May 20, 2025 • 52 mins
Imagine passing your health data down like a family heirloom.
Not just a list of conditions, but a rich personal history — something that could help your children and grandchildren live longer, healthier lives.
But that future depends on what we do now. Right now, governments are pouring billions into electronic health records. But if the data inside them is siloed, inaccessible, or locked in outdated formats — what are we really building?
It’s a bit like building a library, but locking all the books away.
In this episode of Problems Worth Solving, we speak with Professor Rachel Dunscombe, one of the UK’s most influential digital health leaders, about how we can make health data work — for patients, for clinicians, and for the future.
Rachel is CEO at OpenEHR and formerly served as CEO of the NHS Digital Academy and Director of Digital/CIO for Salford/NCA Group. She’s advised the Secretary of State for Health, sat on the UK AI Council, and holds a visiting professorship at Imperial College London — bringing together frontline insight, academic rigour, and strategic vision.
Problems Worth Solving is brought to you by Healthia, the collaborative service design consultancy for health, care and public services.
Find out more about our work at healthia.services.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_00 (00:00):
Imagine passing down your health data as a family
heirloom.
Not just a record of illness,but a rich personal history that
could help your children andgrandchildren live longer,
healthier lives.
But right now, the way we treathealth data is messy.
Governments across the world areinvesting billions in electronic
(00:20):
patient records and digitalinfrastructure.
But if the data itself isinaccessible, siloed or ignored,
what are we really building?
It's a bit like building alibrary and locking all the
books away.
(00:40):
Hello, this is Problems WorthSolving, the podcast where we
meet people transforming healthand care through human-centred
design and digital innovation.
I'm Sam Mentor, founder andmanaging director at Healthier,
the Collaborative Service DesignConsultancy.
If you enjoy listening, you cansubscribe to this podcast and
the accompanying newsletter athealthier.services.
(01:05):
In today's episode, I'm joinedby Professor Rachel Dunscombe,
one of the most influentialfigures in digital health today.
Rachel is Chief ExecutiveOfficer at Open Air
International, also known asOpen EHR, and was previously
Chief Executive of the NHSDigital Academy.
Thank you very much.
(01:45):
why your health data mightoutlive you and why that
matters, how decades of NHS datacould unlock life-saving
insights if we get serious aboutaccess and standards and what it
really takes to build a digitalhealth system that's fit for the
next 100 years.
Rachel, it's a pleasure to haveyou here.
Thank you so much for taking thetime to come and speak with me.
SPEAKER_01 (02:06):
Thank you for inviting me.
It's a privilege to be here.
SPEAKER_00 (02:10):
Before we get into the electronics, I'd like to
start with a bit about yourstory.
I'm wondering what you werecurious about as a young person
and just a potted history of howthat evolved into the work
you're doing today.
SPEAKER_01 (02:21):
I've always been a curious person and I've always
been a person that has wanted toinvestigate things.
So it's not just in this spacethat I feel this way.
I'm curious about things to thepoint where my kids bought me a
Geiger counter for Mother's Daybecause I'd got really
interested in nuclear decay andvarious other things.
And I think that came from mychildhood when my dad, he was
(02:45):
interested in so many differentthings in aviation.
in science, in psychology, ineconomics.
You've been at LSE.
And I was just taught that youshould poke the box and find out
how things work.
And I felt very grateful forbeing given that kind of
upbringing where it allowed meto do that.
Yeah, it's probably a familything, I think, because I see
(03:06):
that in a lot of my familymembers.
SPEAKER_00 (03:08):
It's a bit early in the interview to go off at a
tangent, but you've got a Geigercounter.
Tell me more about that.
I don't know anyone else that'sgot a Geiger counter.
SPEAKER_01 (03:16):
One is broken.
The joke was the cat actuallybroke one of my Geiger counters
by knocking off the sort ofradiator cover.
So I now have one that actuallytells me which isotope is
decaying, which I findfascinating, really fascinating.
But I'll tell you the funnieststory of my Geiger counter.
It was in my bag and we weredriving along the M62, passed a
van with a radiological symbolon it.
(03:38):
And this thing goes off thecharts to the point where I'm
thinking, is that poor driverokay?
So we stay parallel on the M62.
We take the radiation readings,take the name of the company and
contact them and say, I don'tthink your shielding is quite
up.
And it was actually a medicalisotope that was in that van.
But anyway, that's just one ofthe rabbit holes I like going
(03:59):
down because the kids learn aswell about radiation and about
nuclear science.
And if I had another life, Iwould love to go into nuclear
physics or volcanology, I think.
I'd love to be a volcano expert.
I've got a friend that's got aPhD on the volcanic activity on
Mars.
So, yeah.
SPEAKER_00 (04:17):
So, slight tangent.
Bring it back to your patientrecords.
Sorry, tangent at the beginning,yes.
Your curiosity led to youstudying.
Do you have a medicalbackground?
Did you study
SPEAKER_01 (04:27):
medicine?
No, so I was in biomedicalscience originally, but then I
got on the...
It was back in the 90s, thePython train of data.
So I got really engrossed.
Shortly before Linux came out,we were using BSD and things
like that, and I got reallyinterested in the data.
And that's what kind of led tothat overlap of the kind of
(04:47):
scientific background, but alsothe programmatical data
background as well.
I was really interested in whatdata tells us, what we can learn
from it, and starting to seethat beginning of the internet
as well.
We were getting collaborationswith data from different places
and it just felt I could see thefuture emerging at that point.
SPEAKER_00 (05:10):
And how did that evolve into working at the NHS?
SPEAKER_01 (05:13):
So I actually went and I spent a number of years in
Europe working with actuarieslooking at risk with data based
on human health care group riskand that sort of thing.
That was interesting.
And I actually had a personalexperience and a family
experience with the NHS where Ifelt it needed to be improved.
(05:34):
And it was one of those momentsactually where I found a guy
that had worked for me muchearlier in my career who had
moved into the NHS.
I actually met him in a hospitalwhere we were patients.
And he said, yeah, I work in theIT function here.
It's really interesting.
I said, tell me more about it.
And I just had this processwhere I felt it was something
that with the risk piece I'ddone with actuarial, there must
(05:58):
be something we could do inhealthcare that was similar.
I didn't really realise that Iwould be landing in a space that
would be perhaps a long waybehind that, which is the
reality of what I'd landed in.
SPEAKER_00 (06:10):
And you went on to be chief information officer.
SPEAKER_01 (06:13):
Yeah, I did.
SPEAKER_00 (06:15):
Hands-on responsibility for a lot of data
in that role.
SPEAKER_01 (06:19):
Huge amounts of data.
So I have more than one role asa CIO.
And we're talking about hundredsof systems.
We're talking about very complexenterprises here.
Thinking about Salford NCAGroup, you know, nearly 20,000
employees, one and a billionturnover in terms of pounds.
So huge amounts of activitybecause, you know, that size
(06:42):
corresponds to the activity.
They're very complex,mind-blowing environments.
And I think healthcare isprobably the most complex
environment you can work in.
Because there is so much data,but there are also so many
humans and human stories andhuman aspects around that data
as well that need to beconsidered.
It's not, say, aviation, wherean aircraft creates data.
(07:05):
There's also a person behindthis that has thoughts, needs,
values, everything else.
So for me, it was incrediblychallenging, rewarding, but also
in many aspects behind the timesof The Art of the Possible.
SPEAKER_00 (07:21):
How much is that complexity part of the
attraction of working here?
SPEAKER_01 (07:26):
The complexity is both an attraction and a
headache, if you like, becausesometimes it just feels that
there are things you can'tsolve.
It takes time.
The attraction is also humans,as in there are a few other
places where you can improvelives and save lives.
Somebody once said that a dataanalyst or a programmer in
(07:48):
healthcare could...
cause huge amounts of harm tolife or could save a huge amount
of life.
And that is absolutely true.
As a vocation, this is asimportant as doctors and nurses.
SPEAKER_00 (08:06):
Let's go on to talking a bit about electronic
health records.
So I'm hoping we have listenerswho are outside the health
system to Problems WorthSolving.
So I wondered if you couldexplain a bit about EHRs or
electronic patient records,which is another name for them,
what they are, where they'vecome from, what they do and how
do they benefit patients andstaff?
SPEAKER_01 (08:28):
Absolutely.
Electronic patient records,EHRs, started around about 30 or
40 years ago.
And they started really assystems where you could record
things like lab results or putin orders or make notes about a
nursing encounter.
And they have grown in terms ofwhat they do.
(08:51):
They go right the way throughnow to every function you can
think of, billing, managingbloods, managing labs and
pharmacy and everything else.
But the evolution has reallybeen based on the functionality.
So it's all been about how do weadd some more functionality to
(09:11):
this product?
And health systems have reallycompared that functionality,
front-end user functionality, orwhat it can do in terms of
interacting with pharmacy.
And as a result of the fact thatthe market is crept in this way,
you have all these differentvendors who have evolved their
front-end functionality to besomewhat equivalent, but have
(09:35):
bolted things onto this backend, which maybe got its origins
in the 80s or 90s and was neverdesigned to do all of these
things and all of thesefunctions.
You know, some countries they'vebolted on billing.
In the UK, we've bolted onactivity and patient
administration.
We've bolted on bed management.
And as a result, these thingsare architected based on older
(09:59):
technologies and the way thatthey've evolved them has been to
basically just keep the markethappy.
And that has built really atechnical debt into the backend.
So as a CIO, trying to get yourdata out and make sense of it
from one of these things is notthe easiest task.
SPEAKER_00 (10:18):
What sort of data would you be trying to get out?
SPEAKER_01 (10:21):
I'll give you a really simple example.
Say you're doing a study ofpatients that have heart failure
and you just want to pull outtheir blood pressure across
time.
Do a time series, something likethat.
What you find within these EMRsis the blood pressure is stored
in all sorts of different placesin different formats with fields
(10:41):
that don't match.
We call that lossy data becauseyou're losing some of the
context.
And to actually get a timeseries out for my data
scientists, it would have beenat Salford.
You have to do all of thesemappings, which don't fully map.
And the way each of those bitsof functionality that record
blood pressure, one might be adevice here, one might be
nursing here, one might be award around here.
(11:02):
I remember doing an audit of oneEMR that had 47 different ways
of just recording bloodpressure, right?
There's a hell of a lot of work.
And the lossiness and thecontext loss, the provenance
that data is lost because you'vegot all of these different
places where the same concept isstored.
And as soon as you get into thisworld of data science and AI,
(11:22):
you realize the underlying datamodels and the underlying data
are just not built for the nextgeneration of solutions that
we're going to need or startingto need now.
We lovingly used to call some ofthe people that did the work on
the data quality to bring it allin line.
So manually manipulating things,the cardigans, because they
(11:42):
would come in, I'm wearing acardigan today so I can laugh
about this, hang the cardigan orthe jacket on the back of the
chair and do a whole day's workon data quality.
