to narrow the field
2:02
of options , best practices
2:04
and pitfalls to avoid , and thorough vendor
2:06
profiles . This will
2:08
be our second podcast on the topic of
2:10
RWD and I am pleased to be joined
2:12
by Aaron Berger and Ardy Arianpour
2:14
for this discussion about the evolving nature
2:16
of clinical research and therapeutics development
2:18
.
2:19
Hi there , it's great to be here . Thanks , john , for
2:21
inviting us on the podcast .
2:23
Hi everyone , so nice to be here
2:25
. Happy Friday .
2:27
My pleasure . I'm really looking forward to this conversation
2:29
. So for some background
2:32
, aaron Berger is the Senior
2:34
Vice President and Head of Evidence Development
2:36
Solutions for UBC
2:38
that designs
2:40
and executes modernized solutions for
2:43
leveraging forward-leaning technology and
2:45
research designs that deliver fit-for-purpose
2:47
evidence on the safety and effectiveness
2:50
of treatments for patients . Mr
2:53
Berger joined the UBC team in 2003
2:55
and has since served in a range of diverse
2:58
roles , leveraging global capabilities in
3:00
the execution of PERI and post-marketing
3:02
product development programs and real-world
3:04
evidence generation . He is
3:07
responsible for the development of real-world data
3:09
and real-world evidence technology architectures
3:11
and decentralized research solutions to
3:13
drive greater insight into the safety , profile
3:15
and value proposition of medicines . Ardy
3:19
is the CEO and co-founder of Seqster
3:21
, a healthcare technology company breaking down
3:24
health data silos at scale . Its
3:26
enterprise operating system aggregates disparate
3:28
health data sources into a single 360-degree
3:31
view of a patient in real time , producing
3:34
research-ready longitudinal health records
3:36
for use by pharma providers and payers
3:38
. Ardy is a serial
3:40
entrepreneur in life sciences and digital
3:42
health . Prior to starting Seqster
3:44
, Ardy
3:50
launched several clinical and consumer-based genetics tests as chief commercial
3:52
officer of Pathway Genomics and SVP and GM of Genomics at Ambry Genetics , which
3:54
sold to Conica for $1 billion in 2017
3:57
. As a key player
3:59
in the landmark SCOTUS decision scrapping
4:01
gene patents , which would be 2013's
4:04
Association for Molecular Pathology versus
4:06
Myriad Genetics . Ardy played an instrumental
4:09
role in expanding genetic testing access
4:11
with the launch of BRCA testing , and
4:13
benefiting patients and family members across
4:16
the country . He has been recognized
4:18
with many awards over his career , most
4:20
recently as a MedTech World
4:22
Rising Star in 2022 and
4:25
won a Pharma Vo's 100 Inspiring Leaders
4:27
in Life Sciences in 2023
4:29
. All right , so , aaron
4:31
, thank you for being here with us today for this very
4:34
timely and important conversation
4:36
. First off , let's frame
4:38
the whole area a little bit . What
4:40
does your work consist of as UBC's
4:42
Head of Evidence Development Solutions ? Yeah , john , so you kind of mentioned it in the
4:44
intro you did for me .
4:45
Thanks for that . What does your work consist of as UBC's Head of Evidence Development
4:47
Solutions ? Yeah , john , so you kind of mentioned it in the intro . You did for me . Thanks
4:49
for that . You know what we're here
4:52
to do is answer important questions about
4:54
the safety and effectiveness of
4:56
medicines , and oftentimes we're
4:58
doing that within the context of the post-approval
5:01
requirement that's , from a regulator
5:03
, from FDA , from EMA post-approval
5:06
requirement that's from a regulator , from FDA , from EMA . Other times we're doing it within the
5:08
context of the sponsor , of the manufacturer's desire
5:11
to learn more about the
5:13
safety profile and the
5:15
value proposition of their medication
5:18
.
5:19
Can you share either
5:22
yours or UBC's definition of
5:24
real-world data and evidence for our audience
5:26
?
5:28
Yeah , absolutely so . In
5:30
2018 , FDA released
5:32
a real-world evidence framework , and that's
5:34
kind of what the industry has aligned around
5:36
the definitions
5:39
and standards that are outlined there , and
5:42
that document frames it like this Real-world
5:44
data are any data relating to
5:46
the patient's health status and delivery
5:49
of healthcare . They're routinely collected
5:51
through a variety of sources
5:54
, and examples of
5:56
real-world data include electronic health records
5:58
, claims data , data
6:01
that may come from registries or
6:03
data from digital health technologies
6:05
. Real-world evidence
6:07
is the evidence that's
6:10
curated and generated about the usage
6:12
of , or potential benefit of
6:14
, a treatment for a patient that you derive
6:16
from analysis of that real-world
6:18
data .
6:20
And what are some sources of real-world data
6:23
and examples that you can share ? And what are some sources
6:25
of real-world data and examples ?
6:25
that you can share . So you know , here today we're going to mainly talk about
6:28
electronic medical records
6:30
, but you
6:33
know it goes beyond
6:35
that . It also goes into claims . It goes into
6:37
a patient's activity
6:40
and their health status that you may collect
6:42
from wearables . It may relate
6:45
to quality of life or patient
6:47
reported outcomes that a patient is directly
6:50
reporting , but
6:59
the research fundamentally requires that you get access to a patient's
7:01
medical record in order to answer the questions that we're trying
7:03
to answer about the safety and effectiveness of treatments
7:05
.
7:06
Okay , yeah , so that you know
7:08
what's actually happening in the wild , what's happening when
7:10
you have some awareness
7:13
of a patient's full medical
7:15
history , as much as you could collect it . So
7:17
can you tell me how that information has been collected
7:20
historically ? How have you guys gone about
7:22
bringing in , you know , a patient's records and their
7:24
medical history ?
7:29
Yep , definitely . And that kind of really leads us into how I got introduced to Ardy
7:31
and how we started collaborating so classically
7:34
for decades . The
7:36
way that you acquire
7:39
patients' medical records for clinical research or post-marketing
7:41
evidence generation medical records for clinical research or post-marketing
7:43
evidence generation is
7:45
you enlist the healthcare provider . It's
7:48
been very healthcare provider
7:50
mediated , I would say , for
7:53
decades , and
7:55
those healthcare providers are enlisted
7:57
as sites in a clinical research
7:59
study or in a post-approval
8:02
study or registry . The
8:07
healthcare provider has roles such as a study coordinator , and
8:09
those study coordinators and investigators literally
8:13
take data that have been from a patient's
8:15
chart and they would transcribe it into
8:17
a case report form that we use for our research
8:20
purposes . That
8:22
evolved a couple of decades ago and it evolved
8:24
into web-based electronic data capture
8:27
systems and that involves
8:29
keystroking data from a patient's
8:31
chart into the
8:33
case report form , into the study database , if
8:35
you will , and that's how we've
8:37
been conducting research , but it always
8:40
relied upon going through the health
8:42
care provider . So
8:44
I've been very interested for a number
8:46
of years in how do we
8:49
perhaps disintermediate the health care provider
8:51
from that process and really relieve
8:53
them of that burden and
8:56
get directly to the source of
8:58
the real world data that we're looking for in this case
9:00
electronic medical records and do
9:02
that through the patient's
9:04
ability to take agency over their own record and
9:07
decide if they want to contribute it to research
9:09
.
