February 8, 2025 • 12 mins
There's plenty of cynicism and scepticism about the future of AI, but experts are hopeful it could reshape one key industry.
AI is already being utilised by the healthcare sector to process data and analytics, and physicians are using the technology to carry out admin tasks.
However, professor and AI expert John Hirdes says there's room for AI to revolutionise the healthcare sector, particularly in aged care.
"AI systems can help with advanced analytics around that data to let us understand more about the challenges associated with aging and the ways the health system can respond to that."
LISTEN ABOVE
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:06):
You're listening to the Sunday Session podcast with Francesca Rudkin
from News Talks atb.
Speaker 2 (00:12):
Now, when it comes to artificial intelligence, we can be
a bit wary and focus on the negative, can't we?
But there are real benefits to AI. In One industry
where experts see transformative potential is healthcare. This week, the
University of Otago has hosted Distinguished Professor John Herdies from Canada.
John is a world expert on the use of AI
in the healthcare of elderly and he's still in the
(00:33):
country and we've managed to track him down.
Speaker 1 (00:35):
John.
Speaker 2 (00:35):
Good morning, Hi, Hi, good morning. Thank you so much
for being with us. Can you tell me, first of all,
how widely is AI currently being used within the field
of health in general.
Speaker 3 (00:48):
In its early stages, I would say so, it's being
used in diagnostics to some extent. It's certainly being used
in data analytics. It's beginning to be used in primary
care by physicians as a note ticking and clinical summary device.
But it's still early days for now.
Speaker 2 (01:10):
So what is its potential in the future.
Speaker 3 (01:13):
The potential is pretty substantial, I think, you know, the
introduction of AI into our society in general, but into
healthcare specifically, will be a bigger change than going from
landlines to cell phones or from fax machines to computers.
This really can be transformative and it will be every
(01:35):
aspect of the healthcare system. It can affect everything from
person level use by patients and the family to clinical
and management use and use by policymakers.
Speaker 2 (01:46):
Yeah, I was going to ask you about the areas
that it could transform health cany we do you see
it having the biggest impact? Do you think?
Speaker 3 (01:55):
I think the most substantial impact right away will be
around the management and use of complex data systems. So
in my country, Can and New Zealand, we actually have
put in place very sophisticated information systems in the care
home sector and in community based home care services, and
(02:17):
AI systems can help with advanced analytics around those data
to let us understand more about the challenges associated with
aging and the way the health system can respond to that.
So that's the immediate area of impact. Around diagnostics, for sure,
around coding of health information like diagnostic information in hospitals,
(02:41):
I think there's some potential. I'll be a bit more
nervous around some of the clinical applications in care of
older people or persons with disabilities that we can talk
a bit about morphylites.
Speaker 2 (02:52):
Yeah, and I'd like to because your focus is on
older generations, which is fantastic for New Zealand because we
have an aging population. So how can we use this
technology to improve the care and the outcomes of older people.
Speaker 3 (03:07):
Well, the thing that we can do immediately is to
use data systems that we have to better target older
people for specific types of services, to provide a more personalized,
research based approach to care, to identify people at risk
of falls or people who have complex needs that require
(03:27):
support for them and their caregivers. We have very sophisticated
tools we can use for that. We're under utilizing that
information clinically because sometimes all this stuff is a bit
complicated to organize and to communicate, and so using AI
systems to translate all this technical terminology into something that
elderly people in their family can understand and also summarizing
(03:50):
in a useful way for clinicians to act on would
be an easy when to achieve. There are some other
types of technologies that are emerging that could be used
to increase seniors independence or reduce risk. So some sensor
based technology might be helpful and for example, detecting unsteady
(04:12):
gait and risk of falls. But one of the challenges
with sensors is that sometimes elderly people view them as
intrusive to their privacy, and sometimes they don't work all
that well. Yet, an area that I'd be most concerned
about is the substitution of human based care with AI
(04:35):
based technologies. So the example I'll give you four that
would be around loneliness. Loneliness is an important mental health
problem affecting many older people, and there's sometimes an idea that, well,
we could just create a system, an AI companion to
make elderly people feel less lonely. I'd rather find ways
(04:58):
that we could mobilize humans to work with elderly people
to make them feel less lonely.
