You are on page 1of 88

LEADING THE WAY IN HEALTH IT RESEARCH

ANNUAL REPORT
2017 / 2018

RESEARCH . DISCOVERY . INNOVATION


P O W E R S H E A LT H I N T H E D I G I TA L A G E
CONTENTS
The Australian e-Health Research Centre 4

Foreword by the CEO and Chairman 5

Research and Investment Advisory Committee 7

Board of Directors 8

2018 AEHRC Colloquium 10

Management and Research Leadership 12

News and Awards 14

Research Program 15
Health Informatics 16
Biomedical Informatics 30
Health Services 44

The Australian Tele-health Research and Development Group 64

AEHRC Publications 2017-2018 74

AEHRC and e-Health Program Staff, Students and Visitors 81

Special Purpose Financial Report 82


4 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

THE AUSTRALIAN E-HEALTH


RESEARCH CENTRE
An unincorporated joint venture between CSIRO and the Queensland
Government, the Australian e-Health Research Centre (AEHRC)
is the leading national digital health research facility applying
information and communication technology to improve health
service delivery for Australians.
Established in 2003 with initial funding from Through its research program, the AEHRC
the Department of State Development and develops and deploys leading edge
CSIRO, the partnership was extended in information and communication technology
2007 for a further five years with funding innovations in healthcare to:
from CSIRO, Queensland Health and the • improve service delivery in the
Department of Employment, Economic Queensland and Australian health
Development and Innovation. The systems
partnership was extended again in 2012
• g
 enerate commercialisation revenue,
for a further five years with an additional
and
contribution of $15 million from CSIRO
and Queensland Health, supplemented by • increase the pool of world-class
in-kind contributions from the partners, e-health expertise in Australia.
as well as funding from grants, research
The AEHRC’s multidisciplinary team
consulting and commercialisation. The
conducts research across health
partnership was extended again in 2017 for
informatics, biomedical informatics
a further five years.
and health services, and includes
As CSIRO’s e-Health Research Program – internationally prominent researchers,
part of CSIRO Health and Biosecurity – the software engineers and doctoral students,
Australian e-Health Research Centre has dedicated to serving the needs of patients,
grown to a national research centre. In clinicians and health service providers.
2009, the AEHRC established the Australian
Tele-health Research and Development
Group (ATRDG) in Perth in conjunction
with the Western Australia Department
of Health. While the initial focus of this
activity was on telemedicine and ocular
imaging technologies the group has grown
to include telemedicine delivery for a wide
range of conditions. Through further CSIRO
in-kind contributions and other external
funding the AEHRC now has scientists and
engineers in Sydney and Melbourne.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 5

FOREWORD BY THE CEO


AND CHAIRMAN
The past 12 months has been another successful year of growth and
impact for the Australian e-Health Research Centre. As the first year
of the new 2017-2022 Joint Venture Agreement between CSIRO and
Queensland Health, it was pleasing to see more adoption of AEHRC
technologies by Queensland Health and the launch of a number of
new projects and initiatives. This delivery demonstrates the value of
the relationship – with Queensland Health benefiting from the science
and innovation that CSIRO can bring and CSIRO benefiting from the
deep collaborations with Queensland Health clinician and health
services executives.
The AEHRC has continued to grow our Recently the AEHRC received funding
engagement and delivery around Australia, through the Cooperative Research Centre
with our teams in Brisbane, Sydney, (CRC) for Developing Northern Australia
Melbourne and Perth all growing over the to develop and trial referral pathways for
past 12 months. In addition to our strong remote ophthalmology, and a new CRC-P
relationship with Queensland Health through project was recently awarded with Brisbane-
our JV agreement, the AEHRC continues to based startup company Maxwell MRI.
work with Western Australia Health through
our joint Australian Tele-health Research The AEHRC is well positioned to engage
and Development Group in Perth. In Victoria in a number of new precision medicine
the AEHRC leads CSIRO’s involvement in initiatives funded through the Medical
the Melbourne Genomics Health Alliance Research Future Fund and other
project, partners with the Florey Institute on mechanisms. It has also been a key
a number of large Alzheimer’s disease trials member of the Queensland Genomics
and has a number of large projects with the Health Alliance, partnering with QIMR
Department of Health. Our Sydney-based Berghofer, University of Queensland
teams work with a number of partners in (UQ) and Queensland University of
New South Wales, with projects currently Technology (QUT) to lead the Genomic
being developed with e-Health NSW. Information Management Workstream.
This complements our contributions to the
There has been a significant increase Melbourne Genomics Health Alliance and
in the number of trials of AEHRC the Australian Genomics Health Alliance,
technologies with Queensland Health. and to the Global Alliance for Genomics
This includes the launch of a trial of a tool and Health – across all these genomics
for identifying the risk of readmission of projects, we are contributing clinical
patients on discharge. The information it informatics and bioinformatics capability
provides can be used by the health service and technology. Our bioinformatics work,
to plan support services for patients to with Denis Bauer’s team developing new
reduce this risk. There have also been genomic sequence analysis algorithms for
three new mobile health pilot trials over the new cloud computing architectures, is
the past 12 months. The PD-BUDDy receiving international attention.
trial supports patients with home based
peritoneal dialysis, the MoTHER trial We continue to deliver the National Clinical
supports women with gestational diabetes, Terminology Service with the Australian
and the Pain ROADMAP trial aims to Digital Health Agency. The NCTS is a
support people with strategies for chronic key part of Australia’s national digital
pain. There are many other trials going health infrastructure with more than
on with Queensland Health clinicians and 50 organisations licensing the AEHRC
researchers, including in diseases such terminology server, Ontoserver, to support
as Alzheimer’s disease and cerebral palsy. the use of standardised terminology in
The next 12 months will see the AEHRC digital health applications. This approach
undertake a number of new projects with has attracted international attention.
Queensland Health.

aehrc.com
6 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

FOREWORD BY THE CEO


AND CHAIRMAN (continued)

The Biomedical Informatics group The AEHRC was part of the external review
continues to lead medical image analysis of CSIRO Health and Biosecurity conducted
for national trials in areas such as by an external panel in November 2017.
Alzheimer’s disease and cerebral palsy The external review resulted in a strong
and clinical applications such as prostate result for the AEHRC, with the review panel
cancer. assessing the AEHRC as being Benchmark
in the areas of Impact and Innovation
Nationally the AEHRC is delivering Capacity and Strong to Benchmark in
significant projects for the Australian Science. This was a great endorsement of
Government Department of Health. Our the quality of the science and the impact
health data analytics scientists have that the AEHRC is having on healthcare in
developed the risk algorithm for the Australia.
Healthcare Homes program – now being
trialled in 200 clinics around Australia. The AEHRC is now an established part
We were recently awarded funding by the of Australia’s digital health ecosystem
Australian Government Department of – contributing to the digital health and
Health for a Primary Care Data Quality clinical research ecosystem. Going into
Foundations project, to work with industry the new financial year the AEHRC is well
and peak clinical bodies on standards for placed to continue to implement its 2017-
primary care data. These projects are in 2022 strategy and continue the growth of
addition to the Department of Aged Care the past five years.
funded program for our Smarter Safer
Homes Technology. We have also partnered
in the NHMRC-funded Australian Dementia
Network (ADNet).

Richard Royle David Hansen


Chairman, the Australian e-Health Research Centre Chief Executive Officer

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 7

RESEARCH AND INVESTMENT


ADVISORY COMMITTEE
Reporting to the Board of the Australian e-Health Research Centre, the Research and Investment
Advisory Committee (RIAC) performs an advisory function for the Centre’s research activities, and
assists the Board to carry out the functions of the Australian e-Health Research Centre.

RIAC Chair: Dr Michael Steyn A number of AEHRC staff members attended RIAC meetings
The following persons were members of the Research and during 2017-2018 as guest presenters.
Investment Advisory Committee throughout 2017-2018: 8 February 2017
Chair At this meeting the RIAC discussed the research from the
• D
 r Michael Steyn, Director Department of Anaesthesia Health Informatics group, discussing the opportunities for
& Perioperative Medicine, Royal Brisbane & Women’s further work with new initiatives within Queensland Health.
Hospital (2017) • Health data interoperability – Michael Lawley
• P
 rof Keith McNeil, Chief Clinical Information Officer, • Health text processing – Anthony Nguyen
Queensland Health
• Health data analytics – Sankalp Khanna
Members
19 July 2017
• D
 r Andrew Staib, Metro South Health and Hospital
Service (HHS) and e-Health Queensland At this meeting the RIAC discussed the 2017/18 AEHRC
Annual Business Plan, with considerable discussion on the
• M
 s Cathy Ford, Chief Digital Officer, e-Health adoption of and impact from AEHRC research.
Queensland, Queensland Health
• M
 r Michael Drahiem, Chief Information Officer, Metro • D
 iscussion on the AEHRC 2017-2018 project Portfolio –
South Health and Hospital Service David Hansen
• M
 r David Bunker, Executive Director, Queensland • Health services research – Mohan Karunanithi
Genomics Health Alliance • Health informatics research – Michael Lawley
• D
 r David Hansen, CEO, the Australian e-Health Research • Biomedical informatics research – Jurgen Fripp
Centre
• D
 r James Lind, Director, Emergency Medicine Training,
Gold Coast Hospital Emergency Department

aehrc.com
8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

BOARD OF DIRECTORS

Richard Royle Dr Richard Ashby AM Cathy Ford


Chair, the Australian e-Health Dr Ashby is the Chief Executive of Cathy Ford has been acting in the position
Research Centre eHealth Queensland responsible for of Chief Information Officer of the Metro
advancing healthcare through digital North Hospital and Health Service since
Richard is a partner at PwC and is the innovation. December 2017. Cathy’s emphasis during
national and regional digital health lead. this time has been on preparing the
He has over 30 years of senior executive In 2016 Dr Ashby oversaw the successful foundations for the digital transformation
experience in the public, for profit and delivery of Australia’s first large-scale of the health service through targeted
not for profit private hospital sectors in digital hospital, the Princess Alexandra planning and investment activities.
Australia. Richard is a past President Hospital, as the Chief Executive of Metro
of the Australian Private Hospitals South Hospital and Health Service. Dr Cathy’s substantive role is the Chief
Association. Ashby believes that digital healthcare is Digital Officer of eHealth Queensland. In
one of the most important revolutions in this role Cathy was responsible for the
Richard oversaw the successful healthcare – providing highly connected development of both the digital health
implementation of Australia’s first fully and interactive models of care that and innovation strategies as well as the
integrated digital hospital in Hervey support personalised, precise and well- investment roadmap and associated
Bay as the group CEO of UnitingCare informed treatment of patients across oversight activities.
Health in 2014. In 2016 he was asked care settings and care teams.
to be the startup CEO of the newly As a management & ICT professional
established Australian Digital Health Dr Ashby is regarded as one of the Cathy’s focus is to work with
Agency – putting into practice one of state’s most experienced clinicians and organisations to transform their business
his recommendations from a landmark health administrators. In 2010, Dr Ashby by developing strategies to make better
review he was asked to lead in 2013 for was awarded a Member of the General use of people, information, systems &
the Federal Government on digital health Division of the Order of Australia technology.
in Australia. for service to emergency medicine,
medical administration, and a range of With over 25 years in the industry she has
professional associations. He is active worked with many executives and their
across a broad range of areas, including teams on transformation activities that
teaching, research and consultancy. delivered sustained change.

Dr Ashby contributes to a significant Cathy graduated from the University of


number of organisations/committees. Queensland with first class honours,
His roles include: holds a graduate diploma in IT and is a
graduate of the Australian Institute of
• C
 hairman, Queensland Policy and Company Directors.
Advisory Committee on Health
Technology
• C
 hairman, eHealth Executive
Committee
• S
 enior Responsible Owner,
Queensland Digital Hospitals
Program
• B
 oard Member, Australian e-Health
Research Centre
• C
 ouncil Member, Queensland
University of Technology Council.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 9

Richard Symonds
Minutes Secretary

Kelly Tighe
Finance Manager, CSIRO

Rob Grenfell Adrian Turner


Meetings
Dr Rob Grenfell, a public health Adrian Turner is the CEO of Data61
physician, is the Director of CSIRO’s at CSIRO and also co-Chair of the Board Meetings for
Health and Biosecurity business unit. Cybersecurity Growth Centre. Data61
He leads a broad portfolio covering is the datascience arm of the CSIRO 2017/20187 were
nutrition, e-health, medtech and and is focused on solving Australia’s held as follows:
diagnostics, and biosecurity from weeds largest data-driven challenges. Adrian
to Ebola. was previously Managing Director and
28th August 2017
Co-Founder of Borondi Group, a holding
Rob has broad-ranging public health company focused on the intersection of
experience including: pervasive computing, platform economics
20th November 2017
• N
 ational Medical Director at BUPA and traditionally conservative industries
Australia New Zealand and was also co-founder and CEO of
5th March 2018
• N
 ational Director Cardiovascular smart phone and Internet of Things
security company Mocana Corporation.
25th June 2018
Health at the Heart Foundation
Prior to this Adrian had profit and loss
• S
 trategic Health Advisor to Parks
responsibility for Philips Electronics
Victoria
connected devices infrastructure, and
• S
 enior Medical Advisor at the was Chairman of the Board for Australia’s
Department of Health Victoria expat network, Advance.org.
• P
 hysician in charge of travel health
BHP He is regarded as a thought leader on
entrepreneurialism, Internet of Things
• General Practice.
and the impact of network connectivity
He was a member of the Safety and on business economics. He authored
Quality Outcomes Committee of the the eBook BlueSky Mining – Building
Hospital Innovation Reform Council, a Australia’s Next Billion Dollar Industries.
member of the Victorian Quality Council, Adrian is a UTS graduate and has
Chair of General Practice Victoria, completed the Executive Program
and member of the Health Advisory for Managing Growth Companies at
Committee of the National Health and Stanford University, having spent
Medical Research Council. 18 years in Silicon Valley.

aehrc.com
10 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 11

2018 ANNUAL COLLOQUIUM


Over 300 people attended this year's 14th Annual e-Health Research Colloquium, hosted
by the Australian e-Health Research Centre at the RBWH Education Centre in Brisbane on
27 March 2018. The Colloquium was once again a great opportunity for us to hear about
Queensland and national initiatives as well as inform partners about our work.

The first session this year provided an We then had talks from Robert Lee from Attendees had the opportunity to view
opportunity to highlight various digital the Australian Digital Health Agency about posters, discuss our projects and
health initiatives in Queensland. Dr Australia's MyHealthRecord program; technology with our scientists during
Claire Sullivan, Clinical Lead for the A/Professor Andrew Mallett from the the morning tea and lunch breaks,
Digital Health Program Queensland RBWH Nephrology Department about and observe multiple technology
Health, described how Queensland using genomics in the Kidney Clinic; demonstrations.
Health is using digital disruption to and Tim Blake on a new consumer-
transform the state's healthcare – and focussed digital health guide for mobile Feedback from the day was again
some of the great outcomes from apps. The AEHRC's Jim Steel rounded excellent and it was great to see so
the iEMR program. We then had the out the session by discussing a new many of our partners and stakeholders
opportunity hear from Dr Wendy Dutton, Health Informatics on FHIR course being together.
Director of Obstetrics at Redland introduced for third-year IT students at the
Hospital, about the M♥THer trial and University of Queensland in collaboration
how we're working in conjunction with with AEHRC and Qld Health.
Redland Hospital. The M♥THer trial
provides a mobile app to support women The afternoon session provided an
with gestational diabetes app, and opportunity for AEHRC to highlight
this app provides information to their some of our new research. David
clinicians via an Interactive Clinician Ireland spoke about chatbots for people
Portal. The final talk of the morning with autism and other neurological
came from the AEHRC's Dr Sankalp conditions; Natalie Twine presented on
Khanna who provided an update on how machine learning enhances genetic
the Logan Hospital trial of a big data variant discovery for Motor Neurone
analytics-based tool to identify patients Disease; James Doecke spoke on a new
at high risk of re-hospitalisation. blood test for early Alzheimer's disease
diagnosis; Jason Dowling presented
The pre-lunch session highlighted on a phase II prospective trial of MR-
national initiatives. Our key note speaker alone treatment planning for localised
this year was Professor Enrico Coiera, prostate cancer; and finally Alejandro
Chief Investigator NHMRC Centre of Metke and Andrew Patterson provided
Research Excellence in Digital Health use cases for clinical terminology and
(CREiDH) and Professor of Health data when introducing genomics into
Informatics at Macquarie University. clinical care.
Prof Coiera provided a great overview
of the work program of the CREiDH,
of which AEHRC is a core member.
12 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MANAGEMENT AND
RESEARCH LEADERSHIP

Dr David Hansen
CEO, Australian e-Health Research Centre

David Hansen is CEO of the Australian e-Health Research Centre, the national digital
health program of the CSIRO Health and Biosecurity. David leads an e-Health research
portfolio developing information and communication technologies for the healthcare
system. These projects across health informatics, biomedical informatics and health
services research will underpin the e-health architecture in Australia.

Prior to joining CSIRO, David worked for LION bioscience Ltd in the UK, developing
genomic data and tool integration software that was used to publish the first human
genome and used at over 200 pharmaceutical and biotechnology companies and
research institutes worldwide.

Michael Lawley
Group Leader, Health Informatics

Dr Michael Lawley is Senior Principal Research Scientist and Group Leader with the
CSIRO Australian e-Health Research Centre, part of CSIRO’s Health and Biosecurity
business unit. Michael leads the Health Informatics Group with teams in health data
semantics, health statistics, and software engineering.

He has deep expertise in clinical terminology and specifically large scale ontologies
such as SNOMED CT. Work developed by Michael and his team have produced
technologies that have been licensed nationally and internationally by standards bodies,
government organisations and SMEs.

Jurgen Fripp
Group Leader, Biomedical Informatics (May-June)

Jurgen Fripp is the group leader for the Biomedical Informatics group at the AEHRC
and completed his honours thesis and PhD in the area of medical imaging at the
University of Queensland and with the CSIRO. His bachelor degree from the University
of Queensland was in Science (applied mathematics) and Electrical Engineering.
Jurgen’s team conducts research into image analysis algorithms for applications
in Positron Emission Tomography (PET), Magnetic Resonance Imaging (MRI), and
Computed Tomography (CT). Their workflows developed have been applied to support
various large clinical studies, including AIBL (http://aibl.csiro.au/).

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 13

Mohan Karunanithi
Group Leader, Health Services

Mohanraj Karunanithi leads the Health Services group at the AEHRC. Mohan has
a doctorate in Biomedical Engineering from the University of New South Wales.
He has over 10 years of experience in cardiac research and five years’ medical
industries experience. At AEHRC, Mohan manages and coordinates research in
ICT applications in healthcare management, chronic disease and aged care.

Professor Yogesan Kanagasingam


Director, Australian Tele-health Research and Development Group

Prof Yogesan has developed medical technologies from bench to bed and one of his
inventions is used by NASA in the International Space Station. He was an Australian
of the Year finalist from Western Australia (2015) and also Western Australia Business
Leader of the Year finalist in 2014 for his contribution to medical sciences and
prevention of blindness. He is a Visiting Scholar to Harvard University and professor at
the School of Medicine at the University of Notre Dame. He was a Fulbright Scholar to
Stanford University School of Medicine and a NHMRC Research Fellow.

Olivier Salvado
Group Leader, Biomedical Informatics (July-April)

Dr Salvado is the group leader for the Biomedical Informatics group at the AEHRC. He
is adjunct Professor at the University of Canberra, adjunct Associate Professor at the
University of Queensland, and Honorary Research Fellow at the Florey Neuroscience
Institute. His research interests include developing bioinformatics methods for large
multi-scale clinical studies, neuroimaging technologies, and investigating novel multi-
modal clinical imaging biomarkers. Since 2009, he has been co-supervisor or assessor
of 10 PhD students (two current co-supervisions).

Dr Salvado has published over 70 original research papers cited more than 2700 times in
highly vregarded journals including Lancet Neurology, JAMIA, Journal of Neuroscience,
NeuroImage, PLoS One, Annals of Neurology, and Brain. Dr Salvado was the co-chair of
the ISBI conference that took place in Melbourne, Australia in 2017.

aehrc.com
14 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

NEWS AND AWARDS

News and Communications


Our research was again both nationally and internationally well represented throughout the year:

• W
 e took part in a joint media announcement with the Florey • D
 enis Bauer appeared in the “This is my architecture
Institute of Neuroscience and Mental Health to promote our series” on AWS. Her team had numerous blogs published
research about the link between iron in the brain and the on various tech websites, along with a story in CIO online.
speed of Alzheimer’s disease progression. It received national • D
 ana Bradford and David Ireland were interviewed by ABC
media coverage, with stories on the front page of The Herald Radio National about chatbots in the health sector.
Sun, on the ABC and Sky News, and mentions on TV news
• D
 avid Hansen was interviewed by ABC Radio about the
and The Project, reaching more than two million people.
M♥THer app for gestational diabetes.
• T
 he team’s work creating an eye test GPs could use to detect
• J ames Doecke was featured in articles by the ABC and
diabetic retinopathy, and a subsequent grant from Diabetes
The New Daily for his involvement in a project to develop
Research WA, received wide coverage. Yogi Kanagasingam
a blood test that can detect Alzheimer’s disease up to 20
was interviewed for stories in The West Australian, The Echo,
years before symptoms begin.
The Post, ZDNet, Gizmodo, Pulse+IT, IT Wire, Australian
Doctor, Insight News and more, reaching an audience of • T
 he Australian Financial Review published an article
approximately one million. on Cardihab’s commercialisation, and it also received
coverage in tech media.
• D
 avid Silvera’s work with robots to support children on the
autism spectrum was featured across national ABC TV and
radio news, reaching 1.5 million people.

Awards
Our teams were successful again this year winning and placing in a number of awards nationally, including the Queensland and
Victorian iAwards and the Western Australia Information Technology and Telecommunications Alliance Incite Awards (WAITTA).

• Queensland iAwards • T
 he M♥THer Gestational Diabetes Mellitus platform won
−−  erit recipient of the Research and Development
m the Health Round Table Innovation Award in the ‘Improving
Project of the Year award for the Mobile-Pulmonary Patient Centred Care’ stream. It was also a finalist in the
Rehabilitation Platform (m-PR). International Hospital Federation Awards.
−−  inners of the QLD Premier’s iAward for Public
w • C
 ONSULT Neurosurgical Planning project was a finalist in
Sector Innovation for the Mobile-Pulmonary the RBWH Research Excellence Awards.
Rehabilitation Platform (m-PR). • Y ing Xia won the best poster award of the World Federation
−−  erit recipient of the Community Service Markets
m of Nuclear Medicine and Biology conference, for the
award for the PD-BUDDy: Support for Peritoneal Neuroscience competition.
Dialysis Patients. • S
 ajib Saha received a Rising Star Award at the Science on
−−  erit recipient of the Research & Development
m the Swan Conference 2018, for development of artificially
Project of the year for the PD-BUDDy: Support for intelligent methods for the diagnosis of retinal disease.
Peritoneal Dialysis Patients. • S
 haun Frost was included in the list of Top Ten Posters
for the Science on the Swan Conference 2018 for his
• V
 ictorian iAwards: CALD-Assist won the Health Round submission on ‘Retinal morphology for pre-clinical
Table Innovation Award. detection of Alzheimer’s Disease’
• W
 AiTTA (WA Information Technology and
Telecommunications Alliance Incite Awards) 2017-2018:
the team was awarded the winner of “Most Innovative
Enabler in Health Care” on its project titled “AEye
Screening System for Patients with Diabetes” and the
National iAwards Finalist of “Research & Innovation
Project of the Year - Industry Award” on its project titled
“Cloud-based Retinal Vascular AnalySis Platform: VASP”.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 15

RESEARCH PROGRAM
Over the past five years the The AEHRC strategy for the next five years aims to use the research capability of
our three research groups – health informatics, biomedical informatics and health
Australian e-Health Research
services research – to continue to tackle challenges of Australia’s healthcare system
Centre has developed into a and expand the impact of our research. The challenges set out in this strategy are to:
full health and biomedical • increase our science outcomes to be recognised in the top three e-Health
informatics research program. research centres in the world
This program spans health • increase our impact through increased adoption of our technologies
informatics, covering data • d
 evelop new areas of impact in Indigenous health, precision medicine, big data
about patients, services and medical research, healthy ageing and a learning healthcare system

populations; biomedical • increase our commercial outcomes, and

informatics, using patient • continue the growth of the AEHRC around Australia.

genomic and imaging data Our research program is informed through strong partnering with the health industry,
to personalise diagnosis and including clinicians, researchers, health service executives and the health IT vendor
community. With more than half our staff based at the Royal Brisbane and Women’s
treatment; and health services, campus in Brisbane, our scientists and engineers have strong relationships with
the use of technology in Queensland health administrators, clinicians and researchers. As CSIRO’s e-Health
delivering services to patients. research program, we can also access expertise from across CSIRO.

Health informatics Biomedical informatics Health services


The introduction of electronic New medical technologies – The increase in mobile
health and medical records, especially genomic and imaging technologies and high-bandwidth
including the national My technologies – are leading a broadband is changing the way
Health Record, is increasing the revolution in the personalisation of that services are provided in all
demand for clinical information diagnosis and treatment. walks of life – including health
to be shared between health services.
practitioners and with patients. Our biomedical informatics
research develops innovative Our health services researchers
Our health informatics research technologies for the discovery and work with health service providers
develops and applies innovative communication of meaningful to develop internet-enabled
tools and techniques for evidence- patterns from the new medical models of care to overcome the
based solutions and strategies technologies. The aim is to burden being placed by chronic
to support improved health develop techniques to report and disease and aged care. Our
outcomes. Our goal is to unleash visualise complex biomedical teams are trialling technologies
the value in health data, including information for clinical diagnosis to deliver health services through
both electronic health records and screening. This information mobile health, and tele-health
and administrative data sets, to can ensure that diagnosis is technologies for patients with
improve patient outcomes and precise and treatment appropriate conditions such as eye diseases,
health system performance and to reduce unnecessary treatment cardiac diseases, diabetes, stroke,
productivity. and improve outcomes. and hip replacements.

Our scientists and engineers Our scientists and engineers Our scientists and engineers
apply machine learning, natural use the simultaneous use our expertise in mobile
language processing, formal logic, application of statistics, technology, home monitoring,
and statistical and simulation computer programming, and telemedicine, wellbeing and
approaches to the collection, applied mathematics to develop behavioural change to improve
processing, analysis and sharing solutions that communicate health services to urban, rural
of health information for decision insights to clinicians and clinical and remote Australians.
support, systems modelling and researchers.
reporting.

aehrc.com
16 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH INFORMATICS

2017/18 Science and impact highlights


• R
 isk stratification tool was developed to identify patients with chronic diseases who are at risk of hospitalisation. This tool is
now implemented in the National Health Care Homes initiative to enrol patients from GP practices.
• T
 he SnoMAP tool, developed with Metro South Hospital and Health Service (HHS), is being used by all Queensland hospitals
deploying the Cerner iEMR to meet statutory reporting requirements with a resulting measurable increase in data quality.
• T
 he National Clinical Terminology Service, a joint project with the Australian e-Health Research Centre (AEHRC) and the
Australian Digital Health Agency, now has over 50 companies and organisations with licenses to our terminology server,
Ontoserver, for implementation of state-of-the-art support for clinical terminology in their e-health products. The NCTS won
two 2017 Queenslanld iAwards and a National Pitchfest award.
• T
 he Medtex (medical text analytics tool) was deployed within Queensland Health to process and analyse both the historical
and live pathology feeds for cancer notifications reporting from public and private pathology laboratories across Queensland.
• O
 ur Data Interoperability team is contributing to the Australian, Queensland and Melbourne Genomics Health Alliances to capture
accurate clinical phenotypes - a vital part of ensuring that Australia will get full value from investments in genomics medicine.
• W
 e are part of the National Health and Medical Research Council (NHMRC) Centre for Research Excellence in Digital
Health in partnership with Macquarie University Australian Institute for Health Innovation and the University of Melbourne.
AEHRC will be working with scientists and clinicians from around Australia on clinical decision support and automating the
production of systematic reviews.

Health Informatics Group Leader: Michael Lawley


H E A LT H I N F O R M AT I C S

Australia’s healthcare system faces many challenges. One is the increasing demand for clinical information to be shared
between individual health practitioners, healthcare provider organisations and state/territory health departments.

Our health informatics research develops and applies innovative tools and techniques for evidence-based solutions
and strategies to support improved health outcomes. Our goal is to improve the quality of, and unleash the value
in, health data, including electronic health records and administrative data sets, to improve patient outcomes and
health system performance and productivity.

We apply machine learning, natural language processing, formal logic, and statistical and simulation approaches to
problems involving decision support, systems modelling and reporting.

Health Data Interoperability Team Leader: Alejandro Metke


Data is captured about patients in a number of different formats and electronic repositories using many different
terminologies. Our technologies are targeted at understanding the information in data, whether the data is captured
in an electronic health record, coded in a clinical database, captured from sensors, described in medical free text
reports or even captured using imaging technology.

Our team also works in the genomics area, specifically around representing patient phenotype data using standards
and terminologies. Our involvement in several genomics alliances in Australia and internationally has helped us
position ourselves as leaders in this field.

Health Text Analytics Team Leader: Anthony Nguyen


The Health Text Analytics team is focused on deriving value from electronic health data in terms of improving
patient outcomes and health system performance and productivity. The group does this by developing and applying
machine learning, natural language processing, information retrieval and formal logic approaches to deliver and
support meaningful data interoperability and analysis for decision support, analytics, modelling and reporting.

