Research Informatics and Cloud Computing

Report
CLOUD COMPUTING FOR
MEDICAL RESEARCH AND
HEALTHCARE
Yu-Chuan (Jack) Li, M.D., Ph.D., FACMI
Graduate Institute of Biomedical Informatics
College of Medical Science and Technology
Taiwan Medical University
Taipei Medical University (TMU)
• Top private medical university in Taiwan
• 6000 students, 620 faculty members, 7 colleges
• Closest to the world’s highest building – Taipei 101
TMU Healthcare Group
• Largest JCI-Accredited teaching hospitals in Taipei
• 3,150 beds
• Over 10,000 Out-patient visit per day
北醫附醫
萬芳
雙和
3
Taipei Medical University
7 Colleges
13 Departments
16 Graduate
Institutes
Students: 6,059
Alumni: 31,214
3 TMU
Hospital
s
Full-time
Instructor 428
Part-time
Instructor 649
Total Faculty
6,102
4
College of Medical Science and Technology - TMU
• Department of Biomedical Informatics
• 80 master and Ph.D. students
• Department of Medical Technology
• 60 master and Ph.D., 300 undergraduate
students
• Department of Cancer Biology and Drug
Discovery
• Ph.D. only
• Department of Neuro-regenerative medicine
• Ph.D. only
Wellness
Cloud
Citizen
Health
Record
Medical
Cloud
Care Cloud
Long-term Care
NIST Definition v.15
• Cloud computing is a model for enabling
convenient, on-demand network access to a
shared pool of configurable computing resources
(e.g., networks, servers, storage, applications,
and services) that can be rapidly provisioned and
released with minimal management effort or
service provider interaction.
Five Characteristics of Cloud
• On-demand self-service
• Broad network access
• Resource pooling
• Rapid elasticity
• Measured Service
“the kind of service that dry-lab biomedical researchers
would always wanted…”
Other Terms related to Cloud
• Service Model
• Cloud Software as a Service (SaaS)
• Cloud Platform as a Service (PaaS)
• Cloud Infrastructure as a Service (IaaS)
• Deployment Model
• Private cloud  TMUH as an example
• Community cloud
• Public cloud
• Hybrid cloud
Private Cloud in TMUH
HIS主機
HIS主機
連接至第一、二大
樓網路主幹交換器
Fiber 10Gb
Fiber 10Gb
WS-SUP720-3BXL
WS-SUP720-3BXL
DISK 0
DISK 0
PORT 2
CONSOLE
RESET
STATUS SYSTEM ACTIVE
EJECT
SUPERVISOR 720 WITH INTEGRATED SWITCH FABRIC/PFC3BXL
LINK
LINK
PORT 2
CONSOLE
PORT 1
DISK 1
EJECT
PWR
STATUS SYSTEM ACTIVE MGMT
10 Gb
10 Gb
PORT 1
DISK 1
EJECT
PWR
MGMT
RESET
EJECT
SUPERVISOR 720 WITH INTEGRATED SWITCH FABRIC/PFC3BXL
LINK
LINK
LINK
LINK
WS-SUP720-3BXL
DISK 0
WS-SUP720-3BXL
DISK 0
STATUS SYSTEM ACTIVE
STATUS SYSTEM ACTIVE
RESET
RESET
EJECT
SUPERVISOR 720 WITH INTEGRATED SWITCH FABRIC/PFC3BXL
EJECT
SUPERVISOR 720 WITH INTEGRATED SWITCH FABRIC/PFC3BXL
PORT 1
DISK 1
EJECT
PWR
MGMT
PORT 1
DISK 1
EJECT
PWR
MGMT
PORT 2
CONSOLE
1
PORT 2
CONSOLE
1
LINK
LINK
LINK
LINK
LINK
LINK
2
2
WS-X6748-SFP
48 PORT GIGABIT ETHERNET SFP
WS-X6748-SFP
48 PORT GIGABIT ETHERNET SFP
3
Core Switch
Engine
STATUS
3
Core Switch
Engine
STATUS
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
WS-X6748-SFP
48 PORT GIGABIT ETHERNET SFP
WS-X6748-SFP
48 PORT GIGABIT ETHERNET SFP
4
STATUS
4
STATUS
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
2
3
4
5
6
48
5
5
FAN
STATUS
FAN
STATUS
6
6
Power Supply 1
Power Supply 1
Power Supply 2
Power Supply 2
Catalyst 6500 SERIES
Catalyst 6500 SERIES
2 Gb 2 Gb 2 Gb 2 Gb
各機櫃之
Switch
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
13
14
15
16
17
18
19
20
21
22
23
24
Catalyst Express 500 Series
11X
13X
12X
14X
1X
11X
13X
23X
2X
12X
14X
24X
1X
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
13
14
15
16
17
18
19
20
21
22
23
24
Catalyst Express 500 Series
11X
13X
12X
14X
1X
11X
13X
23X
2X
12X
14X
24X
1X
23X
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
13
14
15
16
17
18
19
20
21
22
23
24
Catalyst Express 500 Series
11X
13X
12X
14X
1X
11X
13X
23X
2X
12X
14X
24X
1X
SYSTEM
ALERT
POE
16X
1
23X
1
16X
SETUP
2X
16X
1
2X
16X
16X
100 Mb100 Mb
住院 住院
Client Client
1
16X



