Contenu connexe Similaire à Injecting Intelligence Into Health Care Similaire à Injecting Intelligence Into Health Care (20) Injecting Intelligence Into Health Care1. ACCENTURE
EXECUTIVE SURVEY
ON AI IN HEALTHCARE
MULTI-COUNTRY
INJECTING
INTELLIGENCE
INTOHEALTH
CARE
Survey analysis presented by
Oxford Economics for Accenture
2. • Accenture commissioned Oxford Economics to do a six-
country survey in Europe and Asia-Pacific, among
180 C-level health executives.
• The goal was to assess current and planned use of AI
in their organisations. They were also asked about
benefits expectation, success factors, practical and
clinical challenges related to AI adoption.
• The survey included executives from each of the
following markets: the UK, Spain, Finland, Norway,
Singapore, and Australia.
• Respondents were either in the roles of CIO, CTO, CDO,
CExO, CEO or CCIO or direct reports.
• Respondents were evenly split across health payer
and health provider organisations, as well as public
and private sectors.
• Field work was conducted in summer 2018.
DEMOGRAPHICS
2Copyright © 2019 Accenture. All rights reserved.
3. 3Copyright © 2019 Accenture. All rights reserved.
AIADOPTION
There is no doubt anymore – AI is finally being embraced
by the healthcare industry. It is an encouragingly high priority
among C-level health executives, as is adoption of AI
technology. At the same time, executives are being
appropriately selective about the types of AI they are
choosing to implement.
Our survey shows that an impressive 72 percent of health
leaders polled say they are either piloting or planning AI
adoption. Perhaps even more impressive: 93 percent of
health execs confirmed they have AI projects on their agenda,
with just seven percent saying they are minimally or not at all
focused on AI.
4. THESTATEOFAIADOPTION
• Although 11% say AI is well-integrated into operations,
many may only be using one or a few applications of the
technology.
• Healthcare providers are more likely to say AI is well
integrated into their organisation’s operations (19%, vs. 2%
of payers); another 20% have implemented AI in some
areas of the organisation, vs. 11% of payers.
• Private organisations are more likely to say AI is well
integrated (16%, vs. 6% of public organisations).
• Numbers are directional, but larger organisations are much
more likely to say AI is well integrated into operations
(57%).
• AI is more likely to be implemented in some areas or across
the business among organisations headquartered in the UK
(37%) or Singapore (33%); organisation in Norway (27%),
Finland (23%), Australia (20%), and Spain (17%) are less
likely to say so.
n=180
Copyright © 2019 Accenture. All rights reserved. 4
Adoption: Which of the following statements best describes the current state of AI adoption in your organisation?
11%
16%
42%
30%
1%
AI is well integrated into our
organisation’s operations
We have implemented AI in some
areas of the organisation
We are piloting our use of AI
We are planning our adoption of AI
We are not yet planning to adopt AI
5. MAJORITYAREINTHEPILOTINGAND
PLANNINGPHASE
Adoption: Which of the following statements best describes the current state of AI adoption in your organisation?
Copyright © 2019 Accenture. All rights reserved. 5
n=180 * 3% of Norway respondents said they are not yet planning to adopt AI
17%
20%
37%
27%
7%
10%
47%
37%
10%
17%
30%
40%
10%
13%
60%
17%
13%
20%
37%
30%
7%
13%
47%
33%
AI is well integrated into our
organisation’s operations
We have implemented AI in some
areas of the organisation
We are piloting our use of AI We are planning our adoption of AI
UK Spain Norway Finland Singapore Australia
6. 6Copyright © 2019 Accenture. All rights reserved.
WISDOMOFTHEMEASUREDAPPROACH
While adoption rates are encouragingly high, health
executives are taking an appropriately measured approach to
AI. The popular narrative around benefits of Artificial
Intelligence in Healthcare has largely been around been
focused on making healthcare more effective: better
decisions, better and smarter decisions for doctors, helping
patients navigate their own healthcare.
The reality is that healthcare organisations are being more
focused internally and on cost. For now, healthcare AI is being
implemented mainly in operational back offices areas,
showing that the real transformative opportunity for Artificial
Intelligence is actually on making healthcare more affordable.
7. EARLYGAINSAREINTHEBACKOFFICE
Copyright © 2019 Accenture. All rights reserved.
Value of AI: To what extent have you realised value in the following areas as a result of your application of AI (i.e., the
specific use of technology that can learn by itself without the need for any explicit
programming)? "Substantial/Transformational value” responses
6%
10%
13%
13%
19%
20%
43%
50%
56%
56%
n=124; those who have piloted or implemented AI
7
Increased cybersecurity
Operational efficiency
Improved analytical capabilities (e.g., patient identification
and personalised treatment and management)
Cost savings (e.g., appointments, financial management)
Lower cost of care (with same or better outcomes)
Expanded patient reach (e.g., serving broader population
through digital tools)
Patient satisfaction
Labour savings/reallocation
Better access to care (regardless of outcomes or cost)
Better clinical outcomes
• Healthcare payers are more likely than providers to say they
have realised benefits from AI in terms of operational efficiency
(63% vs. 50%) and cost (52% vs. 35%). Healthcare providers are
more likely to say AI has improved their analytical capabilities
(56% vs. 43%).