9:09
So it sounds like the old process was very labor intensive
9:12
and relied a lot on your ability
9:14
to bring in human capital to actually
9:16
do that transcription and manual abstraction
9:18
. Before we get to Ardy and your guys' kind
9:20
of backstory , I have a couple of quick other questions
9:23
to kind of frame this conversation
9:25
and what you've seen in your career . So
9:27
you've been with UBC since 2003
9:29
, which is surprising loyalty
9:31
to a single company in this day and age . It's
9:34
also the year that I graduated from high school , which
9:36
isn't really dating you because you're only a couple of years older than
9:38
me , but still . I mean just kind of putting it in
9:40
perspective . It was also the year
9:42
that the Human Genome Project was completed
9:44
and we had actually fully sequenced the human genome
9:46
. So it was only a few years
9:48
before that that everyone had been losing their minds about
9:51
Y2K and clearly a lot has changed
9:53
in that period of time across
9:55
many areas that touch clinical research and
9:58
quite notably the first true
10:00
global pandemic in the modern era . You
10:02
know , in a couple generations we haven't seen anything
10:04
like that . So the way that clinical
10:07
research is being done and new therapies are
10:09
being developed has really seen
10:11
a whole host of technological innovation
10:13
, scientific innovations and just
10:15
social disruption that will have touched
10:18
your career in some capacity . So can you
10:20
share a few highlights of the most significant
10:22
changes you've observed in terms of how
10:24
this is practically impacting the work that you guys
10:26
do , and maybe some areas where things
10:28
haven't advanced as much as you would have liked or
10:30
expected to see at this point ?
10:32
Yeah , definitely that's a fun thing
10:34
to think about and reflect on . So
10:36
when I started in this industry in about 2003
10:39
, I had roles called a CRA Clinical
10:42
Research Associate . I had roles called a CRA Clinical
10:44
Research Associate and what that meant was I
10:46
would go fly around to these sites , those healthcare
10:48
providers that I talked about who are listed in in
10:50
a research , in a study , and
10:53
they would have these big
10:55
walls , these big columns of paper charts
10:58
and manila folders and they would
11:00
roll on tracks and you'd use cranks to get to the medical records you're looking
11:02
for and the sites would pull these tracks and you'd use cranks to get to the medical
11:05
records you're looking for and the sites would pull
11:07
these out and we would literally take the paper
11:09
, the patient's chart , and I would review
11:11
what was in the chart and I would review what
11:14
the site had transcribed into the case
11:16
report form , which was ultimately became
11:18
the study database , and to see
11:20
if those things matched up . Then , if
11:22
they did , I would take the case report form
11:24
, which was two part , you know NCR paper
11:26
, take the top page off and I put all
11:28
those pages in a FedEx envelope and
11:31
send it off back to the home office
11:33
where that data would be either
11:35
scanned or manually entered
11:38
into a study database , and that's how we were
11:40
performing , you know , research
11:42
and , remarkably
11:44
, things
11:46
well , things changed . So
11:49
then you had an evolution , about three or
11:51
four years into my career , that the paper went
11:53
away and you started doing that on a web-based
11:55
electronic data capture platform . So we're
11:58
keystroking the data into an EDC
12:00
, directly into the database , and
12:02
that was a big leap forward , to
12:05
be sure , but we're still relying
12:07
on manual transcription . We're still relying
12:09
on sites to do this
12:11
. It has scalability and
12:14
quality issues associated with it
12:17
. And then fast
12:20
forward up to the 21st Century Cures
12:22
Act and the FDA framework on real
12:25
world evidence and disruption
12:27
from technology providers like Seqster
12:29
, and what happens is we now
12:31
are gaining the ability to get
12:33
at those medical records without having
12:35
to sit somebody down at a console , look at two
12:37
screens and punch data in from one
12:39
place to another . And
12:42
so we're just kind of at the dawn of that . We've used
12:44
these types of solutions several times in
12:46
partnership with Ardy , and
12:49
what it allows you to do is
12:51
conduct some research that otherwise would not have
12:53
been possible . I'll give you one example
12:55
when you're trying to do research
12:57
in a rare disease , as
12:59
one of our clients was , and
13:02
you would , literally , if you wanted to do this type of study
13:04
it's a chart review or natural history study you'd
13:06
have to go to find 50 patients
13:08
to get quality data on . You'd have
13:11
to go out to 50 different sites , different
13:13
healthcare providers , and ask them to
13:15
execute that enterprise . I just
13:17
told you about public manual data entry and
13:19
sometimes that's just not economically
13:22
practical and the sponsor is
13:24
going to decide not to do that type of research
13:26
in that setting . But what
13:28
we were able to do once with Ardy
13:30
and Seqster is we
13:32
didn't need to go out to all those patients' healthcare
13:34
providers . We already knew who the patients were . We could go directly
13:37
to the patients and ask them if they'd like to contribute their
13:39
data to research , just as they would in any study that
13:41
they would decide to participate in . And
13:43
if they consented to that and enrolled
13:45
through the CIGSTER portal , we were
13:48
able to gather up their medical records . So
13:50
you're able to do a study that otherwise just would
13:52
not have been very possible . So we think
13:54
it's a really important part of research
13:57
into rare diseases and
13:59
small populations
14:01
that are dispersed . Using
14:04
these types of methodologies are gonna be very
14:06
important . It's also gonna be important
14:08
for chronic disease and high volumes
14:11
of patients . But that's the big evolution
14:14
we've seen . I call it patient agency
14:16
over their medical record . Their ability
14:18
to do so will transform the
14:20
way we execute all
14:23
sorts of research , from pre-registrational
14:25
through post-approval , and we're just starting
14:28
to see some of that .
14:30
Yeah , no , it's fantastic seeing how
14:33
actually having this all digitized and then moved
14:35
to the cloud is opening up new ways for patients
14:37
to have that autonomy and that ownership
14:39
over their own health histories . I'm
14:42
seeing a few different companies that are doing
14:44
various aspects of the real world data solution
14:47
development as part of this research effort that
14:49
I'm currently engaged in , but very
14:51
few are taking that patient forward
14:54
, patient centric approach where they're the ones driving
14:56
the ship and then they also get something out of it on the back
14:58
end . So , Ardy
15:01
, that kind of brings you back into the conversation
15:03
. The first time we spoke
15:05
, you and I bonded around our mutual past
15:07
lives working benchside in genetics labs . You
15:10
have me beat by a few years as well , but we're practically
15:12
contemporaries and have both ended up here in the
15:14
brave new world of modern healthcare informatics
15:16
. Can you give us the quick backstory
15:19
of how your early career in genetics research
15:21
led you to found Seqster later on ?
