00:00hi everyone welcome to day 6 in the
00:03podcast I'm sonal today we're doing
00:05another of our episodes in our series on
00:07healthcare challenges and opportunities
00:09and the topic is all about dark data
00:11access to data patients taking back
00:14control of their data and what else is
00:16possible when such data is brought to
00:18light we therefore also discuss the
00:20interesting nuances and a bit of history
00:22of HIPAA as well as briefly touch on
00:24clinical trials the opioid crisis and
00:27the case of dr. Google is it a bad thing
00:29or even a good thing so joining me for
00:32this episode our special guests are
00:34Susannah Fox who was previously the CTO
00:36at the US Department of Health and Human
00:38Services which oversees all the Health
00:40and Human Services agencies of the
00:42United States from public health abroad
00:44to Medicare and Medicaid at home she's
00:46also an adviser to citizen speaking of
00:49we also have a Neel Sethi founder and
00:51CEO of citizen a patient centered
00:53platform for helping people collect
00:55organize and share their own health data
00:57and last but not least and in fact he's
01:00the first voice you'll hear after mine
01:01and Susanna's we have Vijay Pandey
01:04general partner in a 6 & Z bio so let's
01:07start with this whole problem of dark
01:09data and I don't mean to make it sound
01:10sinister but it is kind of sinister cuz
01:12it's hidden from world view why are we
01:14in this position in the first place
01:16we're in this position because most of
01:18health care has grown up believing that
01:21so professionals generally collect and
01:24hold the clinical data that drives a lot
01:28of clinical care and yet patients now
01:32have an expectation to have access to
01:35data in the modern world plus there's a
01:38whole part of data that isn't even part
01:41of the clinical record and so how might
01:44we bring all of these sources of data
01:46together think about the vestal elements
01:49of where the whole healthcare system
01:50came from it you know origins you know
01:5250 hundred years ago from
01:54fee-for-service like doctors coming to
01:56your house and data then is like writing
01:58something in a notebook we have
01:59electronic medical records now but it's
02:01really not radically different it's
02:03basically a doctor or writing or a
02:04provider taking down some notes it's a
02:06difference in media I'm not actually in
02:08kind what are the structural reasons
02:10first of all I'm coming at this a little
02:11skeptically because AI always hear
02:14health care and I understand it's
02:16important but I'm skeptical because
02:17quite honestly as a user I barely know
02:20when my car registration is due I don't
02:23really think having control of my
02:24healthcare data is really the thing I
02:25need next on my list luckily I'm a
02:27healthy person so maybe that's part of
02:29the reason but still I really can't see
02:31a lot of people really using data so
02:32let's talk about why this matters really
02:34you mention a car analogy let's just
02:36think about the fact that I think we
02:38take better care of our cars and we do
02:39it ourselves okay that's a very you know
02:42I mean you have a fear about data a car
02:44so you have like dashboard lights that
02:46come on when there are problems and then
02:47you actually go to professional and you
02:49give them a full run of what's going on
02:51you have a lot of sense of the car
02:52because you're driving all the time
02:54and it has sensors I get a tire pressure
02:55sensor every three months or six months
02:57or something and so that data that is
02:59sort of the clinical equivalent for the
03:01car is something that the mechanic needs
03:02to be able to improve things what I
03:04think really is missing is sort of
03:06having the alignment of incentives so
03:07you want your car to work you obviously
03:09want your body to work and your provider
03:10wants it be able to help but then how
03:13can that person do that if they don't
03:14have all the information and so I think
03:16really now the question is how can
03:17people who really have the incentive to
03:20take care of themselves finally use the
03:21modern technology to let that data
03:23actually help if we knew four cars we
03:25should be able to four people and you're
03:27absolutely right that most people don't
03:28want to engage in their health most
03:29people don't want to know
03:31what's going on they're not necessarily
03:34ready to look into the mirror of data
03:36I call it standing naked in front of the
03:38mirror of data it's a little too much
03:41truth but then when something happens
03:44you're diagnosed with something or your
03:46child or your mom is diagnosed or
03:48something that will trigger engagement
03:50let me tell you and then people should
03:53have immediate access to the data and
03:56information that they need to make a
03:57decision and that structurally hasn't
04:00been the case and that actually I think
04:02is this moment that we're living through
04:04right now where there's changing
04:06expectations about what we should have
04:07access to and so health care is rather
04:11slow in taking up this new culture of
04:15access but I think it is changing and I
04:19think it has to change
04:20what's the driver for the change one I