Health Data Analytics Team Leader: Rajiv Jayasena


The Health Data Analytics team delivers scientifically robust analytics to improve health outcomes. Our team
improves performance and sustainability of the Australian health system by transforming clinical/operational data
and knowledge with analytics, optimisation, real-time monitoring for decision support, risk stratification tools and
evaluations.

Health Data Engineering Team Leader: Derek Ireland


Our world class Health Data Engineering team is a dedicated team of software engineers who work with scientists
across the AEHRC in delivering solutions to our customers and partners.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 17

PROJECTS

HEALTH DATA INTEROPERABILITY


Successful adoption of standard
terminologies such as SNOMED CT and
the Australian Medicines Terminology
(AMT) is vital for the success of enabling
patient data to move between clinical
systems in a safe way. Many systems
across health organisations such as
Queensland Health will be required to
migrate from other code sets to SNOMED
CT. Our tools help with this migration
and deal with complexities such as the
level of detail in each code and gaps in
the codes while still ensuring that high-
quality data is captured.

We have developed significant national


and international impact through our
tools: the free SNOMED CT and AMT Figure 1. A typical seven days of Shrimp usage.
browser, Shrimp; the terminology
mapping and subsetting tool, Snapper;
the cloud-based terminology server,
Ontoserver; and the reasoning engine,

H E A LT H I N F O R M AT I C S
Snorocket.

A National Clinical Terminology Service


We have worked with the Australian Advantages to this approach include: Through 2017/18 the team finalised
Digital Health Agency to deliver the • P
 roviding terminology server a major upgrade to the next version
National Clinical Terminology Service software ensures consistent of FHIR (STU3), closely engaged with
(NCTS). The AEHRC's Ontoserver is a interpretation of specifications the FHIR community to clarify, refine
key component of the solution and is and that state-of-the-art search and improve details of the Terminology
used to deliver this service through a algorithms are available to all Services subsystem of the HL7 FHIR
nationally hosted service. Technology implementers. Specification, and engaged closely
providers can also license Ontoserver with state jurisdictions and the vendor
• L
 ocal terminology server instances
free of charge for integration into community through a series of very
allow for local autonomy, and
their own health record solutions, successful connnectathons and
local code systems and value sets
with a syndication service keeping the workshops to ensure the resulting
can be supported using the same
standardised terminology content up to service delivers what is needed.
system supporting standard clinical
date. This is a pioneering approach to
terminology like SNOMED CT-AU. The NCTS was the winner of two state
making standard clinical terminology
readily available – going well beyond the • S
 yndication of content ensures that AIIA iAwards: Infrastructure & Platforms
traditional mechanism of providing files every terminology server instance Innovation of the Year and Public Sector
for download along with documentation. can easily remain up-to-date with & Government. NCTS also won the
monthly SNOMED CT-AU releases AIIA Pitchfest in the Public Sector &
without complex and manual update Government category.
processes.
There has also been considerable
• U
 se of the simple and easily interest in the service from NHS Digital
adopted HL7 FHIR API not only in the UK. We have organised a FHIR
means there's no lock-in to one connectathon in London and licensed
proprietary terminology server Ontoserver for evaluation as part of a
implementation, but it is both cloud Proof of Concept project.
and mobile friendly, and paves the
way to broader adoption of what is
promising to be a truly revolutionary
standard for health IT.

aehrc.com
18 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH DATA INTEROPERABILITY (continued)

SNOMED CT in the
Queensland digital hospital
project
The Queensland digital hospital project
introduced SNOMED CT terminology
as part of the implementation of an
integrated Electronic Medical Record.
As part of this project we worked with
the Princess Alexandra Hospital's
(PAH) digital hospital program on
several collaborative projects. An early
outcome was to work with the PAH and
the Australian Digital Health Agency
to release an extension to SNOMED
CT-AU containing more than 100 new
procedure codes required to support
the pioneering use of SNOMED CT in
the deployment of the Cerner Surginet
product. Further work continues to
H E A LT H I N F O R M AT I C S

expand the SNOMED CT reference set


for SurgiNet as the Cerner product is
deployed in other Queensland hospitals. Figure 2. High-level architecture of the National Clinical Terminology Service.

SNOMED CT-encoded data for secondary


reporting purposes
One part of the broader project dealt Injury surveillance and reporting Allied health
with the problem of continued reporting This approach to using SNOMED Terminology work has continued
of emergency department non-admitted CT-encoded data and snoMAP has supporting the development of
patients after the switch from use of now been extended to include specific statewide data collections for the allied
ICD-10-AM to SNOMED CT. The result subsets suited to injury surveillance and health sector. Various allied health
of this was a tool: snoMAP, which reporting. disciplines have unified their approach
extends coverage of SNOMED CT from to data capture and reporting, and
a relatively small subset of clinical Clinical cohorts are standardising using SNOMED CT
findings to all relevant codes clinicians Work continues on developing particular in the Cerner product. At present the
will need to use to document patient subsets of SNOMED CT for analysing legacy termsets have been mapped
records in the emergency department data based on patient cohorts. Early to SNOMED CT (n=2000 terms).
setting. The goal was to re-purpose work has investigated the use of Approximately 250+ new concept
the original SNOMED CT-encoded snoMAP to produce routine reports requests have been generated by the
patient data and maintain its truth for particular patient cohorts, such as allied health discipline groups and
value for clinical care delivery, and to patients presenting with diabetes or have been submitted to the Australian
ensure it complies with and qualifies mental health issues. This work will Digital Health Agency for addition and
for activity-based funding. The digital help preserve the routine and ad-hoc release. Of these, 184 have been added
hospital program has now been rolled reporting at a local hospital level, in the and released in SNOMED CT AU and
out across six Queensland hospitals same way that data analysts previously 77 are currently in process. Work is
and this number is growing. These produced 'dashboard' reports using now underway on the expansion of this
additional hospitals have begun to use ICD-10-AM-encoded data. initial mapset of 2000 concepts which
snoMAP, and with snoMAP content, will increase the expressivity of data
both SNOMED concepts and ICD-10-AM capture in allied health records. A much
codes updated twice per month, PAH richer and descriptive data collection
and other Queensland digital hospitals will enable expanded measures
can submit data for activity-based of the effectiveness and clinical
funding in near real time. outcomes achieved by the allied health
professions, rather than only service
event-based metrics.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 19

RACS MALT goes SNOMED Genomics & Clinical Broadening support for genomics
terminologies
CT-native Phenotypes
Many terminologies used in genomics
During 2017 we continued our During 2017/2018 the AEHRC has formed are distributed in OWL format. FHIR has
collaboration with the Royal a valuable working relationship with a no native support for OWL ontologies
Australasian College of Surgeons number of genomic health alliances. The and some of the characteristics of OWL
(RACS) after the successful transition Australian Genomics Health Alliance are challenging to represent in FHIR
in 2016 of their Morbidity Audit Logbook is a national research collaboration code systems. A generic transformation
Tool (MALT) from ad-hoc terminology to of researchers and clinicians working between OWL and FHIR code systems
using SNOMED CT-AU procedure codes together to provide effective and was designed and implemented. This
natively. This included the development sustainable delivery of genomic medicine allowed supporting terminologies such
and maintenance of aggregation maps in healthcare. The AEHRC is part of as the Human Phenotype Ontology,
to support ongoing reporting processes the Phenotype Capture project, whose the Phenotypic Quality Ontology, the
from the MALT data. During 2016/17 objectives are to enable different sources Foundational Model of Anatomy and
RACS migrated to an NCTS-based of phenotype data to be used for research Orphanet in Ontoserver.
deployment of Ontoserver using FHIR- and clinical applications, and to represent
based valueset and map artefacts, and different data in the same format, by Aligning terminologies
we are now working toward expanding transformation to computer-readable In an effort to commence the
their initial SNOMED CT valueset (from standardised ontologies such as SNOMED standardisation of genomic data collections,
14,000 concepts, to 30,000 procedure CT and the Human Phenotype Ontology. a great deal of work has been done looking
concepts). Work also continues in the at the alignment and the creation of a map
development of new procedure content Standardising research data between the Human Phenotype Ontology
for release in SNOMED CT AU. This One of the main issues in the Australian and SNOMED CT. To date we have found
enhances its usability and relevance in Genomics Health Alliance is that 18% exact match alignment and work in
the Australian surgical domain. To date, flagships (the organisations doing this space continues.

H E A LT H I N F O R M AT I C S
well over 300 new concepts have been clinical research) capture phenotypic
created and released in SNOMED CT AU. data in slightly different ways, despite Defining minimum clinical data sets
RACS is also considering broadening the using the same platform (REDCap). Development has commenced of flagship-
use of SNOMED CT, so MALT will allow This makes it very hard to reuse the specific FHIR value sets based on the
their users to also capture data relevant data or interoperate with other systems. data currently being collected. It is hoped
to diagnoses and outcomes. This problem was solved by designing that these value sets will form the basis
and implementing the FHIRCap for ongoing development of SNOMED
platform, a novel solution based on a CT genomics reference sets for use and
domain specific transformation rules release in the international edition. The
language that allows expressing how value sets have been deployed in an
the REDCap data should be represented Australian Genomics Health Alliance
in standardised format, in this case the Ontoserver instance and can be browsed
FHIR standard. using a value set viewer extension of
Shrimp, our terminology server browser.

aehrc.com
2 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH DATA INTEROPERABILITY (continued)

NHS Digital: evaluating


future options for
classifications
NHS Digital is investigating options
and opportunities for the migration or
implementation of future terminologies
and classifications strategies in light of
World Health Organization developments
of ICD-11 and ICHI. We were tasked with
providing an overview of the current
state, and suggestions for optional
approaches for continuing business-
as-usual during this discovery and
evaluation phase. A report was provided
outlining the scope and scale of issues Figure 3. The value set viewer implemented for the Australian Genomics Health Alliance.
that will need consideration as part of
a discovery work program and provide
necessary evidence for the business case
development for the implementation of a
new classification such as ICD 11.
H E A LT H I N F O R M AT I C S

K
 ey Project Highlights Project Aims for
Collaborators: for 2017/18: 2018/19:
• A
 ustralian Digital Health • D
 evelopment of Ontoserver V5 • E
 xpand and improve
Agency/National eHealth for the NCTS with the latest terminology-enabled data
Transition Authority STU3 version of HL7's FHIR APIs analytics
• R
 oyal Australasian College of and syndication support • D
 evelop enhanced algorithms
Surgeons • C
 ontribution to the refinement for automated analysis of
• P
 rincess Alexandra Hospital, of the HL7 FHIR Terminology terminology quality metrics to
Metro South HHS Services API standard R4 improve data quality
• A
 ustralian Genomics Health • O
 rganised FHIR connectathons • D
 evelop more sophisticated
Alliance. in the UK and negotiated an auto-mapping strategies in
evaluation license of Ontoserver Ontoserver to be utilised via
with NHS Digital Snapper
• Impact and expansion of the • D
 evelop assistive technologies
snoMAP tool as it has been that will support the clinician
taken up at additional sites community to capture accurate,
throughout Queensland codeable data for documentation
• D
 evelopment of FHIRCap for of patient records
the Australian Genomics Health • E
 xtend our automated
Alliance and completion of the techniques for developing
full mapping of the Somatic aggregation-based maps for
Cancer Flagship data reporting and data analytics
• S
 upport for OWL-based • B
 uild on the successful rollout
ontologies in Ontoserver to of Ontoserver as the NCTS
support genomics projects. terminology server platform-
of-choice through international
adoption and licensing.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 2 1

HEALTH TEXT ANALYTICS

Electronic health records (EHR) are


expected to enable better health
outcomes and improved efficiencies
in our health services. The majority of
EHR data is recorded in unstructured
free-text, such as clinical progress
reports, imaging and laboratory
reports, discharge summaries, and
Figure 1: Artificial intelligence capabilities for understanding and reasoning with clinical text data with an
death certificates. This data contains emphasis on standard clinical terminologies for data interoperability and analytics.
information that is valuable for clinical
decision support and secondary
use such as for population health Automating clinical data Reconciling medical
monitoring and reporting. However, the registries records to prevent missed
extraction of clinical information from
Statistics about cancer incidence diagnoses
large volumes of free-text data hinders
manual interpretation. continue to remain several years out of The checking of radiology imaging and
date. The extent of manual processing pathology laboratory reports to ensure
The Health Text Analytics team is of patient records by cancer registry abnormalities or positive results,
developing and applying advanced coders and outdated information respectively, are not missed and that
natural language processing, collection systems are delaying patients receive appropriate follow-up
information retrieval, and machine the delivery of more timely cancer once discharged from the Emergency
learning techniques, along with standard information. Department (ED) is an essential but
clinical terminology (e.g. SNOMED CT)
In partnership with the Queensland laborious task. Due to a busy ED and
semantics and reasoning, to provide
Cancer Control Analysis Team resourcing issues, it can often be days
meaningful and accurate computational
(QCCAT), Queensland Health, the after the patient’s initial presentation
interpretation of clinical free-text.

H E A LT H I N F O R M AT I C S
AEHRC is extracting information about to the ED that this checking process is
Our solutions have been developed cancers from the free-text contents performed. This process results in time
in partnership with healthcare of histopathology reports for cancer inefficiencies with delays in reporting,
practitioners from cancer registries, notifications, synoptic reporting and delays in checking reports, and delays
hospital radiology and emergency cancer staging. This is enabling QCCAT in recalling patients. A timelier and
medicine departments. Working with to build a real-time cancer registry that efficient process is therefore required
health industry stakeholders allows processes new histopathology reports to improve patient outcomes and staff
our health text analytics solutions to as they are available from public and resources.
leverage the wealth of clinical free-text private pathology laboratories across In partnership with the Royal Brisbane
reports and aid in decision support and the state of Queensland. This processed and Women’s Hospital and The
reporting. information provides the capacity to Prince Charles Hospital EDs, we have
support key activities such as cancer developed algorithms and models to
monitoring, health service planning and reliably identify abnormal or positive
research. results from radiology and pathology
This medical text analytic service reports respectively, and link these with
uses the AEHRC Medtex platform to patients’ disposition as recorded in the
automatically read and analyse the Emergency Department Information
pathology reports. System to provide decision support to
the, currently manual, checking process.
Future work will develop software to
demonstrate the clinical and patient
benefits arising from the information
technology-based solution.

aehrc.com
2 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH TEXT ANALYTICS (continued)

Diagnosis coding from


electronic health records
Clinical coders abstract relevant
information from patients’ medical
records and decide which diagnoses and
procedures meet the criteria for coding
as per Australian Coding Standards.
The process mainly relies on manual
inspections and experience-based
judgments from clinical coders, and the
effort required for information abstraction
is extremely labour- and time-intensive
and prone to human errors.

In partnership with the Gold Coast


University Hospital and Health Service,
we have developed algorithms and
models for automating the diagnosis
coding (ICD-10-AM) process from
hospital progress notes. Promising
H E A LT H I N F O R M AT I C S

results were achieved when compared


to what was projected as possible from a
diagnosis code validation study.

Figure 2: Medical free-text analytics applications for three different clinical document classification tasks:
(i) identification of the ICD-10-AM diagnosis codes from progress notes;
(ii) identification of abnormalities from radiology reports; and
(iii) classification of cancer stages from pathology reports.

Health record search and analytics


Search technologies are critical to To support the need for evidence-based
enable clinical staff to rapidly and medicine we have developed a search
effectively access patient information engine providing clinicians easy access
contained in free-text medical records. to the vast and ever-changing body of
Health search is challenging as it medical literature. A key, novel aspect of
suffers from the semantic gap problem: the search engine is that it is specifically
the mismatch between the raw data tailored around the three clinical tasks
and the way a human being interprets of searching for diagnoses, searching
it. Valuable domain knowledge explicitly for treatments and searching for tests.
represented in structured knowledge All results are displayed, and the
resources such as ontologies (e.g. clinician can interact with the system,
SNOMED CT) can be leveraged to according to these three clinical tasks.
support such semantic inferences. The An empirical evaluation of the systems
focus of our research is on health record showed both better quality results and
searching and analytics using text, time savings from the task-oriented
concepts, annotations, and SNOMED CT approach.
subsumption and relation querying.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 2 3

H E A LT H I N F O R M AT I C S
Figure 3: Screenshot of our task-oriented search engine for evidence-based medicine.

NHMRC Centre of Research Excellence in Digital Health


We are part of the NHMRC Centre of aggregate Electronic Health Record Systematic reviews play a key role in
Research Excellence in Digital Health, in (EHR) data, for real-time, personalised evidence-based medicine, informing
partnership with Macquarie University comparative effectiveness information to practice and policy. Existing technology
Australian Institute for Health Innovation empower clinicians in making evidence- to assist with systematic review
and the University of Melbourne, and based decisions in the absence of production has largely ignored the
will be working with scientists from published guidelines. Using this button, search stages of the systemic review.
around Australia on clinical decision clinicians are able to view and evaluate Existing search engine research is also
support and automating the production treatment approaches and outcome often not applicable to the unique task
of systematic reviews. of “patients like theirs” in the context of systematic reviews. This project will
of their hospital or other contributing develop novel search engine technology
Clinicians are often required to make institutions. In particular, we will to significantly improve the process of
decisions in the absence of best- enhance the structured data in EHRs producing systematic reviews. This will
practice evidence. As a result, treatment with text analytics of clinical text to directly impact how systematic reviews
success of individuals with complex better support clinical decision making are produced and, consequently, the
co-morbidities or treatment history may at point-of-care, not only to reduce downstream impact of health decisions
depend on the clinician’s experience. clinical care pathway variations but also made on the basis of these reviews.
The project proposes to use the "Green to provide health outcomes evidence for
Button" concept, which leverages those that do.

aehrc.com
2 4 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH TEXT ANALYTICS (continued)

Collaborators: Project Highlights Project Aims for


for 2017/18: 2018/19:
• Q
 ueensland Cancer Control • D
 eployed Medtex within • A
 utomatically extract important
Analysis Team (QCCAT), Queensland Health to process clinical indicators for cancers
Queensland Health and analyse live pathology to extend the cancer stage and
• D
 epartment of Emergency feeds from public and private synoptic reporting capabilities
Medicine, Royal Brisbane and pathology laboratories across within Queensland Health.
Women’s Hospital Queensland to support the • D
 evelop medical record checking
Queensland Cancer Registry decision support application for
• D
 epartment of Emergency
coding process. clinical interaction.
Medicine, The Prince Charles
Hospital • C
 omparative analysis of • Investigate the application of
conventional machine learning deep learning for developing
• D
 epartment of Emergency
and deep neural network (DNN) a good computational
Medicine, Logan Hospital
transferability across hospitals representation of both
• G
 old Coast Hospital and Health for clinical text classification. the structured data and
Service. Experiments have shown the unstructured free-text in
robustness of DNN models EHRs that could be leveraged
in trivial transfer learning
H E A LT H I N F O R M AT I C S

across a wide range of clinical


strategies that use source-only classification tasks – patient
hospital data. Transfer learning classification, disease risk
strategies using both source stratification and treatment
& target hospital data further outcome.
improve models.
• E
 xtend the health text search
• D
 eveloped a search engine and analytic technology solutions
for evidence-based medicine to other health applications and
tailored to the three clinical report types.
tasks: diagnosis, testing and
treatment.
• D
 eveloped an automated method
to match patients to eligible
clinical trials based on their
electronic patient record.

P
 hD student profile

Harry Scells
AEHRC PhD Top-Up

Scholarship Queensland University of Technology

Improving systematic review creation with information retrieval

Systematic reviews, in particular medical systematic reviews, are time consuming


and costly to produce. The largest contributing factors to the time and monetary costs
are the searching (including the formulation of queries) and screening processes.
These initial processes involve researchers reading the abstracts of thousands and
sometimes hundreds of thousands of research articles to determine if the retrieved
articles should be included or excluded from the systematic review. This research
explores automatic methodologies to reduce the workload relating to the searching
and screening processes.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 2 5

HEALTH DATA ANALYTICS

The Health Data Analytics team develops


and delivers scientifically robust analytics
to improve safety, quality and efficiency of
our healthcare system. These analytics
improve performance and sustainability
of the Australian health system by
transforming clinical/operational data
into knowledge via analytics, optimisation,
real-time monitoring for decision support,
risk stratification tools and evaluations.

The work demonstrates an intimate


knowledge of the Australian health
system and associated datasets as well as
knowledge of the regulatory frameworks
of working with sensitive health data and
potential quality issues associated with
health data. The Health Data Analytics
team includes several statisticians,
engineers and research scientists to
ensure rigour in every analysis.

Patient risk stratification predictive model for Health Risk stratification for

H E A LT H I N F O R M AT I C S
Care Homes hospital avoidance in acute
care
In 2017/18, the team successfully The benefits of such a risk stratification
delivered the Predictive Risk tool include: In 2017/18, partnering with the
Model (PRM) pivotal to the federal Queensland Health’s Healthcare
• r educed level of dependency
government’s Health Care Homes Improvement Unit and Logan Hospital,
and improved health outcomes
initiative aimed at improving quality the team successfully delivered a
for patients due to early clinical
and safety of the primary health care real-time web-based risk stratification
intervention
system. Partnering with Precedence algorithm that can be used to identify
• c ost savings as a result of reduced chronic disease patients with a high
Health Care and Sonic Health Services
readmission rates risk of rehospitalisation while they are
for this project, we were responsible
for the development and validation • reduced acute length of stay still in hospital. Model development and
of the predictive model to be used in • reduced emergency presentations validation involved employing routinely
the program at a GP practice-level to collected administrative and clinical
• improved equality in health care
identify potentially eligible patients, by datasets that were available in real
access based on actual need
stratifying patients according to their time, and a web-based clinical decision
• m
 ore effective and coordinated support tool was built to provide risk
risk of unplanned hospital admission in
health service planning based on groups and individual patient risk profiles
the next 12 months. Patients identified
health needs to care teams. A 12-month evaluation
by our algorithm were then contacted
by the general practice and invited to • facilitation of person based to assess the impact of the risk tool on
undertake an assessment to further health care planning through the chronic disease readmissions and care
assess their eligibility and Health Care provision of risk scores to treating planning processes is underway at Logan
Home tier level. practitioners. Hospital.
To date the tool has been installed in
over 170 GP practices and Aboriginal
Community Controlled Health Services
(ACCHS) around Australia participating
in the Health Care Homes program to
provide services to patients with chronic
and complex conditions.

aehrc.com
2 6 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH DATA ANALYTICS (continued)

Figure 1: Risk of readmission and representation to ED are provided via a dashboard for all patients staying
overnight in hospital.
H E A LT H I N F O R M AT I C S

Generating an evidence
base for informing policy
around after-hours care
There is a need to deliver healthcare
in a manner that maximises outcomes
for patients. In 2017/2018, the team
completed a large state-wide study
commissioned through the Queensland
Clinical Senate to investigate whether
outcomes are different for patients who
Figure 2: Patient-specific factors contributing to risk of readmission are provided for each inpatient via a seek care at public hospitals outside
dashboard. of normal business hours compared to
other times of the week.

The main outcome measure of


Identifying choke points in the ED-ambulance interface interest was mortality due to the
The team undertook a quantitative association between higher levels of wide acceptance of death being the
analysis of Queensland Ambulance ED occupancy, and both the number of ultimate outcome-based quality
Service and Queensland Health data ambulances waiting to transfer care of measure. Other outcomes assessed
to assess the variation in Patient Off patients at the ED and the average POST in the study included ED and inpatient
Stretcher Time and maximum waiting time across most sites. The results length of stay, inpatient readmission
time for transfer to ED care, derive of this analysis can improve resource rate, ED and inpatient costs and ED
a continuous count of ambulances utilisation and assist performance patient experience. The results provide
awaiting patient offload at any time- improvement initiatives in both evidence around variation in patient
point for each site, and model the Queensland Health and QAS. outcomes associated with after-hours
relationship between NEAT, POST, care and have been presented to senior
ambulance arrival rate, and ambulances executives of Queensland Health.
waiting. Among the findings is an Importantly the work is identifying
factors that may influence observed
variation to support improved health
service delivery.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 2 7

Developing an evidence
base to better target efforts
to improve the scheduling
of surgery in public
hospitals
In 2017/18, the team commenced
an engagement with Queensland
Health's Healthcare Improvement
Unit to undertake additional statistical
modelling of surgical and inpatient data
for the major public hospitals across
Queensland. The modelling builds on
previous analysis by the team aimed at
maximising the utilisation of operating
theatres as one of the highest resource
costs of hospitals. The additional HealthLinks Chronic Care evaluation
analysis explores whether the day of the
week that elective surgery is undertaken HealthLinks Chronic Care (HLCC) The HLCC funding model is to provide
affects post-operative length of stay, is a 3.5 year pilot program by the capitation funding (termed "flexible
whether sessions comprising the same Victorian Department of Health and funding") to health services for patients
medical specialty are more highly Human Services (DHHS) and evaluated at risk of multiple unplanned (re)
utilised than sessions with a mix of by CSIRO. It is well established that admissions that the health service
specialties, and whether more frail integrated, co-ordinated care can result may be able to impact through better

H E A LT H I N F O R M AT I C S
patients lead to longer procedure times in better health outcomes for people discharge planning, better streaming to
and turnover times between operations. living with chronic disease. There is existing (clinical and social-economic
These insights into scheduling can be also evidence that current funding driven) programs or new service models
used by administrators in targeting mechanisms can limit health services for community management. A novel
efforts to address current challenges in from providing a cohesive and co- risk algorithm will identify the eligible
surgery scheduling practice. ordinated model of care that integrates cohort of patients who are at high risk
hospital, ambulatory and community- of unplanned readmission to hospital
based services. This project aims to and these patients will be enrolled in
remove some of those barriers, enabling the HealthLinks program. Ten health
health services to trial innovative services in metropolitan Melbourne are
models of integrated care. part of this evaluation.

In 2017/18, additional focus group surveys


and interviews were conducted with
three participating health services to
understand healthcare professionals'
perceptions of how HealthLinks has been
implemented at their health services
and their perceived potential impact
of HealthLinks. Quantitative analysis
compared key outcome measures (e.g.
30-day readmission to hospital and
representation to ED) for patients who
were either enrolled at a flexible funding
site and/or who were streamed into a
particular intervention/model of care to
those patients who were enrolled at a
complimentary control site. Preliminary
results show no significant change in
outcomes measured as a result of flexible
funding. At this stage there is limited data
to make any further inferences on the
impact of flexible funding.

aehrc.com
2 8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH DATA ANALYTICS (continued)

Predicting unexpected
patient deterioration in a
digital hospital setting
In 2017/18, planning commenced for
a new project aimed at reducing the
incidence of unplanned adverse patient
outcomes within hospital using new
applied predictive tools developed
specifically for a digital hospital setting.
Such a system is the future of hospital
health care. The specific outcomes of Source: Clinical Informatics, Metro South Health
focus are to:

1. R
 educe unplanned Intensive Care
Unit admissions from a general
ward
Collaborators:
• Federal Department of Health
2. R
 educe Rapid Response Calls
arising from a non-deteriorating • Victorian Department of Health and Human Services (DHHS)
patient, and • Precedence Health Care/Sonic Health Services
H E A LT H I N F O R M AT I C S

3. R
 educe the rate of serious injury/ • Queensland Ambulance Service
disability or death due to an adverse • Healthcare Improvement Unit, Queensland Health
outcome while in hospital. • Metro South Hospital and Health Service, Queensland Health
Additionally this research will seek to: • Metro North Hospital and Health Service, Queensland Health
• Sunshine Coast Hospital and Health Service, Queensland Health
1. C
 ustomise vital signs thresholds
programmed into a digital hospitals’ • Telstra Health.
Electronic Medical Record that suit
individual patients using historic Project Highlights for 2017/18:
data.
• P
 atient risk stratification model was developed, validated and
2. P
 redict the likelihood of a patient
accepted by the Department of Health for to be used by GPs in the
deteriorating significantly before
National Health Care Homes trial
they become a candidate for critical
or intensive care using a range of • W
 eb based risk stratification algorithm for predicting
data sources. rehospitalisation developed and trial started at Logan Hospital
• R
 eport provided to Queensland Health on patient outcomes around
 wo digital hospitals will be
T after-hours hospital care
participating in this study: the Princess
• F
 irst interim HealthLinks evaluation report delivered to DHHS for this
Alexandra as the training site and the
3.5 year project.
Townsville Hospital as the test site. A
project proposal has been submitted
to the Clinical Excellence Division at Project Aims for 2018/19:
HITEC for potential funding and the
project has employed a post-doc to • H
 ealthLinks Chronic Care evaluation: completion of annual report for
work on the project. year 1 and year 2
• E
 valuation of the web-based risk stratification algorithm at Logan
Hospital
• S
 tart of a new collaboration with Victorian Agency for Health
Information (VAHI) which has oversight of all Victorian hospital data
• S
 tart of a new collaboration with WA Health with a project at Fiona
Stanley Hospital to gain better insight into theatre efficiency
• C
 ontinued support, validation and extension of the Patient Admission
Prediction Tool within Queensland Health.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 2 9

HEALTH DATA ENGINEERING

Our world class Health Data Engineering MoTER chronic disease Delivery of technology into
team is a dedicated team of software
engineers who work with our scientists
platform the health system
across the Australian e-Health Research Many of the projects from the Mobile Our team provides the bridge between
Centre in delivering solutions to our Health group involve testing the research outcomes and deployment
customers and partners. With specialists effectiveness of new care models of technology into the health system
skills in mobile app design and delivered through mobile technologies by productising, system integration,
development and web-based software like smartphones and wearables and deployment and support. As an
development, as well as specialist sometimes passive sensors. Our team example, PAPT has been made into
knowledge in health IT standards such manages a common framework known a product and is deployed into
as HL7, the team contributes to projects as the MoTER platform to deliver these Queensland Health and licensed to
across the AEHRC. projects, attempting to maximise reuse Telstra Health, greatly increasing
while not hindering innovation. The its impact. The Readmission RISK
Over the past 12 months the team has
platform consists of iOS and Android Stratification system is deployed at
continued to develop our MoTER mobile
native applications and a web portal for Logan and being used daily for decision
phone platform to support our mobile
clinicians to review the collected data. support.
health projects, worked with our clinical
terminology specialists to deliver on the
National Clinical Terminology Service, AEHRC on FHIR Clinical trial support
developed a new version of our MedText
medical narrative processing software Activities include: We provide support for clinical trial data
and developed FHIR-based resources for management systems like REDCap and
• p
 articipating in Connectathons and OpenClinica as well as custom data
use across our projects. The team also
Hacking Health to build skills and collection solutions. REDCap has been
continues to support a number of clinical
awareness; using the Medications extended to support using an external
trials with various clinical trial software
Resource to represent medications FHIR-based terminology server for
packages.

