Transcript HL7 2009

HL7 Roadshow 2009
Connected Healthcare
Jon Payne, March 2009
InterSystems
Software for Connected Healthcare
HealthShare – for developing secure
Regional /
electronic health record systems on a
regional and national level
National EHR
TrakCare – an advanced WebInstitutional /
Community EPR
based healthcare information system
that rapidly delivers patient-centric
data and Electronic Patient Record
benefits
Ensemble - #1 interface
Application Level
engine in healthcare (KLAS
2006, 2007, 2008)
Caché - #1 database used
in clinical applications
worldwide
Caché
SQL
Web 2.0
HighODBC JDBC
performance
links to all major
XML
SOAP
COM
EJB
Java
.Net
C++
Perl
InterSystems Caché
Industry-standard
relational access
Caché
Objects
Single Definition
Multi-Dimensional
Storage Manager
Kaiser Permanente and
EPIC Systems
World’s largest civilian
EHR
• 8.7 Million patients
supported in Outpatients
setting
• 2 Million active Personal
Health Record users
• 10 hospitals live with
Inpatients
• >11,000 Physicians
The NICTIZ Project
• « NICTIZ » is an organization of the Dutch
government, in charge of ICT-enabling
the healthcare sector at a national level
(« AORTA » program)
– See www.nictiz.nl (mostly in Dutch, some English)
• In 2005, NICTIZ publishes a tender for building a national
infrastructure for medial information exchange: the « LSP »
(« Landelijk SchakelPunt » or « National switchboard »)
– CSC and InterSystems bid together and are selected
by the end of 2005
General Principles of the Project
• The general goal is patient data interchange
– Complete data, except (for now) imaging, and limited
only by privacy laws
– Includes medical and administrative data
• Patient data are not stored at a central location!
– The central system records the localization of the
information, not the information itself
– It is hence a « meta-index », not a « repository »
• Users of the system: GPs, hospitals, clinics, pharmacists
nationwide
Technological Orientations
(Prerequisites)
• All exchanges are based on HL7 version 3
– « New » version of the HL7 standard, based on XML
– Standard HL7v3 is not finalized: « Dutch variant »
• Communications use SOAP technology on a HTTP/S
communications channel
– Should become one of the largest SOA infrastructures in
the country!
– Security is based on SSL/TLS and X509 certificates
(established industrial standards)
• The LSP was developed 100% on Ensemble
Mechanisms of LSP
GBZ (Goed Beheerd
ZorgSysteem): only certified
clients are allowed to
communicate directly with
the ZIM
In general, a “GBZ” provides
services for several
institutions and/or GPs
The communications
infrastructure between the
final client and the GBZ can
be anything
GBZ
DCN
GBZ
DCN
GBZ
DCN (Data Communication
Network): all
communications transit
through certified networks
LSP
The central “hub” of the LSP
is named the « ZIM » (Zorg
Informatie Makelaar)
DCN
Communication with the LSP
uses HL7v3 and SOAP on
HTTP/S exclusively
GBZ
A hospital , a clinic, even a
GP can be their own « GBZ »
Mechanisms of LSP
GBZ
2. The lab report is transmitted to
the LSP; among other things, it
indicates an allergy to penicillin
GBZ
LSP
3. The ZIM records that a lab
report exists for Mrs Brantjes
at Reinaert clinic in
Maastricht – but it doesn’t
record the lab report itself!
GBZ
GBZ
1. Mrs Brantjes visits
Reinaert clinic in Maastricht
for some medical exams
Mechanisms of LSP
1. Mrs Brantjes, while
visiting Amsterdam, sees a
GP for a rash on the leg that
worries her
GBZ
2. As the GP doesn’t know
Mrs Brantjes, he asks the
LSP for her medical file
5. Naturally, other GBZ can
be involved in answering the
request
6. The ZIM aggregates and
sends the file back to the
requester.