15:28
Yeah , first off , thanks so much , John , for having
15:30
myself and our wonderful partner , UBC , Aaron Berger , here . You
15:33
know I started out when I was 16
15:35
years old and I was lucky enough to grow up
15:37
in the beautiful and finest
15:39
city in America called San Diego
15:41
, and if you know anything about San Diego
15:44
, we have some prestigious
15:46
, amazing universities and
15:49
there's lots of different institutions
15:51
that bring innovation
15:53
out , because of the
15:56
marine biology , I guess , and the
15:58
great , you know , surf that exists in
16:00
La Jolla . Here . La Jolla and Torrey
16:02
Pines has been known as a
16:04
science mecca for quite some
16:06
time . I know we were talking about genomes
16:10
in 2003 . I'm
16:12
one of the first 50 people to actually
16:14
in the world to have my CLIA
16:17
clinically whole genome sequenced . It's
16:19
my 10 year anniversary . There
16:21
you can see my actual flow
16:23
cell where I have
16:25
my blood and my DNA sequenced
16:29
at 100x over 10 years
16:31
ago and that was actually March
16:33
of 2014, . As you can see on the plaque
16:36
here , growing up in San Diego , I was really
16:40
fascinated with science
16:42
and discovery , having
16:44
that first job at the Salk Institute when
16:47
I was 16 years old in a gene
16:49
expression lab probably one of the most prestigious
16:51
gene expression labs with Dr
16:54
Ron Evans , who's won every single
16:56
award other than the Nobel Prize . That's
16:58
when I got introduced to Sanger sequencing
17:00
in the late 90s and
17:02
then from there when I was pre-med at
17:05
UC Irvine and didn't make
17:07
it through and got into commercial
17:09
sales and marketing of biotech
17:11
in my 20s and then became an executive
17:14
in biotech and
17:16
taking next-gen sequencing in the clinic
17:18
with some companies in my late 20s
17:20
. That really opened up my
17:22
eyes to data and
17:25
how data was being siloed . And
17:27
because of my fast track career I was just
17:29
fortunate enough being at the right
17:31
place at the right time and
17:34
seeing where sequencing
17:37
was interoperable . From
17:39
ATCG standpoint , it didn't
17:42
matter what sequencer John
17:44
it came from . Atcg
17:46
standpoint , it didn't matter what sequencer John it came from . What we didn't know is there's
17:49
4,000 plus major healthcare systems and 150,000
17:52
plus providers across our nation
17:54
that are not interoperable
17:57
even if they are on
17:59
the same vendor , such
18:01
as Epic , allscripts or Cerner
18:03
the big three because there's various different
18:05
versions . And so Seqster
18:08
was born out of the fact
18:10
that we can sponsor
18:12
sequencing and stir
18:15
all the other data and
18:17
do some great work for
18:20
both patients and researchers . What
18:22
we didn't know is that we were
18:24
falling on solving
18:26
healthcare's number one problem
18:29
, which is interoperability
18:31
, and what we didn't know also
18:33
was that we were innovating at the same
18:35
time patient centricity
18:38
with what I call patient
18:40
centric interoperability . That's
18:43
what Seqster has pioneered
18:45
and that's our biggest differentiator
18:47
. That's why folks like Aaron
18:49
Berger and UBC have
18:52
solved some problems that they hadn't
18:54
been able to solve before , and
18:57
it's because our operating system
18:59
was built for not
19:01
just patients , but also
19:04
for researchers to have research-ready
19:07
data .
19:09
All right , thank you for breaking that all down for us . I
19:11
can definitely relate to the
19:13
need for that , given my own history in this
19:16
area and working for one of the preeminent
19:18
genetics researchers on the East Coast
19:20
for my first career and learning
19:22
about the importance to try
19:25
to intervene early with a lot of conditions when you can diagnose
19:27
it genetically . So
19:29
, as is the case for many of the entrepreneurial
19:31
minded , you mentioned that you
19:33
lucked out with timing when you were doing the genetics work
19:35
, but I think , based on some of our discussions
19:38
, your initial vision for Seqster was
19:40
, and maybe even is , too early for
19:42
where the market was and might still
19:45
be at the current time . So
19:47
I already know a little bit of this , but could you share
19:49
some background on the original core
19:52
idea of what the company was going to do and some
19:54
of the pivots you've taken along the way to keep the company
19:56
independent and growing
19:58
?
19:59
Well said there , john , and
20:01
it still is the same vision . Actually
20:03
, we haven't changed their vision and mission . It's
20:06
to put the person , the patient
20:08
, at the center of healthcare , to disrupt
20:11
and break down their data
20:13
silos , bringing together their episodic
20:15
electronic health records
20:17
, combining it with their baseline
20:20
genetic material and
20:22
data , as well as adding any
20:24
type of other pieces such as claims
20:27
data , pharmacy data , social
20:29
determinants of health data , medical
20:31
devices , wearables . We're
20:33
connected to 400 plus . We
20:36
have nationwide access 327
20:38
million patient records that we have access
20:40
to with one-click records . If
20:43
you're familiar with fintech
20:45
and finances and
20:48
your net worth , no matter if
20:50
your net worth is a dollar or a
20:52
hundred billion dollars , like Bill Gates
20:54
and others , it doesn't matter . Everyone has a
20:56
network . Well , we created the
20:58
mintcom of healthcare
21:01
and life sciences with that same
21:04
approach by bringing together your
21:06
MD Anderson data , your
21:08
wearable data from Garmin and Apple Watch
21:10
because they don't talk to another , but through Seqster
21:13
it can Four
21:15
different other providers from UCSD
21:17
to Stanford to Emory to NYU
21:20
. We don't know where patients have
21:22
data , and what's important is
21:25
to offer a digital
21:27
front door for both
21:29
patients and researchers to
21:31
connect the dots to
21:33
data . We also pioneered the patient
21:35
mediated method , where we are
21:37
the leaders in that . It's
21:40
not just HIE data that we're bringing
21:42
. We have non-HIE data . It's not just
21:44
FHIR data , fast healthcare interoperability
21:47
resource data it's non-FHIR
21:49
data . We've standardized and harmonized
21:51
every single ICD-9 and ICD-10
21:53
code , rx norm
21:55
and SNOMED codes
21:57
and we built the data refinery
22:00
on the backend . We also have access
22:02
to 90 million plus claims
22:04
data through our partnership with United Healthcare
22:07
Group and we're one of the only
22:09
companies that actually is working
22:11
at that sort of level . We
22:14
have partnerships with six out of the top 10
22:16
pharma companies . What's
22:18
so exciting is that
22:21
Seqster saves lives , and
22:23
one thing I do want to touch base on
22:25
is that it saved my dad's life . We
22:27
ran a tumor board in six hours , got
22:30
his exact sciences Cologuard results
22:33
, got his non-invasive liquid biopsy
22:35
test from Garden Health as well as
22:37
his screening test from Grail , so
22:39
genomic data that is siloed
22:41
, not within the EMRs . He
22:44
had four different health systems . We
22:46
were able to get his data
22:48
on a Saturday , run
22:51
a tumor board with Dan Van Hoff , the
22:53
most prestigious pathologist out
22:55
of TGen in Arizona , with
22:58
his colleagues , and get second opinion
23:00
letters to rush him into surgery
23:02
at Kaiser in six business days
23:04
. If I wasn't the CEO of Seqster
23:06
and if I didn't have access to
23:08
such technology , maybe he wouldn't
23:11
be alive . To be honest with you , he wouldn't
23:13
be alive because how , if you know anything
23:15
about cancer and how cancer metastasizes
23:17
time to intervention is
23:20
key , and that's the essence
23:22
of real world . Data is key , and that's the
23:24
essence of real-world data
23:26
, that's the essence of real-world
23:28
evidence , that's
23:32
the essence of medicine and everything that UBC and Seqster are doing , from a medical record
23:34
release standpoint to utilizing our one-click records
23:37
technology to make
23:39
and accelerate clinical
23:41
trials , decentralized trials , hybrid
23:44
trials and so forth . We couldn't be more
23:46
excited that our technology is
23:48
not being siloed , that is being
23:50
used by , you know
23:52
, incredible people that have , you
23:54
know , decades of experience , like
23:56
Aaron Berger , that understand the
23:59
patient journey but also the researcher
24:01
journey .