H E A LT H I N F O R M AT I C S
from the Australian Medicines coded data fields.
Terminology and other sources;
using ValueSets for medication input
in mobile apps
• e
 xperimenting with the use of FHIR
ConceptMaps for MedText
• investigating the use of FHIR to
represent and exchange clinical
trials data, both as a view of ODM/
CDISC data and natively
• d
 ynamic generation of Angular-
based data entry user interfaces
from FHIR profiles
• a
 n extended version of the HAPI
FHIR Server, Sapphire, that
delegates to Ontoserver for its
terminology subsystem to take
advantage of its advanced SNOMED
capabilities.

A SMART on FHIR reference


implementation has been developed
incorporating OAUTH against third-
party systems (such as Google) and the
AEHRC FHIR server. The first application
to use this reference implementation
will be a genomic test ordering system.

aehrc.com
3 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

BIOMEDICAL INFORMATICS

2017/2018 Science and impact highlights


• T
 he Biostatistics team was involved in the Nature (IF40) publication that described the first blood test for Alzheimer’s disease
with greater than 90% accuracy to detect disease pathology in individuals older than 60 years.
• O
 ur cloud computing platform for image analysis 3x automated image analysis pipelines has processed more than 3000 MRI
and PET scans, including the PET scans used to evaluate the blood test.
• O
 ur cloud computing genomics applications around clustering and genome editing have received broad coverage including
iAwards recognition.
• T
 he first clinical trial using MRI to guide prostate radiotherapy has been progressing well and relies on several important
technologies developed by the Biomedical Informatics Group.
• A suite of machine and deep learning techniques have been deployed in numerous tools.
• M
 ultiple NHMRC clinical studies and clinical trials, including very large multi-year studies on Alzheimer’s. The recognition of
our capabilities in clinical study data management is increasing with multiple studies relying on our support and expertise.

Biomedical Informatics Group Leader:


Jurgen Fripp
The Biomedical Informatics research group
develops innovative medical technologies
to quantify genotyping and phenotyping. By
B I O M E D I C A L I N F O R M AT I C S

developing software and imaging techniques,


we aimed to create value around three main aspects:
• more precise and more affordable diagnosis/screening
• personalised and more effective therapy
• selection of individuals for clinical studies and trials. Figure 1: Our portfolio of technologies covers (blue) a vast areas of disorders (white)

Medical Image Analysis Team Leader: Jason Dowling


The Medical Image Analysis team is leading a paradigm shift in radiology from qualitative to (semi-) quantitative
imaging and the development of a new generation of ‘imaging biomarkers’. The technology developed turns
images into information that is used for earlier detection of diseases and improved the diagnostic accuracy. These
technologies are used in trials across a number of clinical areas including, prostate radiotherapy planning, cartilage
health assessment, characterising cerebral palsy risk in infants, surgical planning and paediatric MR-based lung
assessment.

Neuroimaging Team Leader: Sam Burnham


The Neuroimaging team uses their deep knowledge of medical instrumentation, image processing and machine
learning algorithms to automatically extract and present pertinent information from medical image data both at
the scale of populations and individuals. The team contributes to image-based biomarker analysis for a number
of large studies and supports a range of large Alzheimer’s disease trials around Australia, including the recently
funded Alzheimer Dementia Network (ADNeT). ADNeT is part of Australia’s quest to find cures and prevent and
better manage dementia, involving a registry of clinical trial volunteers to fast-track research and translation. The
team have recently partnered with Maxwell Plus in a CRC-P project to translate CSIRO’s CapAIBL software which
uses artificial intelligence to assess the Amyloid burden in the brain from PET scans. This will involve EC approval
of the software tool so that it will be available for clinical use in nuclear medicine and radiology.

Transformational Bioinformatics Team Leader: Denis Bauer


Our Transformational Bioinformatics team develops clinically usable tools for the incorporation of large, complex
and diverse life science datasets (such as high throughput sequencing, gene expression, proteomics) to facilitate
better patient treatment and improved clinical outcomes.

Biostatistics Team Leader: James Doecke


Behind every large scientific finding seen in the newspaper, on the web, or in scientific journals is a team of
statisticians working to answer the biological questions posed by leading scientists today. Our team of statisticians
is well placed to analyse biomedical data with the view to interpret some of the world’s most important medical
research questions.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 3 1

PROJECTS

MEDICAL IMAGE ANALYSIS


Medical imaging is critical in achieving further improvements in outcomes for patients Neurosurgical planning
and in driving efficiencies across the health system. The medical image analysis team
combines expertise in 3D image processing, machine learning and medical physics with
system: CONSULT
nationwide clinical collaborations to automatically extract and present information for CONSULT aims to reduce rates of
clinicians and researchers, enabling optimal clinical decision making and treatment adverse events in neurosurgery through
delivery. Projects within the team range across the lifespan (pre-term infants to utilisation of cutting-edge brain mapping
diseases of old age) and from head (brain pathways) to toe (musculoskeletal analysis). techniques, including advanced functional
MRI and diffusion MRI tractography.
Cerebral palsy and paediatric neuroimaging The software produces a 3D model of a
patient’s brain and its wiring patterns.
Paediatric neuroimaging of children with Neurodevelopment of preterm-born
Neurosurgeons can then interact with
cerebral palsy infants
this 3D model to plan the safest angle and
Cerebral palsy (CP) remains the world’s Very preterm infants (born before 31 depth to “cut” to cure patients without
most common childhood physical weeks gestation) have a high risk of an inducing critical injuries. The model and
disability secondary to an insult to the adverse neurodevelopmental outcome. plans integrate with surgical-guidance
developing brain. Approximately 700 Approximately 5% of these infants will be technology, providing surgical teams

B I O M E D I C A L I N F O R M AT I C S
babies born each year in Australia are diagnosed with cerebral palsy, and half of with real-time feedback on progress and
later diagnosed with CP. Understanding them have later learning and behavioural potentially imminent safety issues during
the relationship between brain structure difficulties. Diagnosis of cerebral palsy is surgery itself. CONSULT aims to make
and clinical function is crucially important currently made on average at 19 months brain surgery safer, more effective, and
for diagnosis and prognosis. corrected age, diagnosis of learning allow surgeons to treat patients who
and behavioural difficulties even later. would previously have been considered
Using brain MRI, we are developing During the first two years of life, there too high risk. This work is performed
methods to quantify the degree of brain is substantial scope to reduce later in conjunction with clinical partners,
injury and brain developmental status of difficulties, because at this age the brain including radiologists and neurosurgeons
children with CP relative to neurotypical has a large capacity for repair. Earlier at prominent Australian hospitals. The
children. Together with our collaborators diagnosis and prognosis are crucial to research is funded through grants
at The University of Queensland and the identifying infants at high risk so that from Advance Queensland (Research
Cerebral Palsy Alliance, our team is using tailored therapy can be provided at an Fellowship), the Royal Australian and
these methods to predict clinical function early age when it is most effective. New Zealand College of Radiologists, and
and to assess neuroplasticity in response
RBWH Foundation (Diamond Care Grant).
to intervention. One observational study We are developing approaches to model
of children aged 8-10 years is currently normal neurodevelopment of the infant
underway (Predict-CP), with the aim brain using MRI, and identify abnormal
to determine the relationship between development. These approaches include
brain development and clinical function, methods to improve image quality,
as well as two randomised controlled automatically delineate brain structures
trials of early intervention for infants of interest to observe their growth and
aged three months to two years at high changes in their microstructure, and
risk of CP (GAME and REACH). Together the extent to which different parts of
with collaborators at The University of the brain communicate. Together with
Queensland, funding has been secured for our collaborators at The University of
a new randomised controlled intervention Queensland and the Royal Brisbane and
trial for children with CP aged 6 to 18 Women’s Hospital, our internationally
years (HABIT-ILE). Additional funding recognised team is using these methods
has been secured with collaborators at to predict the potential for adverse
Melbourne University for an observational neurodevelopment earlier, and to more
study of the early natural history of CP. accurately identify those in need of therapy.

This research will enable major advances This research will enable major advances
in the management of children with CP in the management of preterm babies
by providing a means to identify those and has the potential to reduce the
therapies that will provide the greatest burden of cerebral palsy on the health Figure 2: CONSULT: Automated delineation of the optic
pathway, including Meyer’s Loop, which is commonly
benefit to the individual child. system, while increasing the integration severed during temporal lobectomy and usually
of these children into society. considered extremely difficult to identify with MRI.

aehrc.com
3 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MEDICAL IMAGE ANALYSIS (continued)

Musculoskeletal image
analysis: ChondralHealth
Osteoarthritis (OA) is a joint disease
that characterised by the breakdown
of joint cartilage and underlying bone.
Traumatic injuries such as a tear of
the anterior cruciate ligament (ACL)
increase the risk of developing OA and
provide a targeted way to investigate
early pathophysiological changes in
Figure 1: Femoroacetabular Impingement in the Hip: An automated MR-based method was developed for
cartilage and intervene in the disease quantitative assessment of the hip joint that evaluates 3D bone morphology and provides 360° calculation of
process. alpha angles around its femoral head-neck (FHN) junction. Analysis of the upper anterior quadrant of the FHN
junction enable to distinguish bone lesions characterised by elevated alpha angles as shown in the figure (red
arrow): hip joints having (a) smaller and (b) larger anterosuperior and anterior alpha angles (left: polar plots
Our current research involves
of alpha angles around the FHN junction, right: 3D bone models of the proximal femur surfaces automatically
developing techniques using advanced segmented from the MR scan).
magnetic resonance imaging (MRI)
and image processing that can be used
to improve the clinical picture of the
MRI-based paediatric lung structure and function
pathophysiological processes preceding assessment
B I O M E D I C A L I N F O R M AT I C S

the development of OA. The algorithms This project is a collaboration between CT scans every two years during which
we have developed can be used for the Lady Cilento Children's Hospital, time untreated, asymptomatic infections
automated segmentation of joint Siemens Healthcare, the Herston can permanently damage their airways.
cartilages that are used for computation Imaging Research Facility and CSIRO. Clinicians are also unable to quickly
of morphological and biochemical The aim is to improve health outcomes and accurately evaluate response
imaging biomarkers of the cartilage for children with Cystic Fibrosis (CF) to treatment. Children with A-T are
damage. and Ataxia-Telangiectasia (A-T) by extremely radio-sensitive and cannot
To evaluate and explore novel developing methods to use MRI to have CT scans. Non-invasive monitoring
biomarkers we are running a clinical provide information on lung status. of disease progression and treatment
trial with patients after knee or hip response is vitally important in
Currently the most informative method managing disease onset and extending
injury (ACL rupture, labral tear) and for lung imaging in these children is
healthy controls. The novel MRI life for these patients.
computed tomography (CT) scanning.
protocols and processing techniques CT scans combine a large number of To address this clinical need we are
are also evaluated on clinical data from x-ray images and repeated CT scanning developing image acquisition methods
our academic and industrial partners increases a child’s cancer risk due to and software to extract quantitative
(Steadman Philippon Research Institute, the radiation dose delivered. This is a disease status metrics from MRI.
USA). Results of these studies have critical consideration, particularly as This work is supported by a CSIRO
been presented in clinical (e.g. European the lifespan of people with CF and A-T OCE postdoc position and external
Journal of Radiology) and technical increases from improved treatment. For funding from the Ataxia Telangiectasia
(e.g. Medical Physics) international this reason children currently only have Children's Project.
journals. The developed software,
ChondralHealth, has been shared
with our industrial partner (Siemens MRI-alone radiation therapy planning for prostate cancer
Healthcare, Germany) who are running
validation studies at multiple centres. In collaboration with the Calvary Mater improvements are also leading to
Newcastle Hospital, and with funding cost savings arising from the reduced
Results of this research will provide support from the Prostate Cancer need to manage side effects. After
tools for medical practitioners to Foundation Australia and the NSW two successful retrospective trials
improve the diagnosis and clinical Cancer Council, we have developed involving 80 men, this work has grown
management of OA worldwide. In the first atlas-based method to map into a Phase II multi-centre prospective
addition we are using the multi- realistic electron densities to MRI scans trial for MRI-alone localised prostate
tissue 3D imaging capabilities of for dose calculations. This method is cancer external beam radiation therapy
magnetic resonance (MR) for non- now being used in the clinic for the first (ANZCTR trial: ACTRN12616001653459).
invasive objective assessments of hip time to refine radiotherapy planning As at 30 June 2018, 15 men have
joint biomechanics, morphology and during treatment to reduce side effects completed treatment for localised
biochemistry in patients. of prostate cancer patients. These prostate cancer with this technology.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 3 3

Improving radiotherapy treatment clinical trial quality assurance


Radiotherapy is a well-established, has been shown to result in reduced the development of the first automated
cost-effective treatment which has patient outcomes. However, manual approach to contouring assessment
an evidence-based indication for review of contouring is resource using four large clinical trial datasets
approximately 50% of cancer patients. intensive, expensive and for advanced with the aim of changing practice for
The weakest link in treatment is treatments unachievable in a timely future studies and enabling consistent
the definition of treatment volumes fashion. This NHMRC-funded project, in assessment in the clinic.
(contouring). Lack of accuracy and collaboration with the major Australian
consistency in clinical trial contouring radiation oncology centres, involves

Collaborators: Project Highlights Project Aims for


for 2017/18: 2018/19:
• Royal Melbourne Hospital • C
 hondralHealth software delivered • C
 linical validation of the
• Q
 ueensland Institute of Medical to Siemens for internal evaluation CONSULT platform for
Research • A
 lex M. Pagnozzi, et al. Best neurosurgical planning
• McCusker Foundation Scientific Poster Award, • E
 valuation of ChondralHealth

B I O M E D I C A L I N F O R M AT I C S
American Academy for Cerebral software in a multicentre study
• Q
 ueensland Cerebral Palsy and
Palsy and Developmental • D
 evelopment of automatic
Rehabilitation Research Centre
Medicine 71st Annual Meeting, quantitative MRI reports for
• Siemens Healthcare September 13-16, 2017, Montreal children with cerebral palsy and
• S
 teadman Philippon Research • L
 ee Reid: People’s Choice Award infants at risk of cerebral palsy
Institute and CSIRO Accelerator Award at • M
 RI-based non-invasive and
• Lady Cilento Children’s Hospital Impact 7, Melbourne zero radiation paediatric lung
• R
 oyal Brisbane and Women’s • International multi-centre structure and function evaluation
Hospital prospective clinical trial for MRI- (for cystic fibrosis and ataxia-
• Q
 ueensland Cerebral Palsy and alone, external beam radiation telangiectasia)
Rehabilitation Research Centre therapy for localised prostate • D
 evelopment and validation of
cancer methods for automatic clinical
• Stella Maris Institute, Pisa
• Initiation of a new project aimed trial quality assurance and
• Ingham Institute for Applied
at MRI-based non-invasive linked information extraction
Medical Research
lung structure assessment for from retrospective radiation
• C
 alvary Mater Newcastle children with cystic fibrosis oncology data sources
Hospital
• D
 evelopment of a method • D
 evelop approaches to predict
• Sir Charles Gairdner Hospital to quantify structures from adverse outcomes from
• L
 iverpool and Macarthur Cancer neonatal MRIs neurosurgery in adults using
Services • D
 evelopment of a method for structural and/or diffusion MRI
• U
 niversities of Queensland, improved regional measures of • S
 tart development of motion
Newcastle, New South Wales, the cartilage from biochemical correction methods to improve
Western Australia, Melbourne, MRIs at 3T/7T the analysis of dynamic image
Sydney, South Australia and • D
 evelopment of a method for data
Wollongong automated detection of small • V
 alidation of methods for
• University of Barcelona, Spain lesions on MRI images contrast synthesis and
• Brisbrain and Spine. • D
 evelopment of a method for segmentation based on deep
automated delineation of the learning
optic radiation using MRI • S
 egmentation tools for
• J ournal papers in top rated retrospective mining of radiation
medical imaging and clinical oncology imaging data.
journals
• S
 uccessful grant funding,
including NHMRC, RANZCR,
RBWH Foundation.

aehrc.com
3 4 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MEDICAL IMAGE ANALYSIS (continued)

P
 hD Student

Miles Seidel Ashley Gillman


University of Queensland University of Queensland

Brain development during the neonatal Patient motion is an important


period plays a decisive role in the consideration in modern PET image
formation of adverse neurological reconstruction. Advances in PET
outcomes associated with very preterm technology, especially the introduction
birth, including motor impairments, of combined PET/MR, mean motion is
cerebral palsy, and cognitive deficits. becoming an increasingly important
While structural brain abnormalities influence on image quality, and motion-
evident on MRI at term equivalent age are induced artifacts can have an adverse
predictive of poor neurodevelopmental effect on clinical outcomes, including
outcomes, definitive diagnosis is rarely missed diagnoses and oversized
made until early childhood and many radiotherapy treatment volumes.
infants without brain abnormalities also
develop neurological deficits. In this PhD, improvements to several
bottlenecks to accuracy in the motion
Second, segmentations of the thalamus compensation pipeline are being
B I O M E D I C A L I N F O R M AT I C S

and hippocampus will be used as seed addressed. Methods for the use of a
regions for delineation of white-matter camera to track head motion, and the
tracts to explore longitudinal changes tracking of head motion directly from
in connectivity measures in associated raw PET data in three dimensions have
tracts. The last aim is to investigate been devised. A particular focus of the
the relationship between structural project is the previous under-researched
measures, diffusion measures, and area of PET motion correction in the
neurological outcomes at three and thoracic area, where novel methods
12 months of age in a large cohort of utilise the capabilities of the newly
neonates born very preterm. available PET/MR scanners.

This work has the potential to benefit


future clinical patients and the wider
scientific community. Clinicians stand
to gain an improved confidence test
results, leading to improved clinical
decision making. Patients will be
able to avoid re-scans where motion
corruption rendered the original scan
unusable. Researchers may be able to
run previously unachievable tests, such
as dynamic modelling of deformable
regions like the abdomen and pelvis.
The software developed is also intended
to foster and accelerate PET/MR motion
correction efforts by the research
community.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 3 5

NEUROIMAGING

Healthy ageing and Alzheimer's disease research Prospective Imaging Study


Neurodegenerative diseases refer to a processing and machine learning
of Ageing (PISA): genes,
group of age-related brain illnesses that algorithms to automatically extract brain and behaviour
result in progressive loss of brain tissue and present pertinent information
PISA studies the interplay between
and cognitive function. Early detection is from medical image data both at the
genetic, epigenetic and environmental
now recognised as a critical component scale of populations and individuals.
factors for dementia, and also aims
to developing effective treatment for We provide automated quantification
to identify risk factors that could be
various forms of neurodegenerative of such biomarkers to collaborators at
modified through intervention – such
diseases, including Alzheimer’s Austin Health, the Florey Institute of
as lifestyle choices. The study is a
disease, as it allows interventions prior Neuroscience, McCusker Alzheimer’s
unique international research resource,
to widespread and irreversible tissue Research Foundation, Edith Cowan
providing new links to studies into
loss. The primary pathway for early University, Macquarie University for the
the causes of dementia, assisting
detection is through identification of Australian Imaging and Biomarker and
clinical trials in dementia prevention,
neuropathology biomarkers derived Lifestyle (AIBL) study, KARVIAH and PISA
and bringing about new possibilities
from neuroimaging. studies. Within these collaborations
for translational research into this
we have provided strong evidence that
The Clinical Imaging team combines important public health issue.
Aβ-amyloid (Aβ) plaque accumulation
knowledge obtained from collaborating commences 10-20 years before clinical PISA uses cutting edge imaging
physicians with a deep knowledge symptoms, highlighting a significant technologies to examine the
of medical instrumentation, image

B I O M E D I C A L I N F O R M AT I C S
window for pre-clinical treatment. neurobiological features associated
with high risk for dementia, and identify
AIBL and the Alzheimer's disease network (ADNet) the changes that lead to a patient’s
transition from high risk to cognitive
As expectations increase for data-driven number of centres, however there has impairment. The combined use of
knowledge and understanding, it will been considerable variability in the genetic risk scores and neurobiological
be necessary to consider data streams exact numbers reported as quantitative markers creates a potential prognostic
in combination with each other, not as outcome measures of tracer retention. marker for dementia development.
silos, to provide improved classification, The centiloid scale was developed by an Outcomes of the study will inform
categorisation of comorbidity burden, international working group to alleviate and establish a platform for future
diagnosis and prognosis at the some of these issues and provide a intervention programs that target
individual level. This will require framework to standardise measures preventing and treating dementia. So
harmonisation of imaging biomarkers to of Aβ burden from PET images. The far, 93 subjects have been recruited in
give the best overall picture. centiloid framework was applied to the PISA.
calibration of CapAIBL; we showed that
An example where we are currently reliable centiloid estimates could be
working on harmonisation is the obtained with our CSIRO pipeline. There
centiloid scale. Amyloid imaging with is still a need to validate the centiloid
positron emission tomography (PET), scale with longitudinal dataset.
has produced remarkably consistent
qualitative findings across a large

aehrc.com
3 6 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

NEUROIMAGING (continued)
B I O M E D I C A L I N F O R M AT I C S

Figure 1: Example quantitative reports for PET analysis (CapAIBL), MR morphometry (CurAIBL) and FLAIR white matter
hyperintensity segmentation (HIST)

Sterling's Dream UK Biobank


Cholinesterase inhibitors (ChEIs) are With the current data revolution and its
a major class of cognitive enhancing application to the medical and health
drugs designed to target the fields, there will be an increasing
symptomatic treatment of Alzheimer’s pull for more and more personalised
disease. This is based on the knowledge health data across populations. This
that cholinergic degeneration of increase in demand will be equally
the basal forebrain is a hallmark applied to imaging markers requiring
pathological feature of Alzheimer's high throughput technologies for image
with specific vulnerability to amyloid- processing. We will implement new
beta (Aβ). Despite widespread use in strategies using deep learning to tackle
clinical practice, in reality only 30-35% this demand. This includes processing
of patients respond to treatment. The of over 50,000 images as part of our
ability to identify patients who will successful proposal to UK Biobank this
respond to ChEI using biomarkers would financial year.
significantly impact treatment and policy
guidelines for the use of these cognitive
enhancing drugs. In terms of brain
scans, Positron Emission Tomography
(PET) and Magnetic Resonance
Imaging (MRI) will be acquired and
from the findings we will better
understand if there are differences in
the characteristics of healthy brains
compared with those with early stages
of Alzheimer’s disease.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 3 7

Project Highlights Project Aims for


Collaborators:
for 2017/18: 2018/19:
• ADNeT team • S
 uccessful CRC-P proposal • E
 valuation of Centiloid
• A
 ustralian Imaging Biomarkers with partners Maxwell Plus calibration in longitudinal PET
and Lifestyle (AIBL) study to commercialise our IP on Amyloid study
automated PET and MRI image • T
 ranslation of CapAIBL and
• Q
 IMR Berghofer Medical
analysis CurAIBL to Maxwell Plus
Research Institute
• C
 o-authored paper in the journal • D
 evelopment of novel deep
• F
 lorey Institute of Neuroscience
Nature (IF40) describing the first learning methods for image
and Mental Health
blood test for Alzheimer’s disease quantification
• N
 uclear Medicine and Centre for with greater than 90% accuracy
PET, Austin Health • H
 armonisation of imaging
to detect disease pathology in
biomarkers across sites for
• University of Melbourne individuals older than 60 years
ADNet
• Edith Cowan University • D
 eployment of the HIST
• V
 alidation of CapAIBL for use
• University of Wollongong white matter hyperintensity
with Tau and FEOBV tracers
segmentation tool on MilxCloud
• University of Sydney • Interpretation of imaging and

B I O M E D I C A L I N F O R M AT I C S
• 1
 7 published journal articles and
• Macquarie University other data to provide clinically
31 conference papers.
• University of Western Australia. meaningful staging criteria for
progressive diseases such as
Alzheimer’s disease.

P
 hD Student

Biting Yu Cathryn McKenzie


University of Queensland University of Western Australia

Magnetic resonance imaging (MRI) is a Cognitive reserve is the hypothesised


widely used medical imaging modality capacity for adapting to physiological
that can be configured to provide changes in the brain, such as those
different contrast between tissues in caused by ageing or disease pathology,
the human body. By setting different in order to maintain normal cognitive
scanning parameters, each MR imaging functioning. Cognitive reserve can be
modality reflects the unique visual assessed as the difference between an
characteristic of the scanned body part, individual’s observed test performance,
which benefits medical image analysis and the performance predicted based
from different perspectives. To utilise on structural brain health (such as
the complementary information from volume of brain matter). Cognitive
different imaging modalities, cross- reserve, as indexed by this residual
modality MR image synthesis has been approach, has been found to be related
sought after and attracted increasing to factors thought to build reserve such
research interest recently. This project as education, cognitive decline and future
aims to develop novel techniques in dementia diagnosis, independently of
deep learning applied to medical image brain health, and increased efficiency
synthesis for the purpose of improving of brain networks as measured by
current image analysis workflows. functional magnetic resonance imaging.
While this approach has shown initial
promise, some aspects that might
improve our understanding of the
cognitive reserve construct remain
unexplored. This project will address
these gaps in the current research.

aehrc.com
3 8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

TRANSFORMATIONAL BIOINFORMATICS

The Transformational Bioinformatics


team develops novel bioinformatics
solutions for research and industry
using the latest in cloud and big data
infrastructure. We specifically focus
on population-scale ‘omics (genomics,
transcriptomics, methylomics) analysis
as well as genome engineering
applications as the two high impact life
science areas. We engage with the rest
of the program on developing algorithms
that can jointly harness information from
diverse sources like genomic profiles,
personal sensing devices, and electronic
health records to build “smart analytics”
systems that are predictive of health
outcomes.

NHMRC Dementia Team


Grant: motor neurone
B I O M E D I C A L I N F O R M AT I C S

disease and dementia


CSIRO is a partner in the Dementia
Team Grant led by Prof Ian Blair at
Macquarie University, which aims to
uncover the molecular mechanisms of
Amyotrophic Lateral Sclerosis (ALS),
the disease Stephen Hawking suffered
from. CSIRO is responsible for the
genomic data analysis of 800 Australian
WGS ALS samples, as well as the
data integration of other ‘omics data
collected through the lifespan of this
five-year project. These samples will
form the Australian contribution to the
international Project MinE consortium.
CSIRO is also engaged in this wider
Figure 1. Novel machine learning methods for disease association (here synthetic phenotype hipster-index)
initiative by analysing this 22,000 strong
case-control cohort.
Cloud-based genome analysis tools
Outcomes
This project develops novel approaches Outcomes
• P
 ublication in Cell Neuron
for analysing genomic data from • F
 eatured as one of
(IF=14.024, cited 4).
population-scale cohorts (i.e. common ComputerWeekly’s Top-10 IT stories
• T
 he work will be presented by CSIRO and rare variants). Our software of 2017 and in an interview on
as a keynote at the International platform VariantSpark has been Future-proofing Companies.
Conference on Frontotemporal demonstrated to save companies 80%
Dementias in Sydney. • M
 achine Learning Algorithm
of cost by requiring fewer samples for
submitted for publication.
• C
 SIRO is developing a relatedness standard case-control studies (GWAS)
tool that improves upon existing to reach significance. It achieves this • C
 SIRO presented VariantSpark as
approaches for detecting distantly by using Apache Spark technology keynote presentation at international
related individuals and accurately for building machine learning models IT conferences (AgileIndia,
assigning the degree of relatedness. capable of analysing the full genome AIDevDays) and will present at
simultaneously. This allows the tool to the AWS Public Sector Summit,
identify genomic locations that jointly Canberra.
contribute to disease as oppose to being
limited to detecting only the strong
individually contributing genes (GWAS).