The GP notices the allergy to
penicillin and adapts the
treatment accordingly
GBZ
LSP
3. The ZIM checks its index
for a list of GBZ that can
provide information relevant
to the request
GBZ
GBZ
4. The ZIM gets the lab
results from Reinaert clinic
as well as all relevant
information from sources
connected to the same GBZ
as Reinaert clinic
Volumes and phasing
• Volumes expected at full deployment:
– 20 000 direct clients for >16 million patients
– 4,312 billion XML messages (30KB each) yearly, that is:
• ~20TB message volume, 4 TB logs
• 200 msgs/second average, 671 msgs/second in peak
– Requirement maximum response time: 5 seconds
• Currently:
–
–
–
–
40 msgs/second average, 130 msgs/second in peak
1000 concurrent sessions
Response times under 3 seconds
Availability 99.982%
Volumes and phasing
• 3 planned phases:
– Phase 1: 5% of final volume, 5Ko average message size,
no redundancy
– Phase 2: the same, with 20% of the final volume
– Phase 3: complete deployment with very high availability
and 30 KB average HL7v3 messages
• The system was recently upgraded to phase « 2++ » level:
– Phase 2
– + very high availability
– + 30KB average HL7v3 messages
Brazil – Federal region of Brazilia
LRC
LRC
CS
CS
CS
CS
CS
CS
Hosp.
Apoio
CS
CS
Regional
HRC
Hospital
CS
CS
CS
CS
CS
CS
LACEN
CS
PSR
PSR CS
CS
CS
PSU
PSU
CS
CS
PSU
PSU
UM
CS
CS
CS
CS
NISa
CS
CS
CS
CS
CS
CS
3
Regional
Hospital
CS
CS
LRGu
LRGu
CS
CS
CS
CS
Regional
Hospital
CS
CS
CS
CS
NIRF
CS
CS
Regional
Hospital
PSU
PSU
PSR
PSR
1
Regional
Hospital
PSU
CS
CS
CS
CS
1
NIG
PSR
PSR PSR
1 PSR
CS
CS
2
1
NINB
ISM
NICd
CS
CS
USP
2
NIP
PSR
PSR
1
NISM
CS
CS
1
PSR
PSR
4
1
1
PSR
PSR
1
1
PSR
PSR
Regional
Hospital
NIBsB
Regional
Hospital
2
PSR
PSR
HUB
CS
CS PSR
1PSU
1
CS
CS
PSR
PSR
NILS
CS
CS
CS
CS
CS
CS
CS
CS
Regional CS
CS
Hospital
CS
CS
CS
CS
HSVP
Regional Project:
• 17 Hospitals
• 134 Health Clinics
• 3 Diagnostic
Laboratories
Central
• 20 Emergency Units
Hospital
• 3 Specialist Diagnostic
Units
• Home Care
• More than 5000 health
professionals using just
one Electronic Health
Record
3
PSR
PSR
PSR
PSR
1
Regional
Hospital
NIGu
1
Regional
Hospital
PSR
PSR
NILN
USP
PSR
PSR
PSU
PSU
CS
CS
PSR
PSR
3
CS
CS COMPP
CS
CS
CBMDF
NITN
CS
CS
NICr
CRT
CRT
CS
CS
PSR
PSR
Regional
Hospital
CS
CS
CS
CS
HFA
CS
CS
CS
CS
Regional
Hospital
Hemocentro
1
NIC
NITS
CS
CS
CS
CS
CS
CS
NIS
FEPECS
CS
CS
CS
CS
CS
CS
CS
CS PSR
PSR
CS
CS
CS
CS
CS
CS
CS
C
CS
S
DISAT
UM
1
CS
CS
1
PSR
PSR1
PSR
PSR1
NIBz
CS
CS
1
1
NIPr
Regional
Hospital
CS
CS
CS
CS
Regional
Hospital
CS
CS
USP
Regional
Hospital
USP
UM
CS
CS
SARAH
2 PSU
PSU
Regional PSU
PSU
2
Hospital
2
PSU
PSU
PSR
PSR
1
1
2
2
Regional
Hospital
PSR
PSR
1
PSR
2
USP
2
10
CS
CS
PSU
PSU
1
Brazila – Citizens portal
USA – Rhode Island
• Having a useful, usable and used
interconnected health information system
that:
 Puts the right information into the
hands of clinicians and their patients
when and where it is needed
 Allows for population health
information to be analyzed and used
for surveillance, health policy, and
research
• Using InterSystems HealthShare for
state-wide master patient index and
clinical data exchange
• Connecting 14 healthcare organizations
across the state (85%+ of all lab and
medication data)
• A total of 1.1Million Residents 3,000
GP’s
• Governor has an “Anytime, Anywhere
Healthcare” agenda
USA – LIPIX
• Long Island Patient Information Exchange
• 5.2Million Residents 8,000 GP’s (Assoc with NS-LIJ Health
System)
• Using InterSystems HealthShare for master patient index and clinical
data exchange.
