introduction of bju-bmr-rg and use case study of applying openehr archetypes to cdr implementation
TRANSCRIPT
Prof.XudongLuZhejiangUniversity,P.R.China
2016.01.22
IntroductionofZJU-BMI-RG&UsecasestudyofApplyingopenEHRarchetypestoCDRImplementation
The Biomedical Informatics ResearchGroup is established in 1996. Through20 years’ development, it covers manyareas in medical informatics, includingEHR, Data Integration, CDSS, mHealth,Medical Language Processing, ClinicalData Mining and TranslationalInformatics.
There are totally around 60 staffsincluding 2 professors, 4 associateprofessors , Ph.D/Master students &software engineers.
BiomedicalInformaticsResearchGroupinZJU
Background of the ResearchØ Part of Project of “Medical Data Integration & Merging”, funded by
Chinese National “863” Program, initiated in 2012.
Ø Research Purpose: An methodology of implementing open and
extensible CDR and a case study in a pilot hospital
Shangxi Dayi Hospital with 2600 beds
What is CDR ?Ø Definition of CDR
• A data store that holds and manages clinical data collected from service encounters at point of service locations (e.g. hospitals, clinics) (from ISO/TR 20514,2005)
Ø CDR has been widely developed and implemented internationally.• 46.7% hospitals in Asia Pacific
• 71.2% hospitals in Middle East
• 67.2% hospitals in Canada
• 94.8% hospitals have implemented CDR in America
0.57%hospitalshaveimplementedCDRinChinauntil2014CDR in China
CDR is particularly important for this stage of the development of medical information in China.
Stage Description 2013(N=2414)
2014(N=2622)
Stage 7 Complete electronicmedical recordssystem,regionalhealth information sharing.
0.04% 0
Stage6 Closed-loop management ofthewhole processofmedical data,advancedmedical decisionsupport.
0.16% 0.19%
Stage5 Unified datamanagement, dataintegrationamongdepartment systems.
0.21% 0.38%
Stage4 Information sharing inhospital, intermediatemedical decision support.
3.89% 5.61%
Stage3 Dataexchange amongdepartments, primarymedical decision support.
13.05% 15.25%
Stage2 Dataexchange withinthedepartment. 22.33% 21.78%
Stage1 Preliminary datacollected within thedepartment 11.10% 10.41%
Stage0 Not formedelectronicmedical recordssystem 49.21% 46.38%
Areal-timeunifieddatabaseofpatientclinicalinformation
Clinical Data Repository
Patient ServiceClinical Support Research
LIS CPOEPACSHMIS EIS OIS
Interoperability
Integration Engine
CIS
CDR in the hospital
8
ThecommonwayofimplementingCDR
It’s over-reliance on vendors and time-consuming and cost-consuming for any extension
Requirementsusers
BIGmodel
BIGschema
Concepts&relationships
datastore
communication
information
GUIApp
Software
OOdevtRDBMSdevt
define
ImplementedinHard-codedin
developer
Problems
Source Items
Requirements 348
ExistingCDR* 93
Diabetes-relatedDataElements
CardiacDataElementsSource Items
Requirements 257
ExistingCDR* 101
ThedataintheCDRarealwaysnotenoughtofittherequirement
*TheexistingCDRaretheoneimplementedinEMR-SinonelargehospitalinChina
Acaseformedicalexpertstoconductalong-termfollow-upstudyondiabetes
AcaseforCVDdepartmenttointroduceadecisionsupportsystem
BiotherapyrelatedDataElementsSource Items
Requirements 103
ExistingCDR* 23
Acaseformedicalexpertstoconductclinicalresearchonbiotherapy
Clinicians Engineers
Clinical Data Repository
HIS LIS PACS …RIS EIS OIS
Data Viewer
DataAnalytics
DecisionSupport 。。。
DataMining
Gap
I want to query the count of patient with CIK therapy plan Researchers
I want to find the relationship between diseases and certain factors
Patients
I want to find the number of patients like me
I want to integrate my new system to CDR
Increasing requirement cannot be filled
Developed software cannot be fully used
TheGapbetweenrequirementandReality
The Ideal Solution
An Open Extensible Information Platform
Let the people with data requirement retrieve & query data themselves
I can configure a simple form for my request
I can get the necessary data as input to analytics software
I can query whatever I needed
I can transfer the data to my own structure
Clinical Data Repository
HIS LIS PACS …RIS EIS OIS
Clinicians EngineersResearchers Patients
OpenEHR methodology ?