And that really, prevention isbetter than cure.
But we're curing the fact thatthese systems have data in such
(12:02):
disparate ways, what we shouldreally be doing is preventing
that by having the right backend data infrastructure.
So yeah, it's been an organicthing and it also hasn't helped
that the NHS has always reallybeen money conscious to the
point where it's always said buythe cheapest system and that has
never considered how good thedata is in the back end.
SPEAKER_00 (12:25):
So as I mentioned earlier, the government has been
funding huge investment intothese EHRs over the past few
years.
Where does all that money go andwhat's the real return on
investment for the system?
SPEAKER_01 (12:36):
So to be clear, you know, these systems do fulfill
the needs of a hospitaltypically, and they allow a
hospital to function.
So the money goes into providingthat incredibly complex
functionality that allows ahospital to operate, along with
a very big infrastructure thatenables it.
(12:56):
What I would say is many ofthese systems are actually US
systems that are reprovisionedfor the UK.
So we're also buying a hugeamount of billing and things
that we don't use in the UK.
And that's always been a bit ofa question for me.
But if you think about everyprocess in a hospital moving
from the analog or from oldsystems through to a single
unified system, that's sort ofwhat the EHRs are.
(13:20):
The nearest analogy, if you'renot from the healthcare
industry, is the ERP and whatthat does for a business in
terms of automating one of theprocesses.
The thing is that these are verycostly, not just because of the
software, because the softwareis costly.
It is training every singlemember of staff.
If you take my old organization,18,000 people to train, and it's
(13:41):
a good few hours training each,and they need updates on that
training.
It's a very big program.
And obviously, you have to trainthem well, because this is about
clinical risk.
It's about clinical safety.
It's about all of those things.
So really, these very bigprojects...
are very big bits of softwarewhich have to be taken in one
(14:02):
lump because they're monolithicsystems, yeah?
You have to take this bigmonolithic system and you have
to implement it.
And I think that's where thesethings get really difficult
because they're massive projectsand programs and it's really,
you have to leap into it.
There's no kind of staging, youknow, your sort of deployment of
these things.
SPEAKER_00 (14:21):
Are we building these systems in the right way?
SPEAKER_01 (14:23):
My view increasingly is is no.
I've done some work actuallywith Jordi Piera in Catalonia to
look at the generations of EHR.
And the first generation waswhere we just built or bought
lots of systems and tried tostrap them together.
That was the 80s, 90s.
So that was best of breed.
(14:43):
You'd buy lots of things and tryand put them together.
Second generation is this bigEMR where you have to leap into
it and do everything.
But now what we know is thatdata is going to be used for so
many different purposes.
If you're born with a chestproblem and you're put into
NICU, high dependency forbabies, that data is going to
(15:05):
bear out for your chest healthfor the rest of your life.
And that data has to beaccessible for decades in a
meaningful format.
And what we're learning now isis that we need a data-centric
architecture because your datais going to be used in
algorithms to help restratifyyou.
It's going to be used to helpmake decisions about you in
(15:26):
different institutions, indeedinternationally, you know, with
the international patientsummary.
It may travel with you so thatyou can be treated abroad when
you need it.
And so this data-centricarchitecture would say you
separate your data fromapplication because one thing I
can bet is that in 10, 15, 20years time, how doctors, nurses,
(15:50):
and those in healthcare systemsoperate will not be via the
keyboard and typing.
It will be via voice, gesture,who knows what, right?
It will be things we haven'teven imagined now.
If you record a really good dateset now, That will last through
that journey.
But what we're doing instead isrecording the data driven by the
(16:13):
interface.
The interface says this, sowe're going to record this data.
What we need to do is, no, weneed to look at human, what we
need to record, and actuallykeep that in a format that is
consistent across time.
And that data-centricarchitecture and separation of
data from application you see inother industries, but it
absolutely makes sense because Ibelieve somebody should be able
(16:36):
to carry their digital twin withthem, which is all of their data
about them.
And you've got a human rightreally to your digital twin
because you can learn aboutyourself from your data.
You can take it with you.
So yeah, so the future is reallyabout the citizen having access
Every actor that they need tohave access in their healthcare
(16:57):
journey should have access.
But it shouldn't preclude peoplefrom actually providing care
because they don't understandthe data.
SPEAKER_00 (17:05):
And this is what you mean by thinking of data as the
foundation, not a byproduct ofthese systems.
SPEAKER_01 (17:11):
Absolutely.
Data is not a capital asset,yeah?
It's something that will havevalue over decades or even 100
years of somebody's life orbeyond.
If you can imagine in thefuture...
100 years of data in your life,you may actually give that to
your family for informing theirhealth care.
I certainly know in my familythere's breast cancer, there's
(17:33):
bowel cancer, there's otherthings.
If that data was available tome, it would help to stratify my
health care.
So this could become somethingthat is a legacy that's passed
on through families andgenerations.
SPEAKER_00 (17:44):
What would it mean to design systematically around
data rather than trying toretrofit the system?
SPEAKER_01 (17:51):
That's very much the space that I'm in now, is
designing what is needed for ahuman's healthcare record for
their life, yeah?
And designing it in anon-duplicative, semantically
harmonized, as we call it, way,which means we take blood
pressure, as we talked aboutearlier, there is one concept of
(18:12):
blood pressure, not 40, aredifferent concepts.
And In order for that to be usedin different places, we use
something called templates,which are a layer that allow it
to be used in different places.
So the idea here iswell-engineered data is a
combination of clinicians andthose that work in the technical
space working together tooptimally work out what data is
(18:36):
needed for a human in the mostefficient way and the most
representative way for it to beused for all of the use cases
today and tomorrow.
SPEAKER_00 (18:45):
So you're the Chief Exec at OpenEHR.
Can you tell me a bit about theorganisation, where it's come
from, and how does it fit intothis landscape and the work
you're doing?
SPEAKER_01 (18:58):
You know, I've not been on the journey with OpenEHR
since its beginning.
It's nearly 24 years that it'sbeen running.
And it started actually as acollaborative between both
clinicians and technologists atUCL in London.
with the input of some of thoseclinicians being from Australia
as well.
So it's an Anglo-Australianinitiative and it's incredibly
(19:22):
powerful.
It took me a while to get thisconcept, I've got to say,
because certainly on my journey,I've not always appreciated how
important the data is.
So 24 years ago, they had thiskind of epiphany and light bulb
moment about data and they setup open air and Professor David
(19:44):
Ingram was pivotal in thishappening at UCL.
And that community has grown andgrown as a not-for-profit
community of interest thatactually defines and develops
these standards.
Around about half or just overhalf of the community are
clinicians.
The rest are technologists andleaders.
(20:05):
And it has a governance whereit's almost a Delphi, if you
like, of people who go throughthese processes and work out
what the optimal dataengineering is for any aspect of
human healthcare.
And as that grows, thatmodeling, as we call it, data
modeling and clinical modelingcontinues so that we can
actually grow the sort of dataarchitecture that is enabling
(20:29):
human healthcare.
SPEAKER_00 (20:31):
So this was what attracted me to invite you onto
the podcast today because it'sall about problems worth solving
I wonder how you how would yousummarize the problem that
you're trying to solve as anorganization
SPEAKER_01 (20:42):
I will summarize this from my experience I think
others will have different viewsbut I was involved in data saves
lives in Manchester and we sawthat as we improved data quality
and access lives were beingsaved and improved it was quite
a scorecard that literally had ahuge massive impact on me And
(21:03):
that made me realize that as wehave a single source of truth
for human life, and as we makethat available as appropriate to
different endpoints, you canhave really big impacts on lives
saved and lives improved andquality of life.
And what I see in the communityis a set of people who have seen
(21:25):
that too.
They've seen that we need toengineer data better and make it
available to particularly to thecitizen and to the healthcare
settings that may not be theprimary healthcare setting to
the citizen, because it is aquality and safety issue.
One thing that I came to, andI'd like to do some research on
(21:47):
this, I haven't had theopportunity yet, is there is a
link between unwarrantedvariation in data and
unwarranted variation in care.
And we know that unwarrantedvariation in care causes harm.
And so this is all about harm,quality and safety.
It's a set of people that cansee an ability to do much better
in health and care.
SPEAKER_00 (22:08):
Imagine you're immensely successful, which of
course you're going to be in thework you're doing.
What does that success look likefor the system?
SPEAKER_01 (22:18):
I think the success for the system for us looks like
every citizen, hopefullyglobally, let's say we're really
successful, let's say globally,every citizen having their data
available to them and the careproviders that they want to
treat them, and those doingresearch, you know, all of the
(22:39):
actors in the system, when theywant it, where they want it, in
a format where you can get speedto value by putting AI,
intelligence, apps on top of it.
So it is really about removingthe friction of data today to
enable a much better healthcaresystem of tomorrow.
And also one personal thing forme, we were talking about EMRs.
(23:00):
EMRs reinforce the hospitalwalls, which pushes more and
more activity into the hospital.
And so we do more in hospitalsnow.
I would really like to seeOpenAir being part of the
solution to allowing a recordthat is available anywhere,
which means we can deliver carein the home via apps, devices,
and citizens can self-care.
(23:21):
So there's this aspect of stopreinforcing the hospital walls
and start allowing care to bedelivered in the most logical
place.
SPEAKER_00 (23:29):
What's the implication of that for
innovation and kind of the widersystem changes?
SPEAKER_01 (23:33):
So if you think about that, There are lots of
things that play against usactually innovating.
And by innovating, I'm going totake that as meaning you models
for operating healthcaresystems.
Yeah.
So if you want to startoperating healthcare systems
where you have distributedservices, you have self-care,
(23:55):
self-management, you havemanagement by exception from
medical devices, you have appsand wearables that support a
citizen to self-care and preventall of that.
requires a single source oftruth about the patient.
And it requires that data can bemanaged across the entire set of
(24:16):
organizations in the experience.
The average patient back inManchester would have been known
to five organizations.
That's just an example.
So for me, for the future, thisubiquitous pathway allows those
innovations because there's notany of the frictions of
stovepipes, data silos, if yougive the citizen the ability to
(24:39):
control their record asappropriate as well, you can
allow things like citizenscience and innovation to happen
in that way, which is a newavenue.
I think that's reallyinteresting, allowing citizens
to participate in trials andstudies and contribute.
I'm seeing some good example ofthat.
And for me, just taking away allof the friction of actually
(25:01):
wrangling this data into theright shape, having it lossy,
not allowing certain actorsaccess to it, we could really
speed up the innovation.
I was just with Lord Darziyesterday and he said the answer
to the next generation of theNHS, the NHS's future, is all in
the data.
SPEAKER_00 (25:25):
There's lots of hype around AI in health at the
moment.
I'm curious about what you thinkwe're skipping over when we go
straight to the shiny stuff.