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 3 9

Computational genome editing services


This project develops computational We also improved accuracy by 30% by Outcomes
solutions that improve the accuracy of tapping into CSIRO’s 25-year experience • P
 ublications in The CRISPR Journal
genome engineering applications (on- in the science of how the 3D organisation and Frontiers in Pharmacology
target scoring, SNP-aware off-target of the genome affects the accessibility (IF=4.4).
search) to enable novel application of the genomic address. Finally the
• C
 SIRO presented GT-Scan on 'This is
areas in high-precision applications machine learning models are built to
my architecture', a youTube channel
such as human health. The task of more precisely fit experimental use
by Amazon Web Services featuring
finding a suitable genome editing spot is cases (e.g. SNP-aware prediction for
international clients with the most
comparable to finding the right grain of wild populations) thereby being able to
outstanding software architecture.
sand on the beach, i.e. it needs to have personalise results to individual patients.
the right shape and colour (properties • C
 SIRO will present the GT-Scan
for CRISPR to bind) and also be unique setup on Alicloud at Alibaba Cloud
compared to all other grains on the Summit in Singapore.
beach (for CRISPR not to accidentally
bind to another gene). This is a very
expensive task computationally. We
reduced the overall runtime for this task
from weeks to seconds by massively

B I O M E D I C A L I N F O R M AT I C S
parallelising the individual search tasks
using a revolutionary new compute
approach called Functions-as-a-Service.

Figure 2. Genome Engineering Symposium in Canberra.

Project Highlights Project Aims for


C
 ollaborators:
for 2017/18: 2018/19:
• Macquarie University • T
 he team has published six • E
 stablish VariantSpark
• C
 hildren’s Medical Research journal papers (three senior internationally as the technology
Institute (CMRI) Westmead author, and two in IF>10) and of choice for large cohort
over 20 conference papers analysis to capture part of the
• Guangzhou Medical University
(seven keynotes at major $19.99 billion (2020) genomics
• National Measurement Institute national and international market (Markets and Markets)
• University of Newcastle conferences with up to 800 • D
 evelop technology for
• Q
 ueensland University of attendees, 14 oral presentations, ImmunoEngineering in
Technology all presenter or senior author). collaboration with CMRI
• Australian National University • O
 ur work was featured in Westmead and Guangzhou
international media articles Medical University to capture
• A
 ustralian Genomics Health
(AWS, ComputerWeekly) and part of USD $7.5 billion (2024)
Alliance
we won the H&B “Outstanding genome editing market (Global
• M
 elbourne Genomics Health Collaboration Award” and Market Insights, Inc.)
Alliance FreshScience Award. • P
 ublish a high impact journal
• G
 raven Institute for Medical • T
 he team has secured $800K publication featuring original
research in competitive funding and research using our technology.
• ProjectMine (Europa) has engaged 14 companies in
customer discovery conversations
(e.g. Illumina, Sonic Health,
Samsung, Sanford Health).

aehrc.com
4 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

TRANSFORMATIONAL BIOINFORMATICS (continued)


B I O M E D I C A L I N F O R M AT I C S

Figure 3. Marc Horlacher (Masters student), Amnon Bleich (Masters student) and Aidan O’Brien (PhD student).

S
 tudent Profile

Aidan O’Brien, PhD student Amnon Bleich, Masters student in Marc Horlacher, Masters student in
co-supervised with ANU. the Bioinformatics Student Exchange the Bioinformatics Student Exchange
Program. Program.
Topic: Genome editing is a new
molecular discipline with transformative Topic: Genome-wide association studies Topic: Genome editing with CRISPR-
impact on human health, environmental (GWAS) have contributed towards finding Cas9 is opening up new avenues in
and agricultural applications. Of disease genes over recent decades. The almost all facets of research. While
particular promise is the ability to traditional GWAS approaches use odds- CRISPR-Cas9 normally induces random
insert synthetic DNA into the genome ratio and defined distributions, like chi- mutations, some control can be exerted
at precise locations as enabled square tests, to compute p-values for by leveraging specific repair pathways
through CRISPR-Cas9 and homology- associating SNPs with traits. While this within the cell. One such pathway is
directed-repair. Being able to do is appropriate and efficient for simple Microhomology Mediated End Joining
this with the precision and efficiency genotype-phenotype patterns, they are (MMEJ), which results in a controlled
requires extensive computational not capable of identifying interacting deletion of a very specific portion of the
optimisation processes. This project SNPs for multi-genic diseases (e.g. DNA. What drives a target to be repaired
builds sophisticated machine learning diabetes, ALS). Machine learning via this mechanism is still unknown,
models that enable researchers to methods, such as Random Forest, can making its use unreliable. This project
identify the optimal genomic location overcome these limitations. This project uses machine learning approaches
for an intervention. Working together investigates how to improve upon to identify features that predict MMEJ
with Australia’s premier research standard Random Forest algorithms to repair, helping researchers design
organisation and CRISPR facility, the be applicable to genomic data analysis. CRISPR-Cas9 applications that take
computational tools will be validated advantage of this pathway and providing
on novel datasets and enable new greater control over the results.
application areas.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 41

BIOSTATISTICS

Australian Dementia Neuropsychology and its Instrumental relationship


Network (ADNeT) disorders with pharmaceutical
Our Biostatistics team combines data Members of the team play a key role in companies
from multiple modalities to answer a further project studying disorders of Our team is working closely with
clinical research questions. This involves cognition and mental health, including international pharmaceutical companies
using statistical methods to combine Parkinson's disease, Alzheimer's Roche and Biogen to assess the
data from imaging, genetics, genomics, disease and cerebral palsy. Providing cognitive trajectory of Alzheimer's
proteomics, neuropsychology and key statistical support and project disease from its very early stages
clinical biomarkers. The team works guidance, team members analyse (prodromal and pre-clinical) through
with internal and external collaborators project data to align with research to late stage clinical AD. Research is
to investigate destructive pathological hypotheses and define novel pathways focused around changes in cognition,
process in Alzheimer's disease. into disease ethology. blood and CSF-based biomarkers and
Our team are key members in the In Alzheimer's disease, the team pathological proteins via PET imaging.
Australian Imaging Biomarkers and conduct research into biomarkers Along with collaborators from world-
Lifestyle (AIBL) study. The core mandate from CSF to align with PET imaging. leading laboratories, our team members
for AIBL is undertaking research to Research from this project has led to are instrumental in round table
identify and validate biomarkers for a real bench-to-bedside outcome, with discussions to discuss the best way
the early detection and treatment results from biomarker studies guiding forward in conducting research across
multiple countries. Figure 1 shows the

B I O M E D I C A L I N F O R M AT I C S
of neurodegenerative disorders and decisions made on disease diagnosis in
psychoses. By bringing together the clinic. In Parkinson's disease, the disease progression model developed in
industry, end users and health care team assesses genomic biomarkers collaboration with partners at the Florey
providers, researchers in the AIBL study that align with the presence of disease Institute for Mental Health, a figure
aim to develop and commercialise our physiology, and in cerebral palsy, the which is widely cited and used in the
research findings in order to deliver team provides statistical guidance on literature.
changes to treatment in medical and MRI data to assess cognitive function
health care practices. Recently, key from children with the disease.
researchers from AIBL were awarded This research has led to six journal
a $20 million grant to start the publications in high-ranking journals.
Australian Dementia Network. This
research project will kick off in 2019,
and will involve recruiting about 4000
participants over the next five years
from around Australia, hosting data
from imaging, biomarkers and lifestyle
similar to the AIBL study.

This year our research in biomarkers


for the early detection of Alzheimer's
disease led to 20 publications and 12
conference presentations. Key research
from the team led to many strong
publications, including papers in top
ranking journals such as Neurology and
Lancet Neurology.

Figure 1: Alzheimer’s disease progression model.

aehrc.com
4 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

BIOSTATISTICS (continued)
B I O M E D I C A L I N F O R M AT I C S

Figure 2: Genome Engineering for Cancer Treatment conference.

Genome Engineering for


Cancer Treatment conference
Team members are part of the
organising team for the CSIRO-
funded Cutting Edge Biosciences
conference. Last year Dr Denis Bauer
(Transformational Bioinformatics) and
Dr James Doecke (Biostatistics) were
awarded $30,000 based on a proposal
to host a conference around the new
genomic technology CRISPR. The
conference has three international
invited speakers, plus a range of
national experts to talk about the latest
breakthroughs in genomic technologies
around CRISPR. The conference was
held in November 2017 at the ANU in
Canberra and attracted 120 participants.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 4 3

Project Highlights Project Aims for


Collaborators:
for 2017/18: 2018/19:
• ADNeT team • R
 eal bench-to-bedside research • Develop new statistical methods
• CRC in Cognition and its Disorders with biomarker studies leading to • Pursue novel science projects
changes in clinical practice
•  ustralian Imaging Biomarkers
A • Engage external collaborators
and Lifestyle (AIBL) study • C
 o-author paper in the journal
• Inclusion as chief investigators
Nature (IF40) describing the first
•  entre for Applied Statistics,
C on successful grant bids with
blood test for Alzheimer’s disease
University of Western Australia external collaborators
with greater than 90% accuracy
• University of Melbourne to detect disease pathology in • G
 row the team through
•  lorey Institute of Neuroscience
F individuals older than 60 years employment of a post-doctoral
and Mental Health scientist
• 2
 0 published journal articles,
•  uclear Medicine and Centre
N including first author papers • A
 nswer pertinent research
for PET, Austin Health in Lancet Neurology (IF22) and questions resulting in peer
Neurology (IF8) reviewed journal publications
•  chool of Medical and Health
S
and conference presentations.
Sciences, Edith Cowan • 12 conference presentations.

B I O M E D I C A L I N F O R M AT I C S
University
• Institute for Future
Environments, QUT
•  risbane Inflammatory Bowel
B
Disease group
• International Inflammatory
Bowel Disease Genetics
Consortium
• MD Anderson Cancer Center
•  epartments of Neurology,
D
Harvard Medical School
•  aurice Wohl Institute for
M
Clinical Neuroscience, Kings
College London
• Institute of Health Informatics,
University College London

aehrc.com
4 4 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH SERVICES

2017/2018 Science and impact highlights


• Research initiatives were made to innovations to address Autism Spectrum Disorder using chatbots and robotics.
• S
 marter Safer Homes project was kicked off to validate the platform with four service providers, and a testbed project was initiated
to evaluating new features for the platform.
• Feasibility study for mobile health delivery of gestational diabetes (M♥THer) completed.
• Remote-I was commercialised to TelemedC.
• Health
 services projects were highly recognised for their innovation, with five Queensland iAwards (including the Premier’s
award) for mobile health projects (kidney disease and pulmonary rehabilitation management); one Victoria iAward for CALD
Assist; and two Western Australia iAwards and one National iAward for Remote-i in the detection of diabetic retinopathy.

Health Services Group Leader: Mohan Karunanithi


Mohanraj Karunanithi has a doctorate in Biomedical Engineering from the University of New South Wales. He has over
10 years of experience in cardiac research and five years' medical industries experience. At the Australian e-Health
Research Centre, he manages and coordinates the Health Services Group undertaking research in translating services
on screening, diagnosis, management and delivery of chronic diseases and aged care to community care settings.

Mobile Health Systems Team Leader: Marlien Varnfield


With the wide uptake of smartphone, internet and health monitoring technologies in people's everyday lifestyles, the
Mobile Health Systems team is translating the delivery of health care from acute care setting into the community to
relieve the undue pressures hospitals face in managing chronic diseases and illnesses. Our team has demonstrated
H E A LT H S E R V I C E S

capabilities as a world leader in providing scientific evidence supporting mobile health. The team's objective is to make
prevention of and management of chronic disease services accessible to all people from their home or community. To
enable this, the team works closely with clinical partners already providing such services, to develop new innovative
care models and technology-based systems and test through clinical trials for evidence base.

Health Internet of Things Team Leader: Qing Zhang


With the advent of wireless sensors, mobile, and health technologies pervasive in everyday use, new and rich
sources of data are now accessible to determine people's lifestyles and their influences in their health and well-
being. The Health Internet of Things (HIoT) team is at the frontier of having developed an innovative home-care
platform that can access and aggregate these data wirelessly from the environment and/or wearable devices,
and mobile or internet devices. We have been developing and exploring smart data analytics (through machine
learning and artificial intelligence) on aggregated data sets to better support the older community and people with
disabilities to live longer in their homes, and also to support their carers and service providers.

Digital Health Engagement Team Leader: Dana Bradford/Jill Freyne


For effective adoption of new health intervention delivery, particularly using technology, consumer/provider driven
design is paramount. Our Digital Health Engagement team aims to provide closely aligned consumer/provider
design and evaluation for new technology-based care delivery systems. The team is equipped with expertise across
human computer interaction, personalisation and recommendations, persuasive technology and neuroscience.
The team designs technology-based interventions that can be used by clinicians to improve workflow and enhance
service delivery, and by individuals to support them in playing an active role in their health management to meet
short or long term health and lifestyle goals. The team collaborates across the health services group, the e-Health
program and other CSIRO business units to contribute to the delivery of consumer design prototypes for testing in
clinical trials.

Australian Tele-health Research and Development Group Lead: Yogi Kanagasingam


The Australian Tele-health Research and Development Group is our partnership with the Western Australian
Department of Health and our Western Australia node of the Australian e-Health Research Centre.

The team, based in Perth, is centred on research that provides digital disease screening and diagnosis using
telemedicine to enable healthcare accessible to rural and remote Australia. The team's main focus is the development
of novel telemedicine technologies to deliver non-invasive ocular imaging techniques for chronic diseases such
as diabetes, neuro-degenerative diseases such as Alzheimer’s disease and stroke, and burns and wound care
management. The team were recognised for WA Information Technology and Telecommunications Alliance (WAiTTA)
Incite Awards 2018 – Winner of Most Innovative Enabler in Health Care, and Achiever of the Year.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 4 5

PROJECTS

MOBILE HEALTH SYSTEMS


Smart mobile devices and wearable Pain ROADMAP: pilot study in chronic pain management
technologies have created a new
communication channel between Chronic pain is estimated to effect 20% The data collected by the user is sent to
healthcare service providers and of Australia's population; arthritis and a online data integration and analysis
patients. The adoption of mobile health back problems account for 40% of forced portal which generates daily graphs,
systems is on the rise, with indications retirements, and the total economic where periods of over-activity can be
they have the potential to alter the way cost of chronic pain is estimated at identified as well as daily and weekly
health services operate to enable better $34 billion. In collaboration with Metro summaries of:
patient care. Benefits of mobile health North HHS, we started a pilot study in • Pain intensity and variation
include direct individualised patient mid-2017 testing a digital intervention
management, enhanced physician • Mean objective activity and variation
for people with chronic pain. Study
efficiency, convenience, improved participants (n=20) were provided with • D
 osage and frequency of medication
multidisciplinary care, enhanced team a mobile application that tracked daily intake
co-ordination and increased information actives, medications and pain levels while • P
 ercentage of time spent on rest,
access and communication. streaming accelerometer data from a productive tasks and leisure/social
wearable device. activities.
In recent years, the Mobile Health
Systems team has designed, developed Examples of the visualisation of the data

H E A LT H S E R V I C E S
and evaluated patient online portals are given in figure 2. This data allows
and mobile applications to improve the collaborating occupational therapist
coordinated care in a variety of chronic to provide a patient-centric intervention
conditions including heart disease, program to lower medication intake
diabetes, lung disease and kidney and pain levels and ultimately increase
disease. Furthermore, the team has wellbeing.
developed a solution for activity pacing
(in pain management) and is also
involved in a project for early detection of
cerebral palsy, using a wearable sensor
system for monitoring body movements
of neonates. Selected projects are
presented below.

Figure 1. Screenshots of the mobile application. The app alerts individuals to enter their pain score every waking hour. Individuals also enter their daily activities and
opioid medication in an electronic diary.

aehrc.com
4 6 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MOBILE HEALTH SYSTEMS (continued)

Figure 2. Examples of the visualisation of the Pain ROADMAP portal that show a user’s day of activities, levels of pain and times of medication intake. Activities are colour
coded to indicate the type of activity.

The results of this pilot study have The aim of this project is to enable major disease at an earlier age. The infant
been extremely encouraging. After the advances in the management of babies at wearable sensors under development
third monitoring period, participants risk of cerebral palsy, reduce the financial can be used to measure and quantify
reported less pain, stress and anxiety, burden of the condition on the health extremity motor characteristics, indicative
with increases in physical activity. system, and increase societal integration of normal or abnormal development.
Most importantly, there was a general and quality of life for children born with Under normal developmental trajectories,
reduction in the amount of pain this difficult-to-diagnose disease. infant movements will be fluent and
H E A LT H S E R V I C E S

medication taken. 100% of participants varied, representative of a healthy and


indicated that specific feedback provided We are developing methods to detect developing brain. In cases where an injury
by the clinician was worth going even subtle brain abnormalities at an or insult may be present in the brain,
through the monitoring procedures. earlier age, allowing earlier intervention the movements are more patterned,
All participants said that they would which will lead to better outcomes for repetitive, and monotonous. Using
recommend Pain ROADMAP to others. patients. The ability to track development unobtrusive sensors placed at the hands,
over time is allowing us to predict the feet, and head, we can measure the motor
Early detection of cerebral extent of improvement for patients using characteristics and identify infants at risk
different therapies, and for the first
palsy time we can examine the association
of cerebral palsy for follow-up testing. In
contrast to existing screening methods,
Babies at risk of developmental between therapy, physical changes within this measurement tool can also be used
disorders can benefit from early the brain, and clinical outcomes. This in rural clinical and home environments,
medical interventions, but methods for knowledge will allow us to develop the expanding the reach of early screening
early diagnosis are nascent, limited, next generation of therapies and tailor methods beyond specialist centres. At-
and often case-specific. them to the individuals. risk infants can theoretically be identified
One of the methods under development is at an earlier age than previously possible
About 27,000 Australian infants born
the use of wearable sensors to augment and referred for additional radiological
preterm every year have a heightened
existing clinical assessments towards testing and intervention therapies.
risk of developmental delay or difficulties,
including cerebral palsy, minor motor higher sensitivity to the presence of
and behavioural difficulties, and cognitive
impairment. Many have subtle brain
abnormalities that are not easily seen
on cranial ultrasound or conventional
magnetic resonance imaging (MRI). Infants
with clinically observable abnormalities
in movement, speech, vision or cognition
typically undergo physical therapy, which
relies on the brain’s innate plasticity to help
them achieve substantial improvements in
clinical performance. These therapies can
significantly increase their quality of life
and lower their lifetime healthcare costs
but can only be undertaken once a positive
diagnosis has been made.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 47

Innovative mobile health


program for COPD
Chronic obstructive pulmonary disease
(COPD) is the fourth leading cause
of death globally. Resources to help
patients self-manage COPD have
long been available, however, patient
adherence to these remains suboptimal.
We are studying whether an innovative
mobile health-enabled care program
(MH-COPD) will improve the patient
self-management and relevant health
outcomes.

A prospective open randomised


controlled trial has been designed.
In the trial, patients with COPD will
be recruited from The Prince Charles
Hospital, Brisbane, Australia. They will
Figure 3. The care model of the MH-COPD program includes the components of health education, electronic
then be randomised to participate in COPD action plan, symptom monitoring, physical activity, smoking cessation, and inhaler technique.
either the MH-COPD intervention group

H E A LT H S E R V I C E S
(n=50 patients), or usual care control
group (UC-COPD) (n=50 patients) for A B C
six months. The MH-COPD program
has been designed to integrate a
mobile health system within a clinical
COPD care service. In the program,
participants will use a mobile health
application at home to review educational
videos, monitor COPD symptoms, use
an electronic action plan, modify the
risk factors of cigarette smoking and
regular physical activity, and learn to use
inhalers optimally. All participants will be
assessed at baseline, three months, and
six months. The primary outcomes will
be COPD symptoms and quality of life;
the secondary outcomes will be patient
adherence, physical activity, smoking
cessation, use of COPD medicines,
frequency of COPD exacerbations
Figure 4. Selected screenshots showing the user interface of the mobile application.
and hospital readmissions, and user
A: Scheduled educational videos preloaded in the app
experience of the mobile app. B: User interface to record symptoms and risk factors
C: User interface showing an assessment of symptoms in the COPD action plan.
The study has integrated an innovative
mobile system with a clinical COPD
service and will evaluate this approach
through a randomised control trial.
The evaluation will provide a unique
opportunity to improve COPD care in
the community through mobile health
innovations.

aehrc.com
4 8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MOBILE HEALTH SYSTEMS (continued)

Mobile Pulmonary
rehabilitation Platform
(m-PR)
Pulmonary rehabilitation (PR) is
an evidence-based and effective
treatment but only 5-10% of eligible
patients are currently attending PR
services. PR services are delivered
in highly specialised hospitals and
this service model is inefficient and
unsustainable. Lack of motivation,
transport issues and clinical patient’s
liability have been signalled as causes
for low uptake and adherence to the
current PR programs, leading to
poorer clinical outcomes and misuse
of resources. The mobile Pulmonary
Rehabilitation (m-PR) platform is a
novel smartphone and Internet-based
interactive solution to support patients
H E A LT H S E R V I C E S

with chronic lung conditions during their


Pulmonary Rehabilitation (PR). m-PR
is a collaborative study with clinicians
from Metro North HHS and New South
Wales health services partnering to
improve health outcomes and quality
of life for these patients by developing
and implementing m-PR to improve
Figure 5. The “mPR” platform: individualised exercise and education.
the access to PR services, potentially
enhancing the effectiveness of the model
of care and the satisfaction level from
the users. The m-PR system integrates
mobile health technologies including a
smartphone app, the Internet, health
measure devices and multimedia
content to support patients on PR.
These technologies enable a productive
interaction between patients and the
PR teams, which may also contribute to
improving quality and cost-effectiveness
to the current model of care.

For this project, the platform will be


pilot trialled in two sites (one in QLD and
one in NSW). The m-PR platform has
recently won 3 awards at the 2018 QLD
iAwards held in Brisbane in June 2018:
QLD Premier’s Award for Public sector
innovation, Award for Public and Sector
Government and Merit Research &
Development project of the year.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 49

PD-BUDDy: app changing lives of renal patients at Metro South Health


PD-BUDDy is a smartphone and PD-BUDDy is in the final stages of being on the beneficial impact this platform
internet-based interactive system that tested in a feasibility study at Logan has had on the service. PD-BUDDy
accompanies and guides peritoneal Hospital, Queensland. The study had has the potential to provide access to
dialysis (PD) patients through every finished recruitment (n=30) patients, healthcare in remote communities that
step of their dialysis treatment by mean age 51± 8 years, by the end of currently lack access to these services
monitoring health measures (including February 2018 and the last participant and we believe that the PD-system
fluid exchanges tracking), providing is due to finish the six-month follow-up will significantly benefit the patients,
educational multimedia, and assisting period in September 2018. Preliminary the staff caring for these patients and
in their appointment scheduling and results indicate a slight decrease in exit ultimately the healthcare system. With
medication management. It also site infections and peritonitis infections minimal modifications, PD-BUDDy has
supports multidisciplinary care delivery since the introduction of the platform. already been identified to be a suitable
by providing shared access to the Patient feedback from initial surveys solution not only to persons undergoing
multidisciplinary care team. The platform are extremely positive, with one patient haemodialysis, but it has also been
was developed in collaboration with calling the app ‘life saving’ with others deemed suitable by nephrologists to
Logan Hospital (Metro South Health). reporting back regularly to staff that the manage their kidney transplant patients.
A smartphone app allows patients to app is extremely valuable and convenient. This signifies utilisation in national and
enter health parameters (e.g. dialysis Use of the portal by staff has become international markets.
fluid intake, weight and blood pressure), routine, with clinicians commenting
medications, exercise, symptoms and exit

H E A LT H S E R V I C E S
site photos, which automatically upload
to a password-protected clinician portal.
The aim of the PD-BUDDy platform is
to assist both patients and clinicians
during their PD procedure, reduce
possible PD-associated complications,
improve quality of life for patients and
reduce the financial burden on the
healthcare system by saving time.
PD-BUDDy leverages from CSIRO’s
existing clinically validated mobile
health digital platform for cardiac
rehabilitation, and through the close
relationship with stakeholders, features
have been incorporated to PD-BUDDy
that are unique to PD patients such as
ultra-filtration volume quantification,
clinician-driven communication and
ability to share photos of catheter exit
sites for potential infection. PD-BUDDy
has extended functionality to connect
Figure 1. A patient from Logan Hospital relating her PD-BUDDy experience.
to wireless physiological monitors and
fits perfectly for new remote health care
delivery models extended to community
and home care.

aehrc.com
5 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MOBILE HEALTH SYSTEMS (continued)

M♥THer: gestational diabetes app helps health workers monitor pregnant women
Gestational diabetes mellitus (GDM) is
an increasing problem among pregnant
women worldwide with a prevalence of up
to 20% depending on diagnosis criteria
and population. The increasing number
of pregnant women being diagnosed with
GDM is leading to snowballing healthcare
costs. M♥THer is an interactive system
designed to support women throughout
diagnosis of GDM to childbirth. It also
improves multidisciplinary care co-
ordination by providing shared access to
the women’s clinical information.

M♥THer was recently tested in


a feasibility study at a hospital in
Queensland, Australia. The study
recruited 40 women (mean age 30 years
old by the end of Jan 2018). The app
was well received by the women as an
H E A LT H S E R V I C E S

alternative to the paper-based blood Figure 1. Redland Hospital Chronic Disease Team CNC Roisine Warwick
glucose level (BGL) recordings. All
respondents agreed that the app was Diversity To address the objective of this study,
user friendly, convenient and helpful in we have engaged with four different
managing their GDM, making them feel In 2014, we validated a mobile countries in Europe starting a pilot
confident in the management of their health home-care model for cardiac study in different CR centres of Sweden,
condition. Treating clinicians reported rehabilitation (CR), which used Belgium, Netherlands and Poland.
improved communication with the smartphones and the internet to deliver Following the results of this study,
women in their care and experienced rehabilitation in the patient’s home a randomised clinical trial with the
an increase in multidisciplinary co- setting to align with their lifestyle. This participant sites will be proposed. For
ordination amongst themselves. The platform was tested in a randomised this study we are utilising Cardihab as
platform enabled early intervention clinical trial, becoming the first clinically commercial partner for delivering of
for a number of women identified with validated mobile health delivery of the cardiac rehab platform, which has
elevated BGL readings in the first week CR, improving uptake, adherence, been translated into Swedish, Dutch and
of using the app, and due to elevated completion rates of participants and with Polish.
fasting BGLs, a number of women were similar health outcomes as that of the
timeously commenced on Metformin or traditional centre-based CR programme.
insulin treatment. Now, in 2018, we want to develop this
concept internationally and test how our
Data input received from the M♥THer technology can be validated irrespective
app include physiological measuring of the intrinsic characteristics of the
outputs and supports decisions on health structure and the culture and
patient management that can impact language of the users.
clinical outcomes and patient care.
M♥THer expands access and can
improve the quality of healthcare for
women in regional and rural areas. It
reduces burdens for these women, such
as travel to receive specialist care, and
improve monitoring, timeliness, and
communications within their care teams.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 51

Highlights Project Aims for


C
 ollaborators:
for 2017/18: 2018/19:
• Q
 ueensland Cerebral Palsy • Queensland iAwards 2018 • T
 o extend the mobile health
Rehabilitation and Research • M
 obile-Pulmonary Rehabilitation platform to incorporate
Centre Platform (m-PR) won three gamification features, to support
• University of Queensland awards: QLD Premier’s iAward for patients in adhering to goals.
• CSIRO’s Data61 Public Sector Innovation; iAward: • S
 ecure extensive external
Public Sector & Government; funding such as NHMRC;
• M
 etro North Hospital and Health
Merit Award: Research & Cerebral Palsy Foundation;
Service (The Prince Charles
Development Project of the Year. Advance Queensland; CRC
Hospital; Royal Brisbane and
• P
 D-BUDDy: Support for Northern Australia.
Women’s Hospital)
Peritoneal Dialysis Patients won • C
 omplete IGMS study 1 on
• M
 etro South Hospital and Health
two Merit Awards: Research & healthy term infants.
Service (Redland Hospital;
Development Project of the Year; • Infant smart toy development
Logan Hospital)
Community Service Markets. and feasibility study.
• D
 avid Ireland was interviewed on • International recognition and
ABC Radio discussing AI, mental collaboration e.g. PD-BUDDy
health and autism spectrum

H E A LT H S E R V I C E S
feasibility study in Czech
disorder. Republic, extending diversity
• T
 he M♥THer platform was study to Kuwait and China,
shortlisted as finalist in the entering the US market.
International Hospital Federation
Awards, with final judging based
on a pitch to be delivered on 10
October 2018 during the World
Hospital Congress in Brisbane.
• D
 r David Ireland and Dr
Christian Redd were successful
in obtaining funding for early
stage projects through Acorn
Awards. These projects are
‘Bots, Bullies and Autism’, and
‘A Universal Smart Toy Sensor
System for Monitoring Child
Development’, respectively.
• T
 he Mobile Health Systems
team has this year been very
engaged in community programs
such as the CSIRO’s Assets
Work Placement for Indigenous
students, work experience, and
Indigenous cadets.

aehrc.com
5 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

MOBILE HEALTH SYSTEMS (continued)

PhD Student

Nazli Ghafouryan
Project title: A mobile based multidisciplinary virtual clinic for patients with Acute Coronary Syndromes: A Randomised
Controlled Trial, MoTER-ACS project

The aim of this research is to develop a The post-discharge virtual clinic on a remote server. Following testing
mobile based multidisciplinary virtual will be offered to patients as an of the feasibility of the mobile-based
clinic based on the existing platform alternative program to a hospital- intervention in the pilot study (N=54),
(MoTER) and to investigate the impact based follow-up. The mobile phones a randomised control trial will be
of such a clinic on health outcomes and are used for providing education and conducted to evaluate its effectiveness.
clinical management of patients with personalised feedback, monitoring We hypothesise that the use of mobile-
Acute Coronary Syndrome (ACS). MoTER physiological data as well as recording based clinics offer effective post-
is a platform designed for home based patients’ self-observations on their discharge management tools for both
monitoring in cardiac rehabilitation health-related behaviour and virtual patients and healthcare professionals.
using smartphone application and consultation via audio or video. All the
the internet. We customised the data is synchronised daily to a portal Collaborators: Metro North Hospital
MoTER platform for delivering care in and Health Service.
patients with ACS as a comprehensive
virtual clinic. This study consists of a
pre-study survey and a pilot testing
H E A LT H S E R V I C E S

and randomised controlled trial. The


objective of the pre-study survey was
to conduct short structured interviews
with small groups of patients (N=30),
one session of focus group (N=10) with
healthcare professionals and a survey
with cardiologists (N=30). The results
of the pre-study survey were used
to develop a model of care for post-
discharge management of ACS patients
using the mobile-based application and Figure 1. MoTER-ACS platform.
its web portal.