• Larger organisations are more likely to have realised a range of
benefits from AI, in terms of operational efficiency, increased
cybersecurity, analytical capabilities, and cost savings.
• Respondents from organisations headquartered in the UK are
more likely than others to say they have realised value in terms
of operational efficiency (73%, vs. 52% of others), cost savings
(59% vs. 39% vs. 16%), and expanded patient reach.
• Numbers are directional, but respondents who say AI is well
integrated into their organisation’s operations are more likely
to have realised benefits in operational efficiency and cost
savings, increased cybersecurity, improved analytical
capabilities, and expanded patient reach.
8. ROUTINETASKSANDCYBERSECURITYMAIN
AIAPPLICATIONAREAS
Copyright © 2019 Accenture. All rights reserved.
n=180
Handling mundane administrative tasks like admissions,
prescriptions, and regular vital-sign testing
Linking AI to structured healthcare data to help transform
care delivery models
AI-assisted routing of healthcare records to patient-
approved doctors
Using AI-assisted alerts to improve cybersecurity
AI-assisted robotic process automation
Providing support to researchers for studies on new fields in
medicine and disease diagnosis and management
Applying AI to unstructured health data to order it & extract
meaningful market intelligence and guide transformation
Advanced analytics to reduce the cost of managing chronic
patients
An AI-assisted clinical coder (also referred to as an
‘Intelligent Chart Review Assistant’)
Providing AI-assisted diagnosis and treatment
recommendations to clinicians
Audio-visual and biometric technology to monitor and report
incidents such as falls or heart attacks among the elderly or infirm
AI-assisted robotic surgery/surgery support
Using chatbots at the patient interface to analyse patient
needs and route them to appropriate clinicians
In progress: Which of the following applications of AI are already in progress at your organisation?
8
44%
31%
29%
29%
26%
47%
53% 24%
22%
20%
18%
17%
3%
9. 9
PLANSANDBENEFITEXPECTATIONS
Copyright © 2019 Accenture. All rights reserved.
AI investments should increase over the next few years;
40 percent of health executives are quite or highly focused
on increasing the use of AI-assisted applications, another
53 percent are moderately focused. While applications of
artificial intelligence are clustered in the back office for now;
patient-facing applications are expected to follow and are
expect to show substantial benefits. The most transformative
benefit of AI though they expect to lie in the operational
application areas for the next three years.
10. GROWTHOFAIISAWIDESPREADIMPERATIVE
• Healthcare providers are more likely to say they are
highly focused (22%, vs. 3% of payers.)
• Private organisations are more likely to say they are
highly focused on increasing the use of AI (17%, vs. 9%
of public organisations); 47% are quite or highly
focused, vs. 33% of public.
• Data on organisation sise are directional (i.e., the
particular group of respondents is not large enough to
be statistically meaningful), but larger organisations are
much more likely to say increasing the use of AI is their
primary goal (65%).
• Respondents from organisations headquartered in
Australia are least likely to say they are quite or highly
focused on increasing their proportion of AI-assisted
applications (23%, vs. 43% of all others).
Importance: To what extent is your organisation currently focused on increasing its proportion of AI-assisted
applications?
13%
27%
53%
6%
1%
Highly focused, it is a
primary/main goal
Quite focused
Moderately focused
Minimally focused
Not at all focused, it is not on
our list of goals
Copyright © 2019 Accenture. All rights reserved. 10
n=180
11. SOMECOUNTRIESMOREFOCUSEDONAI
THANOTHERS?
Importance: To what extent is your organisation currently focused on increasing its proportion of AI-assisted
applications?
20%
10%
13%
10%
13%
10%
23%
40%
20%
33%
33%
13%
43%
50%
33%
43%
46%
23%
UK Spain Norway Finland Singapore Australia
Highly focused, it is a primary/main goal Quite focused
Copyright © 2019 Accenture. All rights reserved. 11
n=180
12. VALUEEXPECTEDTOSHIFTTOWARDPATIENT
CARE
Dark violet bars: n=124; those who have piloted or implemented AI
Light violet bars: n=180
Value of AI: To what extent have you realised value in the following areas as a result of your application of AI?
To what extent do you expect to realise value from AI today? Over the next three years?