Speaker 2 (05:03):
Yeah, that's a good point. Going back to the first
thing you mean to the data, is that a matter
of been able to bring together a lot of information
from different people who may be involved in that care
to to make sure that they're getting the right care.
Is that how that works. That's sort of taking all
that data and pulling it together.
Speaker 3 (05:22):
Yeah. So in New Zealand, In New Zealand and Canada,
an elderly person, for example, getting community based home care
will be assessed with a comprehensive assessment system that my
research team developed. It has about three hundred and fifty
data elements covering all aspects of the person's life. It
tracks them over time, and it's a bit challenging to
(05:44):
make sense of all that. My research team has done
a lot of work to summarize those data to identify
risk of falls or hospitalization, or mortality or cognitive decline,
but sometimes clinicians find it's hard to interpret all those
data elements at once, and we don't do a very
(06:04):
good job at helping older people people and their family
members understand the results of those assessments and what they
can do about it. That's where an AI system I
think could be helpful.
Speaker 2 (06:13):
Yeah, makes sense. Is it these kind of things more
useful in that older demographic or do you think we'll
see widespread benefit from it? I mean, obviously that idea
of using the data can be started at any at
any age count it absolutely.
Speaker 3 (06:29):
You know, we don't have as good data systems in
place for younger individuals or persons with mental health needs
or disabilities as we do with age care. But you know,
the future of healthcare is really a future where we
have data rich environments. So, for example, when it comes
to older persons, we can link their clinical assessment data
(06:51):
to geospatial information about poverty or crime rates, or environmental
characteristics in a neighborhood. If we look at use of
smart watchers, there's a great deal of information that you
get in say an Apple watch or a garment and
watch that could be useful. And then genetic information is
another type of information that's available. Imagine being able to
(07:14):
put all of that together into a single data source
that can be interpreted to help understand immediate and longer
term health risks. That is absolutely possible, but the analytic
task is very challenging.
Speaker 2 (07:28):
John, you mentioned things they're like our smart watches. How
do we know that the information AI provides is going
to be reliable and accurate?
Speaker 3 (07:37):
The short answer is we don't. We should assume that
the AI will have errors attached to it. In my
own experimentation with AI systems in areas where I have expertise,
I find it's right about ninety eight percent of the time.
But what I don't know is about that two percent
(07:57):
of the time. So when I've looked at it in
terms of understanding and describing fall risks or depression or
cognitive impairment. It does a really good job. But I'm
not in a position to know if I was to
ask it, how should I design an airplane that I
can personally used to fly around the country, that it's
going to give me correct information? And so what that
(08:20):
means is that we do need to have people with
expertise that can interpret the evidence from an AI system
and help us make sense of it. A challenge that
we have is there's not one AI system out there.
There's not one massive AI hovering over the earth that
is answering all questions. In fact, there are already hundreds
(08:40):
of different AI solutions out there, and they're not all
created equal. Some are better than others, Some are designed
to try to give answers as clearly as possible, Some
are frankly being designed to give answers that are biased
in a specific direction. So we're in a bit of
a conundrum that AI systems are very useful to experts
(09:03):
like me to very quickly summarize information, but for a
lay person, I would still have concerns about the veracity
of any given AI output.
Speaker 2 (09:15):
Interesting. You mentioned that the sensors are still whiping down
there and your concern around the substitution of human care
for AI. But do you see any other downside to
using AI and healthcare.
Speaker 3 (09:28):
Well, it comes down to what we want the AI
system to do for us. So there's an organization called
the Institute for health Care Improvement that says we should
try to achieve five aims in any intervention in healthcare.