SPEAKER_01 (25:34):
So coming back to the Paris AI Summit back in
February, they said in allsectors, it's about great data
and it's about good governance.
I couldn't agree more.
I'm afraid it's a lot of reallyhard work to get our data right
and to work out how we govern inthe age of AI.
And that's fairly boring stuff.
(25:56):
So risk frameworks in AIgovernance is something I've got
deep into.
It's not going to grab anyheadlines, is it?
But it's the work that needs tobe done.
I would also say that anotherthing that we're missing,
jumping straight to the shiny,is the platform approach to AI.
We're looking at hospitals thathave got tens of different AI
(26:17):
algorithms.
In the US, it may be a lot more.
But for each of thosealgorithms, you have to monitor
it, you have to audit it, youhave to understand how it's
using data, you have to look forbias, you have to look for cases
where it's not been effective.
And if you're going to havehumans do that, that is not
scalable.
So these oversight platforms,which we're starting to see in
(26:40):
the marketplace now, I think arean essential component because
they will help us to govern andensure the safety of these
algorithms.
What we can't do is allowhundreds of algorithms to be out
there without this kind ofoversight platform and
management platform.
And I think that kind of middleground marketplace, if you
(27:01):
almost like, where you canactually monitor those
algorithms and their performanceis going to be an essential part
of the future.
SPEAKER_00 (27:10):
And that would involve having a human in the
loop.
SPEAKER_01 (27:14):
Absolutely.
Human in the loop, but to manageexceptions.
At the moment, we just don't getthat data.
It has to be a human go in andlook at the performance of the
AI and check it on a regularbasis.
Humans in the loop always,right?
The reason you need humans inthe loop are multiple.
One, you need that humanjudgment there.
Two, There will be exceptionsthat are valid.
(27:36):
If you look at certain hospitalsor health systems that deal with
the most problematic cases, youwill see things that are
spurious in terms of outcomesand results there.
It needs that humaninterpretation to actually
understand why things areexceptions.
SPEAKER_00 (27:54):
What do you feel the role of data is in population
level insights and personalisedcare?
SPEAKER_01 (28:00):
I'm a huge fan of social determinants of health.
And the biggest factors thatactually determine your health
over time are your education andwork status over a population.
For me, there is an awful lotmore that we can do to actually
(28:23):
find out how we can help ourpopulations stay well and
intervene and prevent disease.
ill health before it happens.
And a lot of that does reside innon-healthcare data.
And a lot of the questionsaround that are the ethics.
So I certainly know that in theUS, certain health systems have
(28:44):
got agreement to access thefinancial history of their
patients.
They've consented that.
That's going to take a lot ofthinking if we ever wanted to do
that in the UK.
Is that something we would do?
Maybe not.
But there is a whole domain herethat we need to explore in terms
of pulling together that data.
(29:07):
But some, again, coming to theUS, I've just said that, we may
or may not want to do it.
But in terms of actually usingsome of those tools, social
determinants of health and thewider data, I've seen some
excellent examples where they'vetaken people that haven't turned
up for their breast cancerscreening, as an example, for
five years and managed to get65% of them to do that within
(29:29):
six months.
SPEAKER_00 (29:31):
How did they do
SPEAKER_01 (29:32):
that?
They analysed all of this data,found out their motivation,
found out demographics of thesepeople and put nudges out that
were tamered to them, that wouldappeal to them, and would speak
to them about why breast cancerscreening was important.
And that just blew me away.
Really blew me away.
Yeah.
SPEAKER_00 (29:51):
And was that driven by AI?
SPEAKER_01 (29:52):
Partly by AI, partly by the great data, partly by
psychologists, and partly bysome great public health
clinicians, some of whom arefrom the UK.
SPEAKER_00 (30:03):
Is it a false split to say there's health data and
there's social data involved?
Actually, you want to look atthe whole picture.
Yes.
SPEAKER_01 (30:09):
And it's beyond just social data.
It may be lifestyle data.
It may be one thing I'm verykeen on is what is your
preferences as well as a human?
We should be asking that.
That should be put in.
SPEAKER_00 (30:20):
And that surely leads to the responsibility for
ownership of that data needs tobe with the individual.
Yes.
SPEAKER_01 (30:26):
Were possible, it does need to be with the
individual.
And we've seen lovely exampleslike One London at the moment
where citizens can nowcontribute via the NHS app to
their care plans.
We need this to be a partnershipwith our citizens.
They need to be able tocontribute to their record.
And I think that the true futureof healthcare is a partnership
(30:50):
between the NHS and ourcitizens.
Obviously, there'll be somepeople like my late father who'd
had two strokes and had lost hisspeech who wouldn't be able to
do that.
We need arrangements for peoplewhere it's not suitable.
But for those that want to andneed to, especially with the
coming generation of people whoexpect digital services, I think
we need to do this incollaboration together.
SPEAKER_00 (31:12):
The data becomes increasingly valuable when it's
being used in all these otherdifferent ways.
Who should be responsible forgoverning that data?
SPEAKER_01 (31:22):
That's a very good question.
And that's a very interestingquestion when the citizen is
contributing, you know, yourlifestyle, your garment, your
Apple Watch, whatever else.
It becomes a spectrum of datathat is co-owned.
I think for me, anything that iscreated in the clinical domain
(31:45):
needs to be owned by the healthsystem, but also organized.
a copy of that owned by thecitizen.
It's a dual responsibility.
The reason is that the healthsystem needs to own it to make
sure it is safe, it's secure,it's audited, all of those
things.
It has a responsibility to keepthat.
(32:05):
But then there needs to be asynchronized copy in the health
record.
In terms of lifestyle data andpotentially some aspects of
social determinants of health, Ithink there really is an
ownership by the citizen.
If we come to financial data,There are certain patients who
may choose to share that withtheir healthcare system because
(32:27):
they believe it will improvetheir outcomes or it will
improve the population'soutcomes.
I think those sorts of pieces ofdata should be down to the
citizen to choose to share.
And really what I see as alonger-term goal is the ability
for a citizen to go in and say,do you want to share this data?
This is how it will be used.
(32:48):
This is what we intend to dowith the data.
And then a bit like we have withblood products now, if you
donate blood in the UK, you'vegot a lovely text message and
say it's been used in apaediatric hospital or whatever.
You know that you've given back.
I think there should be sort ofpostcards back to the citizen to
say, these are the insightswe've found using your data, or
(33:09):
this is the research we've doneand this has been the outcomes.
And that kind of datadonatorship bit, as you said,
data's got value.
And if you're giving back toothers or you're helping others
with that data, we need to closethat loop.
SPEAKER_00 (33:29):
What do you think health can learn from other
sectors like banking aroundtrust, access and usability?
SPEAKER_01 (33:35):
Oh, trust, access and usability.
I think let's take access.
I think with banking, accessis...
ahead of most industries.
I've worked with BarclaysDigital Eagles and the way in
which they are trying to makethings more accessible to
different groups, olderpopulations, populations that
(33:59):
may not have been digitallyexcluded.
I think there is a huge amountwe can still learn from
financial services in terms ofthat.
Although one thing I would sayis that We're still not seeing
banking in multiple languages,which we need for healthcare.
That is one consideration thatwe have to make and one aim I
know that health systems havegot, which is to make your
(34:20):
information available in yourlanguage of choice eventually,
which is really important.
In terms of trust, I think theFinancial Conduct Authority has
done a very good job of ensuringtrust within the financial
sector.
And I once jokingly said, I'dlove to see the NHS as kind of
(34:42):
having a Bank of England-likefunction that was independent
and could assure things and makedecisions about things.
But I think we need a verysimilar construct.
Some of the things we have seenare what are called data trusts,
and they are senses of thirdsector academia, patient groups.
assuring the use of data onbehalf of citizens.
(35:06):
So if you think about acooperative or a data trust,
might assure the use of yourdata in a geography and kind of
be your appointed party to dothat.
I think we need this kind ofthird entity that helps build
trust because there are so manyactors in the system And while
it is the NHS's job to provethat data is used correctly, I
(35:27):
don't think they can operatethese sorts of groups that cohes
everyone together and helpassure how the wider system is
using data.
And I think that's a constructthat's needed.
Be it a FCA type piece, be itdata trust, there's something
that's needed in the middle.
SPEAKER_00 (35:42):
Is that something that you can help with in your
work?
SPEAKER_01 (35:45):
I'm actually working with one of the leaders of...
one of the data trusts at themoment?
Yes, we absolutely can.
And I think...
The nice thing about open airand everything else is that
being a standard that is usedglobally, it's a lot easier to
put transparency toolings on topto see how it's been used and
(36:05):
what's been done with that data.
Outside of my open air work, I'mhelping chair some research
groups into this, particularlywith patients that have got
chronic long-term conditions,mental health conditions,
looking at how they want theirdata to be used.
So I'm finding, again, the humanaspects of it.
this.
I continue to learn and Icontinue to want to know more
(36:28):
about how we should interactwith citizens and allow them to
really determine theirpreferences for use of data.
SPEAKER_00 (36:37):
What would massive success for open air?
I'm saying it in the right waynow.
I was saying open EHR.
SPEAKER_01 (36:43):
Either way is fine.
SPEAKER_00 (36:44):
I've learned that you're saying open air.
What would massive success ofyour ambitions mean for the big
EHR vendors?
SPEAKER_01 (36:51):
Some of them may well embrace this.
It may be their next generationoffer.
For some of them, they grow atechnical debt or a data debt.
If governments mandate this,then that will be an issue for
them.
My view actually is that thereshould be incentives to help
(37:12):
vendors move towards standardslike this because we have to
recognize that they've done goodservice as EMRs in hospitals and
everything else.
But where we are now, futureforward, there is that
day-to-day in the back end.
And I think certain governmentsmay want to work with those
vendors to incentivize them orsupport them in moving towards
(37:35):
supporting these standards.
Because what you don't want todo is undermine a market that
you have already.
You want to help it evolve withyou.
SPEAKER_00 (37:42):
Do you work quite closely with the vendors in your
work?
UNKNOWN (37:45):
No.
SPEAKER_01 (37:45):
I work with a number of vendors.
Some of the big EMR vendors,yes.
Some of them, no.
So they split into two camps.
Those that feel that this is thefuture and are looking at the
ways that they can start toenable this.
And then some of them that say,no, just buy one big monolithic
system, which I don't think isgoing to be the future.
SPEAKER_00 (38:08):
How do you think we get the balance right between
partnership and dependency?
SPEAKER_01 (38:12):
Yeah.
And I think...
Partnership is key.
Partnership with industry and asymbiosis with industry is
absolutely key.
But dependency, dependency isnot good.
And we have become dependent oncertain vendors because of their
bespoke data and their lock-in.
They become the only people thatcan do certain things of the
(38:32):
data because the way it's beencurated.
Data and dependency, for me, Ithink it's a difficult one
because...