24:04
Okay , thanks for telling us that personal story , Ardy
24:06
. I was going to get to that in a more
24:08
specific question because that was something I wanted to make sure
24:10
you had an opportunity to tell , because
24:12
, you know , making that personal connection always makes
24:14
these founder narratives a lot more compelling
24:16
and a lot more engaging . And you know , for people , personal connection always makes these founder narratives a lot more compelling and a lot more engaging . And you
24:19
know , for people to survive in this
24:21
industry and doing this work , you really have to have
24:23
a real motivation
24:25
, like a personal motivation , to stick
24:27
through all the BS that healthcare throws
24:29
at everyone . I
24:31
don't think I've met a single founder that's been in the industry
24:34
long enough to actually see success , that doesn't have
24:36
a really intense personal reason for doing
24:38
this work . So
24:40
this next question is for both of you , now
24:42
that we've set the foundation how did you two cross
24:44
paths and start working together , and
24:46
what was the problem that you were solving
24:49
? That catalyzes collaboration .
24:52
The way our paths crossed was we
24:55
were attempting to solve a research
24:57
problem for one of our clients , who
25:00
wanted to answer some important
25:03
questions about the safety and effectiveness of
25:05
treatments
25:07
in a certain category of patients
25:09
with the disease , it
25:13
was very difficult to identify where
25:15
these patients received healthcare , so
25:17
that we can enlist their healthcare providers to give
25:19
us the data the real-world data that we needed
25:21
from their electronic medical records . We
25:23
were not getting much traction in
25:26
that respect , and so
25:28
, actually , that particular study
25:30
was about to be canceled , and
25:33
already I had come across
25:35
a podcast , I think or it was maybe
25:38
not a podcast , it was a interview
25:41
that Ardy did . It was online , and
25:45
I heard him describe his solution
25:47
and I
25:49
was thinking you know , that's exactly what we need
25:51
in order to make a study like the one we're
25:53
doing and the others that I envision in the future will be
25:55
necessary come to life
25:57
. I need to be able to get access
25:59
to the patient's medical records when
26:01
that patient decides to participate in
26:03
research and contribute it without having
26:06
to go through that patient's healthcare
26:08
provider . And so
26:10
Ardy and I had had a
26:12
dialogue , we were talking about this and
26:15
we brought the idea to our
26:17
client . We said , hey , we think we
26:19
can do this study . Still , we
26:21
know who the patients are already
26:23
. We can ask them , invite them to participate
26:25
, invite them to consent , and
26:27
we're going to use this new solution
26:29
where they'll go into a portal . It's
26:32
called the Seqster solution . We introduced
26:34
them to Seqster , we had them explain it and
26:36
they said , yes , let's give it a shot . And
26:39
by doing that we were actually able to get
26:41
about 50 patients charts , we
26:43
were able to produce some analysis
26:45
tables and figures and a
26:48
study report and we
26:51
were able to perform some research that we would not
26:53
have otherwise . I've got a couple other examples
26:55
I'll share later and
26:57
where I see this going in the future , because what we
26:59
want to do next is deliver this at
27:02
scale in chronic conditions
27:04
over longer periods of time . But
27:06
that's kind of the origin story of how we met
27:08
and got started .
27:10
Okay . So it was a very clear problem
27:13
and a very specific call
27:15
to action from one of your clients to make this a more
27:18
efficient process . So
27:20
, other than EMR data , what other data sources
27:23
does Seqster bring in that have proven valuable
27:25
to your clients and users ? I know that RD
27:28
rattled off a few of the different ways that they ingest
27:30
and kinds of data that they bring in , but what
27:32
have you seen beyond the EHR
27:34
information ? That's been really valuable .
27:37
So Ardy mentioned something
27:39
that they recently added , which
27:42
was the opt-in claims data
27:44
, and that's something we'll definitely
27:46
be exploring with Ardy how we could
27:48
leverage that . That's an important piece of
27:50
putting together a rich longitudinal
27:53
history and picture of the patient's
27:55
healthcare journey , because it's not
27:57
always just the
27:59
EMR . It's a fundamentally important piece , but
28:02
you often need to fill that in
28:04
with administrative claims
28:06
pharmacy and medical claims so
28:09
classically , what we do for that
28:11
is we will apply
28:13
a process called tokenization and
28:16
that places a unique identifier
28:19
on a patient and you use that same
28:21
algorithm to assign that
28:23
identifier in other
28:25
real-world data sets like pharmacy and medical
28:27
claims . You match the tokens up and
28:30
then you know that this patient is the same as that
28:32
patient in these two different data sets . You license
28:34
the de-identified data and you
28:37
can bring it together while
28:39
respecting data privacy
28:41
considerations . So it's
28:44
very important to bring those types of data
28:46
in sometimes maybe data from
28:48
lab data , sometimes genomic data , and
28:51
we use a lot of tokenization to
28:53
facilitate that . I'll
28:56
let Ardy talk about some of the things that
28:58
are on his roadmap and bringing in additional
29:02
sources into his universe that maybe
29:04
alleviate the need or maybe make it more one-stop
29:07
shop , and that's what we'll be looking to do in
29:09
the future as the
29:11
partnership evolves .
