More Related Content More from Bernard Marr (20) The Amazing Ways Elsevier Uses Artificial Intelligence – Improving Healthcare And Cancer Diagnosis1. The Amazing Ways Elsevier Uses
Artificial Intelligence – Improving
Healthcare And Cancer Diagnosis
2. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Title
Text
IntroductionIntroduction
In my first focus article on Elsevier I talked about how one of the world’s largest
providers of scientific and medical information is managing its digital
transformation.
It started by digitising the vast troves of data it has generated in its 140-year
history through publications such as The Lancet and Cell. These were then
combined with new Big Data sources such as user demographics and behavioural
data.
The Amazing Ways Elsevier Uses Artificial Intelligence –
Improving Healthcare And
Cancer Diagnosis
3. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Title
Text
IntroductionIntroduction
The final step was to build cutting-edge analytics tools capable of aggregating
those datasets and drawing out insights, in order to create products that would
make it easier for scientists, medical professionals, lawyers and researchers to put
the information to work.
In this presentation I will focus specifically on Artificial Intelligence.
The Amazing Ways Elsevier Uses Artificial Intelligence –
Improving Healthcare And
Cancer Diagnosis
4. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Smart Computing Technologies
Just a few years ago Artificial Intelligence might have been a phrase we were
most likely to come across in science fiction. Today, however, it is used to
describe an emerging breed of business software tools which are capable of
ingesting large volumes of data, understanding it, and learning from it how to
become more efficient at generating insights.
What has made this possible is advances in “smart” computing technologies
such as machine learning and deep learning. The theory behind machine
learning – software algorithms which are able to adapt themselves based on
data they are fed – has been around for a while, at least since the 1960s. Its
only recently, however, that there has been enough data, as well as enough
processing power, to put it to use.
5. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Smart Computing Technologies
As soon as those milestones were reached, it was immediately obvious how
potentially powerful and useful this technology was. This is the reason that
ideas which just a few years ago would have been considered outlandishly
futuristic, such as self-driving cars and real-time translation devices – are now a
reality.
6. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Clinical Decision Support
When considering how to make AI work for Elsevier and its customers, as
publishers of 25% of the world’s medical literature, healthcare was an obvious
starting point.
Its mission here is to build what are described as “clinical decision support”
systems. These use information gathered from Elsevier’s archives, combined
with patient medical data and financial claims data, to suggest the best course
of action – known as a “pathway” – for treating specific patients.
The next step of the process is to augment these decision support systems
with machine learning and deep learning. This should mean more accurate
predictions, more efficient treatment and ultimately better patient outcomes.
7. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Clinical Decision Support
John Danaher, Elsevier’s president of clinical solutions, tells me “If you think
about a healthcare enterprise like a hospital, you’ve got radiology
departments, pathology departments – all these different areas which are
running tests, doing studies – they’ve always generated a huge amount of data
but traditionally its all been paper based.
“What we have now is this tremendous penetration of hospital delivery
systems, capturing all of this information digitally – there’s administrative
information, clinical information, insurance claims – then there’s the whole area
of genomic data.
8. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Clinical Decision Support
Danaher tells me that when you aggregate this data the value comes from two
ways:
1. The large amounts of data increase predictability and produce more
accurate models with better inferences.
2. You get a much more accurate understanding of individual patients by
bringing together data their clinical history, their claims data, genomic data,
etc.
Elsevier’s technology is now in use at 2,000 oncology centres at healthcare
facilities across the US, where it informs decisions on patient healthcare every
day, and the adherence rate, i.e. the degree at with doctors follow that advice,
is very high.
9. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Clinical Decision Support
“We get over 85% adherence to our pathways by our clinicians, and when they
do go off pathway, which happens sometimes – patients may have allergies to
certain medicines – we review and look at the reasons for them going off
pathway and if necessary review our decision making,” Danaher says.
Building on this, Elsevier is now building artificial intelligence technology into
its predictive platform.
Based on all their content published in their journal articles, books, etc.,
Elsevier have mapped diseases to symptoms, to create some predictive
models.
10. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Neural Networks
“We were able to create these neural networks of closed loops, and we trained
these predictive models against large patient databases … so we built an
application what generates a differential diagnosis based on that model. It
gives you a weighted differential, it says, with those symptoms in a person of
this age and gender, there’s a 70% chance it’s this, a 35% chance it’s that.”
“Neural networks” refers to a technology designed to mimic the learning
characteristics of the neurons in the human brain. Data is passed between
neurons which essentially each ask a different question about it. The results of
all of those questions are aggregated into a single output – representing the
answer to the question that the neural network is designed to solve.
11. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Neural Networks
For example, in a medical application, a neural network could be designed to
process patient data in order to produce a probabilistic answer to the question
“Does this scan result suggest the patient has cancer?”
Using the patient’s medical records as its input data, different neurons in the
network could ask “what is the age of this patient?”, “what is the sex of this
patient?”, “Does this patient has a previous diagnosis for cancer?”, “what is the
alcohol intake of this patient?”, “is this patient a smoker?”- and many other
questions.
Based on what it knows from other cases it has been trained on – and Elsevier’s
systems are trained on around five million diagnoses as well as insurance
claims – it will assess the likelihood of a particular scan result indicating that a
diagnosis of cancer should be made.
12. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Neural Networks
The really clever part is that a neural network is designed to get smarter over
time, by learning, as it processes more and more data. It learns to assign
greater weight to variables which it discovers are more likely to be indicators
that a diagnosis should be made, or if the patient is lucky, ruled out.
Danaher says “You can see the ramifications for how people will do clinical
research in the future too – its all going to be driven by these analytics.”
The idea is that as well as improving patient outcomes in primary care
scenarios, widespread adoption of this technology will hugely speed up the
time it takes for new treatments and medicines to be made available.
13. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Neural Networks
This is typically a multi-year process – a timespan of 10 years from the first
tests showing that a treatment is effective, to it being made widely available, is
not uncommon.
It’s also hugely expensive, requiring long-term studies carried out by trained
healthcare professionals and genuine patients.
After that, doctors and student doctors have to be taught how to administer
the treatments.
14. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Discovery and Deployment
Machine learning, driven by Big Data, gives us the potential to drastically cut
the lag between discovery and deployment, Danaher tells me.
“So what can happen now, based on aggregated patient data, is you can take
1,000 women treated in one hospital, and 1,000 from another and you can take
that data, normalise it, and run analytics against it.
“So what you’ve done is, in the space of about a day, you’ve got actual patient
data that you can make assertions on – and say, for example, ‘if you use this
therapeutic drug, or do all of those things in a certain sequence, this cohort of
1,000 patients will live 20% longer, and have 15% fewer side effects.”
15. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Discovery and Deployment
Elsevier’s work is a prime example of how, once a business has passed the first
hurdle on the road to digital transformation – the digitisation of its data assets
– those assets can be repurposed through advanced analytics and AI in order
to derive new value.
It is now in talks with leading academic medical centres, including Stanford, to
put the AI component of its decision support platform – known as the
Precision Medicine Differential Diagnosis Tool – to commercial use.
Danaher says “We’ve built the application, it’s generating differential diagnoses
so the two places we’re going are moving into a minimal viable product –
getting it deployed in a clinical setting – and then what we’re doing is
investing in training these engines further against larger patient data sets.”
16. © 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Discovery and Deployment
The long term result they are hoping for is that the technology will both lead
to quicker and more accurate diagnosing, as well as meaning that the journey
from encouraging discovery to delivery of life-saving new treatments will be
hugely decreased.
17. © 2017 Bernard Marr , Bernard Marr & Co. All rights reserved
© 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Bernard Marr is an internationally best-selling author, popular keynote speaker, futurist, and a
strategic business & technology advisor to governments and companies. He helps
organisations improve their business performance, use data more intelligently, and
understand the implications of new technologies such as artificial intelligence, big data,
blockchains, and the Internet of Things.
LinkedIn has ranked Bernard as one of the world’s top 5 business influencers. He is a frequent
contributor to the World Economic Forum and writes a regular column for Forbes. Every day
Bernard actively engages his 1.5 million social media followers and shares content that
reaches millions of readers.
Visit The
Website
© 2017 Bernard Marr , Bernard Marr & Co. All rights reserved
© 2018 Bernard Marr, Bernard Marr & Co. All rights reserved
Bernard Marr is an internationally best-selling author, popular keynote speaker, futurist, and a
strategic business & technology advisor to governments and companies. He helps
organisations improve their business performance, use data more intelligently, and
understand the implications of new technologies such as artificial intelligence, big data,
blockchains, and the Internet of Things.
LinkedIn has ranked Bernard as one of the world’s top 5 business influencers. He is a frequent
contributor to the World Economic Forum and writes a regular column for Forbes. Every day
Bernard actively engages his 1.5 million social media followers and shares content that
reaches millions of readers.
Visit The
Website
18. Title
Subtitle
Be the FIRST to receive news,
articles, insights and event
updates from Bernard Marr & Co
straight to your inbox.
Signing up is EASY! Simply fill out
the online form and we’ll be in
touch!
© 2018 Bernard Marr, Bernard Marr & Co. All rights reserved