6
7
8
9
10
11
12
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
13
14
15
16
17
18
19
20
21
22
23
24
Catalyst Express 500 Series
11X
13X
12X
14X
11X
13X
23X
12X
14X
24X
23X
SYSTEM
ALERT
POE
1
16X
2
24X
16X
SETUP
SETUP
Catalyst Express 500 Series
Catalyst Express 500 Series
SYSTEM
ALERT
POE
2
1
16X
16X
100 Mb100 Mb
SYSTEM
ALERT
POE
門診 門診
Client Client
2
1
16X
16X
2
16X
SETUP
100 Mb100 Mb
SETUP
100 Mb 100 Mb
行政 行政 醫療 醫療
Client Client Client Client
Double-loop system
More Security

5
4
2X
16X
SYSTEM
ALERT
POE
2
SETUP
Save 50-70% IT Cost
Power saving 30%
More efficient and effective
for resource applications
4
3
2
24X
SETUP
Catalyst Express 500 Series
SYSTEM
ALERT
POE

3
2
2X
SYSTEM
ALERT
POE
2
24X
Catalyst Express 500 Series
16X
SETUP
Virtual Machines
More Green
2
1
1X
1X
SYSTEM
ALERT
POE
2
24X
1
16X
1
23X
1
2X
2 Gb 2 Gb 2 Gb 2 Gb
Non-Stop
Intensive OLTP-type
Business
model
(Sustainability)
Service
model
Care
Cloud
Wellness Could
Operation
model
Medical
Cloud
(Duplicability &
profitability)
(Feasibility &
Value)
CARE CLOUD
Care Cloud
• Service model yet to be
determined
• Need more evidence  Systems
Clinical Trial  Insurance
• Useful for the aging population
• HIT development – Taiwan
Experience
Telecare, UK
Example: Telecare Helping Manage Falls Risks
West Lothian
Night
timefall
fall- -bed
bedoccupancy
occupancy
Night time
22 min fall response
4hour Scotland average
Carer
Bed Sensor
Day time fall
Lifeline
PNC
Monitoring Centre
15
Telehealth Service Model
Community
Care
Home Care
Institutional
Care
NHII Information
Platform
0800-008-850
Telehealthcare Service Center
(TSC)
LTC management
center
members
Emergency
care center
Telehealthcare Information Platform
(TIP)
16
17
Telecare Service Center
A Randomized control trial for Cloudbased Blood Pressure Monitoring
Wireless Sphygmomanometer
Fully Integrated with CPOE
• 系統整合-導入醫院與醫令系統
WELLNESS CLOUD
Wellness Cloud
• High possibility of Innovation
• e.g. +Social Networking
• “Hey, it’s cool to stay healthy!”
• e.g. +gym, wellness, travel, sports and
food industry
• e.g. + consumer electronics
Nike+ and iPhone, iPod
Mass Gathering Health Monitor
24
• 101系統圖
25
CLOUD COMPUTING
FOR MEDICAL
RESEARCH
27
Clinical Data Repository (CDR)
• Often a backend of
Electronic Health
Record (EHR) for
efficient patientlevel data access
• High potential for
clinical research
From CC, NIH
From CC, NIH
NHIRDB
• NHIRDB (National Health Insurance
Research Database)
• 12 years of de-identified claim database
•
•
•
•
for 23 million people
Cohort DB (Five 1-million people groups
for 13 years)
Disease-specific DB (16 disease groups)
Random sample DB (outpatient 1/500,
inpatient 1/20)
generates >100 research papers a year
Biostatistics Value-add Cloud
What is MEUC
(pronounced as “Milk”)
• MEUC (MEDICAL END-USER COMPUTING),
an system that provide end-users a simple to use
interface to access “Big Data” for biomedical
research in a Cloud Computing framework.
MEUC (Medical End-User Computing System)
The Size of MEUC
Year
2007
2008
2009
2010
2011
Total
OPD
IPD
ER
Drug
Exam
Lab
Procedure
2,810,380
84,415
164,313
8,377,165
878,392
12,185,057
963,163
3,187,302
102,480
185,755
9,635,517
1,150,302
13,445,714
1,032,721
3,666,632
164,355
252,414
11,370,756
1,368,141
11,435,334
1,274,819
3,058,730
73,759
184,890
10,149,577
694,654
9,440,527
1,370,797
3,580,217
80,253
188,372
10,607,887
769,530
7,929,639
1,442,294
16,303,261
505,262
975,744
50,140,902
4,861,019
54,436,271
6,083,794
16M 505K 975K 50M 4.8M 54M
6M
DDQ-Disease and Disease Q value
DDQ-Disease and Disease Q value
• From NHI Database 2000、2001 and 2002
Data.
• Visits Count more than 230(total
population 1/100,000) Disease , It means
exclude rare diseases , we can get 4,000
disease.
• 16,000,000 Q value of each Group of 20
Group.
Translational Bio-Informatics Road Map
Hospitals
URS
(non-radiology
images, signals
and report)
