Name: The power of AI and digital innovations
Uploaded: 2023-10-31T00:05:50.876Z
Duration: 518.53

MB Edited.mp4

MB Edited.vtt

Subtitles

{{00:00:06}}
You know, I've spent a lot

{{00:00:06}}
of time thinking about, and

{{00:00:08}}
there's a very real risk with

{{00:00:10}}
AI technologies, I think,

{{00:00:11}}
that their introduction

{{00:00:12}}
and their scale, their

{{00:00:13}}
industrial use across the

{{00:00:16}}
entire health system poses

{{00:00:17}}
a very real risk to equity.

{{00:00:19}}
so, hello,  Mitch Burgess,

{{00:00:29}}
my name, I'm Director of

{{00:00:30}}
Strategy, Architecture,

{{00:00:31}}
Innovation and Research

{{00:00:33}}
at Sydney Local Health

{{00:00:34}}
District, which is...

{{00:00:35}}
a health service in the

{{00:00:36}}
inner west of Sydney.

{{00:00:37}}
We look after about five

{{00:00:38}}
hospitals, including RPA

{{00:00:40}}
and Concord, which are major

{{00:00:41}}
tertiary referral hospitals.

{{00:00:43}}
AI does have

{{00:00:44}}
tremendous potential.

{{00:00:45}}
You know, it's one of

{{00:00:46}}
those things where I

{{00:00:47}}
think everybody sees the

{{00:00:48}}
potential in technologies

{{00:00:50}}
like this to revolutionise

{{00:00:51}}
the way we provide care, to

{{00:00:53}}
improve patient experience,

{{00:00:54}}
to make it easier for

{{00:00:55}}
clinicians to do their jobs.

{{00:00:58}}
You know, I think realizing

{{00:00:59}}
that potential is going to be

{{00:01:00}}
one of those things that's a

{{00:01:01}}
great challenge for healthcare

{{00:01:02}}
services over the next 10 or

{{00:01:04}}
even 20 years, you know, it's

{{00:01:05}}
one of those things where I

{{00:01:06}}
think we're going to have to

{{00:01:07}}
be very slow and methodical.

{{00:01:09}}
One area where the district

{{00:01:11}}
has put a lot of time and

{{00:01:12}}
effort over the last five

{{00:01:13}}
years, you know, we've taken

{{00:01:14}}
a very rigorous approach to

{{00:01:17}}
implementing a system for

{{00:01:20}}
use in wound care, which

{{00:01:22}}
uses AI to process images of

{{00:01:24}}
wounds and takes; some of the

{{00:01:28}}
burden away from clinicians

{{00:01:29}}
for calculating the size

{{00:01:31}}
of wounds, for tracking

{{00:01:33}}
how they're improving over

{{00:01:34}}
time or deteriorating as

{{00:01:36}}
the pace may be, analyzing

{{00:01:38}}
the types of tissue

{{00:01:40}}
present in the wound so they

{{00:01:41}}
can get, gauge whether it's

{{00:01:42}}
healing appropriately or not.

{{00:01:44}}
And we're really reaching a

{{00:01:45}}
point now with that use of

{{00:01:47}}
that system that we're scaling

{{00:01:48}}
it out across the district

{{00:01:49}}
and indeed multiple districts

{{00:01:51}}
and that's, you know, been

{{00:01:52}}
demonstrated through rigorous

{{00:01:54}}
clinical trials to have an

{{00:01:55}}
impact, a beneficial impact

{{00:01:57}}
on the quality of care we

{{00:01:59}}
can provide to patients.

{{00:02:00}}
And so now what we want to be

{{00:02:01}}
able to demonstrate is that

{{00:02:03}}
we can use this system at

{{00:02:04}}
scale to really increase the

{{00:02:06}}
productivity of women care

{{00:02:07}}
across the district, but more

{{00:02:09}}
broadly across the health

{{00:02:11}}
service in New South Wales.