P
 ublications:

• B
 ashi N., Fatehi F., Fallah M.,
Walters D., Karunanithi M. Self-
management education via mobile
health (mHealth): A review of
strategies and structures. JMIR
Mhealth Uhealth. 2018 Jun 26. doi:
10.2196/10771. [In print]
• B
 ashi, N., Karunanithi, M., Fatehi,
F., Hang Ding, H., Walters, D.
Remote Monitoring of Patients
with Heart Failure: An Overview of
Systematic Reviews. J Med Internet
Res; 2017;19(1): e18, doi:10.2196/
jmir.6571

Figure 2. Selected screenshots showing educational materials.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 5 3

HEALTH INTERNET OF THINGS

The Internet of Things (IoT) refers to a network of internet connected devices that collect and transmit data through embedded
sensors. The Health Internet of Things team uses IoT to build solutions for health and aged care.

Smarter Safer Homes (SSH) for aged care


CSIRO’s Smarter Safer Homes (SSH) is
a lifestyle-based technology platform
developed to enable older people to
live longer in their own homes. The
platform was designed with consumers
to enable self-management, and engage
assistance from informal and family
carers and clinical services. The platform
uniquely features an individualised
measure of functional independence,
together with health measurements and
social isolation, from an unobtrusive,
sensor-enabled home.

The SSH platform project is led by the


Australian e-Health Research Centre

H E A LT H S E R V I C E S
with multidisciplinary collaboration
including universities, aged care service
providers and local clinicians. It uses
cutting edge pervasive communication
and wireless sensor and monitoring
technology, and features a novel,
personalised measure of functional
independence, indexed through the
“Objective Activity of Daily Living”. The Figure 1. SSH iPad application allows residents to view data derived from the sensors and medical devices.
potential benefits of these technologies
are multiplied where distance separates
families and adds substantial costs to
delivery of health and other services.

Figure 2. Objective-Activity of Daily Living (OADL) scores of the SSH app to represent everyday health and
wellbeing status.

aehrc.com
5 4 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH INTERNET OF THINGS (continued)

Dementia and aged care services: Sustainable, cost effective, Smarter Safer Homes 100
smart assisted independent living – Smarter Safer Homes testbed
This project’s objective is to deploy 100
This project’s objective is to use the To achieve this, our Smarter Safer
Smarter Safer Homes platforms within
Smarter Safer Homes platform to Homes platform will be implemented in
Australia to provide an ongoing testbed
revolutionise aged care services delivered metro, regional and rural service settings
for AERHC research. Its aims are:
to people in their homes, in alignment in four sites that present different
with consumer directed care, through: geographical challenges and work force • T
 o collect longitudinal sensor and
issues. This implementation will be activity data from real residential
• d
 eveloping innovative service models facilitated by national, metro, regional, homes and construct a database
appropriate across a range of and rural service providers as partners to continuously develop, evaluate
service and geographical settings of the consortium. The evaluation will and extend the Smarter Safer
which is effective, client-driven and be conducted among 200 participants, Homes platform (installation of
focused. aged 65 years and over living at home inconspicuous sensors and devices
• e
 nabling a platform that not only and supported by home care service to monitor the activities of daily
tailors to individuals, including those providers, through a pragmatic living of seniors).
with dementia, basic functional and randomised controlled trial which • T
 o foster collaborations with service
health needs, but also supports translates evidence into real practice and providers on a well-developed
carers’ needs. policy. The outcomes of the trial include research platform with continuous
changes in quality of life, functional, inputs of large quantity of real data
psychological and cognitive health, carer to help shape an economic and
H E A LT H S E R V I C E S

burden, and cost-effectiveness. sustainable care delivery system


that can affect policy-makers and
eventually benefit all Australians.

C
 ollaborators: Collaborators:
• S
 ervice providers: Anglicare Australia, integratedliving Australia Ltd, • integratedliving Australia
All About Living. Ltd.

Highlights for 2017/18: Highlights


for 2017/18:
• Project initiation meeting with partners and service providers
• Initial 25 homes installed
• Research protocol developed
to kick off the project.
• Ethics application submitted to CSIRO ethics.
• W
 ork Placement for Indigenous students, work experience, and
Indigenous cadets.
Project Aims
for 2018/19:
Project Aims for 2018/19: • R
 esearch protocol
developed
• Ethics approval granted • Ethics approval granted
• R
 ecruit and randomise participants from three to four geographical • D
 eployment of Smarter
sites specific to the service providers Safer Homes platform to
trial participants’ homes
• Deployment of 100 homes for intervention group
• C
 ommence the
• Commence randomised controlled trial among 200 participants.
longitudinal pilot study.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 5 5

Smart home multi-resident activity recognition


The multi-resident activity recognition
system (MARS) is designed to support
multiple occupants in a smart home with
minimum impact on their lifestyles. The
system uses Google Home and natural
language processing to annotate and
label activities, then adopts deep learning
techniques to capture and predict
activities. We evaluated the system in the
house of a family of three, receiving the
best state-of-the-art results in activity
recognition accuracy.
Figure 3. The MARS system and most labelled activities through voice recognition and natural language
processing (these are real inputs from residents).

Fall detection in smart home using non-wearable radar sensor


Falls are one of the major issues which
can endanger the lives of older adults.
Numerous research studies investigate

H E A LT H S E R V I C E S
the use of wearable technologies to
detect falls in everyday environments.
Although wearable sensor solutions
provide good accuracy and sensitivity
for fall detection, it may not always
be convenient or desirable for older
adults to wear a tag or sensor in home
environments. This project discusses
using non-wearable ultra-wideband
(UWB) radar sensors as a practical,
environmental fall detection solution in
home settings.

Figure 4. The unsupervised fall detection system.

ID sensor for human identification


This project achieves the goal of human
identification through properly processing
and analysing the received signals
from the UWB radar installed in indoor
environments. We employ unsupervised
feature learning techniques to
automatically learn local, discriminative
features that can incorporate intra-class
variations of the same identity, and yet
reflect differences in distinguishing
different human identities. Figure 5. Scattered UWB signals of various people with individual walking styles.

aehrc.com
5 6 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

HEALTH INTERNET OF THINGS (continued)

Immersive augmented
reality: remote clinical
consultation
Using emerging augmented reality
(AR) technologies on mobile devices,
this project developed an AR clinical
consultation system through an iPad
and a Kinect sensor. This low-cost
and highly portable AR consultation
system can be easily deployed in a
patient's home and clinician's office with
minimum footprints on their normal daily
activities. It not only provides a great
and immersive telehealth consultation
experience for patients, but also can
help clinicians to easily explain complex
medical conditions to patients through
Figure 6. AR Doctor system structure.
visualisation and simulation.
H E A LT H S E R V I C E S

Team highlights Team aims for


C
 ollaborators:
for 2017/18: 2018/19:
• C
 onduct a large randomised
• Data 61/Energy Flagship, CSIRO • A
 large randomised control
control trial evaluating how
• Aged care service providers trial supported by Department
information provided by SSH can
of Health is about to start in
• Q
 IMR Berghofer Medical improve quality of care.
Queensland.
Research Institute • C
 onduct two smart home trials
• T
 wo smart home trials
• Washing State University, USA in Brisbane and Geelong to
supported by CSIRO and the
• City University, UK investigate how seniors can use
local community are about to
information provided by SSH to
• U
 niversity of Twente, start in Brisbane and Geelong.
improve their quality of life.
Netherlands • A
 novel ID sensor prototype is
• C
 onduct a smart sensor trial
• Shanxi University, China being developed at CSIRO.
in the PISA study to investigate
• Academy of Science, China • C
 SIRO signed a Memorandum of correlations between sleep
Understanding (MOU) with China patterns and early symptoms of
Telecom Shanghai on research dementia.
collaboration of aged care and
• S
 ubmit patent application of the
smart homes.
novel ID sensor.
• E
 xtend SSH in the Hong Kong
market.

Figure 7. The MOU signing ceremony of CSIRO and


China Telecom.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 5 7

DIGITAL HEALTH ENGAGEMENT

The Digital Health Engagement team Activate TKR: mobile support for orthopaedic surgery
draws on expertise across human-
computer interaction, personalisation Summary
& recommendations, persuasive
technology and neuroscience to be Total knee replacement (TKR) surgeries video-based exercise demonstrations,
a leading team in the design and have increased in recent years. Exercise and tools for self-monitoring of daily
evaluation of technologies to facilitate the programs and other interventions activities including an activity tracker.
provision of equitable health care. The following TKR can facilitate the recovery The app transmits patient-gathered data
team designs technologies that can be process. With limited clinician contact to a purpose-built web portal where
used by clinicians to improve workflow time, patients with TKR have a substantial clinicians can view patient progress and
and enhance service delivery, and by burden of self-management and limited configure exercise programs remotely.
individuals to support them in playing an communication with their care team, Activate TKR is designed to provide
active role in their health management and so they often fail to implement an flexibility in care delivery, and increased
to meet short- or long-term health effective rehabilitation plan. engagement with rehabilitation services.
and lifestyle goals. The team has a
The Australian eHealth Research A randomised control trial began
strong collaborative approach, leading
Centre, together with Johnson & in November 2016 to evaluate the
and contributing to projects across the
Johnson Medical Devices Australia, technology platform. It is being
e-Health program and into several other
are looking at new ways to address the conducted at multiple sites in a number
CSIRO divisions. In 2017/18 the team
challenges of motivating and assisting of states in Australia, with about 150
engaged primarily with industry partners
patients to complete rehabilitation patients undergoing TKR surgery.
to deliver excellence in science.
programs while also reducing the Participants are randomised to either

H E A LT H S E R V I C E S
communication gaps that exist between a control group or the intervention
clinicians and patients. We have group, with both receiving usual care as
designed and developed Activate TKR, provided by their surgeon or hospital.
an orthopaedic support technology The intervention group receives the app
platform, comprised of a mobile app, a and wearable activity tracker in addition
wearable activity tracker, and a clinical to usual care. This trial is investigating
web portal. The technology platform the clinical and behavioural efficacy of
aims to assist patients in managing the app, and the impact of the technology
their surgery preparation and speed platform components in terms of service
up their recovery, and to bridge the satisfaction, acceptance, and economic
communication gaps between clinicians benefits of the provision of digital
and patients. The purpose-built app services. The trial is running for a period
includes behavioural coaching, practical of 13 months for each patient.
hints and tips for surgical preparation,

Figure 1. Activate TKR clinical portal. The web portal allows clinicians (e.g. surgeons, physiotherapists) to view
patient data and configure exercise programs for individual patients.

aehrc.com
5 8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

DIGITAL HEALTH ENGAGEMENT (continued)


H E A LT H S E R V I C E S

Figure 2. Activate TKR app: Home screen. Figure 3. Activate TKR app: Physio Coach Figure 4. Activate TKR app: Today screen. Shows
screen. Exercise programs are configured by data from the activity tracker (daily step, stairs, and
physiotherapists from a library of videos typically sleep) and self-reports (pain and knee ROM).
used for TKR rehabilitation.

• T
 KR user needs paper published
O
 utcomes Outputs to date
in JMIR: van Kasteren Y, Freyne
• Increase patient engagement • D
 evelopment of the technology J, Hussain MS Total Knee
with rehabilitation services platform (patient app and clinical Replacement and the Effect of
(surgical preparation and portal) Technology on Cocreation for
recovery) Improved Outcomes and Delivery:
• R
 andomised control trial launched
Qualitative Multi-Stakeholder
• B
 ridge the communication gaps in November 2016; kicked-off with
Study J Med Internet Res
between clinicians and patients four trial sites in NSW and SA
2018;20(3):e95.
• P
 rovide flexibility in TKR care • First media release in March 2017
delivery, particularly in rural, • S
 howcased project together with
remote, or busy lifestyles, with O
 utputs expected in 2018/19
Johnson & Johnson Medical
the potential to achieve the same Devices at the Bennelong • D
 ata extraction and preliminary
clinical outcomes as normal Innovation Summit at Parliament data analysis (quantitative)
business-as-usual care. House in March 2017
• Qualitative research findings
• R
 esearch protocol paper published
• Interim outcomes workshop with
Impact
 in JMIR: Hussain MS, Li J, Brindal
Johnson & Johnson.
E, van Kasteren Y, Varnfield M,
• J ohnson & Johnson’s interest Reeson A, Berkovsky S, Freyne J
in application to other joint "Supporting the Delivery of Total
replacement studies Knee Replacements Care for Both
• H
 ealth care industry’s Patients and Their Clinicians With
application to health and a Mobile App and Web-Based
wellness solutions. Tool: Randomized Controlled
Trial Protocol" JMIR Res Protoc
2017;6(3):e32

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 5 9

Supporting men’s health


For many older men, there’s a defining Movember Foundation aims to develop and a high level of integrated care and
code around being a man. It’s about implement an agreed set of interventions enhanced health outcomes. In addition,
being in control of any given situation, supported by an innovative online the Movember Foundation provides
of not being seen as weak or emotional, technology platform targeted at improving support and counselling services to men
about having a healthy sexual appetite. the lives of men living with prostate cancer with prostate cancer and their significant
Prostate cancer erodes the very core of in Australia, UK, Canada, the USA and others. This study was undertaken to
this code, through incontinence, fear and New Zealand. The implementation of this determine what needs could be met with
depression, and erectile dysfunction. In online platform is facilitated by a care psychological support services over the
Australia, it accounts for almost a quarter co-ordinator working with the healthcare course of the prostate cancer journey, and
(23.1%) of all new cancer diagnoses, is the team, the individual and the Movember how they could be delivered to facilitate the
second leading cause of all male cancer Foundation team, to ensure men receive greatest uptake.
deaths and carries the highest suicide
risk of all cancers with the greatest risk
in the first year post diagnosis. As there
Outcomes Outputs to date
is a relatively high five-year survival
rate for localised prostate cancer (94%), • F
 ocus groups were conducted
the number of men living with the • B
 radford, Dana. But I’m Healthy:
with men who had participated
psychosocial ramifications of the disease psychological stressors of
in Movember’s TrueNth Prostate
is correspondingly high. prostate cancer and potential
cancer program
strategies for support services.

H E A LT H S E R V I C E S
The prevalence of prostate cancer, the • M
 en spoke candidly about the CSIRO: 2018. EP18233
impact of symptoms on mental wellbeing psychological ramifications of
• B
 radford, Dana. Improving
and the increase in suicide ideation their prostate cancer journey
care for men with prostate
indicate an urgent need for a greater • A
 number of potential strategies cancer. CSIRO: 2017.
understanding of how psychological were identified to enhance the EP176237. https://doi.
services can support men on their provision of psychological services. org/10.4225/08/59b6dd593f5ea
prostate cancer journey. Movember
Foundation has established and funded • B
 radford, Dana. Sunshine
Impact Coast TrueNTH Focus Groups:
the TrueNTH network, a prostate
cancer survivorship initiative which is Improving engagement and care
• U
 nderstanding the trigger points
a collaborative effort among various experience of men living with
for men allows Movember to
global organisations, of which CSIRO is a prostate cancer. CSIRO: 2017.
provide enhanced services when
member. Through the TrueNTH program, EP175572.
they are most required.

CALD Assist: translation support expanded for nursing workforce


The CALD Assist app translates key served by Australian hospitals vary in health disciplines (dietetics, speech
phrases for nursing and allied health staff cultural diversity, with some hospitals, pathology, podiatry, physiotherapy
into 10 common languages (in addition to such as Western Health in Victoria, and occupational therapy) and CALD
English) using pictorial, video, written and serving communities where more than patients when an interpreter is not
voice-over prompts to ensure appropriate 150 languages are spoken. Unmet present. As a response to multiple
and timely care for inpatients. demand for interpreter services means awards (Branko Cesnik Award for Best
that non-English-speaking patients Academic/Scientific paper; ‘Improving
Effective communication in the hospital are sometimes unable to access timely health equality and closing the gap’,
setting is vital to ensure appropriate care, causing inequity in service delivery Victorian Public Healthcare Awards),
care and reduce the risk of adverse and often frustration and anxiety for senior nursing staff expressed a
events. Research indicates a variance patients, their carers and clinical staff. significant need for a nursing-based
in access and quality of healthcare app using a similar platform. In 2017,
provided to people from culturally Researchers from the AEHRC's Digital CALD Assist was expanded to meet the
and linguistically diverse (CALD) Health Engagement team, together specific needs of the nursing workforce.
backgrounds and their English- with clinicians at Western Health, The inclusion of nursing-specific
speaking peers, which can be reduced designed, developed (2014) and material enables nursing staff to better
with the use of interpreters. However, evaluated (2015) CALD Assist - Allied meet the daily care needs of their CALD
demand for these services often Health, a novel communication app patients, reducing variance in practice,
exceeds supply. Interpreter demand is designed to facilitate initial assessment and providing a timely and positive
unsurprising given the communities between clinicians from five allied patient experience.
aehrc.com
6 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

DIGITAL HEALTH ENGAGEMENT (continued)

Working with the Autism


community
Working with Autism Spectrum Australia
(ASPECT), the Autism Hub, Autism
Academy for Software Quality Assurance
and the Autism CRC, we have developed a
body of work with and for the wider autism
community. The work builds on previous
work in social robotics and conversation
agents. Over the next year we will harness
the knowledge gained from these projects
to implement robotics and conversation
agents developed with and for people
with autism and their wider community
to improve social interaction and verbal
participation in early school-aged children
and facilitate participation in tertiary
education in young adults.

Increasingly technology is being


incorporated into the curricula with mobile
H E A LT H S E R V I C E S

Figure 5. CALD Assist app: upon selection of an individual phrase, translated text and appropriate image is devices in almost all classrooms across
displayed on the screen.
Australia. The implementation of these
technologies facilitates skill acquisition
across a range of learning styles and will
have far reaching benefits for students
• 2
 015 Health Informatics Society
O
 utcomes in their future careers. Students on the
Australia, Branko Cesnik Award
autism spectrum tend to show a keen
• A
 new app was developed for Best Academic/Scientific
interest in, and respond positively to,
that includes nursing-specific paper
technology including chat-bots, tablet-
content • 2
 015 Gold ‘Improving health based communication devices, and
• An evaluation trial was finalised equality and closing the gap’, robots. These technologies open the door
Victorian Public Healthcare towards new approaches to education
• It was confirmed that
Awards and learning. We are working with the
interactions between nursing
staff and CALD patients are • 2
 017 The Health Roundtable wider autism community to understand
less successful than those with Innovations Award ‘Improving the roles that innovative technology can
English-speaking patients. the Continuum of Care’ play during successive stages of childhood
• 2
 018 iAwards Victoria Merit development in the home, in the classroom,
Certificate (Public Sector & and during extracurricular activities,
I mpact
Government) with a focus on social participation. In
• Improve the clinicians’ ability to collaboration with the Autism Hub, we
• J . Freyne, D. Bradford, C.
adequately assess and respond conducted a think tank to explore the
Pocock, D. Silvera-Tawil,
to patient needs in a timely directions for the future of technology-
K. Harap, S. Brinkmann.
manner supported education to create better
Developing digital facilitation of
outcomes for students on the spectrum and
• Improve the frequency and assessments in the absence of
the wider autism community.
quality of communication an interpreter: a participatory
between staff and CALD patients. design and evaluation with allied
health groups. JMIR Formative O
 utputs to date
O
 utputs to date Research, 2 (1), 2018.
• D
 . Silvera-Tawil, C. Pocock, • B
 radford, Dana; Ireland, David;
• F
 reyne, J., Pocock, C., Bradford, D. Bradford, K. Harap, S. Silvera, David; Farr-Wharton,
D., Harrap, K., Brinkman, Brinkmann. CALD Assist – Geremy; Xiao, Yi; Cooper, Anabelle.
S. Designing technology for Nursing: Improving Patient Teachnology: Tech reshaping the
assessments of CALD patients. Communication in the Absence face of education in the Autism
HIC Brisbane 4 August 2015 of Interpreters. Journal of community. A report of the Autism
Clinical Nursing, 2018. Technology Think Tank. CSIRO:
2018. EP183925.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 61

Socially assistive robots


O
 utputs to date from our
to support children on the O
 utcomes
robotics work
autism spectrum • O
 ur trials are ongoing, but
results so far suggest that • B
 ruck, S., Silvera-Tawil, D.,
For children on the autism spectrum, Bradford, D. ‘NAO Robot’s new
social interaction and communication social robots are effective tools
to assist during therapy and role - teaching in the classroom,
can be significant challenges. With about Autism in Education.
83 per cent of the 164,000 Australians education of children on the
spectrum • S
 ilvera-Tawil D., and C. Roberts-
living with autism spectrum disorder
• P
 articipants are particularly Yates, “Socially-assistive
under 25 years old, additional support
enthusiastic about humanoid robots to enhance learning
is often needed to help these children
robots. for secondary students with
gain the skills necessary to navigate the
intellectual disabilities and
world independently. Digital technology is • D
 eveloped Max, an Android- autism,” in IEEE Robots Man
often used as a supplement to traditional based user interface to facilitate Systems Interactions Conference,
education, as it provides an environment the use of robots in therapy and 2018.
that allows for self-paced learning and education.
immediate feedback. While considered • S
 ilvera-Tawil, D., C. Roberts-
generally safe and effective, there are Yates and D. Bradford, “Talk
I mpact to me: The role of human-
concerns that a child who is taught to
communicate using interactive technology robot interaction in improving
• O
 bserved improvements
may become dependent on the virtual verbal communication skills

H E A LT H S E R V I C E S
in social interaction,
world and its rewards, while interpersonal in students with Autism or
communication, participation
skills are sacrificed or not generalised to Intellectual Disability,” in IEEE
and self esteem
real world settings. Robots Man Systems Interactions
• B
 enefits observed in students Conference, 2018.
We have developed interactive modules to when they interacted with
support therapy and education for children robots in the classroom are
with autism and intellectual disability. We transferring to interpersonal
are trialling four different social robots: interactions.
NAO, KASPAR (from the University of
Hertfordshire), PARO and ROBOTIS OP3,
including trials in partnership with the
University of New South Wales, Autism
Spectrum Australia (ASPECT) and Murray
Bridge High School in South Australia.
Social robots are novel, animated and
appear to be autonomous, setting
themselves apart from other technology,
and their physical, 3D presence provides
a compromise between the virtual world
and the real world. Robots can provide
complex behaviour patterns, while
appearing much less intimidating than
humans. They can also deliver predictable
behaviours and repetitive feedback, and
they don’t get angry, tired or stressed.
More than anything, robots provide a new
environment where it is fun to learn. To
support our trials, we have developed an
app and compatible platform that can be
used as a user interface to control the
social robots used during our trials.

Figure 6. The humanoid robot NAO and the robot seal PARO at the Murray Bridge High School Disability Unit.

aehrc.com
6 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

DIGITAL HEALTH ENGAGEMENT (continued)

Chat-bots to support ON LaunchCamp: providers and peer mentoring services


to determine how chat-bot technology
people on the autism commercially testing can support education participation and
spectrum chat-bots with the autism education-related health and wellbeing
For several years the development of an community (study-related stress, looking after
yourself on campus). The ultimate aim
artificially intelligent, chat-bot technology In a two-day ON LaunchCamp of the project is to create a chat-bot
has been underway. Similar to Apple's program, we had the opportunity to ‘brain’ that can deliver useful multi-
Siri and Amazon's Alexa, the chat-bot understand the commercial prospects modal resources to support greater
technology Harlie is being developed to for our chat-bot. We interviewed seven participation and wellbeing of people
support people with difficulties in speech, experts who engage with children on the autism spectrum in higher
language and social interaction that are on the autism spectrum, and their education.
often seen in autism spectrum disorder, insights were captured within two key
dementia and Parkinson's disease. problem statements, (I) for clinicians Bots, bullies and autism
and teachers, monitoring development
is challenging, particularly outside of Bullying is a prevalent societal problem
a clinical or classroom setting and (II) which disproportionately impacts
parents and carers need tools tailored to special needs children, particularly
their child’s context that encourage quality those with autism spectrum disorder
family time and inclusion. These findings (ASD). Children with ASD can find
have helped us align the strategic social interaction challenging and
direction for the chat-bot work, enabling frustrating, which can lead to social
H E A LT H S E R V I C E S

direct benefits for those who need this disengagement and result in social
technology the most: individuals on the isolation. Identifying bullying in children
spectrum, and their families and carers. with ASD is hampered by the fact that
children with ASD may have limited
‘Hear’ to help: striving for speech and struggle to communicate
greater participation and their experiences. In these cases, being
bullied can compound social interaction
wellbeing through tailored difficulties and impede communication
chat-bot technology development. Sadly, many children on
the autism spectrum who wish to make
In collaboration with Curtin University,
new friends have an even higher chance
Autism Academy for Software Quality
of being bullied. It can be strongly
Assurance (AASQA), the Queensland
argued there needs to be an ongoing
Department of Education and Training,
and early implementation of programs
and Autism Spectrum Australia, this
to build lifelong resilience against
project will work with people on the
Figure 7. An example of a conversation to identify if a bullying behaviour.
child perceives they are being bullied. autism spectrum, support service
In light of this, seed funding from
CSIRO’s Health and Biosecurity
business unit was provided to develop a
specific bullying module for the Harlie
chat-bot. Harlie’s bullying module is
under development and will be co-
designed by the autism community.
A number of key stakeholders have
expressed interest in being involved
in the evolution of this module for
children, young adults in tertiary
education and young employed adults
with autism. Early content includes an
example interaction of the beginning
of a conversation involving identifying
whether the user is being bullied.
Images are embedded into the
Figure 8. A sample dialogue between Harlie and a person with autism. Green rectangles designate dialog responses to help guide the meaning of
generated from Harlie, blue rectangles designate responses from the user, grey rectangles indicate pauses in
the timeline.
the conversation.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 6 3

Chat-bots for people on the


autism spectrum who may
be non-verbal Outputs to date Collaborators:
A significant portion of individuals
• J ohnson and Johnson Medical
on the autism spectrum are at some • F
 arr-Wharton, Geremy; Ireland,
AU
point categorised non-verbal. An David; Bradford, Dana. Social
individual may later become verbal, Fringe Dwellers: Designing • The Project Factory
however, many remain non-verbal for Chat Bots as Bridges for • Movember Foundation
life. This does not mean they cannot Children with ASD. In: OzCHI • M
 elbourne Genomics Health
communicate, but rather require 2017; November 28th to Alliance
other means. With the ubiquitous use December 1st; Brisbane,
• Western Health
of mobile technologies, software- Australia. SIGCHI; 2017.
based assistive technologies have • Autism Spectrum Australia
• B
 radford, Dana; Ireland,
become increasingly available. In the David; Wiles, Janet. Talk to • Autism Hub
case of augmentative and alternative me: The Chat-bot as a speech • Autism CRC
communication (AAC) systems, the pathology monitoring tool. In: • Curtin University
introduction of tablets was revolutionary Health by Tech; 1 June 2018;
due to their relatively low cost, • University of New South Wales
University of Twente. University
portability, and resulting increase in of Twente; 2018. • Murray Bridge High School.
social acceptance of AAC.
• Ireland, D. Farr-Wharton, G

H E A LT H S E R V I C E S
We have been developing a AAC app and Bradford, D. Social Fringe Project Aims for
with an embedded chat-bot, herein Dwellers: Can chat-bots 2018/19:
referred to as Alex. Alex is designed for combat bullies to improve
participation for children • F
 inalise clinical trial for
use by people on the autism spectrum.
with autism? Journal of Activate TKR program
Programming Alex does not need any
specialist skills and is designed for Community Informatics, • R
 oll out evidence-based
speech therapists, parents and other Special Issue (In press). implementation of robots into
key stakeholders to contribute what • A
 . Cooper and D. Ireland, Queensland schools
they deem relevant. The user is able ‘Designing a Chat-Bot for • D
 evelop and deploy chat-
to practice spontaneous conversation Non-Verbal Children on the bot technology to facilitate
with Alex in a safe, non-judgemental Autism Spectrum’, Studies participation in tertiary
environment. The long-term aim in health technology and education and offer bullying
of this research is to examine the informatics. Volume 252 pp. strategies for children on the
validity of intervention based on 63-68, Health Informatics autism spectrum.
artificial intelligence that augments Conference (HIC), 2018
conventional therapy.

 tudent or postdoc
S
highlight
Yi Xiao initially joined CSIRO as a
Vacation Student from the University
of New South Wales. During this time,
she developed an Android-based user
interface to facilitate the use of robots in
therapy and education. She has a strong
technical background in computer
science with a focus in robotics, artificial
intelligence and machine learning. She
is now doing her Bachelor of Honours
research project with the Digital
Health Engagement team around the
Figure 9. Example of conversation using AAC. The blue box holds the user input, and the green box shows application of social robotics for autism
Alex’s response. and intellectual disability.

aehrc.com
6 4 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

AUSTRALIAN TELE-HEALTH RESEARCH


AND DEVELOPMENT GROUP
Director: Yogi Kanagasingam

The Australian Telehealth


Research and Development
Group (ATRDG) was established
by CSIRO in conjunction with
the Department of Health
Western Australia (DoH) in
June 2012.
Its purpose was to develop a strong
digital health and telehealth research
A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P

and development program to address


pressing and emerging areas of
healthcare delivery, particularly The group has three key research areas:
in respect to the provision of high
quality services to rural and remote Remote delivery of clinical services
populations, and to high-needs groups. This stream will focus on the development and evaluation of remote delivery of clinical
The ATRDG aims to align and work with services to metro, rural and isolated populations. Using the Remote-I system or
service providers and other stakeholders Android-based technology, patient images will be forwarded through existing internet
and assist them in developing systems services to specialist clinicians at major centres for evaluation. Applications in this
and technologies that result in stream include ophthalmology eye screening for aged care facilities and dentistry
better service delivery solutions and (a rural pilot study using an Android-based imaging system), burns (using advanced
preventative health applications. imaging techniques to grade wounds) and emergency medicine (development of an
emergency telehealth service in collaboration with WACHS), amongst others.
The ATRDG strives to be a world-leading
telehealth research and development Chronic disease management
group, and aims to transform the way This stream will include research on how telehealth can be used to assist in the
health services are delivered. The aim management of complex and chronic conditions in the home. An existing trial of
of this research has been to improve automated data collection and alerts for chronic obstructive pulmonary disease
health outcomes in Western Australia patients will be expanded. Clinical decision support and management of patients
and increase the productivity and with chronic health conditions in their homes using mobile technology will be
efficiency of health service delivery in evaluated, mainly those living in rural and remote locations.
the state.
Disease diagnosis and screening technologies
This stream will pursue research around Alzheimer’s disease, hypertension, H.Pylori
and stroke. The Alzheimer’s disease study is designed to evaluate the use of curcumin
(derived from turmeric) as a diagnostic marker which can be detected using high-
resolution retinal imaging. In the stroke study, data from patients after they experience
a first stroke are interrogated with the aim of identifying markers that are predictive of
a second stroke, thereby providing a window for preventive interventions.