The openEHR method has the flexibility and scalability, and archetypes account for them.
TheexistingartefactsofopenEHR community
There are many previous defined archetypes, templates openEHR Clinical Knowledge Manager (CKM), together
with several implementations around the world
Ø InformationModelling:Canarchetypesandtemplatesbe
useddirectlyinChina?Whatshouldbeextended?
Ø CDRImplementation:HowtoimplementaopenEHR-based
systemwhichallowtheexpertsdefine,retrieve&querydata
bythemselves?
14
UseCaseStudyApplyingopenEHR archetypestoCDRimplementation inChina.
Archetypemodeling
Starting from data schemas of existing EMR, to find whether the CDR requirements can be modelled using archetypes in CKM
AnalyzeExistingdatabaseschema
Mergingitemswithsamesemantics
Referringexisitingstandards
AbstractClinicalconcepts
FindCorrespondingarchetypein
CKM
Exist Newarchetype
ModificationCoverconcept
completely
Adoptdirectly
Yes
No
Yes
No
1
2
3
4
5 6
6
6
Analyze data schemas
Analyze the two EMR database schema that is used in more than 200 hospitals in China to collect the basic CDR requirements
Dataschema-1 Dataschema-2 CDRrequirements
PAT_MASTER_INDEX MASTER_PATIENT_INDEX Patient demographics(69items)
MEDREC.DIAGNOSIS DIAGNOSIS Diagnosis information(25items)MEDREC.PAT_VISITOUTPADM.CLINIC_MASTERINPADM.PATS_IN_HOSPITAL
PATIENT_VISITVISIT_IN_HOSPITALVISIT_OUT_PATIENT
AdmissionDischargeTransfer (175items)
ORDADM.ORDERSOUTPDOCT.OUTP_ORDERS
ORDERSORDERS_PERFORM
Order information(92items)
ORDADM.VITAL_SIGNS_REC VITAL_SIGNS_RECORD Vital signsinformation(items)17EXAM.EXAM_MASTEREXAM.EXAM_ITEMSEXAM.EXAM_DATAEXAM.EXAM_REPORT
EXAM_REQUESTEXAM_ITEMEXAM_REPORTEXAM_DATA
Examination information(182items)
LAB.LAB_TEST_MASTERLAB.LAB_TEST_ITEMSLAB.LAB_RESULT
LAB_TEST_REQUESTLAB_TEST_DATALAB_TEST_MASTER
Lab testinformation(112items)
OPERATION_SCHEDULEOPERATION_MASTER
OPERATION_REQUESTOPERATION_REPORT
Operationinformation(200 items)
BLDBANK.BLOOD_APPLYBLDBANK.BLOOD_CAPACITY
Transfusion(36items)
NURSERECORD_SUMMARY Nursinginformation(62items)
CONSULT_MASTER Consultinformation(39items)
NEWBORM_REPORT Newborn information(129items)
EMR.EMR_DOCUMENT EMR_DOCUMENTEMR_DOCUMENT_DETAIL
EMRdocument information(88items)
Total 1226items
17
Items merging
Merge the items from the data schemas with the same semantic into 892 CDR data items.
Number of itemsCDRrequirements Dataschema-1 Dataschema-2 CDRitems
Patient demographics 26 43 31Diagnosis information 12 13 15
AdmissionDischargetransfer
119 56 123
Order information 36 56 40MedicationOrder None None 57Prescription None None 42Therapy None None 21Diet None None 22Dispose None None 22Vital signsinformation 7 10 12Examination information 109 73 63Lab testinformation 48 64 58
Operationinformation 79 121 124Transfusion 36 None 32
Nursinginformation 62 None 30Consultinformation None 39 22
Newborn information None 129 133
EMRdocument information 28 60 45
WS445-2014 CDRrequirements
1)medicalrecordsummary patientdemographics,encounters
2)outpatientandemergencymedicalrecord
imagingexamination
3)outpatientandemergencyprescription
medication
4)examinationandlaboratorytestrecord
imagingexamination,laboratorytest
5)generaltherapyandtreatmentrecord medication
6)deliveryrecordoftherapyandtreatment
Therapy
7)nursingoperationrecords Nursing8)nursingvaluationandplan none9)informinginformation none10)homepageofinpatientmedicalrecord
EMR document
11)homepageofinpatientmedicalrecordsummaryofTCM
EMR document
12)admissionrecord encounters13)inpatientprogressnote imagingexamination14)inpatientorder medication15)dischargedbrief encounters16)transferrecord encounters17)medicalinstitutioninformation Admission
EMRdocument
Standardization
Refer two standards by MOH in China in order to get the standardized representation of data items, totally 553 items.