04:22would think of up the top of my head is
04:23just the mobile phone and things like
04:24push button experiences but I think one
04:27the driver is we've got 10,000 boomers
04:30aging into age 65 and is having a
04:33different sort of the user experience
04:34and because people are living longer
04:38doesn't mean we are living healthy
04:40longer and so they're going to be
04:42dealing with lots of issues that their
04:45forebears didn't have to deal with
04:47because people didn't live this long I
04:49mean a hundred years ago or 120 years
04:51ago the average lifespan was about 48
04:54now we've artificially extended that to
04:57about let's say 83 and these are
04:59approximate numbers but who's to say
05:02that we're really improving the quality
05:03of life so people are going to retire
05:05and become more engaged than their
05:08parents who are in their own health care
05:09and I know my parents go to a doctor and
05:12say here's what I can share with you now
05:15save my life please I think the next
05:17future generations are gonna be much
05:19more engaged when I think about my
05:21parents and how they narrate their oral
05:24histories so much of medicine is about
05:25oral histories and they're immigrants to
05:28this country there's well-educated but
05:30they somehow do this weird being where
05:32they feel like they have to have this
05:33best behavior and you can't actually
05:35report the problems and it drives me
05:37insane you because my dad will call me
05:39maybe like oh yeah I didn't tell the
05:40doctor this why go it didn't come up and
05:42I'm like your job is to report your
05:44history that's why they're asking you
05:46all these questions you don't have to do
05:47this model minority thing you're
05:48behaving well for the doctor for God's
05:50sake and I get really upset because I
05:52love my parents there is a break in the
05:54existing healthcare system where there
05:57is a over reliance on oral histories as
05:59part of the medical experience that I
06:01would think if I could just send them in
06:03with like all of our notes then that
06:06alone would be a huge improvement to the
06:07doctor to have that context well and
06:09there's much less friction for sharing
06:11this type of information so you think
06:13about all the data you can quickly share
06:14on Facebook or LinkedIn or something
06:16like that that's really trivial to do if
06:18you try to do this ten years ago 15
06:20years ago you could come to the doctor
06:21with a bunch of like CDs or DVDs and
06:24they could like load it in but by the
06:25time they've done that thing your
06:27appointments over you have 30 minutes
06:28yeah and so friction is a really
06:30critical thing if you can bring the
06:31friction from 30 minutes to a third of a
06:33second that's when the doctor now is
06:35actually able to use that information
06:36but you know people have under the HIPAA
06:40they've got loads of capabilities that
06:43they don't know how to exercise because
06:45the friction that you're talking about
06:47is still very prominent and I don't
06:50think it's people that are trying to put
06:53friction in front of patients I think
06:55they just haven't been taught or don't
06:58know any better I think most people at
07:00medical records systems and hospital and
07:03h-i-m facilities want to do the right
07:05thing on behalf of patients we were
07:08talking about the car analogy one of the
07:09things that other industries have done
07:11is surface information voluntarily so
07:16that the lights on the dashboard and the
07:17accelerometer and all of that
07:19navigation engagement system they're
07:21meant to be used now imagine if all of
07:24that was dark and you had to open up the
07:27hood and look underneath and know how to
07:29discern that information our healthcare
07:32system is very much like that so you
07:34brought up HIPPA and that's something
07:35that everyone hears about we talk about
07:37the context of privacy but you're saying
07:38it's an opportunity for people to
07:40because of it by they can have data I
07:41usually hear it the other way around
07:42when it comes to startups and HIPAA and
07:44privacy can we quickly talk about what
07:45happens and how it came about
07:47and why it matters here well I'll start
07:49by saying that one of the most important
07:50things to know about the acronym HIPAA
07:52is that the P stands for portability oh
07:54I had no idea what does HIPAA stand for
07:56so it's a health information Portability
07:59act the other important thing to know is
08:01the context that it was actually written
08:03in 1996 so throw your mind back to 1996
08:07when it was about 10 percent of American
08:09adults had access to the Internet this
08:11was a time when we weren't sure what was
08:14really going to be the impact of
08:16technology on healthcare and yet there
08:19were visionary people who said this is
08:22going to be important that people have
08:23access and that people are able to take
08:26that data and have it centered on the
08:29consumer centered on the patient the
08:32idea is that patients should be able to
08:34get access to the data and redirect it
08:37the places that they want it to go so in
08:40the modern world we now talk about that
08:42in terms of an API that somebody could
08:44get access to their data and direct it