• Connecting 15 healthcare organizations
on New York’s Long Island
• Initial LIPIX Deployment
 4 Hospitals
 2 Physician Practices
 Home Care Agency
 Municipal Health Center
 EMS Organization
 Reference Laboratory
 Long Term Care Facility
 Public Health Department
Sweden National Patient Overview
• Population 9 million
• 21 healthcare regions
• National EHR
• Carelink is Government
appointed group to manage the
project.
• InterSystems in partnership
with Tieto Enator.
Sweden National Patient Overview
(NPÖ)
Users
Users
County
Integration
Users
County
Integration
NPO
Users
Users
EN13606
SOAP/SSL
County
Integration
National Security and Integration Framework
Existing
Care
Systems
Central, Federated, Virtual?
View
er
RA
D
Point of Care
LA
B
GP 1
Identity
Consent
GP 2
Hospital
PA
S
Access
GP 3
OP
R
A&
E
Clinic
IC
U
Repository
Clinical
Repository
End-to-End Performance, Scalability
and Flexibility
– Database
• Data source systems / EPRs
• Intermediate Data cache
• EHR repositories
– Applications
• Authentication, Identification & Consent
• Aggregation, Access & Presentation
• Read-only “view” evolving to “transactional” application
– Messaging & Integration
•
•
•
•
•
•
•
HL7 / XML
Security / Encryption HTTPS, SSL
Data capture / Adapters
Orchestration / Workflow
Transformation / Translation / Rules
Audit & Governance / Message Persistence
Resilience / Availability / Recovery
Patient Identity Management
• Identity Policy and Service most commonly centralised
• Use of (existing) national numbers (Social Security Number /
NHS number)
• Identity Hub as clinical record locator
• Scale – large numbers of patients, records, indexes, pointers to
clinical data, network transactions, queries.
• Interaction with User Access and Consent services
• End-to-end performance and scalability
– Response times
– Availability / Manageability
User Provisioning & Consent
– User Access and Consent policy and service is typically centralised
– Access/Consent policy not typically defined / controlled by those who
need to use the systems
– Clinicians may perform different roles at different times in different
facilities.
– Time-based nature of “legitimate user relationship” & “consent”
– Integration of Smartcards and biometrics
– Requirement for tamper-proof audit and governance databases /
message archives
– Policy may drive an unusable architecture
Data Aggregation / Normalisation
• Variable quality of data source systems
– Data Capture – push/pull
– Structured & Unstructured data
– Coded / Non-Coded data – differing coding regimes & purposes
• Numerous (converging) Standards for Normalisation
– CCR, CDA, CCD
– Consume, Expose, Exchange, Act?
• Underlying technology
– Capture, Parse, Persist, Manipulate, Transport, Query
– Scalability, Performance
– Reliability, Manageability, Recoverability
IHE and HealthShare
Cross Community Information Exchange
A variety of IHE profiles certified at Connectathons
XCA Initiating Gateway
XCA Responding Gateway
XDSa Repository
PIXv3 Consumer
PIXv3 Manager (Vienna)
PDQv3 Consumer
PDQv3 Supplier (Vienna)
ATNA Secure application
ATNA Repository
The Trust Integration Engine
CSC selects InterSystems for trust integration
10 Sep 2008
Computer Sciences Corporation is to use InterSystems
Ensemble for the integration of all new applications it
provides to existing trust systems.
CSC, the local service provider for the North East and
Midlands regions in England’s NHS National Programme
for IT, has selected Ensemble as the standard trust
integration engine (TIE).
CSC Data Message Architecture
TIE
TIE
TIE
TIE
TIE
TIE
TIE
Spine
Messages
Data
Message
Hub
Central
Trust DB
CSC Data Message Architecture
Ensemble
Ensemble
Ensemble
Ensemble
Ensemble
Ensemble
Ensemble
TIE
TIE
TIE
TIE
TIE
TIE
TIE
Spine
Messages
Data
Message
Ensemble
Hub
Central
Trust DB
Ensemble TIE
Ensemble now available from CSC
as the standard Trust Integration Engine (TIE)
•Centrally-defined, standard solution for existing systems integration.
•Available to all NHS Trusts scheduled to deploy CSC NPfIT solutions
•Ensemble TIE Licence and annual maintenance is free to Trusts
•Enables connectivity between Trust-side applications and the CSC NPfIT
solutions
•Library of standard departmental systems interfaces
•Trusts can upgrade to a full Ensemble licence
•Trust Integration Forum (TIF) comprises Trusts, CfH, CSC and ISC
Plymouth Hospitals NHS Trust
Ensemble
Questions?
Questions??
• [email protected]
• www.intersystems.co.uk/TIE