We should improve patient outcomes, we should improve equity in outcomes,
we should improve the experience of care when we're receiving care,
(09:53):
we should improve cost effectiveness, and we should improve staff experience.
If we can do all those five things, then we
have the potential to improve quality of care. What we
should be doing is looking at in any value of
an AI systema seed, to what extent does it meet
those objectives. It could meet one but not the other objectives,
(10:13):
and that may not necessarily be helpful. So, for example,
we've developed a targeting mechanism to identify elderly people who
may have a delayed discharge from hospital because of the
complexity of their needs. The intended use of that is
to identify people early so that the hospital discharge planning
staff can intervene early on to put in place the
(10:35):
proper supports to help the person be discharged effectively to
an appropriate level of care. The negative use of it
is that it could be used to deny services. And
it's the same algorithm, it's just a matter of what
is the intended use behind it. So we have to
think about applying guard rails and ethical principles to the
(10:56):
use of the AI system so that they meet the
needs of elderly people and the general population at large.
Speaker 2 (11:01):
Because yeah, and I mean, the main criteria surely would
be that it would increase the quality care.
Speaker 3 (11:07):
Well, you and I agree on that that quality should
be a first priority, but a government may say cost
effectiveness has to be a priority, right, And so if
we look at the case of the delayed discharge, the
perspective of the person really matters. So the elderly person
will say, I want to be able to go home.
(11:28):
I want to get good care. The person's family member
may say, I'm worried about the supports my family members
going to need and whether I can meet all my
competing demands of my time. The hospital may think I've
got to reduce my length of stay and I've got
to empty out some beds so that new patients can
be admitted and people can get access to medical beds.
(11:48):
And the government may say I need to reduce taxes
for my population. Those are competing priorities and we have
to decide whose perspective gets priority.
Speaker 2 (11:56):
John, really nice to talk to you. Thank you so
much for your time today. That was University at the
University of Otago has been hosting John. He is a
distinguished professor from Canada. There talking about the use of
AI in healthcare, especially with the elderly, and he's been
working on this in Canada where he's seen significant improvements
in the lives of older people.
Speaker 1 (12:18):
For more from the Sunday session with Francesca Rudkin, listen
live to News Talks it Be from nine am Sunday,
or follow the podcast on iHeartRadio.
You're listening to the Sunday Session podcast with Francesca Rudkin
from News Talks atb.
Speaker 2 (00:12):
Now, when it comes to artificial intelligence, we can be
a bit wary and focus on the negative, can't we?
But there are real benefits to AI. In One industry
where experts see transformative potential is healthcare. This week, the
University of Otago has hosted Distinguished Professor John Herdies from Canada.
John is a world expert on the use of AI
in the healthcare of elderly and he's still in the
(00:33):
country and we've managed to track him down.
Speaker 1 (00:35):
John.
Speaker 2 (00:35):
Good morning, Hi, Hi, good morning. Thank you so much
for being with us. Can you tell me, first of all,
how widely is AI currently being used within the field
of health in general.
Speaker 3 (00:48):
In its early stages, I would say so, it's being
used in diagnostics to some extent. It's certainly being used
in data analytics. It's beginning to be used in primary
care by physicians as a note ticking and clinical summary device.
But it's still early days for now.
Speaker 2 (01:10):
So what is its potential in the future.
Speaker 3 (01:13):
The potential is pretty substantial, I think, you know, the
introduction of AI into our society in general, but into
healthcare specifically, will be a bigger change than going from
landlines to cell phones or from fax machines to computers.
This really can be transformative and it will be every
(01:35):
aspect of the healthcare system. It can affect everything from
person level use by patients and the family to clinical
and management use and use by policymakers.
Speaker 2 (01:46):
Yeah, I was going to ask you about the areas
that it could transform health cany we do you see
it having the biggest impact? Do you think?