08:46to an app or direct it to where they
08:49want to go or be the conduit that
08:51creates the interoperability between
08:53to clinicians which unfortunately is
08:55sometimes the case well that's actually
08:57the spirit of HIPPA and affordability
09:00yes instead it's been used kind of as
09:02the weapon as a way to say stop what
09:05we're doing let's get the lawyers
09:07involved yet privacy is important but
09:09that's exactly the context I've always
09:12heard it in as a sort of a you can't do
09:13this because of HIPAA but that's exactly
09:15the opposite of what it actually is
09:17there are very few sentences that a
09:19licensed professional the nurse
09:21practitioner anyone can say to a patient
09:23that starts with I can't do that that's
09:26a HIPAA violation there are very few
09:28mental health and sexual health
09:31carve-outs but most of the time I am in
09:35the right when I can ask my doctor and
09:36she can send me all of my information in
09:39any digital format I request as long as
09:42she's got it to my gmail account and
09:44it's not a HIPAA violation I think very
09:46few people know that I have no idea
09:49because it's driven by the patient that
09:52patients are actually outside of HIPAA
09:54so you can ask as a patient a HIPAA
09:58qualified entity to do something for you
10:01and share your information whether it's
10:03through an API or whether it's through a
10:05PDF or email and they have to do it
10:08most people don't know they have that
10:10much is a very important idea if I were
10:12to visualize this structurally you think
10:13about that all these players in the
10:15healthcare system you have insurance and
10:16hospitals and clinics and universities
10:19and I could keep going pharma companies
10:20etc if you picture this whole ecosystem
10:23as a circle right now what you're
10:26talking about is putting the patient at
10:27the center of that circle and therefore
10:29they can pull on all these pieces but
10:31otherwise these entities cannot
10:33necessarily talk to each other directly
10:35what the patient's data patient is at
10:36the center and just doesn't realize it
10:38god at most yeah taking your circle most
10:41of the entities are sharing with each
10:44other but not through the patient so the
10:46longest way around the circle from one
10:48site to another is around the circle the
10:51shortest path if you just do the center
10:53if that's the diameter that's where the
10:55patient is you know once you realize
10:57that the patients at the center of it
10:58you realize there's a huge opportunity
10:59there if companies or startups empower
11:03the patient then the patient can drive
11:05this whole thing they can quarterback if
11:06they can cast for the
11:07data they can send the data now that
11:09question is actually or what can you do
11:11with it and what makes the impact I do
11:12want to talk about what you can do with
11:13it now because going back to the
11:15original question of do people really
11:16need their data and does it really make
11:18a huge difference but I'm thinking of
11:19the analog of what happened with
11:21wearables people have long talked about
11:23wearables and just like they now achieve
11:25the car there are lots of sensors out
11:27there measuring our bodies but the
11:28number one problem when wearables and
11:30data is nobody actually takes that as
11:33actionable data and does anything with
11:35it so it's a case where you have a lot
11:37of data just gathering that's not being
11:39used are we talking about the same
11:41problem here in the case of wearables
11:43you're dealing with something that's
11:44very broad and very shallow and also
11:47often not very clinical like 10,000
11:49steps versus 9000 versus 11,000 that's
11:51largely just made up in terms of its
11:53clinical significance maybe let's turn
11:55upside down which is what are the
11:57clinical areas where you have clinical
11:59data that could create action bla
12:02outcomes for patients maybe you'd start
12:04with what is the greatest need people
12:06maybe that could easily die in a year or
12:08two and you know people that have like
12:10mid to late stage cancer that was an
12:12obvious example and there's data they
12:14have from numerous different areas and
12:16imaging and genomics and all of that and
12:19it's just kind of amazing how much do
12:21you there is now to get and then the
12:23question is how do you gather that and
12:25how can you empower the doctors via the
12:27patient's to actually make a difference
12:28but in the case of cancer aren't the
12:30doctors already sharing that data so why
12:32does the patient need to be at the
12:33center of that so the question we get
12:35asked is what will we do with it and the
12:37answer is you'll be able to share it
12:40with the oncologist who will know how to
12:42operationalize that data and help
12:44treatment for folks who know me my
12:46little sister contracted a number of
12:48years ago late stage metastatic breast
12:50cancer diagnosed and straight into Stage
12:52four because someone missed the
12:54diagnosis and blew at a years earlier
12:56into stage one so the point is in her
12:59last year of life Tonya was seen at 14
13:02facilities all using not all but many of
13:06them using different EHR or medical