{{00:02:13}}
You know, most of the

{{00:02:13}}
technologies, indeed,

{{00:02:15}}
probably 99 percent of them

{{00:02:16}}
at the moment are assistive.

{{00:02:18}}
And so clinicians, when

{{00:02:20}}
they use them, have to be

{{00:02:21}}
very aware of what they are

{{00:02:24}}
capable of doing and what

{{00:02:25}}
they are not capable of doing.

{{00:02:26}}
You know, a lot of the onus

{{00:02:27}}
is still put on clinicians

{{00:02:29}}
to understand what it is the

{{00:02:32}}
system can do, its limitations

{{00:02:34}}
and it's potential uses, you

{{00:02:37}}
know, and so they have to

{{00:02:39}}
be very familiar with the

{{00:02:40}}
capabilities of the technology

{{00:02:41}}
before they can put it into

{{00:02:43}}
practice, you know, and so

{{00:02:44}}
it's really understandable

{{00:02:46}}
that they would be

{{00:02:46}}
conservative about introducing

{{00:02:48}}
these new technologies into

{{00:02:49}}
practice; if they can't

{{00:02:51}}
necessarily interrogate

{{00:02:53}}
how they're working.

{{00:02:54}}
So the so called

{{00:02:54}}
black box problem.

{{00:02:55}}
So what we can do to prepare

{{00:02:57}}
them is to make sure that they

{{00:02:59}}
have opportunities to trial

{{00:03:00}}
these technologies in safe

{{00:03:01}}
spaces, that we rigorously

{{00:03:03}}
evaluate them through research

{{00:03:05}}
projects, that we give them

{{00:03:06}}
opportunities to skill up

{{00:03:08}}
through micro credentialing.

{{00:03:10}}
Through higher degree

{{00:03:11}}
research, through doing

{{00:03:12}}
coursework research, doing

{{00:03:15}}
capstone projects as part

{{00:03:17}}
of their clinical training.

{{00:03:18}}
You know, these are all

{{00:03:19}}
ways that we can help them

{{00:03:21}}
to get to grips with these,

{{00:03:23}}
how these technologies work.

{{00:03:24}}
We also, you know, I

{{00:03:25}}
think it's important that

{{00:03:26}}
we introduce new roles

{{00:03:27}}
into health services.

{{00:03:28}}
We've got one called the

{{00:03:29}}
Digital Patient Navigator,

{{00:03:31}}
specifically to try and

{{00:03:33}}
help patients navigate the

{{00:03:35}}
use of these technologies

{{00:03:36}}
on the patient side.

{{00:03:37}}
So there are ways that we

{{00:03:38}}
can help clinicians better

{{00:03:41}}
come to grips with the use

{{00:03:42}}
of these technologies and

{{00:03:43}}
use them safely and embed

{{00:03:45}}
them into their workflows.

{{00:03:46}}
You know, in health, you

{{00:03:47}}
know, we're very used

{{00:03:48}}
to being within a well

{{00:03:51}}
regulated work setting,

{{00:03:53}}
you know, we have extensive

{{00:03:54}}
regulatory frameworks,

{{00:03:56}}
data governance frameworks

{{00:03:57}}
within which we practice,

{{00:03:58}}
you know, and I think AI

{{00:04:00}}
is another technology.

{{00:04:01}}
It obviously presents some

{{00:04:02}}
unique challenges, but we

{{00:04:04}}
have well established clinical

{{00:04:06}}
governance and data governance

{{00:04:07}}
procedures within health

{{00:04:08}}
services like ours and in

{{00:04:10}}
most health services that

{{00:04:12}}
can absorb the introduction

{{00:04:14}}
of new technologies like

{{00:04:15}}
this, you know, and the work

{{00:04:16}}
that we will need to do.

{{00:04:18}}
is make sure that our

{{00:04:19}}
rigorous, our existing

{{00:04:21}}
rigorous processes can

{{00:04:22}}
adapt to work to regulate

{{00:04:25}}
and to introduce safely

{{00:04:27}}
new technologies and the,

{{00:04:28}}
and the unique challenges

{{00:04:29}}
posed by things like AI.