Highlights:
1. F
 irst to deploy artificial intelligence-based grading system for diabetic retinopathy and other eye diseases into a real-
world clinical setup (Midland GP Superclinic)
2. Developed a automated retinal imaging platform called VASP to measure ocular biomarkers
3. D
 eveloped and validated a mobile app called MICE, Medical Image Communication and Exchange, at the Burns Unit at
Fiona Stanley Hospital. This is expected to be rolled out into hospitals in South Metro Health Service
4. T
 ele-dentistry system has been developed using mobile phones for imaging and evaluated in WA. One PhD (Mohamed
Estai who received Dean’s Award for his PhD on tele-dentistry) has published over 10 papers around this study
5. A
 ttracted major funding from NeuroVision Imaging to conduct further clinical trials around Alzheimer’s disease and
eye imaging.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 6 5

PROJECTS
Artificial intelligence-based
grading and diagnosis of
diabetic retinopathy at GP
clinics
The Australian Tele-Health Research
and Development Group has obtained a
Diabetes Research WA grant to validate
our smart eye-screening system for
early detection of the sight-threatening
condition, diabetic retinopathy (DR). The
aim of the project is to build and validate
a novel artificial intelligence-based

A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P
disease grading and clinical decision
support system for screening and
telemedicine-based diagnosis of DR.
The proposed system will be integrated
with fundus cameras, with real-time
image quality control software for colour
fundus images to produce sensitivity
and specificity over 90% for DR grading.

In the past year, we have completed the


DR screening model training process
and improved the performance of the
Quality Assessment model and DR
grading model. We have developed Figure 1: One example of DR grading performance on training and validation datasets.
a telemedicine system for the DR
screening and built all components
into one integrated system, which is
suitable for GP clinics. Our research
team worked closely with clinicians
from the Midland GP Superclinic, and
integrated the AI and telemedicine
system with GPs' patient management
system. We screened over 200 patients
and two of them were graded as
having severe disease and referred to
ophthalmologists.

aehrc.com
6 6 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

REMOTE DELIVERY OF CLINICAL SERVICES (continued)

Project aims for


Collaborators:
2018/19:
• GP Superclinic, Midland
• C
 omplete clinical trial at GP
• TeleMedC LLC. Superclinic, Midland
• T
 arget screening at least 200
Project highlights more patients
for 2017/18: • Data analysis on clinical data

• C
 ompleted the system • Implement the system at Royal
development Perth Hospital and REACH
A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P

Clinic.
• P
 articipated in “Explore
Diabetes” Expo event
organised by Diabetes
Research WA and showcased
the technology to audiences
• L
 icensed the system to
TeleMedC, and through the
company rolled out the system
in Singapore by collaborating
Figure 2: Smart phone application of the DR with National University
screening system. Hospital Singapore
• W
 AiTTA Incite Awards 2018
- Winner of “Most Innovative
Enabler in Health Care”.

Medical Image
Communication and
Exchange (MICE) app for
burns applications
Using the Medical Imaging and
Communication Exchange App, junior
doctors, specialists and nurses can
now seek expert real-time diagnostic
advice about various conditions by
securely sending burns images through
a mobile device to a burns specialist.
By streamlining the medical image
capture and storage process, the app
introduces a secure way of sending
medical images to the specialist. Images
captured through the app are not saved
on the mobile device and can only be
viewed by the specialists and related Figure 3: MICE app screenshots.
doctors. Medical records and images are
managed through our award-winning
store-and-forward telehealth system
(Remote-I). This system can be accessed
via web or mobile and maintains secure
access rights for all users.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 67

Highlights of MICE Project aims for


Collaborators:
Trial (Apr 2017- 2018/19:
June 2017) • P
 rof Fiona Wood, Director, • R
 oll out the diabetic retinopathy
• 3
 0 users, ranging from iPhone Burns Unit at the Fiona Stanley grading system to other clinics
5 to 7 Hospital • P
 ublish trial outcomes in high
• F
 our user roles (Medical • WA Health impact journals
Illustrations, Specialists, Project • GP Superclinic, Midlands, Perth • A
 pply for funding to expand
Coordinator, Admin) • WA Country Health Service MICE app to other medical
• ~
 40 patients (mainly captured applications (e.g. wounds,
• TeleMedC.
after hours) dermatology)

A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P
• ~
 120 images with ~ 200 image Project highlights • L
 icense MICE to WA Health for
use at different hospitals
tags for 2017/18:
• A
 gile: On-site agile development • F
 urther develop MICE on other
within the hospital – app was • D
 iabetic retinopathy grading mobile platforms and disease
co-developed with hospital staff system evaluation on training conditions
to cater for their direct clinical data set completed • E
 valuate data from various
requirements • Image quality control system clinical trials using VASP
• B
 arcode: Patient record evaluation completed (hypertension, stroke,
integration using barcode • M
 ICE app developed for iOS Alzheimer’s disease).
scanning feature (iPhones)
• R
 eporting: Reporting • M
 ICE app trial completed by the
functionality for hospital systems specialists and junior doctors at
integration Fiona Stanley Hospital and the
• Image tagging: Images taken hospital executive committee
from the app are tagged to is supporting the project for
identify the pictured body part, statewide rollout.
to assist in searchable clinical
images
• E
 ngagement: Engaged
with many other hospital
departments (such as plastics,
orthopedics, business intelligent
unit, security, IT)

Age-related macular degeneration (AMD) grading


Age-related macular degeneration which is time consuming and can learning-based disease classification.
(AMD) is the leading cause of vision loss difficult to maintain intra- or inter-rater Two publicly available AMD datasets
in people over the age of 50 years in agreement. Aiming to achieve population- (ARIA and STARE) were used for the
developed countries, and this number based disease screening, ATRDG aimed experiment along with a private dataset.
is expected to increase by 12- to 20-fold to develop a clinical decision system In the preliminary validation, the
over 10 years due to ageing populations. which can automatically perform AMD proposed system achieved an overall
Therefore, identifying people with “disease/no disease” grading. accuracy of 99.4% on the test dataset.
early signs of the disease and then The proposed colour normalisation
determining their risk based upon their We have developed an automated improves the overall accuracy of the
fundus characteristics is important method for the “disease/no-disease” deep learning framework by 16%.
when considering both the social and grading of AMD by applying deep
economic impacts of AMD. Existing learning techniques. The method
AMD grading systems require a human includes two major steps: 1) image
grader’s subjective manual interaction, illumination correction; 2) deep

aehrc.com
6 8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

REMOTE DELIVERY OF CLINICAL SERVICES


A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P

a) Colour fundus photograph prior to normalisation;


b) colour-normalised photograph; c) ROC curves for
disease/no-disease grading when deep features are
learnt at the last convolutional layer; d) ROC curves
for disease/no-disease grading when deep features
are learnt at the first fully connected layer.

Figure 4: Performance of deep learning model for


AMD disease grading.

P
 ostdoctoral fellow

Sajib Saha, PhD for the detection and classification Dr Saha acts as a reviewer for many
of DR pathologies. He has published journals, IEEE Transactions on
Dr Sajib Kumar Saha is a postdoctoral numerous research articles in top Broadcasting, Biomedical Signal
fellow at the CSIRO Australian e-Health ranked journals including ‘Investigative Processing and Control, Biomedical
Research Centre working towards Ophthalmology & Visual Science (IOVS)’ Physics and Engineering Express,
the development of machine learning journal (IF 3.43), ‘Journal of Medical Computer Methods in Biomechanics and
techniques for the automated detection Systems’ (IF: 2.46), 'Biomedical Signal Biomedical Engineering, Computers in
and progression analysis of sight Processing and Control' (IF: 2.7). Biology and Medicine, Journal of Digital
threatening eye disease, specifically Imaging, and the Journal of Cultural
diabetic retinopathy (DR) and age- Dr Saha has led the establishment Heritage. He is an editorial member of
related macular degeneration (AMD). Dr of a collaboration with the University the journal titled EC Ophthalmology.
Saha joined CSIRO in August 2015 and of California, Berkley and Stanford
during this time he has proposed and University in California to use their
developed several artificial intelligence- retinal image dataset and to develop
based methods for the automated novel techniques for the analysis of
analysis of retinal pathologies. His the disease. Dr Saha is also working
core developments include a novel with researchers from Sankara
retinal image registration method, an Nethralaya, India and Khulna University,
automated image quality assessment Bangladesh. Presently, Dr Saha is
method for DR screening, an automated working at UCLA for six months on
non-uniform/poor illumination applying deep learning methods for
correction method for fundus images, detecting early AMD pathologies from
a novel colour normalisation method OCT images.
to eliminate intersubjective colour
variability to facilitate automated
analysis, a deep learning method for
the "disease/no-disease" grading of
AMD, and a deep learning method

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 69

CHRONIC DISEASE MANAGEMENT

The Australian Tele-Health Research


Development Group (ATRDG) is also
interested in delivering management of . Project aims for
complex and chronic conditions in the Collaborators:
home using telehealth, particularly in
2018/19:
Western Australia. This is undertaken • C
 OPD: Dr Yuben Moodley,
in collaboration with the Mobile Fiona Stanley Hospital • C
 hronic Obstructive
Health team to extend mobile health (Respiratory Physician) Pulmonary Disease (COPD): To
trials currently underway to Western expand from the preliminary
Australian arms. One of the initiatives study, using a new version of
is to undertake a multi-centre trial of Mobile Technology enabled
chronic obstructive pulmonary disease Rehabilitation (MoTeR)
(COPD) with specialists among health developed by the Australian
and hospital services in Perth, Brisbane e-health Research Centre for

A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P
and Melbourne. This work will be an COPD, to include 50 patients
expansion of the pilot study conducted in a follow-on multi-state trial.
previously at the Royal Perth Hospital We plan to submit a grant
in collaboration with Dr Yuben Moodley proposal for a multi-centre
(now posted at the Fiona Stanley study to NHMRC.
Hospital). We are exploring additional
funding to continue a large scale multi-
centre study.

We have also explored using the cardiac


rehab system in remote and rural WA in
collaboration with the West Australian
Centre for Rural Health of the University
of Western Australia, and have applied
for NHMRC partnership funding for this
project.

aehrc.com
7 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

DISEASE DIAGNOSIS AND SCREENING


TECHNOLOGIES

In this stream of work we are exploring Preliminary results have been presented Retinal imaging in resistant
ways to develop biomarkers to screen at the Alzheimer’s Association
and diagnose diseases early, so International Conference, the
hypertension
appropriate treatment and surgery can International Conference on Alzheimer’s This is a collaborative research
be provided in a timely manner. Our and Parkinson’s diseases, and the project between Royal Perth Hospital
focus has been studying the changes Association for Research in Vision Hypertension Clinic and the ATRDG.
related to vascular parameters from and Ophthalmology Annual Meeting. The major aim of the project is to
the retina and the anterior segment in A manuscript communicating these identify novel retinal imaging markers
relation to various disease development results is in preparation. that may closely correlate with best
such as stroke, hypertension, practice blood pressure measurements
Alzheimer’s disease, heart disease and Retinal imaging equipment, protocols
and other signs of hypertensive organ
mental disorders. All data collection and image analysis techniques have
damage in these high risk patients.
has been carried out in partnership with all been improved as a result of these
Such a retinal marker or set of markers
various Perth clinics. studies. The latest technology is being
A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P

may serve as an integral summation


used for recruitment into the A4 AD-
of the blood pressure burden a patient
Ocular biomarkers for therapeutic trial (http://a4study.org/),
has been exposed to at any given time
and being tested in our current clinical
Alzheimer’s disease trial of N=284 participants across Perth
and provide a simple, non-invasive and
inexpensive test for accurate prediction
We have been developing ocular and Melbourne AIBL sites (N=250
of cardiovascular risk in these patients.
biomarkers for early detection of completed).
Furthermore, longitudinal assessment
Alzheimer’s disease. The goal of the and the corresponding changes in
We have also recently published papers
study is to see if a non-invasive and retinal markers may serve as an
in the Journal of Ophthalmology
inexpensive eye test can detect people indicator of adequate/inadequate anti-
and Current Alzheimer Research
on the pathway to the disease. hypertensive treatment.
reporting on changes in the way
The trial involves two visits by volunteers the pupil responds to light and the
Data collection has been completed for
to the Australian Alzheimer’s Research optical properties of retinal vessels
N=150 participants.
Foundation, where they have their eyes in Alzheimer’s disease. These ocular
tested using retinal image fluorescence changes are also evident in pre-clinical Additionally, another trial on the effects
photography. Between appointments, Alzheimer’s participants, suggesting of Tropicamide pupil dilating eye drops
volunteers take a curcumin supplement, that eye testing could be useful for on the retinal vasculature has been
which is a natural ingredient used detecting Alzheimer’s many years prior completed (N=41). Preliminary analysis
in cooking and also gives the spice to symptoms, allowing earlier testing of reveals some changes and a publication
turmeric its fluorescent yellow colour. interventions. is in preparation.
We use curcumin to light up the
Our new Cloud Based Multi-modality
amyloid-beta plaques in people’s
Intelligent Retinal Vessel AnalysiS
retinas. If what we see in the eye tests
Platform (VASP) has been patented and
correlates with what is occurring in their
utilised to search for further retinal
brains, we will have the makings of a
markers of Alzheimer’s.
screening tool for Alzheimer’s. It may
enable us to identify early development
of the disease, which could enhance
our ability to intervene and stop or delay
Alzheimer’s progression.

Clinical data collection has been


completed for the following studies:

• N=200 Nidek device


• N=100 Longitudinal study
• N=20 Young controls
• N=250 New Eidon/Retia device.

Figure 1: Retinal image analysis software interface


- VASP.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 7 1

Retinal biomarkers to predict stroke Retinal imaging and


Improvements in acute care of stroke progression. This pilot study investigated
nutrition
patients have decreased mortality, but the utility of a novel handheld, non- This is a collaborative research project
survivors are still at increased risk of mydriatic retinal camera in the stroke between CSIRO Food and Nutrition
future vascular events and mitigation of ward, and explored associations and ATRDG. Retinal imaging is being
this risk requires thorough assessment between retinal vascular features conducted in the Adelaide SAHMRI
of the underlying factors leading to and stroke risk factors. This camera centre. Recent findings include lower
the stroke. The brain and eye share circumvented the practical limitations retinal AVR, previously reported to be
a common embryological origin and of conducting retinal photography in the predictive of future stroke or CVD, is
numerous similarities exist between the stroke ward setting. also associated with lower plasma
small vessels of the retina and brain. carotenoid concentration. Simple,
Recent population-based studies have The study results were published in the
non-invasive retinal photography
demonstrated a close link between Journal of Stroke and Cerebrovascular

A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P
may have utility both in assessing
retinal vascular changes and stroke, Diseases in 2017. A positive correlation
stroke/CVD risk and monitoring the
suggesting that retinal photography was found between carotid disease
vascular impact of dietary intervention
could have utility in assessing and both mean width of arterioles (r
to increase blood carotenoids.
underlying stroke risk factors and = 0.40, p = 0.00571) and venules (r =
Additionally, in malnourished
prognosis after stroke. 0.30, p = 0.0381). The results provide
populations, this technique offers
further evidence that retinal vascular
potential to track carotenoid status as
The retinal microvasculature is highly features are clinically informative about
part of interventions aimed to reduce
accessible with modern imaging underlying stroke risk factors, and
vitamin A deficiency. A publication is in
equipment facilitating precise demonstrate the utility of handheld
preparation.
measurement and monitoring of retinal photography in the stroke ward.
vascular features. However, use of this
equipment is a challenge in the stroke With the RPH Stroke ward closing, the Retinal imaging to
ward setting as patients are frequently ATRDG is now pursuing contacts at detect H.Pylori related
unable to maintain the required seated Sir Charles Gairdner Hospital and the inflammation
position and pupil dilatation is often not Perron Institute for Neurodegenerative
feasible as it could potentially obscure Diseases to develop a new stroke This is a collaborative research project
important neurological signs of stroke research program. between Hollywood Private Hospital
(HPH) Nobel prize winner Professor
Barry Marshall and ATRDG. Retinal and
anterior eye image collection has begun
at HPH investigating ocular signs of
inflammation co-morbid with H.Pylori
infection, which results in increased
incidence of glaucoma. This is a pilot
study, with five patients so far, and it will
continue until we reach 20 patients.

Figure 2: Retinal fluorescence image - identifying AD plaques in retina and how they increase over time.

aehrc.com
7 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

DISEASE DIAGNOSIS AND SCREENING


TECHNOLOGIES (continued)

VASP: Vessel Analysis


Software Program
VASP is a cloud-based Vessel Analysis
Software Program developed in
2016/17. VASP was developed to fill
the gap of disease risk analysis from
the eye, can measure more than 80
vascular parameters and enables
retinal biomarker research. The findings
from research using VASP enables
disease diagnosis and preventative
management. There is a provisional
A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P

patent filled for VAS, and it was awarded


at the WAITTA iAwards (a Merit award
under the &D category).

VASP is supported by the state-of-the-


art performance algorithms, and has
user-intuitive web visualisation and user
interactions. Data is stored in a central
database, and novel graph-based vessel Figure 3: Retinal optical coherence tomography (OCT) identifying inflammation and subsequent atrophy in
topology structures are implemented. retinal layer thickness.
VASP is the first system deployed into
real-world clinics to assist disease
risk analysis. Some of the research
techniques include deep learning,
machine learning, image processing,
computer vision and graph theory.
VASP generates more than 80 retinal
parameters from each retinal image,
including central retinal equivalent
calibers, A/V width ratio, fractal
dimension, lacunarity, tortuosity, central
reflex and branching parameters. VASP
is unique in that it is the first cloud-
based retinal vessel analysis platform.
The intelligent algorithms in VASP
provide hints for user interaction. VASP
can also process multiple modality
images, in less than three minutes
including user interaction.

Figure 4: VASP image processing modules detect vessel tree structures, classify artery/vein, and measure
vessel thickness at various regions.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 7 3

A U S T R A L I A N T E L E - H E A LT H R E S E A R C H A N D D E V E L O P M E N T G R O U P
Figure 5: VASP system architecture is based on client-server web framework. The dynamic data representations and user
interactions are translated into data, saved at central database, and communicated to the Automatic Image Processing Unit.

Postdoc highlight
Project aims for
C
 ollaborators:
2018/19: Shuang Yu, PhD

• N
 eurovision Imaging, • C
 omplete new AD trial with As a Postdoctoral Research Fellow
Sacramento, California new imaging technology and at CSIRO, Dr Shuang Yu joined the
protocols Australian e-Health Research Centre in
• H
 ollywood Private Hospital
September 2015. Her research focussed
(Ramsay Health Care) • P
 ublish AD, tropicamide and
on the automatic analysis of retinal
• Royal Perth Hospital SAHMRI study results
vascular structures with graph theory
• University of Western Australia • D
 evelop RPH studies into and machine learning algorithms. This
additional clinical areas year her work led to a software patent
Project highlights • C
 omplete HPH trial on and numerous improvements to the way
H.Pylori infection and retinal retinal image analysis is conducted. Dr
for 2017/18: inflammatory markers Yu’s work was presented at ARVO 2017
• C
 linical trial nearing and EMBC 2017.
• A
 s the stroke clinic at Royal
completion - external funding
Perth Hospital has been closed
from Janssen, next-generation
and moved to Sir Charles
retinal fluorescence imaging
Gairdner Hospital we will be
in AD
seeking to establish similar
• V
 ASP software patented - study with a new cohort
Cloud Based Multi-modality
• R
 edefine the stroke study to
Intelligent Retinal Vessel
explore people with secondary
Analysis Platform for stroke,
stroke.
hypertension and H.Pylori
studies
• T
 ropicamide study completed -
manuscript under preparation.

aehrc.com
74 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

AEHRC PUBLICATIONS 2017-18


1. S
 . Ayton, A. Fazlollahi, P. 7. B
 . Celler, M. Varnfield, S. 12. J. Dowling, J. Wyatt, C. Kelly, 19. J. George, S. Fiori, J. Fripp,
Bourgeat, P. Raniga, A. Ng, Y. Nepal, R. Sparks, J. Li, R. J. McKenna, E. Johnstone, R. K. Pannek, A. Guzzetta,
Lim, I. Diouf, S. Farquharson, Jayasena, “Impact of at Speight, A. Henry, P. Greer, M. David, R. Ware, S.
J. Fripp, D. Ames, J. Doecke, home telemonitoring on H. McCallum, “Investigating Rose, P. Colditz, R. Boyd,
P. Desmond, R. Ordidge, C. health services expenditure the generalisation of an “Relationship between very
Masters, C. Rowe, P. Maruff, and hospital admissions atlas-based synthetic-CT early brain structure and
V. Villemagne, O. Salvado, in patients with chronic algorithm to another centre neuromotor, neurological
A. Bush, the AIBL Research conditions: a before and and MR scanner for prostate and neurobehavioral
Group, “Cerebral quantitative after control intervention MR-only radiotherapy”, function in infants born <31
susceptibility mapping analysis”, Journal of Medical Physics in Medicine and weeks gestational age”, Early
predicts β-amyloid-related Internet Research, vol. 5, pp. Biology, vol. 62, pp. 548-560, Human Development, vol.
cognitive decline”, Brain, vol. e29, Sep 2017. Nov 2017. 117, pp. 74-82, Jan 2018.
140, pp. 2112--2119, Aug
2017. 8. P
 . Chatterjee, K. Goozee, L. 13. Q. Duche, H. Saint-Jalmes, 20. J. George, K. Pannek, S.
Chai K., I. James, K. Shen, K. O. Acosta Tamayo, P. Raniga, Rose, R. Ware, P. Colditz, R.
2. S
 . Ayton, I. Diouf, A. Jacobs, H. Sohrabi, T. Shah, P. Bourgeat, V. Dore, G. Boyd, “Diagnostic accuracy
Bush, “Evidence that iron P. Asih, P. Dave, C. ManYan, Egan, O. Salvado, “Partial of early magnetic resonance
accelerates Alzheimer’s K. Taddei, D. Lovejoy, R. volume model for brain imaging to determine
pathology: a CSF biomarker Chung, G. Guillemin, R. MRI scan using MP2RAGE”, motor outcomes in infants
study”, Journal of Martins, “Alterations in Human Brain Mapping, vol. born preterm: a systematic
Neurosurgery Neurology and serum kynurenine pathway 38, pp. 5115–5127, Oct 2017. review and meta-analysis”,
Psychiatry, vol. 0, pp. 1-5, metabolites in individuals Developmental Medicine and
Sep 2017. with high neocortical 14. F. Fatehi, A. Smith, A. Child Neurology, vol. 60, pp.
amyloid-β load: A pilot Maeder, V. Wade, L. Gray, 134-146, Feb 2018.
3. A
 . Ball, D. Rye, D. Silvera, study”, Scientific Report, vol. “How to formulate research
M. Velonaki, “How should a 8, pp. 8008, May 2018. questions and design studies 21. S. Ghose, P. Greer, J. Sun, P.
robot approach two people?”, for telehealth assessment Pichler, D. Rivest-Henault,
Journal of Human-Robot 9. S
 . Cutler, J. Doecke, I. and evaluation”, Journal of J. Mitra, H. Richardson, C.
Interaction, vol. 6, pp. 71-91, Ghazawi, J. Yang, L. Griffiths, Telemedicine and Telecare, Wratten, J. Arm, L. Best,
Dec 2017. K. Spring, S. Ralph, A. vol. 23, pp. 759-763, Oct 2017. J. Dowling, “A Regression
Mellick, “Novel STAT binding and Statistical Shape and
4. D
 . Bauer, A. Zadoorian, L. elements mediate IL-6 15. F. Fatehi, A. Menon, D. Bird, Appearance Model based
Wilson, T. Kahlke, N. Thorne, regulation of MMP-1 and “Diabetes Care in Digital Substitute CT Generation
“Evaluation of computational MMP-3 in colon cancer.”, Era: a Synoptic Review”, for Standard MRI Sequence
programs to predict HLA Nature Scientific Reports, Current Diabetes Reports, vol. Alone External Beam
genotypes from genomic vol. 7, pp. 1-12, Aug 2017. 18, pp. Online, May 2018. Radiation Therapy Planning”,
sequencing data”, Briefings Physics in Medicine and
in Bioinformatics, vol. 19, pp. 10. H. Ding, R. Jayasena, A. 16. S. Frost, L. Robinson, Y.
Kanagasingam, “Evaluation Biology, vol. 62, pp. 8566-
179–187, Mar 2018. Maiorana, A. Dowling, 8577, Oct 2017.
S. Chen, M. Karunanithi, of cholinergic deficiency
5. B
 . Brown, S. Rainey-Smith, J. Layland, I. Edwards, in pre-clinical Alzheimer’s 22. A. Gillman, S. Rose, J.
N. Castalanelli, N. Gordon, “Innovative telemonitoring disease using pupillometry”, Smith, P. Thomas, N.
S. Markovic, H. Sohrabi, enhanced care program Journal of Ophthalmology, vol. Dowson, “A review of
M. Weinborn, S. Laws, J. for chronic heart failure 2017, pp. 8 p, Aug 2017. practical techniques for
Doecke, K. Shen, R. Martins, (ITEC-CHF) to improve motion correction in PET”,
J. Peiffer, “Study protocol of 17. S. Frost, J. Doecke, L.
guideline compliance and Macaulay, Y. Kanagasingam, Medical Physics, vol. 44, pp.
the Intense Physical Activity collaborative care: protocol e430-e445, Dec 2017.
and Cognition (IPAC) study: “Modulation of Retinal
of a multicenter randomized Arteriolar Central Reflection
The effect of high-intensity controlled trial”, BMJ Open, 23. S. Gulliford, S. Ghose,
exercise training on cognitive by APOE genotype”, Current M. Ebert, A. Kennedy, J.
vol. 7, pp. n/a, Oct 2017. Alzheimer Research, vol. 14,
function in older adults”, Dowling, J. Mitra, D. Joseph,
Alzheimer’s & Dementia: 11. J. Doecke, A. Rembach, V. pp. 916-923, Sep 2017. J. Denham, “Radiotherapy
Translational Research and Villemagne, S. Varghese, S. 18. J. George, S. Fiori, J. Fripp, dose-distribution to the
Clinical Interventions, vol. 3, Rainey-Smith, S. Sarros, L. K. Pannek, J. Bursle, A. perirectal fat space (PRS) is
pp. 562-570, Nov 2017. Evered, C. Fowler, K. Pertile, Guzzetta, A. Coulthard, R. related to gastrointestinal
R. Rumble, B. Trounson, K. Ware, S. Rose, P. Colditz, control-related
6. C
 . Budgeon, K. Murray, Taddei, S. Laws, L. Macaulay, complications”, Clinical
B. Turlach, V. Villemagne, R. Boyd, “Validation of
A. Bush, K. Ellis, R. Martins, an MRI Brain Injury and and Translational Radiation
D. Baker, S. Burnham, D. Ames, B. Silbert, H. Oncology, vol. 7, pp. 62-70,
“Constructing longitudinal Growth Scoring System
Vanderstichele, C. Masters, in Very Preterm Infants Nov 2017.
disease progression curves D. Darby, Q. Li, S. Collins,
using sparse, short-term Scanned at 29- to 35-
t. AIBL Research Group, Week Postmenstrual
individual data with an “Cerebrospinal fluid T-tau/
application to Alzheimer’s Age”, American Journal of
Aß42 ratio demonstrates Neuroradiology, vol. 38, pp.
Disease”, Statistics in strongest concordance
Medicine, vol. 36, pp. 2720- 1435-1442, Jul 2017.
with Alzheimer’s Disease
2734, Jul 2017. pathology independent of
assay platform”, Journal of
Alzheimer’s Disease, vol. 61,
pp. 169-183, Jan 2018.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 7 5