13:08record systems across multiple states
13:10which have their own transmission of
13:13health data across state lines issues
13:15and she was deemed by 23 oncologists so
13:18that sets the groundwork every time she
13:21new oncologist either there was a
13:23restart a frustration factor which she
13:26had to explain everything but because
13:28she had her electronic health record
13:30imagine a LinkedIn of your health
13:32profile she had that she had that okay
13:34most people don't even have that her big
13:36brother was in the industry so of course
13:38Tanya used glimps which is a former
13:40company Apple acquired it in a
13:42transaction and some of the health
13:44records that we see them releasing it to
13:46the world is some of that technology and
13:48so she had the ability to share her
13:52profile her portable health record with
13:55all her meds all her labs all her
13:57genomic information what we had done is
14:00we had built a depth of health record
14:03that could operationalize her cancer
14:06care this is in contrast to others who
14:09think that a mile wide and an inch deep
14:11and data is the way to go we actually
14:13think in order to operationalize health
14:15data you have to find self incented
14:18people and that is cancer patients lupus
14:21HIV autoimmune disease
14:23we found that we needed to add imaging
14:26and genomic which glimps didn't have we
14:28want to democratize that across all
14:30seven billion people it shouldn't just
14:31be my little sister so one of the
14:34promises I made to my little sisters I
14:36would go on and continue this work and
14:39so we are starting with cancer moving
14:42towards autoimmune all chronic diseases
14:44they all have in common it's actually
14:46the most frequent touch points into the
14:48next system and you don't need to
14:49convince them of the value they know it
14:52yes one of the chronic condition those I
14:54know that yeah yeah but the three things
14:56that we have found that we absolutely
14:57need to solve for patients is lower the
15:00friction of three touch points so three
15:03long poles in the tent one is how does
15:05someone actually request their records
15:08be released and it should be eventually
15:11something as easy as making a payment at
15:15Whole Foods with Apple pay just as an
15:16example someone should send you an
15:18invoice and you should be able to
15:20adjudicate the financial transaction by
15:22touching your thumb and boom it's paid
15:24we'd like to help health records release
15:26to that easily so that's one of the long
15:28poles the request point is when the
15:30release of information or request point
15:32is one that's where your HIPAA right of
15:34access gives you all the power of the
15:36federal government and the regs to say I
15:38can get this from you patients can do
15:41what hospitals and insurers and everyone
15:42else can't yes absolutely and they can
15:45help other physicians by being part of
15:47the referral network because patients
15:49are in the middle the second thing is
15:51most of the information that's released
15:53is not in any coded form its documents
15:56and PDFs and XML files is it the dark
15:58data part of it really efficient
16:00readability and a lot of the cancer
16:02information is in the pathology report
16:04which doesn't come out of API yet we got
16:07a fingers crossed so we need a long pole
16:10in the tent about a data refinery that
16:12takes this crude oil of documents and
16:15converts it to data it's like turning a
16:17Word document into Excel in Excel we
16:20know how to apply a feature or a
16:21function that operate on it you're
16:23basically taking unstructured too
16:24structured but you're actually adding a
16:25step even before the instructor which is
16:27making that data machine readable in the
16:28first place especially in the case of a
16:29PDF correct then the final question is
16:32so what now I have my structured data
16:34what can I do with it
16:35well you can certainly share it like you
16:37share a LinkedIn profile with your
16:38physician that physician he or she might
16:41say thank you God for bringing all this
16:43because I never get to see this sort of
16:45back but beyond sharing it with people
16:46you can also share it through an API as
16:49Suzanne has said to other app developers
16:52who can then say look citizen you've
16:54done a great job producing this fuel
16:56this is computable on your data refinery
16:58yes but we'll take it from here and
17:01we'll run a calculation that might be a
17:03cardiac calculator or will aggregate
17:06populations together and do a population
17:09survey or ideally because we can
17:12automate clinical trial inclusion
17:14exclusion you could have a series of
17:16these algorithms running in the cloud
17:19all the time and every time a patient
17:22becomes a citizen in our parlance a
17:25trial match can be detected on their
17:28behalf that's the way we reduce friction
17:31as Vijay said so you know one of the
17:33things that I think it gets sort of
17:35covered almost too fast is this
17:37unstructured is structured because
17:38structure could mean lots different
17:40things really like if you can go all the
17:42way deep with ontology x' and a true
17:45semantic structure you
17:46go from just words to understanding and
17:49understanding is like the Holy Grail for