29:15
Okay , thanks for defining tokenization there
29:17
for me too . I was going to put that into the
29:19
questions in my planning for this , but I wasn't sure
29:21
if it made sense to get that technical . So thanks for
29:23
addressing that directly . Ardy
29:25
, did you want to add anything to
29:27
that ? Aaron kind of mentioned that you might have some pipeline
29:30
things that you're ready to disclose or talk
29:32
to or anything else you might just want to
29:34
contribute there .
29:35
Yeah , I think it's really important to know
29:37
that we can connect
29:40
to any data source and
29:43
every use
29:46
case , every disease
29:48
, every cohort , every
29:50
company , every investigator
29:53
, every observational study , to
29:55
preclinical , to phase one to phase four
29:58
. They have their own inclusion exclusion
30:00
criteria and so
30:02
at Seqster we don't decide that
30:04
. But what we do
30:06
really well is
30:09
we can go fetch
30:11
it if it doesn't even
30:13
exist in our operating
30:15
system . For example , if
30:18
there's some new medical device
30:20
that just came out and someone's
30:22
going to run a remote patient monitoring
30:25
use case and they have
30:27
a thousand patients that
30:29
have this new RPM device to track
30:31
XYZ data and
30:34
we're bringing in electronic
30:37
medical record data , which is really the bread
30:39
and butter . But , as both of you
30:41
just stated , it's not just
30:44
EMR , ehr data that's
30:46
important for real world data . Real
30:48
world data really is what
30:51
patients have
30:53
, and every patient is
30:56
different , from rare disease to
30:58
cancer , oncology , to
31:00
autoimmune , to a healthy individual
31:02
that has a lot of trackers . All
31:05
that data is characterized
31:07
by us at Seqster , from our
31:09
opinion , as real world data . But
31:12
our interoperability engine allows
31:15
you to
31:17
actually request a
31:20
specific data source if
31:22
we're not already connected to
31:24
it . I'll use an example . Let's
31:26
just use Jawbone
31:28
as an example , people
31:39
that used to have a jawbone device that's tucked away in someone's drawer , like the old
31:41
palm devices before the iPad came out decades ago . Right
31:43
, it doesn't mean that that data
31:46
is worthless . There's
31:48
some data maybe in
31:50
there that is valuable
31:52
, but no one's collecting
31:54
that data because no one is using it . But
31:56
how awesome would it be to
31:59
look at that data from 2001
32:01
and compare it to Fitbit's
32:04
newest device data . I'm
32:06
just using this as an example where
32:08
you know , we didn't create Seqster
32:11
to be connected to Jawbone and
32:13
we don't know if patient has
32:16
a Jawbone medical device
32:18
. Or are they seeing a dermatologist
32:21
in small town Arkansas with
32:23
a very , you
32:25
know , rare EMR system ? That's
32:27
not one of the top 10 EMRs
32:30
because we're plugged into 28 top EMRs directly . But
32:32
if we're not plugged in EMRs directly , but
32:34
if we're not plugged in , we've
32:36
also built a statistical learning
32:39
tool or
32:46
some people would call it AI . I wouldn't call it AI , but I would
32:48
call it more statistical learning where we can recognize , from
32:51
the request of the individual
32:53
, our patient mediated method to
32:56
bring in any data source and
32:58
fill in the gaps of data , whether
33:00
that be claims data with their EMR
33:03
data , whether that be social
33:05
determinants of health data , whether
33:08
that be you know some clinical
33:10
diagnostic data that just doesn't exist
33:12
, not your regular 23andMe and
33:14
Ancestry data . Maybe
33:16
it's a whole exome file that
33:19
you had run on your
33:21
child that had some sort
33:24
of rare disease and you wanted
33:26
a diagnostic odyssey finding
33:30
of some sort .
33:31
So that's really where that is Fantastic
33:34
. Thanks for going into that detail . That's a lot of
33:36
ways that you can kind of pull in that patient-generated
33:39
information . What about non-biometric
33:42
or non-fitness tracker data ? Do
33:45
you do anything with patient-reported outcomes in
33:48
any capacity ?
33:50
Absolutely yeah , built
34:03
with actually , abbvie , one of our clients , one of the most prestigious ePro modules
34:05
. So we have an ePro module built into the Seqster operating
34:07
system where our partners
34:09
, our customers , can actually
34:12
launch that with one click to
34:15
one patient or millions
34:17
of lives . And what's
34:19
nice about that is , while you're
34:21
ingesting all of this
34:24
data whether it's coming from the
34:26
EMRs or your genome or your
34:28
wearables or your medical devices or
34:30
whatever that may be the
34:32
researchers get a research
34:34
admin portal where
34:37
they are looking at the de-identified data
34:40
. Number one , number two we
34:42
were able to run analytics on that
34:45
and query the database
34:47
of data that they're collecting
34:50
and their patients for that specific
34:52
data lake that they're building for
34:54
whatever study . And , more
34:56
importantly , they
34:59
can customize the
35:01
ePro to fit not
35:03
only their particular study , but
35:06
they can customize it so they can
35:08
launch a brand new study . I'll give you
35:10
a real world example . We
35:12
ran a migraine observational
35:15
study . They wanted
35:17
50 patients . We got
35:19
them 5,433
35:21
. And in 2026
35:24
, they were going to launch some endometriosis
35:26
projects . But because
35:28
of the data that we got and the scale that we
35:30
got it in less than three months and
35:32
with our ePro module , that
35:34
endometriosis study was
35:37
launched within four months and
35:39
and not like three
35:41
years . And the reason being
35:44
is because there's nothing
35:46
more valuable and powerful
35:48
than real time , real
35:50
world data .
35:52
Yeah , I mean that's a pretty clear ROI , being able to expedite
35:54
pipeline and , you know , really speed up that
35:56
trial recruitment process by three years
35:58
. I mean that can be a massive gain
36:01
to pharma because , as we all know , if they have
36:03
a molecule or a new therapy
36:05
that's ready to go into humans , that patent's
36:07
already a couple of years old . So saving them three years on
36:09
a patent could be billions of dollars , depending on
36:11
the indication , you know , many billions of dollars
36:14
, while something's still protected .
36:16
Exactly , absolutely , and I think it's
36:18
the fact that you know the drug development process
36:20
we shorten
36:22
significantly because
36:25
of real world data and and the work
36:27
that UBC is doing with their
36:29
customers utilizing Seqster's
36:31
technology is a great example of that .
36:34
So that actually gets us to the next question , which is
36:36
, Aaron , what are some of the early results
36:38
you've seen from working with your clients ? Can you
36:40
share any other specific metrics or outcomes
36:43
beyond the example that Ardy just gave us
36:45
?