WS363-2011 CDRrequirements
1)identification patientdemographics,encounters,medication,imagingexamination,laboratorytest
2)demographicsandsocialeconomicscharacteristics
patientdemographics
3)healthhistory EMR documents4)healthriskfactor EMR documents,operation5)chiefcomplaintandsymptom Diagnosis, EMRdocuments6)physicalexamination Operation
EMRdocumentsOrders
7)assistantexamination imagingexamination8)laboratoryexamination patientdemographics,
laboratorytest9)diagnosis encounters10)medicalassessment encounters11)medicalplanandintervention encounters,medication12)healthexpenditure Orders,13)healthcareorganization patientdemographics,
encounters,medication,imagingexamination,laboratorytest
14)healthpersonnel Nursing,Therapy15)drugandmaterial medication16)healthmanagement Nursing, EMRdocuments
CDR requirements and WS 445-2014 CDR requirements and WS 363-2011
Concept acquisition
Guided by Information Model of openEHR, based on the clinical practice in China, classified 62 clinical concepts
20
Mappingrules
Mapping concepts to archetypes in CKM based on the mapping rules.
Result offinding Category operation
Exist correspondingarchetype
Covered byarchetypecompletely
Useddirectly
Needtomodifydescription, translation,extendthevaluesets.
Revision
Need tospecialize thearchetype, addmoreconstraint.
Specialisaztion
Need toaddnewitems inthedefinition section andkeepcompatibility.
Extension
Modification thatmakethearchetypeisincompatible withoriginalarchetype.
New version
Nocorrespondingarchetype
New
22
Results
45 new archetypes, 15 modification , 13 existing archetypes used directly.
New archetypes(45)Modification and extended(15) No changed(13)
Discussion 1
Revision, Specialisation and New version are included in openEHR specification, while Extension is omitted.
Extension
Revision
Specialisation
Newversion
MODIFICATION
compatibility
Modify description; Expand attributes, range of value sets, terminology.
Customize an general purpose archetype.
Modify definition part, add new object nodes that no need to narrow than the original.
official
24
Discussion 2
Mismatches exist between metadata-level modelling and data-level modelling which happen in candidate archetypes and the CDR requirements.
Data-level
Metadata-level
25
Discussion 3
Problems of the granularity and relationship representation.
Request
Requestitem
Result
Report
DICOMStudy
Image
1
N
1 1
1
1
1
1
1
1
NN
N
N
Request
Requestitem
Result
Report
DICOMStudy
Image
1
N
1 1
1
1
1
N
N
N
Request
Requestitem
Result
Report
DICOMStudy
Image
1
N
1 1
1
1
1
1
1
1
NN
N
N
Requirements Archetypes in CKM After modification
2archetypes 3 archetypes
Image examination data relationship
Ø The methodology of openEHR could be used in China, but
extension to existing archetypes is necessary,
Ø The modelling results so far are still coarse, need to be re-
thinking, discussion with clinicians and align with CKM.
Ø Translation and a convenient editing tool are necessary if we
want clinician to be involved.
Ø The next step would be further diving into broader and deeper
area in the special biomedical domain like cardiovascular,
health management data, and omics data.
What we learned from archetyping?
ØInformationModelling:Canarchetypesandtemplatesbe
useddirectlyinChina?Whatshouldbeextended?
ØCDRImplementation:HowtoimplementaopenEHR-based
systemwhichallowtheexpertsdefine,retrieve&querydata
bythemselves?
27
UseCaseStudyApplyingopenEHR archetypestoCDRimplementation inChina.