17:51machine learning and AI right now it's a
17:53hard thing to do and when you can
17:55finally use ontology another things that
17:57people have driven now the data actually
17:59really becomes useful and I think Excel
18:01it's a very natural one and the key
18:02thing is that it's not just even in a
18:04spreadsheet it's in this retreat where
18:05that computer actually knows what each
18:06of these things mean it's right and
18:08that's gonna be key because a doctor
18:09does not have time to sort of go through
18:12pages of things understand he or she
18:14wants to be able to go very rapidly just
18:16get a sense of the leyland and from that
18:18say you know where we are and what we
18:20need to do and I think you know we
18:21talked about friction there's friction
18:23in each of these levels friction at the
18:25patient getting the data of friction at
18:27sort of what the data is and friction at
18:29the doctor side I think the ideal is to
18:31reduce all three so take me up in level
18:33then beyond the individual patient
18:35experience and talk about this at a
18:36structural healthcare system systemic
18:39level what does this mean for insurers
18:40for hospitals for researchers for
18:42clinics drug testing to boil the ocean
18:45here but how does this play into that
18:47there's so many different ways that data
18:49can inform us you know so just one off
18:51the top of my head is that if you're a
18:52pharmacist will come Pinilla
18:53you want to be able to understand how
18:56things are going first patients may be
18:57in a clinical trial may be in a baseline
18:59and you need to be able to get a large
19:01number of patients that are the right
19:02ones so this is not necessarily a
19:04hundred million people this is maybe
19:05thousands of the right ones that
19:06couldn't give you the dates you need and
19:08well I think we'll start to see is
19:10especially as new statistical methods
19:12come online that real-world evidence
19:14will be very useful in some ways will be
19:16more useful than what you could do in a
19:18clinical trial both do the power and due
19:20to the fact of real life is different
19:22than the clinical trial right it's like
19:23an NC to experience in vivo experience
19:25but the real world is the laboratory
19:27instead of the actual laboratory yeah
19:28definitely and and this is useful for
19:30Pharma but frankly it's also useful for
19:31payers in a world where things have
19:34gotten past clinical trials but really
19:36now the new barrier is not in came past
19:38the FDA the new barriers reimbursement
19:39the payer will want to know with
19:41real-world evidence like this is really
19:43helping and this is something that you
19:45obviously can't answer without data and
19:47you can't answer without the right type
19:48of data structure in the right way data
19:50is fuel for an ecosystem and so when we
19:53talk about the structure of the
19:56healthcare system as it stands now we
19:58talk about pharmo we talk about
20:00the payers we talk about hospitals but a
20:03big part of this are the patients who
20:05don't appear on anybody's organizational
20:07chart no that's such a good point and
20:08yet they have so much insight to share
20:10and we need to make sure that we are
20:15pushing the power out to the edges of
20:17the network that's where expertise lives
20:19that we don't even know about we don't
20:23yet know what will really happen when we
20:26free the data and allow people to create
20:30the dashboards that they really need we
20:32don't yet know how different patient
20:36groups are gonna create something really
20:37useful how an entrepreneur is going to
20:39look at this opportunity and say I could
20:42create something that really helps
20:44people and by the way it could be a
20:46small group but have a significant
20:48I actually the analogy that comes to
20:50mind for me when I hear and that is I'm
20:51a big historian of the history of
20:53computing tech internet and one of my
20:55favorite themes is the idea
20:57permissionless innovation and what I
20:58love about this is what you're
21:00describing because if you think about it
21:01what happened with the internet and
21:02permissionless innovation allowed people
21:03to build on top of the platform that is
21:05the Internet if you think of data as a
21:07platform and what people can build you
21:09cannot predict the use case they're
21:11second-order and third-order effects
21:12that nobody the designer of a system can
21:14never predict upfront so what I love
21:16about this is this is permissionless
21:17innovation in a permissioned way where
21:20the permission is actually coming from
21:21the patient because essentially the
21:23patient is saying you have my permission
21:24to move this data around this
21:26portability and then to your point who
21:29knows what that can unleash and that is
21:30a really exciting thing and it's also
21:32it's really American like we're totally
21:37true we're such a country of rugged
21:39individualists for good or for else
21:41right so a lot of health care depends on
21:43whether you have the wherewithal or
21:46whether somebody in your family has the
21:48wherewithal and we as a country win when
21:51we make it easier for everyone to
21:53participate so speaking of something