36:46
Yeah . So that's a
36:48
great segue because here's
36:50
a good example of how a
36:53
solution like Seqster and collaboration
36:55
and plugging it into clinical
36:57
research can accelerate drug
37:00
development . So at the start of of any
37:02
clinical trial you
37:04
need to identify patients who are
37:06
eligible for that clinical trial . One of
37:09
the ways you're going to uh , but
37:11
it's fundamental to determining
37:13
eligibility is looking at
37:16
the patient's medical record . What medications
37:18
are they taking that may make them ineligible
37:21
for the trial ? What types of things in
37:23
their medical history and confirmations and
37:26
diagnosis do they have that make them
37:28
eligible or not eligible ? You
37:30
have to be able to look at their records . When a patient
37:32
shows up to a site to participate
37:34
in research and that site already has all their medical
37:37
records , that process will go very quickly and they'll
37:39
make a determine of eligibility and that patient will be enrolled or not enrolled
37:41
. But if that patient comes to a site and they do not have all the medical records at their fingertips
37:43
for that patient will be enrolled or not enrolled . But if that patient comes to a site and
37:45
they do not have all the medical records
37:47
at their fingertips for that patient
37:49
, they have to be gathered and
37:52
classically this process can
37:54
take a long time and it's very manual . You're
37:56
talking about signing medical information release forms
37:58
, sending them around to the places where
38:01
that patient receives healthcare , waiting for the records
38:03
to come back . Sometimes those records requests
38:05
are fulfilled in 30 days
38:08
or something like that , and it just slows
38:10
down the entire effort
38:12
of trying to get eligible patients
38:14
into your study . So we see it as
38:16
a really vital solution there that
38:18
can be applied in site-based research could also
38:20
be applied as we look at studies
38:23
that maybe don't have to go through sites and you're going
38:25
directly to the patients and recruiting them , kind
38:27
of like the examples I cited earlier . You still
38:30
have to gather that patient's medical records if we're
38:32
going direct to patient and these
38:34
tools can be used to do that
38:36
. So we see it playing an
38:38
important role in accelerating
38:40
clinical research . We
38:42
also see it in an important role
38:44
in accelerating real-world
38:46
evidence generation . I'll
38:49
give one more example of that . Let's say
38:51
you're a pharma company and you're bringing a new
38:53
treatment to market in ASCVD
38:56
or atopic dermatitis , and
38:58
one of the initiatives that is pretty common
39:01
for a lot of these companies that are doing
39:03
that is that they may want to conduct
39:05
what's called the disease registry and they just
39:07
want to study patients with a certain disease
39:09
and be able to look back at their
39:11
healthcare journey a certain number of years and then follow
39:13
it going forward . And the reason
39:16
they want to do that is they
39:18
want to understand the treatment landscape , the treatment
39:20
patterns , and then , as their new product
39:22
is introduced into the market , publish
39:25
data on different comparisons and
39:27
looking at patients who are taking their product versus
39:29
others . And we call this
39:31
discretionary evidence generation to drive publications
39:33
and convey potential value
39:35
proposition of one treatment over another
39:38
. In order to do this type of research and this is what
39:40
UBC does , a lot of these large
39:42
disease registries Classically
39:44
in the past , you're enrolling hundreds of
39:47
sites across the country and paying
39:49
them to do the things I talked about earlier manual data
39:52
entry . That becomes very economically impractical
39:54
, but it's not necessary in today's world
39:57
. In today's world , we can go direct to patients
39:59
and we can find patients with atopic dermatitis
40:01
or who are at risk of ASCVD and
40:03
we can ask them directly if they'd like
40:05
to join this disease registry , and
40:07
if they do , then we can
40:09
gather their medical records at scale . We can get
40:11
thousands of patients , tens of thousands of patients
40:13
, if we need to , into a study like
40:16
this , and now
40:18
we're generating evidence that we
40:20
could not have otherwise , and
40:23
so this is going to help
40:25
clinicians and patients understand the
40:28
safety profile and the
40:31
potential efficacy profile of a certain medication
40:33
in ways that , in the past
40:35
, would have taken much longer and been much
40:38
more costly to yield
40:40
that type of evidence .
40:42
Yeah for sure . I think that's one
40:44
of the more apparent and
40:47
discussed use cases for this is that
40:49
clinical trial recruitment , in particular A
40:51
lot of the RWD companies I've been talking to , it's
40:54
identifying where there might be pockets of
40:56
individuals with these conditions that
40:59
haven't really been targeted for clinical
41:01
trials in the past , either because they weren't part of
41:03
an AMC or they're too far away from
41:05
a larger urban center where there was just
41:07
a higher density of those types of patients
41:09
. So it's definitely interesting seeing how
41:11
this can , in theory , democratize
41:15
and open up the
41:17
access to clinical
41:19
trials , both on the provider and the care organization
41:21
side , as well as on that patient side
41:24
. You know , giving this a more open
41:27
access , patient driven
41:29
framework . It does create
41:31
a lot of additional opportunity
41:33
to expand research efforts into
41:35
new pockets , which I think is really exciting . However
41:37
, I was recently speaking with somebody in the CRO
41:40
world who was saying that pharma
41:42
as a whole is a very
41:44
conservative group when it comes to decision-making
41:46
and thinking about the future , which
41:49
means they won't be all-in early adopters
41:51
of anything until someone else proves it
41:53
as an accepted standard way to practice
41:55
or regulations force a change
41:57
. Do you have any examples where
41:59
you've seen a
42:01
client or seen somebody taking
42:03
a more aggressively forward thinking or
42:06
innovative approach to this kind
42:08
of work , where that has ended up being a competitive
42:10
advantage .
42:11
Yeah , you're absolutely right . You
42:14
know it is a , you know , traditionally
42:16
a very conservative group
42:19
and what
42:21
has to happen is , you
42:24
know , you usually do
42:26
have to be a pilot study , a proof of concept
42:29
, done in small settings , with
42:31
lower stakes , and
42:36
once that's proven , then it can be applied and it's
42:38
slowly adopted by the organization
42:40
, by pharma , and applied
42:42
in other places , the
42:45
. You know I won't go through it again
42:47
, but but it really is the example I
42:49
cited earlier where we were doing a chart
42:51
review , a natural history study , important
42:54
research , but you know we're not
42:56
in that setting . It wasn't a
42:58
pivotal trial for for for registration
43:00
. So it's a different
43:02
stakes in terms of that and and once it's
43:04
proven and it works . And . So so you know different stakes in terms of that and and once it's it's proven
43:07
and it works and we prove , hey , we
43:09
actually were able to get this patient's medical
43:11
records . We didn't
43:13
have to pay this healthcare provider
43:16
to keystroke it all into a
43:18
web-based EDC . Once you
43:20
show that that works , and then
43:23
word starts to spread
43:25
, it then becomes
43:28
, it will become more
43:30
part of what people do , just like my
43:32
earlier example where we were using
43:34
paper for a long time
43:36
, just transcribing pieces of data from
43:38
a patient's paper chart into a piece
43:41
of paper that went into a study database . ultimately
43:43
, paper chart into a piece of paper that went
43:45
into a study database . Ultimately we got the web based CDCs . Somebody
43:47
proved that it worked and then it became the
43:51
universal standard . This is
43:53
already starting to proliferate . We
43:55
are seeing more and more of our pharma
43:57
clients come to us already
43:59
We'll raise the idea if they
44:01
haven't . But they are coming to us and saying
44:03
you know we'd like to go with a solution
44:06
. They call it different things patient mediated medical
44:08
record release , patient authorized medical
44:10
record release , hie medical
44:13
record release . They're saying we'd like to
44:15
apply that technology here
44:17
. We think it can help us in our
44:19
efforts to identify eligible patients for
44:21
a clinical trial . We think it can help us in our efforts
44:23
to enroll a
44:25
massive registry , starting
44:28
to hear about it and understand
44:30
it and ask for it . So the snowball
44:32
is starting to build and grow
44:34
, so to speak .