24. K. Harrington, Y. Lim, D. 30. B. Koopman, G. Zuccon, 35. R. Martins, V. Villemagnen, 40. A. Nakamura, N. Kaneko,
Ames, J. Hassenstab, S. P. Bruza, “What Makes an H. Sohrabi, P. Chatterjee, T. V. Villemagne, T. Kato,
Laws, R. Martins, S. Rainey- Effective Clinical Query and Shah, G. Verdile, P. Fraser, J. Doecke, V. Dore, Q. Li,
Smith, J. Robertson, C. Querier?”, Journal of the K. Taddei, V. Gupta, S. R. Martins, C. Rowe, T.
Rowe, O. Salvado, V. Dore, Association for Information Rainey-Smith, E. Hone, S. Tomita, K. Matsuzaki, K.
V. Villemagne, C. Masters, Science and Technology, vol. Pedrini, W. Lim, I. Martins, S. Ishii, K. Ishii, Y. Arahata, S.
P. Maruff, “Amyloid-β 68, pp. 2557–2571, Sep 2017. Frost, S. Gupta, S. O’Bryant, Iwamoto, K. Ito, K. Tanaka,
associated cognitive A. Rembach, D. Ames, K. C. Masters, K. Yanagisawa,
decline in the absence of 31. J. Lan, A. Galindo, J. Doecke, Ellis, S. Fuller, B. Brown, S. “High performance plasma
clinical disease progression C. Fowler, R. Martins, S. Gardener, B. Fernando, P. amyloid-β biomarkers
and systemic illness”, Rainey-Smith, O. Cominetti, Bharadwaj, S. Burnham, S. for Alzheimer’s disease”,
Alzheimer’s and Dementia, L. Dayon, “A systematic Laws, A. Barron, K. Goozee, Nature, vol. 554, pp. 249–254,
vol. 8, pp. 156-164, Aug 2017. evaluation of the use of E. Wahjoepramono, P. Asih, Jan 2018.
human plasma and serum J. Doecke, O. Salvado, A.
25. S. Holmes, I. Esterlis, C. for mass spectrometry- Bush, C. Rowe, S. Gandy, 41. A. Nakamura, N. Kaneko,
Mazure, Y. Lim, D. Ames, S. based shotgun proteomics”, C. Masters, “Alzheimer’s V. Villemagne, T. Kato, J.
Rainey-Smith, C. Fowler, K. Journal of Proteome Disease: A Journey Doecke, V. Dore, C. Fowler,
Ellis, R. Martins, V. Dore, O. Research, vol. 17, pp. 1426- from Amyloid Peptides Q. Li, R. Martins, C. Rowe,
Salvado, V. Villemagne, C. 1435, Feb 2018. and Oxidative Stress, to T. Tomita, K. Matsuzaki, S.
Rowe, S. Laws, C. Masters, Biomarker Technologies Iwamoto, K. Ito, K. Tanaka,
R. Pietrzak, P. Maruff, 32. S. Laws, S. Gaskin, A. C. Masters, K. Yanagisawa,
Woodfield, V. Srikanth, D. and Disease Prevention
“Trajectories of depressive Strategies-Gains from AIBL “High performance plasma
and anxiety symptoms Bruce, P. Fraser, T. Porter, P. Aβ-amyloid biomarkers
Newsholme, N. Wijesekara, and DIAN Cohort Studies”,
in older adults: a 6-year Journal of Alzheimer’s for Alzheimer’s disease”,
prospective cohort study”, Q. Li, S. Burnham, V. Dore, Nature, vol. 554, pp. 249–254,
P. Maruff, C. Masters, S. Disease, vol. 62, pp. 965–992,
Geriatric Psychiatry, vol. 33, Mar 2018. Feb 2018.
pp. 405–413, Feb 2018. Rainey-Smith, C. Rowe, O.
Salvado, V. Villemagne, R. 36. A. Menon, L. Gray, F. 42. A. Neubert, K. Wilson, C.
26. H. Huang, D. Bauer, P. Martins, G. Verdile, “Insulin Fatehi, D. Darssan, D. Engstrom, R. Surowiec,
Lelliott, M. Dixon, L. Tilley, resistance is associated Bird, D. Bennetts, A. A. Paproki, N. Johnson,
B. McMorran, S. Foote, G. with reductions in specific Russell, “A comparison of S. Crozier, J. Fripp, C. Ho,
Burgio, “Ankyrin-1 gene cognitive domains and characteristics of patients “Comparison of 3D bone
exhibits allelic heterogeneity increases in CSF tau in seen in a tertiary hospital models of the knee joint
in conferring protection cognitively normal adults”, diabetes telehealth service derived from CT and 3T MR
against malaria”, G3, vol. 7, Scientific Reports, vol. 7, pp. versus specialist face-to- imaging”, European Journal
pp. 3133-3144, Sep 2017. 9766, Aug 2017. face outpatients”, Journal of of Radiology, vol. 93, pp.
Telemedicine and Telecare, 178–184, Aug 2017.
27. M. Kholghi, L. Sitbon, G. 33. H. Leroux, A. Metke
Zuccon, A. Nguyen, “Active Jimenez, M. Lawley, vol. 23, pp. 842-849, Nov 43. A. Pagnozzi, N. Dowson, J.
learning reduces annotation “Towards achieving semantic 2017. Doecke, S. Fiori, A. Bradley,
time for clinical concept interoperability of clinical 37. S. Mohammad Mohaimin, R. Boyd, S. Rose, “Identifying
extraction”, International study data with FHIR”, S. Saha, A. Mahamud relevant biomarkers of brain
Journal of Medical Journal of Biomedical Khan, A. Mohammad Arif, injury from structural MRI:
Informatics, vol. 106, pp. 25- Semantics, vol. 8, pp. Article Y. Kanagasingam, “An Validation using automated
31, Oct 2017. 41, Sep 2017. Automated Method for the approaches in children with
Detection and Segmentation unilateral cerebral palsy”,
28. M. Kholghi, L. De Vine, 34. Y. Lim, S. Rainey-Smith, PLoS ONE, vol. 12, pp.
L. Sitbon, G. Zuccon, Y. Lim, S. Laws, V. Gupta, of Drusen in Color Fundus
Image for the Diagnosis e0181605, Aug 2017.
A. Nguyen, “Clinical T. Porter, P. Bourgeat, D.
Information Extraction Ames, C. Fowler, O. Salvado, of Age-related Macular 44. K. Pannek, J. Fripp, J.
Using Small Data: An Active V. Villemagne, C. Rowe, Degeneration”, IET Image George, S. Fiori, P. Colditz, R.
Learning Approach Based on C. Masters, X. Zhou, R. Processing, vol. 12, pp. 919- Boyd, S. Rose, “Fixel-Based
Sequence Representations Martins, P. Maruff, “BDNF 927, May 2018. Analysis reveals alterations
and Word Embeddings”, Val66Met in preclinical 38. G. Mokhtari, N. Bashi, Q. in brain microstructure
Journal of the Association Alzheimer’s disease is Zhang, G. Nourbakhsh, and macrostructure of
for Information Science associated with short-term “Non-wearable human preterm-born infants at term
and Technology, vol. 68, pp. changes in episodic memory identification sensors for equivalent age”, NeuroImage:
2543–2556, Nov 2017. and hippocampal volume smart home environment: a Clinical, vol. 18, pp. 51-59,
but not serum mBDNF”, review”, Sensor Review, vol. Jan 2018.
29. L
 . Koivula, M. Kapanen, T. International Psychogeriatrics,
Seppala, J. Collan, J. Dowling, 38, pp. 391-404, Dec 2017. 45. A. Paproki, C. Engstrom,
vol. 29, pp. 1825-1834, Nov
P. Greer, C. Gustafsson, A. 2017. 39. G. Mokhtari, A. Anvari- M. Strudwick, K. Wilson, R.
Gunnlaugsson, L. Olsson, L. Moghaddam, Q. Zhang, Surowiec, C. Ho, S. Crozier,
Wee, J. Korhonen, “Intensity- M. Karunanithi, “Multi- J. Fripp, “Automated T2-
based dual model method residential localization and mapping of the menisci from
for generation of synthetic activity recognition in Smart magnetic resonance images
CT images from standard Home Using Bluetooth in patients with acute knee
T2-weighted MR images – Low Energy Technology”, injury”, Academic Radiology,
generalized technique for International Journal of vol. 24, pp. 1295-1304, Oct
four different MR scanners”, Medical Informatics, vol. 18, 2017.
Radiotherapy and Oncology, pp. 908, Mar 2018.
vol. 125, pp. 411-419, Dec
2017.

aehrc.com
76 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

AEHRC PUBLICATIONS 2017-18 (continued)

46. S. Pedrini, V. Gupta, E. Hone, 52. D


 . Roach, M. Jameson, 58. V. Villemagne, C. Rowe, 64. C. Zupanc, G. Wallis, A. Hill,
J. Doecke, S. O’Bryant, I. J. Dowling, M. Ebert, P. K. Barnham, R. Cherny, R. Burgess-Limerick, S.
James, A. Bush, C. Rowe, Greer, A. Kennedy, S. Watt, M. Woodward, S. Pejoska, Riek, A. Plooy, M. Horswill,
V. Villemagne, D. Ames, L. Holloway, “Correlations O. Salvado, P. Bourgeat, M. Watson, H. De Visser,
C. Masters, R. Martins, t. between contouring similarity K. Perez, C. Fowler, A. D. Conlan, D. Hewett,
AIBL Research Group, “A metrics and simulated Rembach, P. Maruff, C. “Assessing colonoscopic
blood-based biomarker treatment outcome for Ritchie, R. Tanzi, C. Masters, inspection skill using a virtual
panel indicates IL-10 and prostate radiotherapy”, “A randomized, exploratory withdrawal simulation: a
IL-12/23p40 are jointly Physics in Medicine and molecular imaging study preliminary validation of
associated as predictors Biology, vol. 63, pp. 035001- targeting Aβ-amyloid with performance metrics”, BMC
of β-amyloid load in a AD 035015, Jan 2018. a novel 8-OH quinoline in Medical Education, vol. 17, pp.
cohort.”, Nature Scientific Alzheimer’s disease (The 10, Jul 2017.
Reports, vol. 7, pp. 1-12, Oct 53. C. Rowe, V. Dore, G. Jones, PBT2-204 IMAGINE Study).”,
2017. D. Baxendale, R. Mulligan, Alzheimer’s & Dementia: 65. Y. van Kasteren, J. Freyne,
S. Bullich, A. Stephens, S. Translational Research & S. Hussain, “Total Knee
47. T. Porter, S. Burnham, De Santi, C. Masters, L. Clinical Interventions, vol. 3, Replacement and the Effect
V. Dore, G. Savage, P. Dinkelborg, V. Villemagne, pp. 622–635, Nov 2017. of Technology on Cocreation
Bourgeat, K. Begemann, L. “18F-Florbetaben Ab- for Improved Outcomes and
Milicic, A. Gozt, P. Maruff, Amyloid PET Expressed 59. V. Villemagne, V. Dore, P. Delivery: Qualitative Multi-
S. Rainey-Smith, D. Ames, in the Centiloid Scale.”, Bourgeat, S. Burnham, R. Stakeholder Study”, Journal
C. Masters, C. Rowe, R. European Journal of Nuclear Mulligan, S. Laws, J. Fripp, of Medical and Internet
Martins, D. Groth, G. Verdile, Medicine, vol. 44, pp. T. Cummins, O. Salvado, C. Research, vol. 20, pp. e95,
V. Villemagne, S. Laws, 2053–2059, Nov 2017. Masters, C. Rowe, “The Tau Mar 2018.
“KIBRA is associated with MeTeR composites for the
accelerated cognitive decline 54. S. Saha, D. Xiao, S. generation of continuous 66. A. van der Vegt, G. Zuccon,
and hippocampal atrophy in Frost, Y. Kanagasingam, and categorical measures B. Koopman, P. Bruza, “A
APOE β4-positive cognitively “Performance Evaluation of tau deposits in the brain”, Task Completion Framework
normal adults with high Aβ- of State-of-the-art Local Journal of Molecular Medicine to Support Single-Interaction
amyloid burden”, Scientific Feature Detectors and and Therapy, vol. 1, pp. 25-32, IR Research”, Journal of
Reports, vol. 8, pp. 2034, Feb Descriptors in the Context of Dec 2017. Documentation, vol. 74, pp.
2018. Longitudinal Registration of 289–308, Jan 2018.
Retinal Images”, Journal of 60. V. Villemagne, V. Dore,
48. Project MinE Consortium, D. Medical Systems, vol. 42, pp. S. Burnham, C. Rowe, 67. S. Saha, A. Fletcher, D.
Bauer, N. Twine, “Genome- 57, Feb 2018. C. Masters, “Imaging Xiao, Y. Kanagasingam, “A
wide Analyses Identify KIF5A proteinopathies in Novel Method for Automated
as a Novel ALS Gene”, 55. C. Tomadesso, V. de La Alzheimer’s disease and Correction of Non-uniform/
Neuron, vol. 97, pp. 1268- Sayette, R. De Flores, P. other conditions”, Nature Poor Illumination of Retinal
1283, Mar 2018. Bourgeat, V. Villemagne, Reviews Neurology, vol. 14, Images without Creating
S. Egret, F. Eustache, G. pp. 225–236, Apr 2018. False Artifacts”, Journal of
49. L. Reid, M. Sale, R. Chetelat, “Neuropsychology Visual Communication and
Cunnington, J. Mattingley, and neuroimaging profiles 61. L. Wilson, D. Reti, A. Image Representation, vol. 51,
S. Rose, “Brain changes of amyloid-positive versus O’Brien, R. Dunne, D. pp. 95-103, Feb 2018.
following four weeks of negative mild cognitive Bauer, “High activity target-
unimanual motor training: impairment patients”, site identification using
Evidence from fMRI-guided Neuroimage, vol. 10, pp. 269- phenotypic independent Conference
diffusion MRI tractography”, 277, Mar 2018. CRISPR-Cas9 core 3 Publications and
Human Brain Mapping, vol.
56. N
 . Twine, L. Harkness,
functionality”, The Crispr Presentations
38, pp. 11, Sep 2017. Journal, vol. 1, pp. 182-190,
J. Adjaye, A. Aldahmash, Apr 2018. 1. P
 . Bourgeat, V. Villemagne, V.
50. L. Reid, A. Pagnozzi, S. M. Wilkins, M. Kassem, Dore, C. Masters, D. Ames,
Rose, “Brain changes “Molecular Phenotyping of 62. D. Xiao, J. Vignarajan, Y. C. Rowe, O. Salvado, J. Fripp,
following four weeks of Telomerized Human Bone Kanagasingam, “Content “PET-Only 18F-AV1451 Tau
unimanual motor training: Marrow Skeletal Stem Cells design and system Quantification”, ISBI 2017,
Evidence from behavior, Reveals a Genetic Program of implementation of a Melbourne, Australia, 18-21
neural stimulation, cortical Enhanced Proliferation and teleophthalmology system April 2017.
thickness, and functional Maintenance of Differentiation for eye disease diagnosis
MRI”, Human Brain Mapping, Responses”, JBMR Plus, pp. and treatment and its 2. D
 . Bradford, S. Berkovsky,
vol. 38, pp. 15, Sep 2017. 1-11, Mar 2018. practice in Guangdong S. Rodriguez, D. Ireland, D.
China”, Telemedicine and Bauer, C. Gaff, “Interacting
51. M. Remika, D. Raffelt, T. 57. M. Varnfield, S. Nepal, R. e-Health, vol. 23, pp. 948- with Genomic Data: Clinician
Dhollander, D. Vaughan, Sparks, J. Li, R. Jayasena, 954, Dec 2017. Requirements and Prototype
J. Tournier, O. Salvado, A. B. Celler, “Impact of at Design”, HIC, Brisbane,
Brodtmann, C. Rowe, V. home telemonitoring on 63. S. Yu, D. Xiao, Y. August 2017.
Villemagne, A. Connelly, health services expenditure Kanagasingam, “Machine
“Fibre-specific white matter and hospital admissions Learning Based Automatic 3. J . Doecke, Q. Li, C. Fowler,
reductions in Alzheimer’s in patients with chronic Neovascularization Detection S. Collins, V. Dore, C. Rowe,
disease and mild cognitive conditions. Results of the on Optic Disc Region”, IEEE C. Masters, O. Salvado, V.
impairment”, Brain, vol. 141, CSIRO Australian National Journal of Biomedical and Villemagne, “CSF P-tau is
pp. 888-902, Mar 2018. Telehealth Trial”, BMC Public Health Informatics, vol. 22, correlated with tau PET, while
Health, vol. 5, pp. e29, Sep pp. 886-894, May 2018. Aβ PET correlates with Aβ1-42
2017. and the T-tau/Aβ1-42 ratio”,
AAIC, London, July 2017.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 7 7

4. J . Doecke, Q. Li, P. Bourgeat, 11. N. Hearn, V. Dore, J. Fripp, 18. H. Ngo, A. Metke Jimenez, A. 24. H. Scells, G. Zuccon, B.
C. Fowler, S. Collins, V. Dore, M. Grothe, S. Teipel, C. Nguyen, “Knowledge-based Koopman, A. Deacon, S.
C. Rowe, C. Masters, V. Masters, C. Rowe, O. Feature Engineering for Geva, “A Test Collection
Villemagne, O. Salvado, “CSF Salvado, E. Coulson, V. Detecting Medication and for Evaluating Retrieval
Total Tau as a biomarker Villemagne, “Global tau Adverse Drug Events from of Studies for Inclusion in
for neuronal injury in burden correlates with Electronic Health Records”, Systematic Reviews”, SIGIR,
Alzheimer’s Disease: basal forebrain atrophy in NLP Challenges for Detecting Tokyo, Japan, 7-11 August
Aligning rates of CSF change healthy aging subjects”, Medication and Adverse Drug 2017.
with rates of hippocampal World Federation of Nuclear Events from Electronic Health
and cortical gray matter Medicine and Biology, Records (MADE1.0), USA, May 25. H. Scells, G. Zuccon, B.
atrophy”, AAIC, London, July Melbourne, 20-24 April 2018. 4, 2018. Koopman, A. Deacon,
2017. L. Azzopardi, S. Geva,
12. S. Hussain, B. Nicholson, 19. T. Nguyen, Q. Zhang, D. Le, “Integrating the Framing of
5. J . Doecke, E. Whittle, V. J. Freyne, “Tune your sun M. Karunanithi, “Dirichlet Clinical Questions via PICO
Villemagne, C. Masters, right: Persuasive app Process Gaussian Mixture into the Retrieval of Medical
S. Varghese, B. Roberts, towards healthy & safe sun Model for Activity Discovery Literature for Systematic
“Graphical network analyses exposure behavior”, The 29th in Smart Homes with Reviews”, CIKM 2017,
informs PET Aβ-amyloid Australian Conference on Ambient Sensors”, EAI Singapore, November 6-10,
biomarker discovery via Human-Computer Interaction International Conference 2017.
quantification of individual (OzCHI), Brisbane, Australia, on Mobile and Ubiquitous
peptide connections”, AAIC, 28th Nov - 1st Dec, 2017. Systems: Computing, 26. Z. Shahabi Kargar, S.
London, July 2017. Networking and Services, Khanna, A. Sattar, J. Lind,
13. Jimmy, G. Zuccon, B. Melbourne, Australia, “Improved Prediction of
6. R
 . Finnegan, J. Dowling, L. Koopman, “Choices 6-9/11/2017. Procedure Duration for
Holloway, J. Otton, E. Koh, in Knowledge-Base Elective Surgery”, HIC 2017,
C. Luo, P. Atluri, S. Tang, Retrieval for Consumer 20. A. Pagnozzi, N. Dowson, Brisbane, Australia, 6-9
G. Delaney, V. Batumalai, Health Search”, European J. Doecke, A. Bradley, S. August.
D. Thwaites, “Feasibility Conference on Information Fiori, R. Boyd, S. Rose,
of Automatic Cardiac Retrieval, Grenoble, France, “Optimization of MRI-based 27. K. Shen, V. Dore, S. Rose,
Segmentation in Planning 26-29 March 2018. scoring scales of brain J. Fripp, K. McMahon, G.
CT”, Engineering and Physical injury severity in children de Zubicaray, N. Martin,
Sciences in Medicine, Hobart, 14. B. Koopman, G. Zuccon, J. with unilateral cerebral P. Thompson, M. Wright,
29 October 2017. Russell, “A Task-oriented palsy”, American Academy O. Salvado, “Genetic
Search Engine for Evidence- for Cerebral Palsy and Correlation between cortical
7. S
 . Frost, Y. Kanagasingam, based Medicine”, SIGIR, Developmental Medicine, gray matter thickness and
M. Koronyo-Hamaoui, Tokyo, August 7-11, 2017. Montreal, 13 - 16 September white matter connections”,
Y. Koronyo, D. Biggs, S. 2017. MICGen - MICCAI Workshop
Verdooner, K. Black, R. 15. B. Koopman, L. Cripwell, on Imaging Genetics, Munich,
Martins, t. AIBL Research G. Zuccon, “Generating 21. S. Pedrini, V. Gupta, E. Hone, Germany, 9 October 2015.
Group, “Retinal morphology Clinical Queries from Patient J. Doecke, S. O’Bryant, I.
in pre-clinical Alzheimer’s Narratives: A Comparison James, A. Bush, C. Rowe, 28. H. Sohrabi, S. Gardener,
Disease”, ARVO Annual between Machines and V. Villemagne, D. Ames, C. M. Weinborn, K. Shen, S.
Meeting - The Association Humans”, SIGIR, Tokyo, 7-11 Masters, R. Martins, t. AIBL Rainey-Smith, K. Taddei,
for Research in Vision and August 2017. Research Group, “IL-10 J. Doecke, O. Salvado, V.
Ophthalmology, Honolulu, and IL-12/23p40 are jointly Villemagne, G. Savage,
16. A. Largent, J. Nunes, D. Ames, C. Masters, C.
Hawaii, April 2018. S. Herve, A. Simon, N. associated as predictor of
β-amyloid load in a broader Rowe, R. Martins, t. AIBL
8. T
 . Fuangrod, A. McNeilly, P. Perichon, A. Barateau, Research Group, “Cognitive
C. Hervé, L. Caroline, J. blood-based biomarker
Pichler, J. Dowling, P. Greer, panel”, AAIC, London, July super-aging vs. typical aging
“Verification of accuracy of Dowling, R. de Crevoisier, in community-dwelling
O. Acosta, “Pseudo-CT 2017.
substitute CT for MRI-only older adults: Longitudinal
prostate IMRT using transit Generation by Conditional 22. L. Reid, A. Pagnozzi, S. Rose, trajectories in global cortical
dosimetry”, Engineering and Inference Random Forest J. Fripp, “Self-Organising thickness over six years”,
Physical Sciences in Medicine, for MRI-based Radiotherapy Maps Enable Global AAIC, London, July 2017.
Hobart, 29 October 2017. Treatment Planning”, Searching For Patch Based
European Signal Processing Segmentations”, ISMRM, 29. Y. Thong, T. Nguyen, Q.
9. P
 . Greer, P. Pichler, H. Conference (EUSIPCO) 2017, Paris, 2018. Zhang, M. Karunanithi,
Richardson, P. Hunter, Greece, August 28 2017. L. Yu, “Predicting Food
C. Wratten, J. Dowling, 23. D. Roach, M. Jameson, Nutrition Facts Using
“MRI-only prostate cancer 17. A. Neubert, C. Engstrom, J. Dowling, M. Ebert, A. Pocket-size Near-Infrared
treatment – the HIPSTER I. Croci, S. Chandra, B. Kennedy, P. Greer, L. Sensor”, 39th Annual
study”, Engineering and Schmitt, S. Crozier, J. Fripp, Holloway, “Development and International Conference of
Physical Sciences in Medicine, “Automated evaluation validation of a CT-MR pelvic the IEEE Engineering, Korea,
Hobart, 29 October 2017. of T2 relaxation time atlas for clinical trial quality 11/7/2017.
measurements in the knee assurance”, Engineering and
10. T. Hadwen, V. Smallbon, Q. cartilage at 3T”, ISMRM Physical Sciences in Medicine,
Zhang, M. D’Souza, “Energy Annual Meeting, Paris, Hobart, 29 October 2017.
Efficient LoRa GPS Tracker France, 16-21 June 2018.
for Dementia Patients”,
39th Annual International
Conference of the IEEE
Engineering in Medicine
and Biology Society, Korea,
11/7/2017.

aehrc.com
7 8 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

AEHRC PUBLICATIONS 2017-18 (continued)