Archetype-basedCDRsystem– mainideas
Archetype/Template
Data persistence
Data application
Data manipulation
Modelofdatastoragegeneratedfromarchetype
Fullfeatureddatamanipulationlanguageonarchetype
Userinterfacegeneratedfromarchetype/TemplateStructureddataqueryandentry
Domainexpertsmanagethearchetypes
Archetype-based flow chart of CDR platform
Start
Archetypeedit
DBDeploy
APIDeploy
APPEdit APPDeploy
Database
Application
API
TemplateArchetype Templateedit
Experts
Based on the established flow, user(s) can acquire database schema, API and application they want by data modeling, while the CDR platform generates them automatically with archetype-driven method.
Archetype-driven data storage
Archetype Model
openEHR
Data Storage
Medical Knowledge
Data Requirement Reference
Model
ArchetypeTemplate
Experts
TRM
Schema
Rules
+
L.Wang,L.Min,R.Wang,X.Lu,andH.Duan,"Archetyperelationalmapping-apracticalopenEHRpersistencesolution,"BMCmedicalinformaticsanddecisionmaking,vol.15,p.1,2015.
Data persistence
xml database
Basic serialization
XML databaseNode+path
1. Performance slower than conventional RDB2. Not suite to answer complex query
Take into consideration that almost all the hospitals in China adopt relational database, the relational database persistence with openEHR approach is necessary.
Hybrid serialization
TRM Data persistence
Mapping archetype model into multiple tables, meanwhile, mapping leaf nodes into field name of relation database table.
(Instruction)PK
(Observation)PK
(Evaluation)PK
(Composition)PK
Template-driven data persistence
Rulesformappingtemplatetoentity
Archetype
OET
TRMConfig
Class
-status_Value
-reservedOrder_Value-……
Class
-memberName-memberName-……
Template Object Mapping(TOM)
JPA entity object
Object Relational Mapping(ORM)
Template relational Mapping
(TRM) RDB
+
+
Performance evaluation Query IV(ms) ARM(ms) Node+Path (ms)
Query1.1 80(+74%) 46 5017
Query1.2 91(+54%) 59 5121
Query1.3 196(+15%) 170 5358
Query2.1 221(+16%) 191 24866
Query2.2 219(+17%) 187 25094
Query2.3 474(+129%) 207 26158
Query3.1 242 270(+12%) 294774
Query3.2 224 299(+33%) 297388
Query3.3 254 411(+62%) 362950
Query4.1 198(+13%) 176 127547
Query4.2 254(+32%) 193 128508
Query4.3 1249(+57%) 797 129901
Query5.1 113 186(+65%) 328181
Query5.2 125 205(+64%) 329097
Query5.3 139 239(+72%) 388727
Query6.1 14596(+5150%) 278 5746
Query6.2 16340(+5293%) 303 6029
Query6.3 16453(+5140%) 314 6984
Ø A comparison study among the conventional relational database, the generated ARM database and the Node + Path database.
Ø Five data-retrieving tests (Query 1.1- Query 5.3)
Ø Two patient-searching tests (Query 6.1 – Query 7.2)
Ø The ARM achieve similar performance as the conventional relational databases.
Ø The Node + Path database requires far more time than the other two databases.