Tier 1 & 0
Tier 2 & 3
LIS
(Lab Information System)
Bio-repository
Bio-repository
CTM
PIS
CBIS
(Pathology Information System)
(Bio-repository
Information
System)
CPOE
(Computer Physician Order Entry)
EMR
(Electronic
Medical Record)
MEUC
(Medical End-User
Computing)
PCS
(Proactive Consent System)
IAS
(Intelligent
Acquisition
System)
Bioinformatics
Platform
EMR Integration
Gateway
PACS
(Radiology
images
and report)
Drug-Discovery
Platform
from Prof. Yu-Chuan (Jack) Li, 2009-12, last revised 2012-07-24
Translational Bio-Informatics Road Map
Translational
Research Cloud
from Prof. Yu-Chuan (Jack) Li, 2009-12, last revised 2012-07-24
NCI
National Cancer Institute
CaBIG Gateway
Therapeutic
Information DB
Exome, SNP, Microarray,
Proteomic 2D page…etc.
Lab & Exam DB
Secondary Cancer
Research Database
Master Patient Index
(Pseudo ID)
Radiology, Pathology(MTA...)
Nuclear Medicine Image
Bio-signal DB
Systems Epidemiology
for Cancer Research
Data Mining Tool Box
Privacy & Security
Firewall
NHIRDB Cancer Data
Tissue Bank Data
Electronic Medical Record
Cancer Registry Data
8 Center of Excellence
for Cancer Research
Cancer Screening Array Data
Researchers
Challenges
• ELSI and privacy issues
• IRB or no IRB or a Joint IRB
• More complex with more partners
• Lack of data standard for EMR data
• caBIG, caDSR (Cancer Data Standards Registry and
Repository) for hosting and managing metadata)
• Scale and complexity of EHR
• High demand of computational resource to maintain
multiparty private computation (shared results
without sharing data)
• Not RCT clinical trials (but much more available)
CLOUD COMPUTING
FOR HEALTH CARE
Medical Cloud
Cloud Computing for Major Hospitals
• Virtualization (IaaS)
• Cut the maintenance cost in half
• Much less server room space and much greener
• More flexibility and portability of services
• Service-Oriented Architecture (SaaS)
• An overhaul of the basic system architecture
• Highly flexible and efficient new architecture
• Fast deployment of new applications
• Web-friendly
Cloud Computing for Healthcare
Future
From Service Provider
Current
Cloud Computing for Personalized Health
Care
• PHR or ePHR (electronic
Personal Health Record) is
at the the core of the Next-Generation Health IT
• PHR =
• Personal part of the EMR from all the providers
• + self-measured bio-signals
• + self-entered health related information like family history, exercise,
food consumption, food allergies, OTC drugs, cigarette
consumption…etc.
• PHR will be the basis of
Personalized Healthcare
EMR vs. PHR
EMR
PHR
BP,
Glucose, ..
etc.
others
Intake
record
Exercise
record
Personalized Medicine
• It is estimated in 2014, a personal
Genome can be sequenced under
$1,000 USD
• 3 billion DNA and 33K genes
 more than 100K proteins
 metabolic pathways
all the functions of body
Some estimated 4GB to store all
the short-reads (before compression)
Personalized Medicine (cont.)
• Need a place to
• store it
• review it
• make sense out of it
by linking to a person’s
health information
• PHR on the Cloud will
be the ideal place
Cloud Computing for Rural Heath Care
• The lack of IT resource of rural health stations and
small hospitals in China (>16,000)
• Pioneered since 2007 by Steve Chan (designer of
Cray-2)
• Deployed into two medium-sized cities in the
western part of China (200 health stations, 800 care-givers)
• A new sustainability model for HIT in
developing countries and resource-poor
areas
Conclusion
• Cloud Computing will change the face of Biomedical
Research Data Service
• Cloud Computing, with privacy-enhanced, could change
the future of HIT delivery in developing countries and
resource-poor areas
• Fits the needs of many healthcare sectors due to flexibility
and cost-effectiveness
• Cloud Computing will be a “liberator” for
scalability/accessibility limitations
Thank you for your attention

similar documents