30. S. Tran, Q. Zhang, M. 37. J. Yin, S. Tran, Q. Zhang, 45. D. Bauer, A. O’Brien, O. Luo, 53. P. Bourgeat, V. Villemagne,
Karunanithi, “Hidden Markov “Human Identification via N. Twine, A. Bayat, L. Wilson, V. Dore, C. Masters, D.
Models for Multi-resident Unsupervised Feature L. Langit, P. Szul, “Genomics Ames, C. Rowe, O. Salvado,
Activity Recognition in Smart Learning from UWB Radar from a Bioinformatician and J. Fripp, “18F-AV1451 Tau
Homes”, IJCAI 2017 AI4IoT Data”, Pacific-Asia Knowledge a Developer”, YOW! Brisbane, Quantification without MRI”,
Workshop, Melbourne, Discovery and Data Mining, Brisbane, 4 - 5 December AAIC 2017, London, UK, 16-
August 2017. Melbourne, Australia, 3-6 2017. 20 July 2017.
June 2018.
31. S. Tran, Q. Zhang, V. 46. D. Bauer, A. O’Brien, O. Luo, 54. P. Bourgeat, V. Dore, D.
Smallbon, M. Karunanithi, 38. T. Young, D. Thwaites, J. N. Twine, A. Bayat, L. Wilson, Ames, C. Masters, J. Fripp,
“Multi-Resident Activity Dowling, G. Liney, R. Rai, P. Szul, “How novel compute O. Salvado, V. Villemagne,
Monitoring in Smart P. Greer, L. Holloway, “The technology transforms C. Rowe, “Implementing the
Homes: A Case Study”, Effect of MRI Sequence life science research”, Centiloid transformation
SmarterAAL Workshop at Variations on Substitute Cloudera Government Forum, for 18F-Florbetaben
IEEE International Conference CT Generation for MR-Only Canberra, 5 December 2017. and 18F-NAV4694 using
on Pervasive Computing and Planning”, Engineering and CapAIBL”, AAIC 2017,
Communicaitions, Athen, Physical Sciences in Medicine, 47. D. Bauer, A. O’Brien, O. Luo, London, UK, 14-20 July 2017.
Greece, 19-23 March 2018. Hobart, 29 October 2017. N. Twine, A. Bayat, L. Wilson,
L. Langit, P. Szul, “Genomics 55. P. Bourgeat, V. Dore, P.
32. V. Villemagne, V. Dore, N. 39. B. Yu, L. Zhou, L. Wang, J. from a Bioinformatician and Raniga, J. Fripp, C. Masters,
Okamura, D. Baxendale, Fripp, P. Bourgeat, “3D cGAN a Developer”, YOW! Sydney, C. Rowe, V. Villemagne,
R. Harada, R. Mulligan, S. Based Cross-Modality MR Sydney, 7-8 December 2017. O. Salvado, “MilxCloud: a
Furumoto, O. Salvado, K. Image Synthesis for Brain web-based platform for
Yanai, C. Masters, C. Rowe, Tumor Segmentation”, ISBI 48. D. Bauer, A. O’Brien, O. Luo, PET and MR quantification”,
“To tau or to MAO-B? Most 2018, Washington DC, USA, N. Twine, A. Bayat, L. Wilson, 12th Congress of the World
of the 18F-THK5351 signal is 4-7 April 2018. P. Szul, “How novel compute Federation of Nuclear
blocked by selegiline”. technology transforms life Medicine and Biology,
40. S. Yu, D. Xiao, Y. science research”, Linux Melbourne, Australia, 20-24
33. K. Williams, N. Twine, E. Kanagasingam, “Exudate Conference, Sydney, 22 April 2018.
McCann, A. O’Brien, D. Detection for Diabetic January 2018.
Bauer, I. Blair, “Discovery Retinopathy With 56. P. Bourgeat, V. Villemagne,
of previously unknown Convolutional Neural 49. D. Bauer, A. O’Brien, O. Luo, V. Dore, D. Ames, R. Martins,
relationships between ALS Networks”, Annual N. Twine, A. Bayat, L. Wilson, C. Masters, O. Salvado, C.
patients increases power International Conference P. Szul, “How novel compute Rowe, J. Fripp, “18F-AV1451
to identify causal disease of the IEEE Engineering in technology transforms life TAU Quantification Without
genes”, International Medicine and Biology Society science research”, Agile India MRI”, 12th Congress of the
Symposium on ALS/MND, (EMBC), Jeju Island, Korea, - Digital Transformation Day, World Federation of Nuclear
Boston, 8-10 December 11 July 2017. Bengaluru, India, 8 March Medicine and Biology,
2017. 2018. Melbourne, Australia, 20-24
41. G. Zuccon, B. Koopman, April 2018.
34. D. Xiao, S. Yu, J. Vignarajan, “SIGIR 2017 Tutorial on 50. D. Bauer, A. O’Brien, O. Luo,
M. Tay-Kearney, Y. Health Search (HS2017): A N. Twine, A. Bayat, L. Wilson, 57. P. Bourgeat, V. Dore, S.
Kanagasingam, “Retinal Full-day from Consumers “Going Server-less for Web- Williams, D. Ames, R.
hemorrhage detection by to Clinicians”, SIGIR, Tokyo, Services that need to Crunch Martins, C. Masters, O.
rule-based and machine Japan, 7-11 August. Large Volumes of Data”, Salvado, V. Villemagne, C.
learning approach”, IEEE Agile India - Continuous Rowe, “Implementing the
EMBC 2017, Jeju, Korea, 11 42. E. Bariola, N. Good, J. Li, Delivery and DevOps Day, Centiloid transformation
July 2017. M. Varnfield, R. Sparks, Bengaluru, India, 9 March of 18F-Florbetaben and
D. Rolls, A. Reeson, R. 2018. 18F-NAV4694 using
35. Y. Xie, S. Khanna, N. Good, Jayasena, “Reducing CapAIBL”, 12th Congress
J. Boyle, “Weekly hospital unplanned hospital re- 51. D. Bauer, A. O’Brien, O.
Luo, N. Twine, A. Bayat, L. of the World Federation of
workforce data: a data admissions: Evaluation of Nuclear Medicine and Biology,
visualisation exercise”, alternative models of care”, Wilson, P. Szul, “How AI and
‘Big Learning’ transforms Melbourne, Australia, 20-24
HIC (health information HIC Conference, Brisbane, April 2018.
conference), Brisbane, 6-9 9th August. life science research”, AI Dev
August 2017. Days, Bengaluru, India, 9 58. D. Bradford, “Healthy
43. D. Bauer, A. O’Brien, O. Luo, March 2018. Mates: Collaborations
36. Y. Xie, N. Good, S. Khanna, N. Twine, A. Bayat, L. Wilson, to enhance Indigenous
J. Boyle, “Exploratory L. Langit, P. Szul, “Cloud 52. G. Bischof, P. Bartenstein,
H. Berthel, B. van Berckel, health autonomy”, NHMRC,
Multivariate Analysis of Data Pipelines for Genomics Brisbane, November 2017.
Hospital Admissions Data in from a Bioinformatician and V. Dore, T. Van Eimeren,
Conjunction With Workforce a Developer”, YOW! Data, N. Forster, J. Hammes, 59. D. Bradford, M. Barber,
Data”, The 39th Annual Sydney, Sept 2017. A. Lammertsma, S. C. Robinson, D. Smith, G.
International Conference Minoshima, C. Rowe, O. Farr-Wharton, “Leveraging
of the IEEE Engineering in 44. D. Bauer, A. O’Brien, O. Luo, Sabri, J. Seibyl, K. Van Technologies to Improve
Medicine and Biology Society N. Twine, A. Bayat, L. Wilson, Laere, R. Vandenberghe, Workplace Wellbeing for
(EMBC’17), Jeju Island, L. Langit, P. Szul, “Genomics V. Villemagne, I. Yakushev, Aboriginal and Torres
Korea, 11-15 July 2017. from a Bioinformatician A. Drzezga, “Comparing Strait Islander Employee”,
and a Developer”, YOW! Amyloid PET Tracers and OzCHI, Brisbane, Australia,
Melbourne, Melbourne, 30 Interpretation Strategies: November 28th - December
Nov - 1 Dec 2017. Results from the CAPTAINs 1st.
Study”, SNMMI, Philadephia,
June 2018.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 7 9

60. D. Bradford, D. Ireland, 67. G. Farr-Wharton, D. Ireland, 74. H. Leroux, P. Raniga, “FHIR 81. K. Pannek, J. Welsh, J.
J. Wiles, “Talk to me: D. Bradford, “Social Fringe as an integrative platform Fripp, J. George, P. Colditz,
The Chatbot as a speech Dwellers: Designing Chat Bots for reproducible biomedical R. Boyd, S. Rose, “Removal
pathology monitoring tool”, as Bridges for Children with imaging research”, INCF or correction of volumes
Health by Tech, University of ASD”, OzCHI 2017, Brisbane, Neuroinformatics 2017, Kuala affected by bulk motion:
Twente, 1 June 2018. Australia, November 28th to Lumpur, Malaysia, August impact on DTI and NODDI
December 1st. 20-21, 2017. metrics”, Joint Annual
61. D
 . Bradford, S. Brinkmann, Meeting of ISMRM and
“Speaking my language: 68. A. Fazlollahi, S. Ayton, I. 75. H. Leroux, K. Fung, I. Priebe, ESMRMB, Paris, France, 16-
technology to facilitate hospital Diouf, P. Bourgeat, P. Raniga, M. Buckley, A. Ovenden, E. 21 June 2018.
assessment of culturally and V. Dore, A. Ng, J. Fripp, D. Berton, K. Surinya, J. Wei,
linguistically diverse patients”, Ames, C. Masters, C. Rowe, M. Thomas, J. Moore, R. 82. D. Rolls, J. Boyle, S.
Health by Tech, Netherlands, 1 V. Villemange, A. Bush, Singh, A. Ruszkiewicz, L. Khanna, N. Good, “Practical
June 2018. O. Salvado, “Quantitative Cosgrove, “Using Clinical considerations for using
susceptibility mapping of Research Data to Evaluate administrative data
62. S. Burnham, P. Bourgeat, V. the hippocampus predicts a Blood-Based Biomarker and risk stratification
Dore, S. Laws, O. Salvado, hippocampal atrophy in Panel for the Detection of to reduce unplanned
C. Masters, C. Rowe, V. Abeta+ elderly controls Colorectal Cancer”, Health rehospitalisations”, HIC 2017,
Villemagne, “Implementing and Alzheimer’s disease Data Analytics, Brisbane, Brisbane, August 6-9, 2017.
the ATN classification in patients”, Alzheimer’s Queensland, Australia,
AIBL”, 12th Congress of the Association International October 11-12, 2017. 83. C. Rowe, V. Dore, R.
World Federation of Nuclear Conference, London, Mulligan, P. Bourgeat, O.
Medicine and Biology, England, 16-20 July 2017. 76. J. Li, J. Freyne, D. Bradford, Salvado, C. Masters, V.
Melbourne, Australia, 20-24 “Streamlining Surgeries: Villemagne, “Tau Imaging
April 2018. 69. S. Gardener, S. Rainey- Digital Preference Card for in Alzheimer’s Disease with
Smith, K. Shen, P. Bourgeat, Surgical Item Management”, 18F-MK6240, a Second
63. S. Burnham, S. Laws, K. Taddei, D. Ames, C. HIC2017, Brisbane, 6-9 Generation Selective Tau
V. Dore, P. Bourgeat, O. Masters, C. Rowe, O. August 2017. Tracer”, 12th Congress of
Salvado, C. Masters, C. Salvado, R. Martins, the World Federation of
Rowe, V. Villemagne, “Mediterranean diet 77. E. McCann, J. Fifita, K. Nuclear Medicine and Biology,
“The impact of APOE e4 adherence is associated with Williams, S. Chan, K. Melbourne, Australia, 20-24
carriage and gender on attenuated cortical thinning Zhang, S. Freckleton, R. April 2018.
the age of onset of amyloid in an Australian study of Lyu, N. Twine, D. Bauer,
deposition”, 12th Congress ageing”, AAIC, London, UK, D. Rowe, G. Nicholson, I. 84. C. Rowe, S. Amadoru, V.
of the World Federation of 16-20 July 2017. Blair, “Discovery of new Dore, C. McLean, F. Hinton,
Nuclear Medicine and Biology, disease genes in motor C. Shepherd, G. Halliday,
Melbourne, Australia, 20-24 70. J. George, R. Ware, S. Fiori, neuron disease utilising C. Leyton, J. Hodges, V.
April 2018. J. Fripp, K. Pannek, S. next generation sequencing Villemagne, “Correlation
Rose, P. Colditz, R. Boyd, technologies and genetic of Amyloid PET in Centiloid
64. T. Cummins, A. Elias, M. “Early structural MRI linkage”, FightMND, units with neuropathological
Hopwood, J. Rosenfeld, V. and concurrent clinical Melbourne, March 2018. findings in Alzheimer’s
Dore, F. Lamb, R. Williams, assessment in very preterm disease”, SNMMI,
L. Margison, O. Salvado, C. infants: relationship with 78. R
 . Mito, T. Dhollander, D. Philadelphia, June 2018.
Masters, V. Villemagne, C. 12 month motor outcome”, Raffelt, Y. Xia, O. Salvado,
Rowe, “Assessing Aβ & tau Australasian Academy A. Brodtmann, C. Rowe, V. 85. C
 . Rowe, V. Dore, R. Mulligan,
pathology in Vietnam war of Cerebral Palsy and Villemagne, A. Connelly, F. Lamb, P. Bourgeat, T.
veterans with chronic Post- Developmental Medicine, “Investigating microstructural Cummins, O. Salvado, C.
Traumatic Stress Disorder”, Auckland, New Zealand, 21- heterogeneity of white matter Masters, V. Villemagne,
SNMMI, San Diego, June 24 March 2018. hyperintensities in Alzheimer’s “Tau Imaging in Alzheimer’s
2016. disease using single-shell Disease with 18F-MK6240, a
71. N. Hearn, V. Dore, J. Fripp, 3-tissue constrained spherical second generation selective
65. V. Dore, P. Bourgeat, R. M. Grothe, S. Teipel, C. deconvolution”, Joint Annual tau tracer”, SNM, Miami, 22-
De Flores, R. La Joie, Masters, C. Rowe, E. Meeting ISMRM-ESMRMB, 25 June 2018.
A. Fazlollahi, J. Fripp, Coulson, V. Villemagne, Paris, France, 16-21 June,
V. Villemagne, C. Rowe, “Global tau burden 2018. 86. S. Smith, H. Leroux, J.
G. Chetelat, O. Salvado, correlates with basal Carrano, D. Gorse, P. Croll,
“Selective age-association forebrain atrophy in healthy 79. A. O’Brien, D. Bauer, M. Santamaria, L. Schaper,
of hippocampus subfields in aging subjects”, AAIC, “DevOps 2.0: Evidence-based “A thematic analysis of
cognitively healthy elderly”, London, July 20117. evolution of serverless Health Data Sharing
AAIC, London, July 2017. architecture through Barriers and Information
72. K. Lai, L. Wilson, D. Reti, automatic evaluation of Blocking from the HISA
66. V. Dore, S. Bullich, C. D. Bauer, “Improving “infrastructure as code” Survey for the Productivity
Rowe, P. Bourgeat, S. predictability of CPRISPR- deployments”, YOW! Data, Commission”, Health
Konate, A. Stephens, Cpf1 activity”, AGTA 2017 Wesley Conference Centre, Informatics Conference,
J. Fripp, C. Masters, O. Annual Conference, Hobart, Sydney, 14-15 May 2018. Brisbane, QLD, August 6-9,
Salvado, V. Villemagne, 29/10/2017 - 01/11/2017. 2017.
S. De Santi, “Comparison 80. A. Pagnozzi, L. Reid, R.
of 18F-florbetaben 73. H. Leroux, M. Lawley, Boyd, S. Rose, “Measuring
quantification results using S. Gibson, “REDCap neuroplasticity in cerebral
MR-based and MR-Less under FHIR: Enhancing palsy: what cohort sizes are
CapAIBL: Validation against electronic data capture with needed for MR imaging?”,
histopathology”, AAIC, FHIR capability”, Health American Academy
London, July 2017. Informatics Conference, for Cerebral Palsy and
Brisbane, QLD, Australia, Developmental Medicine,
August 6-9 2017. Montreal, Canada, 13-16
September, 2017.
aehrc.com
8 0 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

AEHRC PUBLICATIONS 2017-18 (continued)

87. P. Szul, D. Bauer, O. Luo, A. 93. J. de Frutos-Lucas, B. 4. D


 . Bradford, “Improving 12. J. Li, D. Bradford,
O’Brien, N. Twine, R. Dunne, Brown, T. Porter, L. Milicic, care for men with prostate “Introducing Mobile Apps
“VariantSpark – Apache S. Rainey-Smith, P. Maruff, cancer”, CSIRO Internal for the Management and
Spark for Bioinformatics”, P. Bourgeat, V. Villemagne, Report Number: EP176237, Use of Surgeons’ Preference
Spark Summit Europe 2017, K. Taddei, C. Masters, C. Aug 2017. Cards: A Study of Staff’s
Dublin, Ireland, 25-26 Rowe, D. Ames, R. Martins, Perceptions at Cabrini
October 2017. S. Laws, “Fighting Inactivity 5. D
 . Bradford, K. Harrap, D. Hospital - Malvern”, CSIRO
to Prevent Cognitive Decline: Ireland, D. Hansen, “Digital Internal Report Number:
88. N. Twine, P. Szul, A. The Role of Dopamine in technology to support the EP178155, Oct 2017.
O’Brien, I. Blair, K. Modulating Physical Activity Work It Out Program. A
Williams, D. Bauer, “Using Levels in Older Adults”, 2nd report on the Indigenous 13. D. Silvera, “Virtual Reality
Big Data technologies to HBP Student Conference - Health Portal Project. to Enhance Learning for
uncover genetic causes Transdisciplinary Research Project ID 71075”, CSIRO Secondary Students with
of Amyotrophic lateral Linking Neuroscience, Brain Internal Report Number: Intellectual Disabilities and
sclerosis”, Health Data Medicine and Computer EP176009, Jan 2018. Autism”, CSIRO Internal
Analytics (HISA), Brisbane, Science, Ljubljana, Slovenia, Report Number: EP175862,
11 October 2017. 6. D
 . Bradford, “But I’m Jul 2017.
February 14-16, 2018. Healthy: psychological
89. N. Twine, P. Szul, I. stressors of prostate cancer 14. D. Silvera, “Immersive
Blair, K. Williams, D. Book Chapters and potential strategies for Virtual Reality to Enhance
Bauer, “Understanding support services”, CSIRO Learning for Secondary
neurodegenerative 1. B
 . Celler, M. Varnfield, R. Internal Report Number: Students with Intellectual
disease with BigData Jayasena, “What Have We EP18233, Mar 2018. Disabilities and Autism User
machine learning”, Sydney Learned from the CSIRO Needs Report and Evaluation
Bioinformatics Research National NBN Telehealth 7. D
 . Bradford, D. Ireland, Framework”, CSIRO Internal
Symposium, Sydney, 13 June Trial?”, M.E.L. van den Berg D. Silvera, G. Farr- Report Number: EP178156,
2017. and A.J. Maeder (Eds.), IOS Wharton, Y. Xiao, A. Oct 2017.
Press, 2018, Feb 2018, pp. Cooper, “Teachnology:
90. V. Villemagne, S. Burnham, 1-17. Tech reshaping the face 15. D. Silvera, C. Pocock, D.
P. Bourgeat, V. Dore, of education in the Autism Bradford, S. Brinkmann,
S. Laws, O. Salvado, R. 2. M
 . Karunanithi, Q. Zhang, community. A report of the “CALD Assist - Nursing
Martins, C. Masters, C. “An Innovative Technology Autism Technology Think Impact Analysis Report”,
Rowe, “Refining the natural to Support Independent Tank”, CSIRO Internal Report CSIRO Internal Report
history of global and regional Living: he Smarter Safer Number: EP183925, Jun Number: EP178791, Nov
Aβ-amyloid deposition Homes Platform”, Maayken 2018. 2017.
in sporadic Alzheimer’s E.L. van den Berg, Anthony
disease”, 12th Congress J. Maeder, IOS Press, Mar 8. S
 . Khanna, D. Rolls, J. 16. D. Silvera, “Socially Assistive
of the World Federation of 2018, pp. 102-110. Boyle, R. Jayasena, Y. Xie, “A Robots to Enhance Learning
Nuclear Medicine and Biology, Predictive Risk Model for the for Secondary Students with
Melbourne, Australia, 20-24 Health Care Homes Trial: Intellectual Disabilities and
April 2018.
Reports Development & Validation Autism”, CSIRO Internal
using Private Health Report Number: EP18114,
1. M
 . Barber, D. Bradford,
91. Y. Xia, N. Yassi, P. Bourgeat, Insurance Data”, CSIRO Jan 2018.
C. Robinson, D. Smith, K.
P. Raniga, P. Desmond, J. Internal Report Number:
Baumann, M. St Clair, D. 17. Y. Xia, P. Raniga, P.
Fripp, D. Ames, P. Maruff, EP177689, Oct 2017.
Hansen, “Organisational Bourgeat, S. Rose, J. Fripp,
C. Fowler, R. Martins, V.
factors that have the 9. S
 . Khanna, D. Rolls, J. “Evaluating early cholinergic
Villemagnen, C. Masters,
potential to influence the Boyle, R. Jayasena, Y. Xie, degeneration in preclinical
C. Rowe, O. Salvado,
ability of Aboriginal and “A Predictive Risk Model Alzheimer’s disease:
“Cerebrovascular pathology
Torres Strait Islander for the Health Care Homes Milestone 2 - MRI and PET
in Alzheimer’s disease:
employees to flourish in the Trial”, CSIRO Internal Report imaging and analysis”,
findings from the Australian
workplace”, CSIRO Internal Number: EP178994, Nov CSIRO Internal Report
Imaging, Biomarkers and
Report Number: EP176231, 2017. Number: EP186055, Feb
Lifestyle study of aging”,
Sep 2017. 2018.
12th Congress of the World 10. S. Khanna, D. Rolls,
Federation of Nuclear 2. E
 . Bariola, N. Good, R. D. Ireland, J. Boyle, Y. 18. Y. Xie, S. Khanna, N. Good, D.
Medicine and Biology, Jayasena, “HealthLinks: Xie, “Risk Stratification Rolls, J. Boyle, “Reviewing
Melbourne, Australia, 20-24 Chronic Care System Level Algorithms to Quantify Risk Equitable Access to Care
April, 2018. Evaluation Patient Survey of Rehospitalisation”, CSIRO outside of Hours - Final
Protocol v4”, CSIRO Internal Internal Report Number: Report”, CSIRO Internal
92. Y. Xia, N. Yassi, P. Bourgeat,
Report Number: EP178324, EP179547, Dec 2017. Report Number: EP174110,
P. Raniga, P. Desmond, J.
Aug 2017. Aug 2017.
Fripp, D. Ames, P. Maruff, 11. J. Li, D. Bradford,
C. Fowler, R. Martins, V. 3. D
 . Bradford, “Sunshine “Introducing Digital 19. Y. Xie, J. Boyle, S. Khanna,
Villemagne, C. Masters, Coast TrueNTH Focus Preference Cards for the “Improving the scheduling
C. Rowe, O. Salvado, “MRI Groups: Improving Management of Surgical of surgery in Queensland
based assessment of engagement and care Items: A Study of Nurses’ public hospitals - Interim
cerebrovascular pathology experience of men living Perceptions at Calvary John Report”, CSIRO Internal
in Alzheimer’s disease”, with prostate cancer”, CSIRO James Hospital”, CSIRO Report Number: EP184356,
ISMRM-ESMRMB 2018, Paris, Internal Report Number: Internal Report Number: May 2018.
France, 16 June 2018. EP175572, Aug 2017. EP177694, Sep 2017.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


AEHRC A N N U A L R E P O R T 2 0 17/ 2 0 18 8 1

AEHRC AND E-HEALTH PROGRAM


STAFF, STUDENTS AND VISITORS
Staff Mr Alex Zahnleiter, PhD Student Dr Christian Redd, Postdoctoral Fellow
Dr David Hansen, Chief Executive Officer Ms Ying Xia, Postdoctoral Fellow Dr David Ireland, Postdoctoral Fellow
Ms Marianne Chalk, Program Support Officer/ Dr Pierrick Bourgeat, Research Scientist Dr Hang Ding, Research Scientist
Office Manager Dr Jeremy Beaumont, Postdoctoral Fellow Dr Vasu Rani, Postdoctoral Fellow
Mrs Janet Fox, Business Development Dr Biting Yu, Postgraduate Student Dr Manuel Gonzalez-Garcia, Postdoctoral
Manager Mr Jason Wood, Research Assistant Fellow
Ms Erica Wallace, Visiting Scientist Miss Vanessa Smallbon, Experimental
Health Informatics
Mr Ashley Gillman, Postgraduate Student Scientist
Dr Michael Lawley, Group Leader and
Mr Matthew Chesini, Research Assistant Dr Son Tran, Postdoctoral Fellow
Principal Research Scientist
Dr Susmita Saha, Postdoctoral Fellow Dr Jill Freyne, Team Leader and Senior
Dr Rajiv Jayasena, Team Leader and Senior
Mr Anthony Paproki, Postdoctoral Fellow Research Scientist
Research Consultant
Mr Roger Baker, Visiting Scientist Prof Celler Branko, Adjunct Science Fellow
Mr Derek Ireland, Senior Software Engineer
Mr Caleb Stewart, Postgraduate Student Prof K Yogesan, Research Director, WA
Dr Anthony Nguyen, Team Leader, and Senior
Dr Kerstin Pannek, Research Scientist AEHRC
Research Scientist
Dr Ales Neubert, Postdoctoral Fellow Dr Shaun Frost, Biomedical Scientist
Dr Alejandro Metke, Team Leader, and Senior
Dr Dale Roach, Postgraduate Student Dr Farhad Fatehi, Postdoctoral Fellow
Research Scientist
Ms Angel Kennedy, Visiting Scientist Dr Di Xiao, Research Scientist
Ms Kate Ebrill, Product Manager, Clinical
Mr Alex Pagnozzi, Postdoctoral Fellow Mrs Maryam Mehdizadeh, Senior Research
Terminology
Mr Lee Reid, Postdoctoral Fellow Technician
Dr Hugo Leroux, Research Scientist
Dr Shekhar Chandra, Visiting Scientist Dr Sajib Saha, OCE Postdoctoral Fellow
Ms Donna Truran, Clinical Terminologist
Dr Ibrahima Diouf, Experimental Scientist Mr Janardhan Vignarajan, Software Engineer
Mr Andrew Patterson, Research Scientist
Mr Zi-mim Weng, Industrial Trainee Dr Shuang Yu, Postdoctoral Fellow
Dr Hoa Ngo, Postdoctoral Fellow
Mr Chih-Chia Chen, Postgraduate Student Dr Cirous Dehghani, Postdoctoral Fellow
Ms Madonna Kemp, Clinical Terminologist
Dr Parnesh Raniga, Research Scientist Mr Liam Robinson, Research Assistant
Mr Anton van der Vegt, Postgraduate Student
Dr Marcela Cespedes, Postdoctoral Fellow Dr Cheryl Chen, Experimental Scientist
Mr Harrisen Scells, Postgraduate Student
Mr Fan-Yu Yen, Industrial Trainee Ms Shreya Somasundaran, Industrial Trainee
Dr Hamed Hassanzadeh, Postdoctoral Fellow
Mr Charley Budgeon, Postgraduate Student
Dr Bevan Koopman, Research Scientist Future Science Platform
Dr Amir Fazlollahi, Postdoctoral Fellow
Mr Norm Good, Senior Experimental Scientist Prof Stephen Rose, Future Science Platform
Dr Saba Momeni, Postgraduate Student
Ms Kate Birch, Principal Research Consultant Leader
Dr Michael Vacher, Postdoctoral Fellow
Dr Justin Boyle, Principal Research Scientist
Ms Julie Trinder, Research Assistant PhD Students
Dr Sankalp Khanna, Research Scientist
Ms Sara Hetzel, Postgraduate Student Ms Nazli Bashi, University of Queensland
Dr David Rolls, Research Scientist
Dr Natalie Twine, Postdoctoral Fellow Ms Amy Chan, The University of Queensland
Dr Yang Xie, Postdoctoral Fellow
Mr Arash Bayat, Postdoctoral Fellow
Ms Philippa Niven, Senior Experimental Vacation Scholarship Students
Dr Emily McCann, Postgraduate Student
Scientist David Lawson, University of Queensland
Dr Oscar Luo, Research Scientist
Mr Ming Zhang, Software Engineer Annabelle Cooper, University of Queensland
Mr Aiden O’Brien, Postgraduate Student
Mr Jim Steel, Senior Research Technician Pavin Wu, Monash University
Dr Laurence Wilson, Postdoctoral Fellow
Mrs Deanne Ukovich, Senior Engineer Charlie Kang, University of Queensland
Prof Shoba Ranganathan, Visiting Scientist
Mr Simon Gibson, Senior Software Engineer Ben Hill, University of Queensland
Mr Amnon Bleich, Postgraduate Student
Ms Karen Harrap, Software Engineer Jason Wood, Griffith University
Mr Marc Horlacher, Postgraduate Student
Mr John O’Dwyer, Senior Engineer Nachiketa Mudgil, University of Sydney
Ms Anita Sathyanarayanan, Postgraduate
Mr John Grimes, Principal Research Ian Brettell, Australian National University
Student
Consultant Luke Doyle, RoyaL Melbourne Institute of
Ms Denise Thiel, Postgraduate Student
Mr David Conlan, Senior Software Engineer technology
Health Services Charlie Bradford, University of New South
Biomedical Informatics
Dr Mohan Karunanithi, Group Leader and Wales
Dr Jurgen Fripp, Group Leader and Principal
Principal Research Scientist Daniel Todd, Curtin University
Research Scientist
Mr Simon McBride, Senior Research Kevin Saric, Griffith University
Dr Olivier Salvado, Group Leader and
Consultant Ruqian Lyu, University of Melbourne
Principal Research Scientist
Dr Marlien Varnfield, Team Leader and
Dr Jason Dowling, Team Leader and Visitors
Research Scientist
Research Scientist Dr Mark Braunstein, Visiting Scientist
Ms Anna Kubiak, Research Technicial
Miss Jessica Bugeja, PhD Student
Dr Qing Zhang, Team Leader and Senior Honours Student
Dr Farnaz Akther, Postdoctoral Fellow
Research Scientist Yi Xiao, UNSW Honours Student
Dr James Doecke, Team Leader and Senior
Dr Amir Aminbeidokhti, Postdoctoral Fellow
Research Scientist Support staff
Dr David Silvera Tawil, Research Scientist
Ms Denis Bauer, Team Leader and Senior Finance - Kellie Tighe
Dr Geremy Farr-Wharton, Postdoctoral Fellow
Research Scientist HR - Laurie Mackenzie
Dr Sazzad Hussain, Postdoctoral Fellow
Mrs Sam Burnham, Team Leader and Finance Support - Katie Forrestier
Dr Jane Li, Research Scientist
Statistician Contract Support - Sandy Fairnworth
Dr Dana Bradford, Senior Research Scientist
Dr Christine Cong Guo, Visiting Scientist HSE Support - Shane Casson
Mr Leonardo Iwaya, Postgraduate Student
Dr Vincent Dore, Research Scientist Business Development Support - Dr Peter
Ms Ziggy O’Reilly, Postgraduate Student
Dr Kaikai Shen, Postdoctoral Research Fellow Kambouris
Mr Jamie Exton, Life Scientist

aehrc.com
8 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


(An unincorporated joint venture)

SPECIAL PURPOSE FINANCIAL REPORT


30 JUNE 2018

Detailed financial information from pages 84-91 have been deliberately omitted from this report.

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 8 3

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


DIRECTORS DECLARATION

The directors have determined that the unincorporated joint venture is not a reporting entity and that this special
purpose financial report should be prepared in accordance with the terms of the joint venture agreement and the
accounting policies outlined in Note 1 to the financial statements.

The directors declare that the accompanying Statement of Comprehensive Income, Statement of Financial
Position, Statement of Cash Flows, Statement of Changes in Joint Venture Funds and Notes to the Financial
Statements present fairly the unincorporated joint venture’s financial position as at 30 June 2018 and its
performance for the year ended on that date in accordance with the terms of the joint venture agreement and the
accounting policies described in Note 1 to the financial statements.

This declaration is made in accordance with a resolution of the Board.

Director Director

Brisbane Brisbane
Date: Date:

Director Director

Brisbane Brisbane
Date: Date:

aehrc.com
9 2 A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18

THE AUSTRALIAN E-HEALTH RESEARCH CENTRE


A E H R C A N N U A L R E P O R T 2 0 17/ 2 0 18 9 3

aehrc.com
The Australian e-Health Research Centre
Level 5, UQ Health Sciences Building 901/16
Royal Brisbane and Women's Hospital
Herston QLD 4029 Australia

P +61 7 3253 3600


F +61 7 3253 3690
E enquiries@aehrc.com
aehrc.com

Acknowledgments
Editorial: Dr David Hansen, the Australian e-Health Research Centre
Marianne Chalk, the Australian e-Health Research Centre

You might also like