Archetype-driven Data manipulation
Archetype Model
Information source
Data Storage
Medical Knowledge
Data Requirement TRM
rules
ArchetypeTemplate
Experts
+SQL
API
Clause Key wordSELECT SELECT
FROMWHEREORDER BY
INSERT INSERTUPDATE UPDATE
SETWHERE
DELET DELET FROMWHERE
AQLData Storage
TQL Query Engine
Developer
Archetype Query Language (AQL) A complete data manipulate language, including data select, update, delete, and insert functionClause Key word ParameterSELECT SELECT Attribute identify path in archetype
FROM Archetype nameWHERE Attribute identify path in archetype operator (>,
>=, =, <, <=, !=) condition valueORDER BY Attribute identify path in archetype
INSERT INSERT INTO Archetype instances in the format of dADLVALUES Attribute identify path in archetype and assigned value
UPDATE UPDATE Archetype nameSET Attribute identify path in archetype operator (=)
condition valueWHERE Attribute identify path in archetype operator (>,
>=, =, <, <=, !=) condition valueDELET DELET Attribute identify path in archetype
FROM Archetype nameWHERE Attribute identify path in archetype operator (>,
>=, =, <, <=, !=) condition value
SELECT o/data[at0001]/events[at0006]/data[at0003]/items[at0004]/value,
o/data[at0001]/events[at0006]/data[at0003]/items[at0005]/value
FROM OBSERVATION o [openEHR-EHR-OBSERVATION.blood_pressure.v1]
WHERE o/data[at0001]/events[at0006]/data[at0003]/items[at0004]/value >= 140 OR
o/data[at0001]/events[at0006]/data[at0003]/items[at0005]/value >= 90
INSERT INTO
OBSERVATION o [openEHR-EHR-OBSERVATION.blood_pressure.v1]
VALUES o/uid/value = newUID(),
o/data[at0001]/events[at0006]/data[at0003]/items[at0004]/value = 140,
o/data[at0001]/events[at0006]/data[at0003]/items[at0005]/value = 90
Ø SELECT
Ø INSERT
Archetype Query Language examples (AQL)
Ø UPDATE UPDATE OBSERVATION o [openEHR-EHR-OBSERVATION.blood_pressure.v1]
SET o/data[at0001]/events[at0006]/data[at0003]/items[at0004]/value = 140
WHERE o/data[at0001]/events[at0006]/data[at0003]/items[at0005]/value >= 90
Ø DELETE DELETE o/data[at0001]/events[at0006]/data[at0003]/items[at0004]/value,
o/data[at0001]/events[at0006]/data[at0003]/items[at0005]/value
FROM OBSERVATION o [openEHR-EHR-OBSERVATION.blood_pressure.v1]
WHERE o/data[at0001]/events[at0006]/data[at0003]/items[at0004]/value >= 140
OR o/data[at0001]/events[at0006]/data[at0003]/items[at0005]/value >= 90
•AQL grammar
•ANTLR grammar analyzer
•Abstract grammar tree
Grammar analysis
•Archetype
•Variable
•Path
Legality verification •HQL
•Multiple SQLs
Query execution
•XML format dADL
•Gzip compression
Result capsulation
AQL – Execution process
Performance comparison
Queryserial number
Records count
API(ms) AQL(ms)
1 1 5 6
2 1 9 6
3 1 5 6
4 1 5 7
5 1 5 5
6 1 5 5
7 1 5 5
8 1 6 13
9 1 6 5
10 1 5 4
Average 1 5.6 6.2
Queryserial number
Recordscount
API(ms) AQL(ms)
1 209 10 20
2 1209 21 71
3 2847 41 150
4 56 5 8
5 1221 19 72
6 1971 28 106
7 1337 24 74
8 7 5 5
9 279 15 20
10 532 15 33
Average 966.8 18.3 55.9
Retrievingpatientinformationbypatientidentifier Retrievingimageinformationbyexamidentifier
The execution time is similar between AQL query and API query. On account of package for dADL, the AQL average performance is little slower than API.
Archetype Model
Information source
Data Storage
Medical Knowledge
Data Requirement TRM
rules
ArchetypeTemplate
Domain experts
+ UI layout
Clinician experts
Drag
DropData
bindingAttributes
edit
Data entry UI
Archetype-driven data application
archetype/templateeditor
template
archetype/templatedatabase
Applicationdatabase
Create/modify
Domainexpert
user
developer
Adjust/organize
archetype
application
Domain expert can use the GUI generator to define GUI whatever he likes
Methods for application development
archetype/templateeditor
template
archetype/templatedatabase
Applicationdatabase
create
TRM
expert
AQL
“WYSIWYG”editor Generalsoftwareframework
Dataapplicationtemplate
design
archetype
Application development with user-control approach
user
application
In order to achieve user-control application development, this study proposes archetype-driven approach using application template and general software framework.