{{00:04:31}}
And, you know, they

{{00:04:32}}
have very real privacy,

{{00:04:34}}
potential impacts on

{{00:04:35}}
things like privacy.

{{00:04:36}}
You know, they're

{{00:04:37}}
very data hungry.

{{00:04:38}}
The tendency will be for us to

{{00:04:41}}
harness patient data in ways

{{00:04:43}}
that, you know, probably are

{{00:04:44}}
not generally well understood

{{00:04:46}}
by patients when they consent.

{{00:04:48}}
to undertaking care

{{00:04:49}}
within the health service.

{{00:04:50}}
And so, you know,

{{00:04:50}}
there's an onus.

{{00:04:52}}
services to make sure we're

{{00:04:54}}
safeguarding the data of

{{00:04:57}}
our communities, you know,

{{00:04:58}}
and making sure we're

{{00:04:59}}
operating within what sort

{{00:05:00}}
of social license  and we're

{{00:05:03}}
adhering to the procedures

{{00:05:04}}
and protocols of our data

{{00:05:05}}
governance frameworks.

{{00:05:06}}
You know, that's not,

{{00:05:07}}
I mean, that's easier

{{00:05:08}}
said than done, though.

{{00:05:09}}
You know, I think the rapidity

{{00:05:11}}
with which the technology

{{00:05:13}}
evolves and the sort of the

{{00:05:16}}
unique capabilities associated

{{00:05:18}}
with AI pose real challenges

{{00:05:19}}
for our existing regulatory

{{00:05:21}}
and governance processes

{{00:05:22}}
and that's where we need to

{{00:05:23}}
put a lot of time and effort

{{00:05:23}}
into trying to make sure

{{00:05:25}}
they're fit for purpose.

{{00:05:26}}
And I think that's a huge

{{00:05:28}}
potential for technologies

{{00:05:30}}
like AI, to make sure

{{00:05:32}}
we can better tailor the

{{00:05:34}}
experiences within our health

{{00:05:36}}
services to the unique needs

{{00:05:38}}
of our patients and our

{{00:05:38}}
carers and their families.

{{00:05:40}}
You know, in Sydney

{{00:05:41}}
Local Health District,

{{00:05:41}}
we're doing a couple of

{{00:05:42}}
really important things.

{{00:05:44}}
to really realize that

{{00:05:46}}
potential one is our

{{00:05:47}}
Florence program, which is

{{00:05:49}}
really looking to put the

{{00:05:51}}
control back in the hands of

{{00:05:52}}
patients about booking their

{{00:05:53}}
appointments, rescheduling

{{00:05:56}}
their appointments,

{{00:05:56}}
managing their referrals

{{00:05:57}}
to outpatient clinics.

{{00:05:59}}
And that's really, you know,

{{00:06:01}}
a front door for the health

{{00:06:02}}
service where patients can get

{{00:06:05}}
all the information they need

{{00:06:06}}
to navigate through the health

{{00:06:07}}
system, you know, which can

{{00:06:08}}
be incredibly confusing for

{{00:06:10}}
patients and their families.

{{00:06:11}}
So that's one key way.

{{00:06:12}}
Another key way is the

{{00:06:13}}
development of a custom app

{{00:06:15}}
called MyCare, where we're

{{00:06:17}}
really trying to enable

{{00:06:19}}
patients to use their own

{{00:06:20}}
devices to enable us to

{{00:06:23}}
remotely monitor them in

{{00:06:24}}
their own homes, you know,

{{00:06:25}}
and that's a critical way

{{00:06:26}}
we're putting the power

{{00:06:27}}
back in patients hands  to

{{00:06:30}}
participate in their own care.

{{00:06:31}}
You know, I've spent a lot

{{00:06:32}}
of time thinking about, and

{{00:06:33}}
there's a very real risk

{{00:06:35}}
with AI technologies, I

{{00:06:36}}
think, that they'll, their

{{00:06:37}}
introduction and their

{{00:06:38}}
scale, their industrial use.