If you don't want to bedependent on vendors for
decades, you don't build in datalocking.
You don't build in data usestandards.
(38:53):
However, what I will say is theflip side of that is that the
health system needs to be theintelligent client and be able
to describe what it wants to buyand procure and how it will
work.
And they also need to recognizecustomers.
As I said earlier, that there isan existing market that you want
to take on a journey with you.
So you have to influence theexisting vendors.
(39:16):
And I think in certaingeographies, we're seeing
intelligent clients emerge inhealthcare.
They're really understandingthis and they're really driving
it forward.
But perhaps in some othergeographies, less so, which
means that they're not reallybuilding for a well-engineered
data future.
they really don't understand theimportance of that asset or how
(39:37):
to build it for their geography.
SPEAKER_00 (39:39):
And what about getting NHS support and then
global support for thesestandards?
SPEAKER_01 (39:45):
So we have a huge amount of global support at the
moment.
You know, increasingly I'm inwith governments from different
countries, working with the AIDHin Australia at the moment, and
they invited me onto theirdigital health advisory board.
Ireland is moving to open airfor its integrated record.
We've got Greece moving in thesame direction.
(40:07):
We've got other countries Ican't mention because the
governments aren't ready to talkabout it, but we have the
Nordics who've traditionallyused open EHR within their
standards.
You've then got countries likeJamaica that are using this and
some of the African countries.
So, you know, Jamaica's datamodel is open air.
This is moving forward, butObviously, with Opener being
(40:29):
very small, not for profit, veryfrugal, this isn't done by
selling anything.
This is just done by talkingthrough the logics of why you
would do it and the successesthat other people are having and
just really getting peopleinvolved in a global community.
It's really amazing that we'vegot as far as we've got and
we're continuing to go onbecause we are a not-for-profit
(40:53):
open source community.
Anyone can take this and use itfor free.
It's a global good, a bit likeLinux.
It's very much an open sourcegood.
And so for me, it is thecommunity that permeates the
future.
It is the community.
It is the word of mouth.
It is an ophthalmologist in onecountry talking to an
ophthalmologist in another.
It's somebody in one ministrytalking to a ministry in another
(41:15):
country and saying, look, we'vedone this work.
We found that a data asset hasgot value for us.
We are going to actuallyarchitect our data well.
I'm going to base it on OpenAirand these other standards for
the future.
And actually, coming back tothis, she talked about the NHS.
There was work done around about2019, I think, that valued the
(41:36):
data of the NHS.
This is not selling the data.
This is not giving the dataaway.
This is the value of being ableto use the data and deliver the
services effectively within theNHS.
That was billions.
It was around about six billiona year, potentially up to nine
billion a year.
Data has value.
(41:57):
And people are realizing thatnow.
SPEAKER_00 (42:04):
It's coming across that you think about data from
the human perspective.
Yes.
I was wondering how much youthink about data as a tool for
social change and how we shiftfrom seeing data as just a side
effect of the activities.
SPEAKER_01 (42:17):
So that's why I mentioned earlier the science
collaboratives with thecitizens, because citizen
science, and I'm thinking aboutone example I've been involved
with where we have doctors, wehave patients, we have data, and
they all collaborate around thedata and the outcomes.
I think that this is a hugeopportunity for social change.
(42:41):
On the other side of it, if welook at...
I've just been reading a book onFacebook and how they use data
globally to get the outcomesthey want from people.
If we use that science of beingable to find out what people
want or need from the data andactually nudge them, coming back
to this, getting an uptick inbreast screening, the things
(43:02):
that have been used to market tous from the internet as a kind
of dark science...
could be turned into a lightscience of helping people and
nudging them and supporting themwith what they need to see,
hear, or act on that'sappropriate for them, yeah?
And this whole piece for me,social change, we have to be
(43:23):
transparent about how we'redoing all of this.
I think absolutely transparent.
But the idea that we canactually allow citizens to drive
more of this, it reallyinterests me actually that in
Finland, they have citizenassemblies that drive their
healthcare system that see dataand decide, make the difficult
(43:43):
decisions about haves andhave-nots with the budget.
But is our future healthcaresystem going to be data-driven
in that way, where you have atransparency of what there is, a
transparency of the problems,transparency of the
opportunities, tools that showyou where you could put your
money and what outcomes youcould get, and you can make
decisions as a set of citizens.
I think there's an awful lotthat could be done.
(44:06):
And I'm also seeing global raredisease groups who are
aggregating their data andtaking action as a citizen group
and approaching life sciencescompanies about finding
different repurposing of drugs,whatever else.
And you think healthcare reallytraditionally going back to the
(44:27):
old NHS had the patient in thechild state.
And the clinicians in the parentstate, and it was a
parent-child, you did what youwere told, you turned up when
you turned up, you were givenwhat you were given.
Medicine of the Future, I think,is about the adult, which is
about an adult relationshipbetween the clinician and the
patient and roles for both ofthem.
SPEAKER_00 (44:49):
What are you most optimistic about right now?
SPEAKER_01 (44:52):
There are a few things I'm optimistic about, and
particularly today, becauseyesterday I was with a cohort of
graduates from the DigitalHealth Leadership Programme.
I am really optimistic that wecan use data, technology, and
human change to create the nextgeneration of the NHS.
(45:12):
I sometimes work with boardleaders and I say, okay, let's
sit down and imagine when youretire 20 odd years time, what
does your healthcare look likein 20 odd years time?
You're living by yourself,potentially, what do you want
around you that helps you to beindependent and feel confident?
And we just describe what thewhat their ideal health system
looks like.
(45:33):
And then we look at how they getthere, right?
And I think people are startingto see what that future health
system looks like and arestarting to see how we get
there.
So I'm feeling quite heartenedthat actually we have the
ingredients.
We've got to wait for this wholechange with NHS England and the
Department of Health andeverything else to play through.
But if there's one thing I wouldsay to the great leaders out
(45:55):
there that I've seen, you willbe part of delivering that
future and we really need you.
SPEAKER_00 (46:00):
What do you think that future health system looks
like?
SPEAKER_01 (46:03):
Again, going down rabbit holes.
I talked about going down rabbitholes.
One that I've been down recentlywith a couple of former
colleagues is this idea of youhave the capacity of the NHS and
you manage the capacity anddemand across the whole system.
And there's no such conceptanymore of referrals into a
(46:25):
pathway.
It's always next best action.
using the whole capacity of asystem.
And I think we can get to apoint where we can leverage
every part of the NHS and knowwhat capacity we need as a
system, know what actions weneed to take, and can bring in
that efficiency across the wholesystem.
(46:46):
So I've got this view that thisintelligence that we're talking
about for the future will reallyallow us to optimize what we
have today?
Because there is an opportunityto optimize, but that will also
allow us to invest in thetechnologies for the future.
One thing's for certain, we willalways need our wonderful NHS
staff because this is about aservice that is really
(47:10):
human-faced, has empathy, canput things into context.
AI is never going to put acancer diagnosis into context
for you as an individual, right?
that human contact, that humanemotion, you need that.
But what it will do is be ableto augment our doctors, our
nurses.
And another thing I've beenlooking at is agentic AI.
(47:31):
So what if you can actually takeall of the knowledge that's out
there, all the procedures withinan organization, and for a nurse
that's actually working on anight shift, he or she is able
to actually ask it questions andget the knowledge of a huge body
of knowledge, but instantly, youknow, via a headset or via
(47:53):
whatever else.
Those sorts of things, I think,will really assist the NHS in
revolutionising.
I don't quite know how it'sgoing to play out, but all I
know is that in that 15, 20 yeartimeframe, we're going to have
to have got there and the toolsare there to get there.
And that I see enough goodleaders working and potential
(48:13):
leaders for that journey to betaken.
SPEAKER_00 (48:16):
And what about more immediately?
So if you could wave a magicwand right now and change one
thing about the way the systemworks, what would you change?
I
SPEAKER_01 (48:23):
think leadership needs to change.
I don't mean take leaders away.
What I mean is how people lead,how people in leadership are
trained.
So for me, just a basic example,if you go to
youraverageboard.com in ahealthcare organization, quite
(48:44):
traditionally you'll have anestates director on there
because that's something thatdelivers healthcare.
You won't have somebody on therethat's a CIO or equivalent
because they're still thinkingin terms of bricks and mortar
for everything.
I think the leadership needs achange of mindset and I have
some hope that the future NHSwill do this.
(49:06):
The basis says we're going todeliver healthcare a new target
operating model of care.
So if there's one thing I couldchange, it would be changing the
leadership's mindset to say, weneed to look at operating model
2.0 for the NHS.
SPEAKER_00 (49:20):
What's next for you, Rachel?
SPEAKER_01 (49:23):
I think more open eye.
Lots and lots to do.
That said, I still do help theNHS.
I still do spend time with theNHS.
I'm very passionate that I willhelp the NHS going forward.
But I think for me, I think Iwill just keep going down rabbit
holes that I find.
I keep finding problems and needto find out more.
(49:43):
And so for me, this is anemerging journey and being
engaged with data, AI,intelligence and the human
aspects of that are somethingthat are going to keep me busy
for the next couple of decadesat least.
SPEAKER_00 (49:59):
I'd like to end on something thought-provoking.
So what's One belief that youhold that others might not agree
with, but you believe is true.
SPEAKER_01 (50:08):
Does this have to be to do with healthcare or not?
SPEAKER_00 (50:13):
No, just anything.
SPEAKER_01 (50:14):
Okay, I'm just going to be slightly out there then.
But again, I've been watchingsome of the hearings from
Congress in the US and itappears that they have disclosed
that there's non-humanintelligence.
SPEAKER_00 (50:32):
That's really interesting.
I've also been following thosehearings and it does seem
incredible.
SPEAKER_01 (50:37):
Absolutely, but hang on.
In Congress, multiple peoplehave said there is non-human
intelligence.
SPEAKER_00 (50:43):
I'm slightly mystified as to why this hasn't
had more coverage.
SPEAKER_01 (50:46):
Right.
SPEAKER_00 (50:47):
It seems like quite a big deal.
SPEAKER_01 (50:48):
And there is emerging evidence from huge
numbers of military people inthe US, that is true.
Now, bear with me, this doesrelay back to healthcare.
Because what does it mean to behuman?
If there are others, what doesit mean to be human?
And what does it mean to behuman?
We also have to consider inhealthcare in terms of how we
(51:09):
want humans to live goingforward with the AI and
increasing automation.
We've got questions about thefuture of work, what happens
when jobs are automated.
I think the two things cometogether for me to say, we
really need to consider what itmeans to be human in this
current environment.