Copyright © 2019 Accenture. All rights reserved. 12
54%
52%
61%
59%
58%
61%
86%
79%
89%
91%
6%
10%
13%
13%
19%
20%
43%
50%
56%
56%
Today In three years
Increased cybersecurity
Operational efficiency
Improved analytical capabilities (e.g., patient identification &
personalised treatment & management)
Cost savings (e.g., appointments, financial management)
Lower cost of care (with same or better outcomes)
Expanded patient reach (e.g., serving broader population through
digital tools)
Labour savings/reallocation
Patient satisfaction
Better access to care (regardless of outcomes or cost)
Better clinical outcomes
13. 13
SUCCESSFACTORS
The early successes for AI among health executives
seem to beg the question: “Why aren’t even more AI
projects underway?” Part of the answer may lie buried
in a second major conclusion of the survey: people
skills matter most to AI success, and insufficient skills
within the health organisation workforce represent
the top barrier to AI adoption and implementation.
Clear guidelines and legislation, access to mature
technology and sufficient funds for investing are a
must, too.
Copyright © 2019 Accenture. All rights reserved.
14. PEOPLEMATTERMOSTTOAISUCCESS
Success factors: Which of the following factors are most important to successful AI project implementation at your
organisation?
1%
1%
2%
1%
3%
2%
4%
26%
19%
23%
21%
1%
1%
7%
2%
9%
8%
24%
23%
27%
1%
1%
1%
3%
4%
7%
10%
7%
19%
18%
25%
Staff buy-in
C-suite leadership
Patient/customer buy-in
Clinician buy-in
Ability to free up clinician time
Clear benefit articulation
Access to successful healthcare case studies
Integration into business as usual/workflows
Budget to hire consultants/acquire technology
Maturity of technology
Clear guidelines and legislation
Sufficient staff training/expertise
Rank 1 Rank 2 Rank 3
n=180
Copyright © 2019 Accenture. All rights reserved. 14
• Healthcare providers are more likely to rank maturity of the technology as important to successful AI project implementation (69%, vs. 54% of payers); payers are more
likely to rank integration into business as usual/workflows (29% vs. 17%).
• Public organisations are more focused on clear guidelines and legislation (69%, vs. 59% of private organisations) and maturity of technology (70% vs. 53%); private
organisations are more focused on the budget to hire consultants or acquire technology (57% vs. 26%).
15. 15
BARRIERS
Copyright © 2019 Accenture. All rights reserved.
In addition to being the most important
success factor, skills (or the absence thereof)
also constitute the biggest obstacle to
effective AI projects in healthcare. Ethical
and data privacy concerns are ranked high
too. Too a much lesser extent it was feared
that distrust from clinicians or distrust from
patients can hinder AI take off in healthcare
organisations.
16. WORKERSKILLSTOPROADBLOCKSTO
ADOPTION
n=180
Obstacles: To what extent are the following factors obstacles to the broad-scale implementation of AI at your
organisation? "Meaningful/Major obstacle" responses – Part 1
63%
44%
40%
37%
29% 29%
26%
An insufficiently
skilled workforce
Ethical concerns Data privacy concerns Insufficiently
developed/ tested
technology
Lack of clear
legislative framework
Lack of strategic
alignment within the
organisation
Lack of focus on
innovation/ new
technologies
Copyright © 2019 Accenture. All rights reserved. 16
17. 18% 17%
16% 15% 15%
11% 11%
8%
Systemic
structural/
organisational
factors (e.g., lack
of available data)
Distrust from
clinicians about
effectiveness of
technology
Overly ambitious
early goals (e.g.,
focusing on
complex tasks
rather than
mundane
functions)
Clinician concerns
over potential job
loss or job changes
The absence of
structured
healthcare
databases
Legal liability/ risk
to health outcomes
Distrust from
patients
A lack of industry
standards
PATIENT,PROFESSIONALCONCERNSNOT
(YET?)AMAJORROADBLOCK
• Payers are more likely to cite ethical concerns (50%, vs. 39% of providers), lack of clear legislative framework
(38% vs. 21%), and lack of strategic alignment within the organisation (37% vs. 21%) as obstacles to AI implementation.
n=180
Obstacles: To what extent are the following factors obstacles to the broad-scale implementation of AI at your
organisation? "Meaningful/Major obstacle" responses – Part 2
Copyright © 2019 Accenture. All rights reserved. 17
18. CONCLUSIONS
18
The survey has shown that executives have a strong
focus on AI, and are already realising concrete value
from implementations mainly in the back office.
Executives absolutely need to focus on increasing AI
skills among staff, and will likely need to turn to
external talent to make rapid progress and exploit
AI’s huge potential.
With this measured approach, health organisations
will position themselves well to lead the market with
respect to AI in healthcare.
AI´s most transformative aspects lie in making
healthcare more affordable and productive, as
opposed to just making healthcare better.