Speaker 3 (01:55):
I think the most substantial impact right away will be
around the management and use of complex data systems. So
in my country, Can and New Zealand, we actually have
put in place very sophisticated information systems in the care
home sector and in community based home care services, and
(02:17):
AI systems can help with advanced analytics around those data
to let us understand more about the challenges associated with
aging and the way the health system can respond to that.
So that's the immediate area of impact. Around diagnostics, for sure,
around coding of health information like diagnostic information in hospitals,
(02:41):
I think there's some potential. I'll be a bit more
nervous around some of the clinical applications in care of
older people or persons with disabilities that we can talk
a bit about morphylites.
Speaker 2 (02:52):
Yeah, and I'd like to because your focus is on
older generations, which is fantastic for New Zealand because we
have an aging population. So how can we use this
technology to improve the care and the outcomes of older people.
Speaker 3 (03:07):
Well, the thing that we can do immediately is to
use data systems that we have to better target older
people for specific types of services, to provide a more personalized,
research based approach to care, to identify people at risk
of falls or people who have complex needs that require
(03:27):
support for them and their caregivers. We have very sophisticated
tools we can use for that. We're under utilizing that
information clinically because sometimes all this stuff is a bit
complicated to organize and to communicate, and so using AI
systems to translate all this technical terminology into something that
elderly people in their family can understand and also summarizing
(03:50):
in a useful way for clinicians to act on would
be an easy when to achieve. There are some other
types of technologies that are emerging that could be used
to increase seniors independence or reduce risk. So some sensor
based technology might be helpful and for example, detecting unsteady
(04:12):
gait and risk of falls. But one of the challenges
with sensors is that sometimes elderly people view them as
intrusive to their privacy, and sometimes they don't work all
that well. Yet, an area that I'd be most concerned
about is the substitution of human based care with AI
(04:35):
based technologies. So the example I'll give you four that
would be around loneliness. Loneliness is an important mental health
problem affecting many older people, and there's sometimes an idea that, well,
we could just create a system, an AI companion to
make elderly people feel less lonely. I'd rather find ways
(04:58):
that we could mobilize humans to work with elderly people
to make them feel less lonely.
Speaker 2 (05:03):
Yeah, that's a good point. Going back to the first
thing you mean to the data, is that a matter
of been able to bring together a lot of information
from different people who may be involved in that care
to to make sure that they're getting the right care.
Is that how that works. That's sort of taking all
that data and pulling it together.
Speaker 3 (05:22):
Yeah. So in New Zealand, In New Zealand and Canada,
an elderly person, for example, getting community based home care
will be assessed with a comprehensive assessment system that my
research team developed. It has about three hundred and fifty
data elements covering all aspects of the person's life. It
tracks them over time, and it's a bit challenging to
(05:44):
make sense of all that. My research team has done
a lot of work to summarize those data to identify
risk of falls or hospitalization, or mortality or cognitive decline,
but sometimes clinicians find it's hard to interpret all those
data elements at once, and we don't do a very
(06:04):
good job at helping older people people and their family
members understand the results of those assessments and what they
can do about it. That's where an AI system I
think could be helpful.
Speaker 2 (06:13):
Yeah, makes sense. Is it these kind of things more
useful in that older demographic or do you think we'll
see widespread benefit from it? I mean, obviously that idea
of using the data can be started at any at
any age count it absolutely.
Speaker 3 (06:29):
You know, we don't have as good data systems in
place for younger individuals or persons with mental health needs
or disabilities as we do with age care. But you know,
the future of healthcare is really a future where we
have data rich environments. So, for example, when it comes
to older persons, we can link their clinical assessment data
(06:51):
to geospatial information about poverty or crime rates, or environmental
characteristics in a neighborhood. If we look at use of
smart watchers, there's a great deal of information that you
get in say an Apple watch or a garment and
watch that could be useful. And then genetic information is
another type of information that's available. Imagine being able to
(07:14):
put all of that together into a single data source
that can be interpreted to help understand immediate and longer
term health risks. That is absolutely possible, but the analytic
task is very challenging.