21:55very American in a sad way because I do
21:57agree with you that permissionless
21:58innovation is incredibly American is why
22:00I'm a capitalist but a sad reality of
22:02American life today especially it's
22:04something that we talk about a lot in
22:06healthcare is the opioid crisis and I
22:09for one would never ever say it's
22:11something that can be solved through
22:12technology alone because it's a
22:14problem but in this context how would
22:17something like this play a role in a
22:19public health crisis like the opioid
22:21crisis the opioid crisis is one place
22:25where data is the canary in the coal
22:27it is a early warning system where if
22:30you digitized the data you could
22:32literally have an Excel function and I'm
22:35you know I'm trivializing it that is
22:37scamming for populations and where it
22:40sees a concentration of certain things
22:43happening and population health folks
22:45have turned this syndromic surveillance
22:47that CDC calls this the ability to look
22:51over large populations and see trends
22:53and patterns because of data we started
22:56seeing signs of the opioid crisis in
22:58death certificates and it was actually
23:01public health researchers who started
23:03looking at data and seeing wow that
23:05we're really seeing an uptick in this
23:07sort of death among my addiction to OB
23:11well the problem was that it was
23:12unstructured data so people started
23:15looking at the death certificates and
23:17started understanding what was happening
23:19what I'm passionate about is learning
23:21lessons from the past so that in the
23:24future we can take the temperature of
23:26the country more accurately and more
23:28quickly to solve those problems so how
23:31might we create a dashboard for the
23:33country so that we see something like
23:35the opioid crisis happening everybody
23:37plays a role in the opioid crisis the
23:40pharmaceutical companies played a role
23:42public health agencies played a role
23:44payers played a role why did payers play
23:46a role really quick oh because of the
23:48way people were being reimbursed or not
23:51reimbursed for pain management all
23:53across the world pain is managed using
23:56therapies other than drugs so there's
23:59all kinds of ways here in the United
24:01States we encourage the development of
24:04drugs and through a complicated history
24:06which I won't get into it became more
24:09than norm in the fashion to reimburse
24:12for drugs to prescribe drugs for pain
24:14said of alternative therapies like tens
24:16and various other things exactly and so
24:18how does data enter into this data can
24:21tell us when a crisis is happening it
24:23can show you also where we're actually
24:25gaining ground we're seeing that
24:28gaining ground in Ohio against the
24:30opioid crisis data is telling that story
24:32so that tells me how it informs public
24:35health and people thinking about this
24:37but how can something like what we're
24:39talking about weird this dark data
24:40unveiling actually solve I'm not again
24:43trying to advocate for a solution istic
24:44view it's a larger bigger problem beyond
24:46technology but how can it help yeah I
24:49mean there's different issues with the
24:51opioid crisis one is that often the
24:53first intervention is opioids while you
24:55wait to see the back surgeon or you know
24:57musculoskeletal surgeon that's a
24:59accessibility issue and then then once
25:01that starts then the problem is that the
25:03opiates are more accessible and then the
25:04other solutions and so now you have
25:06people sort of doing a doctor or opioid
25:09arbitrage between places and so
25:11hopefully we can understand first why
25:12this is happening and then why it starts
25:15to sprout and then what's facilitating
25:17it and having the records in one place
25:18we'll do it I think it's a little
25:19challenging because if it's driven by
25:21the patient the patient's gonna want to
25:23have to obscure yeah and forget
25:25passenger so all of these diseases we've
25:27been talking about so far cancer lupus
25:29autoimmune disorders I have a chronic
25:31condition nothing to worry about not to
25:33scare my listeners but in the context of
25:35I have to see a doctor regularly etc
25:36these are all cases where you have
25:38multiple touchpoints in the system and
25:39often longitudinal data helps how would
25:42the longitudinal data and having a
25:44patient at the center now what does it
25:46do epsilon22 Donald Dale is going to
25:48become ever more important as compared
25:50to episodic data and that's because
25:53we're moving from acute care to chronic
25:55care if we had a pillar or procedure we
25:58could do something in a hospital so
26:00that's an acute episode the fact that
26:02it's chronic the condition is chronic
26:04means we don't have a pillar procedure
26:06it's gonna take place over a long period
26:09of time and people are increasingly
26:11mobile so their health data portability
26:14problem is exacerbated as we go forward
26:17unless we address it now there's been
26:20really some classic examples recently
26:21where the beauty of having time series
26:24data is that you're comparing you
26:25against your previous self unless you
26:26have that time series data all you can
26:28do is compare you against the population