SPEAKER_00 (51:27):
rachel it's been fascinating talking to
SPEAKER_01 (51:29):
you i didn't think i'd be talking about that there
you go
SPEAKER_00 (51:33):
we must talk more but it's been great to talk to
you thanks so much for takingthe time to come and share your
thoughts and i've learned lotsabout the work that you're doing
and open air and hopefully ourpaths will cross at some point
in the future
SPEAKER_01 (51:44):
thanks for the opportunity
SPEAKER_00 (51:49):
you Problems Worth Solving is brought to you by
Healthier, the collaborativeservice design consultancy for
transformation in health careand public services.
Find out more about how we canhelp you deliver user-centered
change at healthier.services.
Imagine passing down your health data as a family
heirloom.
Not just a record of illness,but a rich personal history that
could help your children andgrandchildren live longer,
healthier lives.
But right now, the way we treathealth data is messy.
Governments across the world areinvesting billions in electronic
(00:20):
patient records and digitalinfrastructure.
But if the data itself isinaccessible, siloed or ignored,
what are we really building?
It's a bit like building alibrary and locking all the
books away.
(00:40):
Hello, this is Problems WorthSolving, the podcast where we
meet people transforming healthand care through human-centred
design and digital innovation.
I'm Sam Mentor, founder andmanaging director at Healthier,
the Collaborative Service DesignConsultancy.
If you enjoy listening, you cansubscribe to this podcast and
the accompanying newsletter athealthier.services.
(01:05):
In today's episode, I'm joinedby Professor Rachel Dunscombe,
one of the most influentialfigures in digital health today.
Rachel is Chief ExecutiveOfficer at Open Air
International, also known asOpen EHR, and was previously
Chief Executive of the NHSDigital Academy.
Thank you very much.
(01:45):
why your health data mightoutlive you and why that
matters, how decades of NHS datacould unlock life-saving
insights if we get serious aboutaccess and standards and what it
really takes to build a digitalhealth system that's fit for the
next 100 years.
Rachel, it's a pleasure to haveyou here.
Thank you so much for taking thetime to come and speak with me.
SPEAKER_01 (02:06):
Thank you for inviting me.
It's a privilege to be here.
SPEAKER_00 (02:10):
Before we get into the electronics, I'd like to
start with a bit about yourstory.
I'm wondering what you werecurious about as a young person
and just a potted history of howthat evolved into the work
you're doing today.
SPEAKER_01 (02:21):
I've always been a curious person and I've always
been a person that has wanted toinvestigate things.
So it's not just in this spacethat I feel this way.
I'm curious about things to thepoint where my kids bought me a
Geiger counter for Mother's Daybecause I'd got really
interested in nuclear decay andvarious other things.
And I think that came from mychildhood when my dad, he was
(02:45):
interested in so many differentthings in aviation.
in science, in psychology, ineconomics.
You've been at LSE.
And I was just taught that youshould poke the box and find out
how things work.
And I felt very grateful forbeing given that kind of
upbringing where it allowed meto do that.
Yeah, it's probably a familything, I think, because I see
(03:06):
that in a lot of my familymembers.
SPEAKER_00 (03:08):
It's a bit early in the interview to go off at a
tangent, but you've got a Geigercounter.
Tell me more about that.
I don't know anyone else that'sgot a Geiger counter.
SPEAKER_01 (03:16):
One is broken.
The joke was the cat actuallybroke one of my Geiger counters
by knocking off the sort ofradiator cover.
So I now have one that actuallytells me which isotope is
decaying, which I findfascinating, really fascinating.
But I'll tell you the funnieststory of my Geiger counter.
It was in my bag and we weredriving along the M62, passed a
van with a radiological symbolon it.
(03:38):
And this thing goes off thecharts to the point where I'm
thinking, is that poor driverokay?
So we stay parallel on the M62.
We take the radiation readings,take the name of the company and
contact them and say, I don'tthink your shielding is quite
up.
And it was actually a medicalisotope that was in that van.
But anyway, that's just one ofthe rabbit holes I like going
(03:59):
down because the kids learn aswell about radiation and about
nuclear science.
And if I had another life, Iwould love to go into nuclear
physics or volcanology, I think.
I'd love to be a volcano expert.
I've got a friend that's got aPhD on the volcanic activity on
Mars.
So, yeah.
SPEAKER_00 (04:17):
So, slight tangent.
Bring it back to your patientrecords.
Sorry, tangent at the beginning,yes.
Your curiosity led to youstudying.
Do you have a medicalbackground?
Did you study
SPEAKER_01 (04:27):
medicine?
No, so I was in biomedicalscience originally, but then I
got on the...
It was back in the 90s, thePython train of data.
So I got really engrossed.
Shortly before Linux came out,we were using BSD and things
like that, and I got reallyinterested in the data.
And that's what kind of led tothat overlap of the kind of
(04:47):
scientific background, but alsothe programmatical data
background as well.
I was really interested in whatdata tells us, what we can learn
from it, and starting to seethat beginning of the internet
as well.
We were getting collaborationswith data from different places
and it just felt I could see thefuture emerging at that point.
SPEAKER_00 (05:10):
And how did that evolve into working at the NHS?
SPEAKER_01 (05:13):
So I actually went and I spent a number of years in
Europe working with actuarieslooking at risk with data based
on human health care group riskand that sort of thing.
That was interesting.
And I actually had a personalexperience and a family
experience with the NHS where Ifelt it needed to be improved.
(05:34):
And it was one of those momentsactually where I found a guy
that had worked for me muchearlier in my career who had
moved into the NHS.
I actually met him in a hospitalwhere we were patients.
And he said, yeah, I work in theIT function here.
It's really interesting.
I said, tell me more about it.
And I just had this processwhere I felt it was something
that with the risk piece I'ddone with actuarial, there must
(05:58):
be something we could do inhealthcare that was similar.
I didn't really realise that Iwould be landing in a space that
would be perhaps a long waybehind that, which is the
reality of what I'd landed in.
SPEAKER_00 (06:10):
And you went on to be chief information officer.
SPEAKER_01 (06:13):
Yeah, I did.
SPEAKER_00 (06:15):
Hands-on responsibility for a lot of data
in that role.
SPEAKER_01 (06:19):
Huge amounts of data.
So I have more than one role asa CIO.
And we're talking about hundredsof systems.
We're talking about very complexenterprises here.
Thinking about Salford NCAGroup, you know, nearly 20,000
employees, one and a billionturnover in terms of pounds.
So huge amounts of activitybecause, you know, that size
(06:42):
corresponds to the activity.
They're very complex,mind-blowing environments.
And I think healthcare isprobably the most complex
environment you can work in.
Because there is so much data,but there are also so many
humans and human stories andhuman aspects around that data
as well that need to beconsidered.
It's not, say, aviation, wherean aircraft creates data.
(07:05):
There's also a person behindthis that has thoughts, needs,
values, everything else.
So for me, it was incrediblychallenging, rewarding, but also
in many aspects behind the timesof The Art of the Possible.
SPEAKER_00 (07:21):
How much is that complexity part of the
attraction of working here?
SPEAKER_01 (07:26):
The complexity is both an attraction and a
headache, if you like, becausesometimes it just feels that
there are things you can'tsolve.
It takes time.
The attraction is also humans,as in there are a few other
places where you can improvelives and save lives.
Somebody once said that a dataanalyst or a programmer in
(07:48):
healthcare could...
cause huge amounts of harm tolife or could save a huge amount
of life.
And that is absolutely true.
As a vocation, this is asimportant as doctors and nurses.
SPEAKER_00 (08:06):
Let's go on to talking a bit about electronic
health records.
So I'm hoping we have listenerswho are outside the health
system to Problems WorthSolving.
So I wondered if you couldexplain a bit about EHRs or
electronic patient records,which is another name for them,
what they are, where they'vecome from, what they do and how
do they benefit patients andstaff?
SPEAKER_01 (08:28):
Absolutely.
Electronic patient records,EHRs, started around about 30 or
40 years ago.
And they started really assystems where you could record
things like lab results or putin orders or make notes about a
nursing encounter.
And they have grown in terms ofwhat they do.
(08:51):
They go right the way throughnow to every function you can
think of, billing, managingbloods, managing labs and
pharmacy and everything else.
But the evolution has reallybeen based on the functionality.
So it's all been about how do weadd some more functionality to
(09:11):
this product?
And health systems have reallycompared that functionality,
front-end user functionality, orwhat it can do in terms of
interacting with pharmacy.
And as a result of the fact thatthe market is crept in this way,
you have all these differentvendors who have evolved their
front-end functionality to besomewhat equivalent, but have
(09:35):
bolted things onto this backend, which maybe got its origins
in the 80s or 90s and was neverdesigned to do all of these
things and all of thesefunctions.
You know, some countries they'vebolted on billing.
In the UK, we've bolted onactivity and patient
administration.
We've bolted on bed management.
And as a result, these thingsare architected based on older
(09:59):
technologies and the way thatthey've evolved them has been to
basically just keep the markethappy.
And that has built really atechnical debt into the backend.
So as a CIO, trying to get yourdata out and make sense of it
from one of these things is notthe easiest task.
SPEAKER_00 (10:18):
What sort of data would you be trying to get out?
SPEAKER_01 (10:21):
I'll give you a really simple example.
Say you're doing a study ofpatients that have heart failure
and you just want to pull outtheir blood pressure across
time.
Do a time series, something likethat.
What you find within these EMRsis the blood pressure is stored
in all sorts of different placesin different formats with fields
(10:41):
that don't match.
We call that lossy data becauseyou're losing some of the
context.
And to actually get a timeseries out for my data
scientists, it would have beenat Salford.
You have to do all of thesemappings, which don't fully map.
And the way each of those bitsof functionality that record
blood pressure, one might be adevice here, one might be
nursing here, one might be award around here.
(11:02):
I remember doing an audit of oneEMR that had 47 different ways
of just recording bloodpressure, right?
There's a hell of a lot of work.
And the lossiness and thecontext loss, the provenance
that data is lost because you'vegot all of these different
places where the same concept isstored.
And as soon as you get into thisworld of data science and AI,
(11:22):
you realize the underlying datamodels and the underlying data
are just not built for the nextgeneration of solutions that
we're going to need or startingto need now.
We lovingly used to call some ofthe people that did the work on
the data quality to bring it allin line.
So manually manipulating things,the cardigans, because they
(11:42):
would come in, I'm wearing acardigan today so I can laugh
about this, hang the cardigan orthe jacket on the back of the
chair and do a whole day's workon data quality.
And that really, prevention isbetter than cure.
But we're curing the fact thatthese systems have data in such
(12:02):
disparate ways, what we shouldreally be doing is preventing
that by having the right backend data infrastructure.
So yeah, it's been an organicthing and it also hasn't helped
that the NHS has always reallybeen money conscious to the
point where it's always said buythe cheapest system and that has
never considered how good thedata is in the back end.
SPEAKER_00 (12:25):
So as I mentioned earlier, the government has been
funding huge investment intothese EHRs over the past few
years.
Where does all that money go andwhat's the real return on
investment for the system?