Speaker 2 (07:28):
John, you mentioned things they're like our smart watches. How
do we know that the information AI provides is going
to be reliable and accurate?
Speaker 3 (07:37):
The short answer is we don't. We should assume that
the AI will have errors attached to it. In my
own experimentation with AI systems in areas where I have expertise,
I find it's right about ninety eight percent of the time.
But what I don't know is about that two percent
(07:57):
of the time. So when I've looked at it in
terms of understanding and describing fall risks or depression or
cognitive impairment. It does a really good job. But I'm
not in a position to know if I was to
ask it, how should I design an airplane that I
can personally used to fly around the country, that it's
going to give me correct information? And so what that
(08:20):
means is that we do need to have people with
expertise that can interpret the evidence from an AI system
and help us make sense of it. A challenge that
we have is there's not one AI system out there.
There's not one massive AI hovering over the earth that
is answering all questions. In fact, there are already hundreds
(08:40):
of different AI solutions out there, and they're not all
created equal. Some are better than others, Some are designed
to try to give answers as clearly as possible, Some
are frankly being designed to give answers that are biased
in a specific direction. So we're in a bit of
a conundrum that AI systems are very useful to experts
(09:03):
like me to very quickly summarize information, but for a
lay person, I would still have concerns about the veracity
of any given AI output.
Speaker 2 (09:15):
Interesting. You mentioned that the sensors are still whiping down
there and your concern around the substitution of human care
for AI. But do you see any other downside to
using AI and healthcare.
Speaker 3 (09:28):
Well, it comes down to what we want the AI
system to do for us. So there's an organization called
the Institute for health Care Improvement that says we should
try to achieve five aims in any intervention in healthcare.
We should improve patient outcomes, we should improve equity in outcomes,
we should improve the experience of care when we're receiving care,
(09:53):
we should improve cost effectiveness, and we should improve staff experience.
If we can do all those five things, then we
have the potential to improve quality of care. What we
should be doing is looking at in any value of
an AI systema seed, to what extent does it meet
those objectives. It could meet one but not the other objectives,
(10:13):
and that may not necessarily be helpful. So, for example,
we've developed a targeting mechanism to identify elderly people who
may have a delayed discharge from hospital because of the
complexity of their needs. The intended use of that is
to identify people early so that the hospital discharge planning
staff can intervene early on to put in place the
(10:35):
proper supports to help the person be discharged effectively to
an appropriate level of care. The negative use of it
is that it could be used to deny services. And
it's the same algorithm, it's just a matter of what
is the intended use behind it. So we have to
think about applying guard rails and ethical principles to the
(10:56):
use of the AI system so that they meet the
needs of elderly people and the general population at large.
Speaker 2 (11:01):
Because yeah, and I mean, the main criteria surely would
be that it would increase the quality care.
Speaker 3 (11:07):
Well, you and I agree on that that quality should
be a first priority, but a government may say cost
effectiveness has to be a priority, right, And so if
we look at the case of the delayed discharge, the
perspective of the person really matters. So the elderly person
will say, I want to be able to go home.
(11:28):
I want to get good care. The person's family member
may say, I'm worried about the supports my family members
going to need and whether I can meet all my
competing demands of my time. The hospital may think I've
got to reduce my length of stay and I've got
to empty out some beds so that new patients can
be admitted and people can get access to medical beds.
(11:48):
And the government may say I need to reduce taxes
for my population. Those are competing priorities and we have
to decide whose perspective gets priority.
Speaker 2 (11:56):
John, really nice to talk to you. Thank you so
much for your time today. That was University at the
University of Otago has been hosting John. He is a
distinguished professor from Canada. There talking about the use of
AI in healthcare, especially with the elderly, and he's been
working on this in Canada where he's seen significant improvements
in the lives of older people.
Speaker 1 (12:18):
For more from the Sunday session with Francesca Rudkin, listen
live to News Talks it Be from nine am Sunday,
or follow the podcast on iHeartRadio.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
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.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com