{{00:06:41}}
across the entire health

{{00:06:42}}
system poses a very

{{00:06:43}}
real risk to equity.

{{00:06:45}}
You know, I think that we

{{00:06:47}}
don't really understand

{{00:06:48}}
well enough the risks to

{{00:06:49}}
equity posed by some of

{{00:06:50}}
these technologies, you

{{00:06:51}}
know, particularly things

{{00:06:52}}
like misrepresentation,

{{00:06:54}}
you know, there's well

{{00:06:55}}
known that clinical trials

{{00:06:57}}
have very, very significant

{{00:06:59}}
issues with bias in their

{{00:07:00}}
findings, and we're building

{{00:07:01}}
a lot of our AI systems based

{{00:07:03}}
on clinical trial findings

{{00:07:06}}
that sort of perpetuate and

{{00:07:07}}
amplify disparities that are

{{00:07:10}}
embedded within those trials.

{{00:07:11}}
And so we have to be much

{{00:07:12}}
more conscious of that.

{{00:07:13}}
The principles, you know, that

{{00:07:16}}
exist within the new sort of

{{00:07:17}}
responsible AI frameworks are

{{00:07:19}}
all well and good, but the

{{00:07:21}}
challenge will be actually

{{00:07:22}}
implementing them in practice,

{{00:07:24}}
operationalizing them.

{{00:07:25}}
You know, that's going to

{{00:07:26}}
be a monumental effort.

{{00:07:27}}
You know, and the example I

{{00:07:29}}
always use is, you know, in

{{00:07:31}}
cyber security, where there is

{{00:07:34}}
absolutely no question about

{{00:07:35}}
the criticality of something

{{00:07:36}}
like Making sure your

{{00:07:38}}
organization is cyber secure

{{00:07:39}}
and there are well established

{{00:07:41}}
standards and frameworks in

{{00:07:42}}
that space like ISO 27001.

{{00:07:44}}
But health services like ours

{{00:07:46}}
still struggle to implement

{{00:07:49}}
even those well known, well

{{00:07:51}}
established cyber security

{{00:07:52}}
frameworks, you know.

{{00:07:53}}
And so, I think we have

{{00:07:54}}
to face up to the reality

{{00:07:56}}
of the gap between very

{{00:07:59}}
ambitious ethical principles.

{{00:08:02}}
around the responsible use of

{{00:08:03}}
AI and the actual challenge of

{{00:08:04}}
making sure we realize those

{{00:08:06}}
in practice and in operations.

{{00:08:07}}
You know, that's gonna require

{{00:08:08}}
a lot, a lot of focus, you

{{00:08:10}}
know, and I think the other

{{00:08:11}}
key risk is that the drive

{{00:08:13}}
to increase productivity

{{00:08:14}}
and efficiency in health

{{00:08:15}}
care will mean we're less

{{00:08:18}}
conscious of the risks to

{{00:08:20}}
patient quality to patient

{{00:08:22}}
experience and to equity as

{{00:08:24}}
we sort of resolve things

{{00:08:26}}
like staff workforce shortages

{{00:08:28}}
with AI you know, I think

{{00:08:29}}
We still have to be attuned

{{00:08:31}}
to those risks to equity.

Expert Insight

Don't miss this insightful interview with Mitchell Burger, Director of Strategy, Architecture, Innovation, and Research at Sydney Local Health District (LHD). 

In this video, Mitchell delves into the possibilities and obstacles of integrating AI into healthcare. He sheds light on AI's capacity to revolutionize patient care and optimize clinical processes, spotlighting their achievements in wound care with plans for a broader district-wide implementation.


healthcare

digital health

thumbnail_1601009970966.jpg

xsmall_1601009970966.jpg

1601009970966.jpg

Mitchell Burger

Director of Strategy, Architecture, Innovation, and Research

Revolutionizing healthcare with innovative AI Applications

Empowering healthcare providers with AI for enhanced decision-making and diagnostics