SPEAKER_01 (12:36):
So to be clear, you know, these systems do fulfill
the needs of a hospitaltypically, and they allow a
hospital to function.
So the money goes into providingthat incredibly complex
functionality that allows ahospital to operate, along with
a very big infrastructure thatenables it.
(12:56):
What I would say is many ofthese systems are actually US
systems that are reprovisionedfor the UK.
So we're also buying a hugeamount of billing and things
that we don't use in the UK.
And that's always been a bit ofa question for me.
But if you think about everyprocess in a hospital moving
from the analog or from oldsystems through to a single
unified system, that's sort ofwhat the EHRs are.
(13:20):
The nearest analogy, if you'renot from the healthcare
industry, is the ERP and whatthat does for a business in
terms of automating one of theprocesses.
The thing is that these are verycostly, not just because of the
software, because the softwareis costly.
It is training every singlemember of staff.
If you take my old organization,18,000 people to train, and it's
(13:41):
a good few hours training each,and they need updates on that
training.
It's a very big program.
And obviously, you have to trainthem well, because this is about
clinical risk.
It's about clinical safety.
It's about all of those things.
So really, these very bigprojects...
are very big bits of softwarewhich have to be taken in one
(14:02):
lump because they're monolithicsystems, yeah?
You have to take this bigmonolithic system and you have
to implement it.
And I think that's where thesethings get really difficult
because they're massive projectsand programs and it's really,
you have to leap into it.
There's no kind of staging, youknow, your sort of deployment of
these things.
SPEAKER_00 (14:21):
Are we building these systems in the right way?
SPEAKER_01 (14:23):
My view increasingly is is no.
I've done some work actuallywith Jordi Piera in Catalonia to
look at the generations of EHR.
And the first generation waswhere we just built or bought
lots of systems and tried tostrap them together.
That was the 80s, 90s.
So that was best of breed.
(14:43):
You'd buy lots of things and tryand put them together.
Second generation is this bigEMR where you have to leap into
it and do everything.
But now what we know is thatdata is going to be used for so
many different purposes.
If you're born with a chestproblem and you're put into
NICU, high dependency forbabies, that data is going to
(15:05):
bear out for your chest healthfor the rest of your life.
And that data has to beaccessible for decades in a
meaningful format.
And what we're learning now isis that we need a data-centric
architecture because your datais going to be used in
algorithms to help restratifyyou.
It's going to be used to helpmake decisions about you in
(15:26):
different institutions, indeedinternationally, you know, with
the international patientsummary.
It may travel with you so thatyou can be treated abroad when
you need it.
And so this data-centricarchitecture would say you
separate your data fromapplication because one thing I
can bet is that in 10, 15, 20years time, how doctors, nurses,
(15:50):
and those in healthcare systemsoperate will not be via the
keyboard and typing.
It will be via voice, gesture,who knows what, right?
It will be things we haven'teven imagined now.
If you record a really good dateset now, That will last through
that journey.
But what we're doing instead isrecording the data driven by the
(16:13):
interface.
The interface says this, sowe're going to record this data.
What we need to do is, no, weneed to look at human, what we
need to record, and actuallykeep that in a format that is
consistent across time.
And that data-centricarchitecture and separation of
data from application you see inother industries, but it
absolutely makes sense because Ibelieve somebody should be able
(16:36):
to carry their digital twin withthem, which is all of their data
about them.
And you've got a human rightreally to your digital twin
because you can learn aboutyourself from your data.
You can take it with you.
So yeah, so the future is reallyabout the citizen having access
Every actor that they need tohave access in their healthcare
(16:57):
journey should have access.
But it shouldn't preclude peoplefrom actually providing care
because they don't understandthe data.
SPEAKER_00 (17:05):
And this is what you mean by thinking of data as the
foundation, not a byproduct ofthese systems.
SPEAKER_01 (17:11):
Absolutely.
Data is not a capital asset,yeah?
It's something that will havevalue over decades or even 100
years of somebody's life orbeyond.
If you can imagine in thefuture...
100 years of data in your life,you may actually give that to
your family for informing theirhealth care.
I certainly know in my familythere's breast cancer, there's
(17:33):
bowel cancer, there's otherthings.
If that data was available tome, it would help to stratify my
health care.
So this could become somethingthat is a legacy that's passed
on through families andgenerations.
SPEAKER_00 (17:44):
What would it mean to design systematically around
data rather than trying toretrofit the system?
SPEAKER_01 (17:51):
That's very much the space that I'm in now, is
designing what is needed for ahuman's healthcare record for
their life, yeah?
And designing it in anon-duplicative, semantically
harmonized, as we call it, way,which means we take blood
pressure, as we talked aboutearlier, there is one concept of
(18:12):
blood pressure, not 40, aredifferent concepts.
And In order for that to be usedin different places, we use
something called templates,which are a layer that allow it
to be used in different places.
So the idea here iswell-engineered data is a
combination of clinicians andthose that work in the technical
space working together tooptimally work out what data is
(18:36):
needed for a human in the mostefficient way and the most
representative way for it to beused for all of the use cases
today and tomorrow.
SPEAKER_00 (18:45):
So you're the Chief Exec at OpenEHR.
Can you tell me a bit about theorganisation, where it's come
from, and how does it fit intothis landscape and the work
you're doing?
SPEAKER_01 (18:58):
You know, I've not been on the journey with OpenEHR
since its beginning.
It's nearly 24 years that it'sbeen running.
And it started actually as acollaborative between both
clinicians and technologists atUCL in London.
with the input of some of thoseclinicians being from Australia
as well.
So it's an Anglo-Australianinitiative and it's incredibly
(19:22):
powerful.
It took me a while to get thisconcept, I've got to say,
because certainly on my journey,I've not always appreciated how
important the data is.
So 24 years ago, they had thiskind of epiphany and light bulb
moment about data and they setup open air and Professor David
(19:44):
Ingram was pivotal in thishappening at UCL.
And that community has grown andgrown as a not-for-profit
community of interest thatactually defines and develops
these standards.
Around about half or just overhalf of the community are
clinicians.
The rest are technologists andleaders.
(20:05):
And it has a governance whereit's almost a Delphi, if you
like, of people who go throughthese processes and work out
what the optimal dataengineering is for any aspect of
human healthcare.
And as that grows, thatmodeling, as we call it, data
modeling and clinical modelingcontinues so that we can
actually grow the sort of dataarchitecture that is enabling
(20:29):
human healthcare.
SPEAKER_00 (20:31):
So this was what attracted me to invite you onto
the podcast today because it'sall about problems worth solving
I wonder how you how would yousummarize the problem that
you're trying to solve as anorganization
SPEAKER_01 (20:42):
I will summarize this from my experience I think
others will have different viewsbut I was involved in data saves
lives in Manchester and we sawthat as we improved data quality
and access lives were beingsaved and improved it was quite
a scorecard that literally had ahuge massive impact on me And
(21:03):
that made me realize that as wehave a single source of truth
for human life, and as we makethat available as appropriate to
different endpoints, you canhave really big impacts on lives
saved and lives improved andquality of life.
And what I see in the communityis a set of people who have seen
(21:25):
that too.
They've seen that we need toengineer data better and make it
available to particularly to thecitizen and to the healthcare
settings that may not be theprimary healthcare setting to
the citizen, because it is aquality and safety issue.
One thing that I came to, andI'd like to do some research on
(21:47):
this, I haven't had theopportunity yet, is there is a
link between unwarrantedvariation in data and
unwarranted variation in care.
And we know that unwarrantedvariation in care causes harm.
And so this is all about harm,quality and safety.
It's a set of people that cansee an ability to do much better
in health and care.
SPEAKER_00 (22:08):
Imagine you're immensely successful, which of
course you're going to be in thework you're doing.
What does that success look likefor the system?
SPEAKER_01 (22:18):
I think the success for the system for us looks like
every citizen, hopefullyglobally, let's say we're really
successful, let's say globally,every citizen having their data
available to them and the careproviders that they want to
treat them, and those doingresearch, you know, all of the
(22:39):
actors in the system, when theywant it, where they want it, in
a format where you can get speedto value by putting AI,
intelligence, apps on top of it.
So it is really about removingthe friction of data today to
enable a much better healthcaresystem of tomorrow.
And also one personal thing forme, we were talking about EMRs.
(23:00):
EMRs reinforce the hospitalwalls, which pushes more and
more activity into the hospital.
And so we do more in hospitalsnow.
I would really like to seeOpenAir being part of the
solution to allowing a recordthat is available anywhere,
which means we can deliver carein the home via apps, devices,
and citizens can self-care.
(23:21):
So there's this aspect of stopreinforcing the hospital walls
and start allowing care to bedelivered in the most logical
place.
SPEAKER_00 (23:29):
What's the implication of that for
innovation and kind of the widersystem changes?
SPEAKER_01 (23:33):
So if you think about that, There are lots of
things that play against usactually innovating.
And by innovating, I'm going totake that as meaning you models
for operating healthcaresystems.
Yeah.
So if you want to startoperating healthcare systems
where you have distributedservices, you have self-care,
(23:55):
self-management, you havemanagement by exception from
medical devices, you have appsand wearables that support a
citizen to self-care and preventall of that.
requires a single source oftruth about the patient.
And it requires that data can bemanaged across the entire set of
(24:16):
organizations in the experience.
The average patient back inManchester would have been known
to five organizations.
That's just an example.
So for me, for the future, thisubiquitous pathway allows those
innovations because there's notany of the frictions of
stovepipes, data silos, if yougive the citizen the ability to
(24:39):
control their record asappropriate as well, you can
allow things like citizenscience and innovation to happen
in that way, which is a newavenue.
I think that's reallyinteresting, allowing citizens
to participate in trials andstudies and contribute.
I'm seeing some good example ofthat.
And for me, just taking away allof the friction of actually
(25:01):
wrangling this data into theright shape, having it lossy,
not allowing certain actorsaccess to it, we could really
speed up the innovation.
I was just with Lord Darziyesterday and he said the answer
to the next generation of theNHS, the NHS's future, is all in
the data.
SPEAKER_00 (25:25):
There's lots of hype around AI in health at the
moment.
I'm curious about what you thinkwe're skipping over when we go
straight to the shiny stuff.
SPEAKER_01 (25:34):
So coming back to the Paris AI Summit back in
February, they said in allsectors, it's about great data
and it's about good governance.
I couldn't agree more.
I'm afraid it's a lot of reallyhard work to get our data right
and to work out how we govern inthe age of AI.
And that's fairly boring stuff.
(25:56):
So risk frameworks in AIgovernance is something I've got
deep into.
It's not going to grab anyheadlines, is it?
But it's the work that needs tobe done.
I would also say that anotherthing that we're missing,
jumping straight to the shiny,is the platform approach to AI.
We're looking at hospitals thathave got tens of different AI
(26:17):
algorithms.