26:29and people are just so different and
26:32such high variances and overlapping
26:33distribution the well-known example
26:35recently is and still are had prostate
26:37cancer but as PSA level actually never
26:39went high from a popular
26:41standpoint it just went high from his
26:43own baseline uh-huh that's actually a
26:44much stronger signal much much stronger
26:46signal and so that's why his doctors
26:48actually could tell he had prostate
26:50cancer even though if he just didn't
26:51want off there'd be no reason to think
26:52that right it's actually a lot like new
26:54moms and the pediatricians always
26:56telling them don't worry about the
26:57normed curve track the kids curve
26:58because you just care about them growing
27:00and gaining weight but the otherwise a
27:01new moms lose sleep when their kid is
27:03like in the 25th percentile and so the
27:0575 percentile it goes with that same you
27:07just watch the deltas right exactly the
27:09other thing that people can use their
27:12longitude in the health history for is
27:14to participate in clinical trials most
27:17clinical trials upwards of 90% or worse
27:20don't get filled and it's because
27:22there's no frictionless way for a
27:25clinical trials inclusion exclusion
27:26criteria to detect a candidate patient
27:30but if you have on the one side a
27:33digital health summary in cyberspace
27:36on the other side you have a inclusion
27:38exclusion let's say match made in heaven
27:41and we think we can move the really move
27:43the needle for both pharma who wants to
27:45detect patients and patients who want to
27:48be matched up with targeted clinical
27:50trials yeah anything about the current
27:52state of coming up with new therapeutics
27:54new drugs cost of clinical trials is
27:56really high and too many things fail and
27:59there's different reasons for failing
28:00one reason for failing is actually not
28:02getting the right patient cohort and
28:03designing the trial to run such that you
28:06would have a successful outcome at the
28:07end and that's really a data problem and
28:09the opportunity is actually with that is
28:11that more drugs could get through and
28:13even certain things in principle could
28:16that would actually radically shape how
28:18these therapies get to market I'm gonna
28:19go even more radical and say that when
28:21patients all have access to their data
28:24they could form a coalition and ask for
28:27a certain clinical trial that's
28:28beautiful and actually in a sense
28:29there's a sort of centralized versions
28:32of that with like you know the cystic
28:33fibrosis foundation with vertex and so
28:35on but you're talking about it more like
28:38life yeah something where that you don't
28:41need the army there that you could band
28:42together the other thing with chronic
28:44care is that because it's such a
28:46longitudinal time horizon an
28:50eighty-year-old person if they had their
28:54record there's no single institution
28:56that would keep the record on file for
28:58that long so it turns out that patient
29:00portals and api's give you a smaller
29:03window than in eighty year old history
29:06some of the information available is
29:08aged out over a couple of years so you
29:12can't rely on a single interface to keep
29:14all your data one because you're getting
29:16treatments every other place and two
29:18because they're not responsible for
29:21keeping it and so it becomes incumbent
29:23on the patient to manage at least the
29:25longitudinal aggregation of the if not
29:28the interpretation of the data so what
29:29I'm hearing overall is a theme here is
29:31that when you get the horizontal data
29:33that connects all the players in the
29:35healthcare system that can now
29:36communicate to each other through the
29:37patient at the center and then there's a
29:39vertical piece which is the history of
29:41the patient and the past moving forward
29:43etc now this gets us to the idea of big
29:46data because now you have a lot of data
29:47to work with we've been talk about the
29:49canary in the coal mine we've been
29:50talking about all this stuff that people
29:52can do on top of this data honestly it's
29:55a buzzword I hear all the time like big
29:56data in health care what's the big
29:58picture here on that front so here's the
30:00dirty little secret about big data and
30:03there isn't any and the salt a big data
30:07in health care is small data that means
30:09what Suzanna said earlier the small data
30:12is at the edges that's where patients
30:14live so there's a moral ethical and
30:16technical imperative to work with and
30:19through the patient and really for the
30:21patient if you took the top few vendors
30:25for electronic health records and you
30:27reduce the entire US population down to
30:30those vendors those vendors in aggregate
30:33will hold about 6% of all data being
30:37generated digitally on you today where's
30:39the other 94 percent it's in the imaging
30:42systems of course images are large so
30:44there's a lot more data it's in the
30:46microbiome it's in the genome whether
30:48it's the full genome or just you know a
30:50portion of it it turns out that the
30:53electronic health record systems weren't
30:56built to handle all of this other
30:58explosion of data and even if there's a
31:00consolidation of vendors in the