44:36
That's good to hear . I do think the
44:38
data technologies are becoming more robust
44:40
. Obviously , one of the big issues for a long time
44:42
was just constantly the data quality
44:44
issue and the interoperability problem of trying
44:46
to reconcile disparate data from disparate
44:49
underlying infrastructure would
44:51
cause complications .
44:53
And that's a good point , because that's still that's
44:56
a really good point , because that still is the
44:58
part that needs to be solved a bit . You know
45:00
we're going to get this data
45:02
, okay , but what formats are we
45:04
getting it in ? What level of harmonization do we
45:06
still need to do ? How do we address missingness
45:09
, which shouldn't be a problem in real world
45:11
data ? You know there's , there's certain you know
45:13
procedures that were just not performed , but you have to know
45:15
was it not performed or
45:17
are we missing it and exists somewhere else
45:20
? So there's still a lot of Devils
45:22
in the details to be solved in that
45:24
respect , and that's
45:27
the next step change
45:29
Makes sense .
45:31
So already you know don't want to
45:33
belabor this too much because I kind of mentioned
45:35
it briefly right before , but I think it'd be interesting
45:37
to hear your perspective on
45:39
what you might be seeing at the kind of
45:41
macro level at Seqster around
45:44
the interest in these types of tools for
45:47
increasing access to trials and improving
45:49
equity in trial recruitment , and
45:52
if you actually are seeing any uptick or
45:54
if you're collecting any metrics around that .
45:57
Yeah , so 2016,
45:59
. Nobody wanted to talk to me . 2017
46:02
, we got maybe a couple phone
46:04
calls from investors that
46:06
were interested and we raised some
46:08
funds after
46:10
I put the first big
46:13
chunk in myself because no one wanted to fund
46:15
the dirty engineering . 2018
46:18
, we come out of stealth . Dr
46:20
Eric Topol tweets that he
46:22
brought his data from 1985
46:24
to present four different health systems , four
46:27
different EMRs , his 23andMe
46:29
data , his genetic data , his Fitbit data
46:31
, his MyFitnessPal
46:34
nutrition data . We didn't know that he was
46:36
even tracking that information and
46:38
in less than 24 hours he said step
46:41
in the right direction . Same
46:43
time , lots of folks
46:46
find out about us because Dr Topol
46:48
is so famous and his tweet
46:50
got us 6 million plus hits and broke
46:52
our consumer model . And
46:55
then we meet Bill Gates . Bill
46:57
Gates tells us to build
47:00
an operating system and take an enterprise . I
47:03
listened to him . That was the best advice that I got
47:05
to this day on the business
47:07
still from any investor or anybody
47:09
. And then we
47:12
do some things around Alzheimer's . With
47:14
Boston University , we
47:17
build the research portal , which was known
47:19
as the SRP , which is the Seqster
47:21
Research Portal , which now
47:23
became our admin portal
47:25
or staff portal for researchers
47:28
and pharma sponsors and CROs
47:31
. 2019 , we're
47:33
trying to figure things out . But pharma starts knocking
47:35
on our door . It was the first year . At JP
47:38
Morgan Healthcare Conference , right before
47:40
the pandemic , patient centricity was
47:42
all over the walls in downtown
47:44
San Francisco . At every single event
47:46
that you went to March 2024
47:48
years ago , the pandemic hits . Everything
47:51
shuts down , everything goes decentralized
47:54
, everything goes hybrid kind
47:56
of , but more so decentralized
47:58
trials , dcts kind of take off
48:00
. Pharma starts getting
48:02
interested in how they can bring patients
48:04
and data from home , and home health and
48:07
telehealth explode 2021
48:09
, 2022, . Pharma starts
48:11
partnering with us and finally
48:14
we see the light at the end of the
48:16
tunnel . It took like six years . I
48:18
think I told you when we were showing you the
48:20
Seqster demo , john , how long
48:22
it took six , six and a half years to really
48:24
hit that growth
48:27
stage . Cro started calling
48:29
us end of 2021 . We strike
48:31
our partnership February of 2022
48:34
with Aaron Berger and friends at UBC
48:36
. We've been working with them for over two years
48:38
now on multiple different studies , getting
48:40
them the data that they need to run
48:42
their RWE generation . And
48:53
then you know , 2023 was the breakout
48:55
year . We exploded . I mean , we got two dozen
48:57
deals . Our technology was built by Pharma , mckinsey
48:59
, deloitte , accenture , cros , patients
49:02
, innovators like Dr Eric
49:04
Topol and his friends . It's pretty
49:06
amazing . And now , in 2024
49:08
, I think there's
49:10
more opportunity that we can handle
49:12
, and before we were
49:14
behind the eight ball , the wind was
49:16
always in our face . I'm here to tell
49:19
you today on your podcast this
49:21
is my 64th or 65th podcast
49:23
doing this sort of thing that
49:26
it's the first time I feel
49:28
like the wind is behind our back and
49:31
the only reason is because we actually
49:33
solve not one problem , not
49:36
two problems , not three problems
49:38
. We solve a multitude
49:40
of problems across healthcare and
49:42
life sciences , and it's not just
49:44
about medical record data . We
49:47
have multiple different tools
49:50
that we have embedded into
49:52
our operating system that relates
49:54
to payers , pharma , cros
49:57
, consumer brands and above Super
50:01
excited .
50:02
Yeah , I can tell , I can hear that coming through . Okay
50:06
, so in closing , I have one
50:08
more question that I want to get to , which is getting
50:10
your guys' perspective on the future
50:12
. So where do you see the industry in
50:15
the next five years , the next 10 years
50:17
? And then can you point to any
50:19
specific inflection points on
50:21
the horizon that
50:23
you can foresee , given recent trends
50:25
and regulatory action ?
50:27
Yeah , I can jump in on that and I
50:30
saw in our pre-show notes
50:32
. You really can't do a podcast on
50:34
anything these days without maybe talking a little
50:36
bit about .