In the US, it may be a lot more.
But for each of thosealgorithms, you have to monitor
it, you have to audit it, youhave to understand how it's
using data, you have to look forbias, you have to look for cases
where it's not been effective.
And if you're going to havehumans do that, that is not
scalable.
So these oversight platforms,which we're starting to see in
(26:40):
the marketplace now, I think arean essential component because
they will help us to govern andensure the safety of these
algorithms.
What we can't do is allowhundreds of algorithms to be out
there without this kind ofoversight platform and
management platform.
And I think that kind of middleground marketplace, if you
(27:01):
almost like, where you canactually monitor those
algorithms and their performanceis going to be an essential part
of the future.
SPEAKER_00 (27:10):
And that would involve having a human in the
loop.
SPEAKER_01 (27:14):
Absolutely.
Human in the loop, but to manageexceptions.
At the moment, we just don't getthat data.
It has to be a human go in andlook at the performance of the
AI and check it on a regularbasis.
Humans in the loop always,right?
The reason you need humans inthe loop are multiple.
One, you need that humanjudgment there.
Two, There will be exceptionsthat are valid.
(27:36):
If you look at certain hospitalsor health systems that deal with
the most problematic cases, youwill see things that are
spurious in terms of outcomesand results there.
It needs that humaninterpretation to actually
understand why things areexceptions.
SPEAKER_00 (27:54):
What do you feel the role of data is in population
level insights and personalisedcare?
SPEAKER_01 (28:00):
I'm a huge fan of social determinants of health.
And the biggest factors thatactually determine your health
over time are your education andwork status over a population.
For me, there is an awful lotmore that we can do to actually
(28:23):
find out how we can help ourpopulations stay well and
intervene and prevent disease.
ill health before it happens.
And a lot of that does reside innon-healthcare data.
And a lot of the questionsaround that are the ethics.
So I certainly know that in theUS, certain health systems have
(28:44):
got agreement to access thefinancial history of their
patients.
They've consented that.
That's going to take a lot ofthinking if we ever wanted to do
that in the UK.
Is that something we would do?
Maybe not.
But there is a whole domain herethat we need to explore in terms
of pulling together that data.
(29:07):
But some, again, coming to theUS, I've just said that, we may
or may not want to do it.
But in terms of actually usingsome of those tools, social
determinants of health and thewider data, I've seen some
excellent examples where they'vetaken people that haven't turned
up for their breast cancerscreening, as an example, for
five years and managed to get65% of them to do that within
(29:29):
six months.
SPEAKER_00 (29:31):
How did they do
SPEAKER_01 (29:32):
that?
They analysed all of this data,found out their motivation,
found out demographics of thesepeople and put nudges out that
were tamered to them, that wouldappeal to them, and would speak
to them about why breast cancerscreening was important.
And that just blew me away.
Really blew me away.
Yeah.
SPEAKER_00 (29:51):
And was that driven by AI?
SPEAKER_01 (29:52):
Partly by AI, partly by the great data, partly by
psychologists, and partly bysome great public health
clinicians, some of whom arefrom the UK.
SPEAKER_00 (30:03):
Is it a false split to say there's health data and
there's social data involved?
Actually, you want to look atthe whole picture.
Yes.
SPEAKER_01 (30:09):
And it's beyond just social data.
It may be lifestyle data.
It may be one thing I'm verykeen on is what is your
preferences as well as a human?
We should be asking that.
That should be put in.
SPEAKER_00 (30:20):
And that surely leads to the responsibility for
ownership of that data needs tobe with the individual.
Yes.
SPEAKER_01 (30:26):
Were possible, it does need to be with the
individual.
And we've seen lovely exampleslike One London at the moment
where citizens can nowcontribute via the NHS app to
their care plans.
We need this to be a partnershipwith our citizens.
They need to be able tocontribute to their record.
And I think that the true futureof healthcare is a partnership
(30:50):
between the NHS and ourcitizens.
Obviously, there'll be somepeople like my late father who'd
had two strokes and had lost hisspeech who wouldn't be able to
do that.
We need arrangements for peoplewhere it's not suitable.
But for those that want to andneed to, especially with the
coming generation of people whoexpect digital services, I think
we need to do this incollaboration together.
SPEAKER_00 (31:12):
The data becomes increasingly valuable when it's
being used in all these otherdifferent ways.
Who should be responsible forgoverning that data?
SPEAKER_01 (31:22):
That's a very good question.
And that's a very interestingquestion when the citizen is
contributing, you know, yourlifestyle, your garment, your
Apple Watch, whatever else.
It becomes a spectrum of datathat is co-owned.
I think for me, anything that iscreated in the clinical domain
(31:45):
needs to be owned by the healthsystem, but also organized.
a copy of that owned by thecitizen.
It's a dual responsibility.
The reason is that the healthsystem needs to own it to make
sure it is safe, it's secure,it's audited, all of those
things.
It has a responsibility to keepthat.
(32:05):
But then there needs to be asynchronized copy in the health
record.
In terms of lifestyle data andpotentially some aspects of
social determinants of health, Ithink there really is an
ownership by the citizen.
If we come to financial data,There are certain patients who
may choose to share that withtheir healthcare system because
(32:27):
they believe it will improvetheir outcomes or it will
improve the population'soutcomes.
I think those sorts of pieces ofdata should be down to the
citizen to choose to share.
And really what I see as alonger-term goal is the ability
for a citizen to go in and say,do you want to share this data?
This is how it will be used.
(32:48):
This is what we intend to dowith the data.
And then a bit like we have withblood products now, if you
donate blood in the UK, you'vegot a lovely text message and
say it's been used in apaediatric hospital or whatever.
You know that you've given back.
I think there should be sort ofpostcards back to the citizen to
say, these are the insightswe've found using your data, or
(33:09):
this is the research we've doneand this has been the outcomes.
And that kind of datadonatorship bit, as you said,
data's got value.
And if you're giving back toothers or you're helping others
with that data, we need to closethat loop.
SPEAKER_00 (33:29):
What do you think health can learn from other
sectors like banking aroundtrust, access and usability?
SPEAKER_01 (33:35):
Oh, trust, access and usability.
I think let's take access.
I think with banking, accessis...
ahead of most industries.
I've worked with BarclaysDigital Eagles and the way in
which they are trying to makethings more accessible to
different groups, olderpopulations, populations that
(33:59):
may not have been digitallyexcluded.
I think there is a huge amountwe can still learn from
financial services in terms ofthat.
Although one thing I would sayis that We're still not seeing
banking in multiple languages,which we need for healthcare.
That is one consideration thatwe have to make and one aim I
know that health systems havegot, which is to make your
(34:20):
information available in yourlanguage of choice eventually,
which is really important.
In terms of trust, I think theFinancial Conduct Authority has
done a very good job of ensuringtrust within the financial
sector.
And I once jokingly said, I'dlove to see the NHS as kind of
(34:42):
having a Bank of England-likefunction that was independent
and could assure things and makedecisions about things.
But I think we need a verysimilar construct.
Some of the things we have seenare what are called data trusts,
and they are senses of thirdsector academia, patient groups.
assuring the use of data onbehalf of citizens.
(35:06):
So if you think about acooperative or a data trust,
might assure the use of yourdata in a geography and kind of
be your appointed party to dothat.
I think we need this kind ofthird entity that helps build
trust because there are so manyactors in the system And while
it is the NHS's job to provethat data is used correctly, I
(35:27):
don't think they can operatethese sorts of groups that cohes
everyone together and helpassure how the wider system is
using data.
And I think that's a constructthat's needed.
Be it a FCA type piece, be itdata trust, there's something
that's needed in the middle.
SPEAKER_00 (35:42):
Is that something that you can help with in your
work?
SPEAKER_01 (35:45):
I'm actually working with one of the leaders of...
one of the data trusts at themoment?
Yes, we absolutely can.
And I think...
The nice thing about open airand everything else is that
being a standard that is usedglobally, it's a lot easier to
put transparency toolings on topto see how it's been used and
(36:05):
what's been done with that data.
Outside of my open air work, I'mhelping chair some research
groups into this, particularlywith patients that have got
chronic long-term conditions,mental health conditions,
looking at how they want theirdata to be used.
So I'm finding, again, the humanaspects of it.
this.
I continue to learn and Icontinue to want to know more
(36:28):
about how we should interactwith citizens and allow them to
really determine theirpreferences for use of data.
SPEAKER_00 (36:37):
What would massive success for open air?
I'm saying it in the right waynow.
I was saying open EHR.
SPEAKER_01 (36:43):
Either way is fine.
SPEAKER_00 (36:44):
I've learned that you're saying open air.
What would massive success ofyour ambitions mean for the big
EHR vendors?
SPEAKER_01 (36:51):
Some of them may well embrace this.
It may be their next generationoffer.
For some of them, they grow atechnical debt or a data debt.
If governments mandate this,then that will be an issue for
them.
My view actually is that thereshould be incentives to help
(37:12):
vendors move towards standardslike this because we have to
recognize that they've done goodservice as EMRs in hospitals and
everything else.
But where we are now, futureforward, there is that
day-to-day in the back end.
And I think certain governmentsmay want to work with those
vendors to incentivize them orsupport them in moving towards
(37:35):
supporting these standards.
Because what you don't want todo is undermine a market that
you have already.
You want to help it evolve withyou.
SPEAKER_00 (37:42):
Do you work quite closely with the vendors in your
work?
UNKNOWN (37:45):
No.
SPEAKER_01 (37:45):
I work with a number of vendors.
Some of the big EMR vendors,yes.
Some of them, no.
So they split into two camps.
Those that feel that this is thefuture and are looking at the
ways that they can start toenable this.
And then some of them that say,no, just buy one big monolithic
system, which I don't think isgoing to be the future.
SPEAKER_00 (38:08):
How do you think we get the balance right between
partnership and dependency?
SPEAKER_01 (38:12):
Yeah.
And I think...
Partnership is key.
Partnership with industry and asymbiosis with industry is
absolutely key.
But dependency, dependency isnot good.
And we have become dependent oncertain vendors because of their
bespoke data and their lock-in.
They become the only people thatcan do certain things of the
(38:32):
data because the way it's beencurated.
Data and dependency, for me, Ithink it's a difficult one
because...
If you don't want to bedependent on vendors for
decades, you don't build in datalocking.
You don't build in data usestandards.
(38:53):
However, what I will say is theflip side of that is that the
health system needs to be theintelligent client and be able
to describe what it wants to buyand procure and how it will
work.
And they also need to recognizecustomers.
As I said earlier, that there isan existing market that you want
to take on a journey with you.
So you have to influence theexisting vendors.
(39:16):
And I think in certaingeographies, we're seeing
intelligent clients emerge inhealthcare.
They're really understandingthis and they're really driving
it forward.