UIController Datamodule Database
WebApplications
Generalframework
GUIinformation
Databinding
Controlinformation
Applicationtemplate
<body> <li style="position: absolute; height: 32px; width: auto; cursor: pointer; top: 187px; left: 15px;"> <img ng-class="UIData.label.picType" class="CLUSTER">
<span> <b class="ng-binding">饮酒史</b></span></li><li style="position: absolute; height: 32px; width: auto; cursor: pointer; top: 187px; left: 89px;"><select name="Flag_Alcohol drinking history" id="/items[at0005]" template="openEHR-EHR-
COMPOSITION.first_interview.v1" > <option value=""> </option>
<option value="无">无</option>
<option value="有">有</option></select></li><li style="position: absolute; height: 32px; width: auto; cursor: pointer; top: 216px; left: 40px;"> <img ng-class="UIData.label.picType" class="DV_CODED_TEXT">
<span class="ng-binding">种类:</span> <select name="Type_Alcohol drinking history" id="/items[at0001]" template=" openEHR-EHR-
COMPOSITION.first_interview.v1" style="width:80px;" >
…… </select></li><li style="width: auto; position: absolute; height: 32px; cursor: pointer; top: 216px; left: 200px;" ><img class="DV_QUANTITY">
<span class="ng-binding">饮酒量: </span>
…… <button class="btn btn-default" type="button" ng-click="save()">保存</button>
</body><script> function save( scope ){ var aql=”insert into”;
foreach (item in scope){ aql=aql+item.tempate;
……} }
</script>
Dataoperation function
controls
openEHR template
Application template
Applicationtemplatefile
The development flow
Createarchetype/template
Chooseandexpandtemplate
AddcontrolsandadjustpagecontentSavepagesLoadapplication
ChoosedataitemsandgenerateHML
• Archetype/ Template Editor
• Archetype –relational mapping
• AQL-based Data Access Interfaces
• Archetype driven UI configuration
Archetype-based CDR system – architecture
CDR Implementation in Chinese Hospital
HIS LIS PACS …RIS EMR CIS
Archetype template repository
Data access service
Diabetes follow-up Integrated data viewClinical decision support
Research Data Query Quality Data analysis Data mining and analysis
Application based on the CDR
Plan to implement
Have been implemented
CDR outcome
643665
34664534
1355345
14727482
32676 528600 115371 673619 760687
9236476
0
5000000
10000000
15000000
20000000
25000000
30000000
35000000
40000000Data statistics
Examinationrequest Examinationdata Labtestrequest Labtestdata
Operationrequest Diagnosis Healthexamination Patient
Encounter Order
Ø 89 archetypes, 85 tables and 142 APIs.Ø Data volume (2012.8—2015.12)
• Medical data 120G• 673619 patients and 760687 encounters
Integrated viewer application based on CDR
Integrated viewer application allows clinicians to view the demographic, imaging examination, laboratory test and orders based on the CDR platform and services.
Diabetes Follow-up data management application based on archetype
Ø DomainexpertsdraganddroptheattributesfromthetemplatestocontroltheUI.
Ø Thecontroltypeautomaticgeneratedbasedonthearchetypetemplate,ARMandReferencemodel.
Ø ManipulatingtheDatabindingwitharchetypemodelaccordingtheclinicalpractice.
Diabetes Follow-up data management application can be build by the domain experts with the archetype-driven method, including data persistence, data access and UI.
Ø openEHR is a promising methodology to provide a open
and highly extensible solution for building CDR
Ø Although more applications is necessary for verifying the
feasibility and performance of Archetype/Template driven
approach, the use case show its feasibility.
Ø To fully use the advantages of openEHR still need the
support from the communities.
What we learned from implementation?
Ø Not too many groups are working with openEHR, while HL7
things are very popular.
Ø The information interoperability is not good among different
institutions and even among different systems in one
institution. HL7 don’t solve the problem as well as they
promised to.
Ø BIG DATA and data sharing become more and more
important, it’s just the time for promoting openEHR! But the
engagement of clinicians and vendors is still a problem
The situation of openEHR in China now
Ø Open Speech to introduce openEHR concepts in forum of CMEF’16
in April 17-20
Ø Initiate the openEHR Chapter under China Association of Medical
Devices Industry Medical Software?
Planned activities to promote openEHR--make more people know openEHR
Ø Use openEHR methodology in national projects
Ø Enhancing the CDR use case in Shangxi Dayi Hospital.
Ø Possible linkage to Regional Healthcare Data Center in Ningxia
Province
Ø Working closely with openEHR community and learn
from successful localization samples like Japan.
Planned activities to promote openEHR--create successful stories of using openEHR