31:02marketplace which we predict there will
31:03be and I think that's going to be a good
31:05thing the consolidation is going to
31:07outstripped by the runaway fragmentation
31:12in digital data the only person who is
31:14legally ethically morally incentive to
31:17pull it all together is the patient when
31:20we think about the patient at the centre
31:22of the circle and how all the big
31:24players are sharing data around the
31:26circle and for them the most useful
31:29share might be to a community of fellow
31:32patients it might be to create a
31:35longitudinal record that they share with
31:38other people who have lupus other people
31:42who have cystic fibrosis yes this
31:44actually reminds me of a website name
31:45that I used to be obsessed with patients
31:49because it's essentially like the long
31:50tail of the internet to find like-minded
31:52people suffering with whether it's a
31:54dysfunctional uterine bleeding which is
31:55a weird category or you know some
31:58titrate there's like a million things
31:59that you don't know eustachian tubes
32:00collapse there's a million specific
32:02things I love that aspect of people
32:04being able to send her and create
32:06community around in these and patients
32:08like me is a great example it's like a
32:10time machine that you can travel
32:11backward and forward in your own record
32:13and in other people's record and what's
32:15essential is that people are creating
32:18this small data for themselves and what
32:22is the opportunity is to create an
32:27version yes I agree that makes me think
32:30about the Google doctor problem
32:31the googling problem where patients
32:33think they're doctors because they're
32:35informing each other and googling things
32:37and it actually creates more problems
32:38for a lot of doctors in that space when
32:41you say industrial production side it to
32:43me means taking that data and putting it
32:45in a more rigorous system than one
32:48that's just so informal folksy I
32:50wouldn't say that patients like me is
32:51folksonomy I would say that they are a
32:54serious taxonomy of people's own
32:57tracking and when I think about
32:59industrial-strength health data I think
33:02about the data that is currently being
33:04held by the clinical system that most
33:06people don't have access to and when
33:09people get access to that where are they
33:12going to direct it and the choices that
33:14they have for many people the joke is
33:16they go to doctor Google what does 300
33:18you know cholesterol I mean you know and
33:21I think the opportunity here's something
33:22much grander if you finally have the
33:24data in a portable way you can actually
33:27just ship this off and then get
33:29information from a real doctor who now
33:31has everything all in one place I'm
33:35really passionate about looking at the
33:38expertise that patients really can have
33:40so there are expert patients out there
33:42who really know their disease they
33:45really know their condition and can
33:48bring that expertise and what I want to
33:50see is that everybody operates at the
33:52top of their license and by the way I
33:53think patients can operate at a pretty
33:55high level and so yes there's a danger
33:58of dr. Google yes there's a danger of
34:00amateur pathology there's a lot of
34:03amateur dermatology but there's also the
34:07possibility of everybody being educated
34:09everybody raising their game in
34:11healthcare including patients and
34:13caregivers what I would like to see is
34:15an ecosystem flourish around the
34:18possibility of access to industrial
34:21strength and health data so that we can
34:23see we have no idea what's going to
34:26happen we have no idea what kind of
34:28engagement can happen around health data
34:30because we don't yet have access to it
34:32again we're leaving half the team on the
34:34bench by not giving patients access to
34:37their own data much less access to each
34:40other which i think is really going to
34:42unleash well-being I love this idea
34:45because there's going to be two buckets
34:47there's going to be the very acute
34:49highly clinical bucket of information
34:52and you want a licensed professional to
34:54weigh in and a quarterback but I as a
34:56migraine sufferer I have shared a recipe
34:59I had with whoever will listen to me and
35:01in the moment I do anything for
35:02allergies and I'd love to put that out
35:05yeah me too because it's a solve for me
35:07and I have the protocol that I follow
35:10and it's a chronic condition and the
35:13clinical establishment to said you don't
35:16go home you're fine but I still haven't
35:18solved my migraine so that's where
35:20peer-to-peer health care which Suzanne
35:22has been promulgating for a long time it
35:25really that's the outer circle that we
35:27don't know how that will be shaped
35:29that's an ecosystem play that we're very
35:31excited to help power I'm actually
35:32hearing all three of you say the same
35:34in different ways which is it's about
35:36empowerment and if you put the P in
35:38power and that should be the P and HIPAA
35:40actually not just affordability and
35:41patience but that's what we're talking
35:43about is empowering the patient thank
35:45you guys for joining a system Z podcast
35:47yeah thank you thank you