50:37
AI so .
50:38
I know that that's on
50:40
your mind , perhaps on the listeners'
50:42
minds as well . You know
50:44
, what we talked about today is a
50:46
pretty straightforward use case , is being able
50:48
to go direct a patient to do
50:50
things like confirm eligibility and collect
50:53
data at scale for evidence generation
50:55
. I've talked a lot about the way we've used it
50:57
and we're talking mostly about structured
51:00
fields that are in the EMR and
51:02
things you can do with that , and then things
51:04
that you can do with the unstructured field , the source
51:07
notes . At this point , you know , rely mostly
51:09
the way we've used it people reading
51:12
those source notes and extrapolating pieces
51:14
of information . So the
51:16
next crank of the wheel and it's already you
51:18
know it's upon us is is applying natural
51:21
language processing to
51:23
reading those source notes and
51:26
moving far
51:28
ahead in terms of what type of evidence
51:30
can we derive by looking at , you know
51:32
, tens of thousands , hundreds of thousands
51:35
of patients worth of source notes
51:37
with a given disease and being able to
51:39
do that at scale , which you
51:41
know natural language processing can do
51:44
that humans cannot , and
51:46
that will really supercharge the
51:50
types of research we're able to perform , the questions
51:52
we're able to ask and answer from
51:54
reading tens of thousands
51:56
of patients worth of source notes
51:59
and just
52:01
generate richer and more valuable evidence
52:03
. So it's using AI , machine learning
52:06
, to
52:08
read those notes and do some of
52:10
this research . We're gonna see that start to
52:12
happen a lot in the future , but
52:14
Seqster is the . We
52:16
need solutions to obtain those medical records
52:18
in the first place before we can apply some
52:21
language processing and LLM
52:24
. So that's how we obtained
52:26
the records .
52:27
Fantastic and Ardy . What about you
52:29
? Any thoughts on the future ?
52:32
Yeah , you know , I've always had this dream of
52:34
bringing
52:36
the Netflix
52:40
, amazon one-click
52:42
experience to healthcare . It's
52:45
kind of why I started Seqster without even
52:47
knowing it , and it took a long time
52:49
working with wonderful people like
52:51
yourselves to understand that we
52:54
are on the brink of
52:56
making that happen . At
52:58
the time , I wanted to empower 7.7
53:00
billion people to collect
53:03
, share and own their data , and
53:05
now there's 8 billion people to the planet
53:07
. So 300 million plus have been added
53:09
since our little journey here in
53:11
eight and three months and
53:13
change running Seqster
53:15
here , but that
53:18
still is , I think , attainable
53:22
. It's a lot closer than what it was because
53:24
we were early , like you stated , john
53:26
, but we stuck at it , and the advantage of
53:28
being early is that we know what not to
53:30
do . I think we're
53:33
getting closer to
53:35
folks really
53:38
understanding their health journey
53:40
in a way that was non-imaginable
53:43
or not just for the innovators
53:45
, and I believe
53:48
that every single person
53:50
let's just take it looking
53:53
at just the United States should
53:55
have access to being enrolled in a clinical
53:58
trial , and I
54:00
think our technology is
54:02
well positioned to make that
54:04
happen for CROs , for
54:06
pharmacies , for payers , for
54:09
pharma companies and all
54:12
the above , because it is
54:14
an experience . It's
54:16
not just data in , data out . This
54:18
isn't just an API . Those
54:21
things are actually worthless to me and
54:23
that's why we focused on the patient
54:26
. The patient centricity
54:28
and the innovation is not going to
54:30
go away and we're
54:32
really excited about the
54:35
next phase of
54:37
where our clients are going to take our technology
54:40
. Of course , ai
54:42
is meaningless without data .
54:44
Very true . I
54:47
mean that's . The other thing with all this data is that healthcare generates more data
54:49
than any other enterprise sector . Right now , I
54:52
think it's growing at roughly 30% a year . Part
54:54
of that's also because it's late to the game , but
54:57
it doesn't change the fact that it's also generating massive
54:59
amounts of new data every year , and
55:02
so , as we
55:04
look at the future of healthcare
55:06
, it has to be data driven . It's the only thing that we haven't
55:08
really been able to do at scale until the modern
55:10
era , and it'll be very . I'm
55:12
just really excited to see what comes of all that , as
55:14
we start to identify new patterns , and
55:16
these new AI tools that we're bringing
55:19
to market are good at pattern recognition
55:22
, right , so they're going to be able to identify the patterns that
55:24
we haven't been able to see before because we haven't had that
55:26
perspective , and so population
55:28
health and public health could have some very
55:30
significant changes to how that's
55:33
perceived in the coming decade or
55:35
two . All right
55:37
, so thank you so much , both of you , for
55:39
joining us today as we explore one of the primary
55:42
use cases for real-world data and
55:44
real-world evidence , which is the clinical
55:46
research process . If any of
55:48
our listeners want to learn more , how should they go about
55:50
reaching out to either of ?
55:52
you ? Yeah , sure , aaron Berger at ubccom is
55:55
probably the
55:57
best way to reach me and
56:00
, yeah , I'm happy to talk . We're
56:02
passionate about , as I
56:04
mentioned in the beginning , helping
56:07
pharma , biotech answer
56:09
important questions about the safety and
56:11
effectiveness of medicines . We're
56:13
always really excited about having
56:16
collaborative thought
56:18
, partnership discussions on what is the best way to
56:20
design a study to do that , which
56:22
technologies should we bring
56:24
in , and
56:27
those are fun discussions . So if anybody would like
56:29
to have that discussion , yeah , definitely reach out to me
56:31
at my email address .
56:33
Okay , and Ardy .
56:34
Yeah , and I'm Ardy
56:36
at Seqster . com S-E-Q-S-T-E-R dot com
56:40
, and you can also connect with me on LinkedIn
56:43
. I'm very active on LinkedIn and
56:45
so I'm a LinkedIn junkie and
56:48
it's probably a better way , while I'm
56:50
traveling , to connect with me if
56:52
you can't catch me on email or if I don't get back
56:54
to you , but I usually get back to people
56:56
and love having any type of
56:58
conversation with anyone that resides
57:01
around data , interoperability
57:03
, trials , innovation and
57:05
all the above . Thanks so much , john , for having
57:07
me , and Aaron .
57:10
Yeah , thanks , john for having me .
57:12
Yeah , this has been fantastic . So thank
57:14
you so much for being part of this and joining us today and
57:16
sharing your expertise with our community . To
57:20
continue following along with these conversations
57:22
, be sure to subscribe to the Chilcast
57:24
on your favorite podcasting app and follow
57:26
our work on LinkedIn and or via
57:28
our newsletter . Feel free to reach out
57:30
with any questions and suggestions for future
57:33
guests or topics to feature on the show
57:35
, and thank you for tuning in .
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More