But perhaps in some othergeographies, less so, which
means that they're not reallybuilding for a well-engineered
data future.
they really don't understand theimportance of that asset or how
(39:37):
to build it for their geography.
SPEAKER_00 (39:39):
And what about getting NHS support and then
global support for thesestandards?
SPEAKER_01 (39:45):
So we have a huge amount of global support at the
moment.
You know, increasingly I'm inwith governments from different
countries, working with the AIDHin Australia at the moment, and
they invited me onto theirdigital health advisory board.
Ireland is moving to open airfor its integrated record.
We've got Greece moving in thesame direction.
(40:07):
We've got other countries Ican't mention because the
governments aren't ready to talkabout it, but we have the
Nordics who've traditionallyused open EHR within their
standards.
You've then got countries likeJamaica that are using this and
some of the African countries.
So, you know, Jamaica's datamodel is open air.
This is moving forward, butObviously, with Opener being
(40:29):
very small, not for profit, veryfrugal, this isn't done by
selling anything.
This is just done by talkingthrough the logics of why you
would do it and the successesthat other people are having and
just really getting peopleinvolved in a global community.
It's really amazing that we'vegot as far as we've got and
we're continuing to go onbecause we are a not-for-profit
(40:53):
open source community.
Anyone can take this and use itfor free.
It's a global good, a bit likeLinux.
It's very much an open sourcegood.
And so for me, it is thecommunity that permeates the
future.
It is the community.
It is the word of mouth.
It is an ophthalmologist in onecountry talking to an
ophthalmologist in another.
It's somebody in one ministrytalking to a ministry in another
(41:15):
country and saying, look, we'vedone this work.
We found that a data asset hasgot value for us.
We are going to actuallyarchitect our data well.
I'm going to base it on OpenAirand these other standards for
the future.
And actually, coming back tothis, she talked about the NHS.
There was work done around about2019, I think, that valued the
(41:36):
data of the NHS.
This is not selling the data.
This is not giving the dataaway.
This is the value of being ableto use the data and deliver the
services effectively within theNHS.
That was billions.
It was around about six billiona year, potentially up to nine
billion a year.
Data has value.
(41:57):
And people are realizing thatnow.
SPEAKER_00 (42:04):
It's coming across that you think about data from
the human perspective.
Yes.
I was wondering how much youthink about data as a tool for
social change and how we shiftfrom seeing data as just a side
effect of the activities.
SPEAKER_01 (42:17):
So that's why I mentioned earlier the science
collaboratives with thecitizens, because citizen
science, and I'm thinking aboutone example I've been involved
with where we have doctors, wehave patients, we have data, and
they all collaborate around thedata and the outcomes.
I think that this is a hugeopportunity for social change.
(42:41):
On the other side of it, if welook at...
I've just been reading a book onFacebook and how they use data
globally to get the outcomesthey want from people.
If we use that science of beingable to find out what people
want or need from the data andactually nudge them, coming back
to this, getting an uptick inbreast screening, the things
(43:02):
that have been used to market tous from the internet as a kind
of dark science...
could be turned into a lightscience of helping people and
nudging them and supporting themwith what they need to see,
hear, or act on that'sappropriate for them, yeah?
And this whole piece for me,social change, we have to be
(43:23):
transparent about how we'redoing all of this.
I think absolutely transparent.
But the idea that we canactually allow citizens to drive
more of this, it reallyinterests me actually that in
Finland, they have citizenassemblies that drive their
healthcare system that see dataand decide, make the difficult
(43:43):
decisions about haves andhave-nots with the budget.
But is our future healthcaresystem going to be data-driven
in that way, where you have atransparency of what there is, a
transparency of the problems,transparency of the
opportunities, tools that showyou where you could put your
money and what outcomes youcould get, and you can make
decisions as a set of citizens.
I think there's an awful lotthat could be done.
(44:06):
And I'm also seeing global raredisease groups who are
aggregating their data andtaking action as a citizen group
and approaching life sciencescompanies about finding
different repurposing of drugs,whatever else.
And you think healthcare reallytraditionally going back to the
(44:27):
old NHS had the patient in thechild state.
And the clinicians in the parentstate, and it was a
parent-child, you did what youwere told, you turned up when
you turned up, you were givenwhat you were given.
Medicine of the Future, I think,is about the adult, which is
about an adult relationshipbetween the clinician and the
patient and roles for both ofthem.
SPEAKER_00 (44:49):
What are you most optimistic about right now?
SPEAKER_01 (44:52):
There are a few things I'm optimistic about, and
particularly today, becauseyesterday I was with a cohort of
graduates from the DigitalHealth Leadership Programme.
I am really optimistic that wecan use data, technology, and
human change to create the nextgeneration of the NHS.
(45:12):
I sometimes work with boardleaders and I say, okay, let's
sit down and imagine when youretire 20 odd years time, what
does your healthcare look likein 20 odd years time?
You're living by yourself,potentially, what do you want
around you that helps you to beindependent and feel confident?
And we just describe what thewhat their ideal health system
looks like.
(45:33):
And then we look at how they getthere, right?
And I think people are startingto see what that future health
system looks like and arestarting to see how we get
there.
So I'm feeling quite heartenedthat actually we have the
ingredients.
We've got to wait for this wholechange with NHS England and the
Department of Health andeverything else to play through.
But if there's one thing I wouldsay to the great leaders out
(45:55):
there that I've seen, you willbe part of delivering that
future and we really need you.
SPEAKER_00 (46:00):
What do you think that future health system looks
like?
SPEAKER_01 (46:03):
Again, going down rabbit holes.
I talked about going down rabbitholes.
One that I've been down recentlywith a couple of former
colleagues is this idea of youhave the capacity of the NHS and
you manage the capacity anddemand across the whole system.
And there's no such conceptanymore of referrals into a
(46:25):
pathway.
It's always next best action.
using the whole capacity of asystem.
And I think we can get to apoint where we can leverage
every part of the NHS and knowwhat capacity we need as a
system, know what actions weneed to take, and can bring in
that efficiency across the wholesystem.
(46:46):
So I've got this view that thisintelligence that we're talking
about for the future will reallyallow us to optimize what we
have today?
Because there is an opportunityto optimize, but that will also
allow us to invest in thetechnologies for the future.
One thing's for certain, we willalways need our wonderful NHS
staff because this is about aservice that is really
(47:10):
human-faced, has empathy, canput things into context.
AI is never going to put acancer diagnosis into context
for you as an individual, right?
that human contact, that humanemotion, you need that.
But what it will do is be ableto augment our doctors, our
nurses.
And another thing I've beenlooking at is agentic AI.
(47:31):
So what if you can actually takeall of the knowledge that's out
there, all the procedures withinan organization, and for a nurse
that's actually working on anight shift, he or she is able
to actually ask it questions andget the knowledge of a huge body
of knowledge, but instantly, youknow, via a headset or via
(47:53):
whatever else.
Those sorts of things, I think,will really assist the NHS in
revolutionising.
I don't quite know how it'sgoing to play out, but all I
know is that in that 15, 20 yeartimeframe, we're going to have
to have got there and the toolsare there to get there.
And that I see enough goodleaders working and potential
(48:13):
leaders for that journey to betaken.
SPEAKER_00 (48:16):
And what about more immediately?
So if you could wave a magicwand right now and change one
thing about the way the systemworks, what would you change?
I
SPEAKER_01 (48:23):
think leadership needs to change.
I don't mean take leaders away.
What I mean is how people lead,how people in leadership are
trained.
So for me, just a basic example,if you go to
youraverageboard.com in ahealthcare organization, quite
(48:44):
traditionally you'll have anestates director on there
because that's something thatdelivers healthcare.
You won't have somebody on therethat's a CIO or equivalent
because they're still thinkingin terms of bricks and mortar
for everything.
I think the leadership needs achange of mindset and I have
some hope that the future NHSwill do this.
(49:06):
The basis says we're going todeliver healthcare a new target
operating model of care.
So if there's one thing I couldchange, it would be changing the
leadership's mindset to say, weneed to look at operating model
2.0 for the NHS.
SPEAKER_00 (49:20):
What's next for you, Rachel?
SPEAKER_01 (49:23):
I think more open eye.
Lots and lots to do.
That said, I still do help theNHS.
I still do spend time with theNHS.
I'm very passionate that I willhelp the NHS going forward.
But I think for me, I think Iwill just keep going down rabbit
holes that I find.
I keep finding problems and needto find out more.
(49:43):
And so for me, this is anemerging journey and being
engaged with data, AI,intelligence and the human
aspects of that are somethingthat are going to keep me busy
for the next couple of decadesat least.
SPEAKER_00 (49:59):
I'd like to end on something thought-provoking.
So what's One belief that youhold that others might not agree
with, but you believe is true.
SPEAKER_01 (50:08):
Does this have to be to do with healthcare or not?
SPEAKER_00 (50:13):
No, just anything.
SPEAKER_01 (50:14):
Okay, I'm just going to be slightly out there then.
But again, I've been watchingsome of the hearings from
Congress in the US and itappears that they have disclosed
that there's non-humanintelligence.
SPEAKER_00 (50:32):
That's really interesting.
I've also been following thosehearings and it does seem
incredible.
SPEAKER_01 (50:37):
Absolutely, but hang on.
In Congress, multiple peoplehave said there is non-human
intelligence.
SPEAKER_00 (50:43):
I'm slightly mystified as to why this hasn't
had more coverage.
SPEAKER_01 (50:46):
Right.
SPEAKER_00 (50:47):
It seems like quite a big deal.
SPEAKER_01 (50:48):
And there is emerging evidence from huge
numbers of military people inthe US, that is true.
Now, bear with me, this doesrelay back to healthcare.
Because what does it mean to behuman?
If there are others, what doesit mean to be human?
And what does it mean to behuman?
We also have to consider inhealthcare in terms of how we
(51:09):
want humans to live goingforward with the AI and
increasing automation.
We've got questions about thefuture of work, what happens
when jobs are automated.
I think the two things cometogether for me to say, we
really need to consider what itmeans to be human in this
current environment.
SPEAKER_00 (51:27):
rachel it's been fascinating talking to
SPEAKER_01 (51:29):
you i didn't think i'd be talking about that there
you go
SPEAKER_00 (51:33):
we must talk more but it's been great to talk to
you thanks so much for takingthe time to come and share your
thoughts and i've learned lotsabout the work that you're doing
and open air and hopefully ourpaths will cross at some point
in the future
SPEAKER_01 (51:44):
thanks for the opportunity
SPEAKER_00 (51:49):
you Problems Worth Solving is brought to you by
Healthier, the collaborativeservice design consultancy for
transformation in health careand public services.
Find out more about how we canhelp you deliver user-centered
change at healthier.services.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Special Summer Offer: Exclusively on Apple Podcasts, try our Dateline Premium subscription completely free for one month! With Dateline Premium, you get every episode ad-free plus exclusive bonus content.