Transcript HCLS$$CSHALS2009$$Tutorial$Kashyap

Clinical Observations Interoperability (COI):
How can Semantic Web Technologies Help?
Vipul Kashyap
[email protected]
http://esw.w3.org/topic/HCLS/ClinicalObservationsInteroperability
CSHALS 2008
February 25, 2009
Cambridge, MA
Acknowledgments: Helen Chen, Eric P and Holger Stenzhorn for COI Demo!
Parsa Mirhaji for providing the real world clinical data!
Outline
•
W3C Task Force on Clinical Observations Interoperability
•
Healthcare and Life Sciences (HCLS): A Taxonomy
•
HCLS Ecosystem: Current and Goal State
•
Use Cases and Functional Requirements
•
Use Case Demo Step Through
•
Advantages of Semantic Web Technologies
•
Next Steps
W3C Task Force on Clinical Observations
Interoperability
•
Goals and Objectives
— Establish a collaboration between Providers, Pharma and other HCLS
stakeholders for re-use of EMR data in Clinical Research
— Establish the key stakeholders and respective value proposition
— Create consensus on a common use case, needs statements and functional
requirements
— Develop Proofs of Concept by implementing key use cases
•
Participants
— Healthcare Providers
•
Partners, Cleveland Clinic, Intermountain Healthcare, Mayo Clinic,
VA/Regenstrief
— Pharmaceutical Companies
• Eli Lilly, Astra Zeneca, Novartis, Pfizer, Bristol Myers Squibb
— Consortia
• W3C, CDISC, HL7
What is Translational Medicine (TM)?
Biomedical
Research
Outcomes and Utilization
Research
Biological
Translational
Research
Risk and Cost
Assessment
Clinical
Clinical
Research
Clinical
Practice
Research
Practice
Personalized
Medicine
HCLS Ecosystem: Current State
Characterized by silos with uncoordinated
supply chains leading to inefficiencies in the system
Patients
National Institutes
Of Health
Patients,
Public
FDA
Pharmaceutical
Companies
Hospitals
Center for
Disease
Control
Payors
Universities,
Academic Medical
Centers (AMCs)
Biomedical Research
Clinical Practice
Clinical Research
Organizations (CROs)
Hospitals Doctors
Patients
Clinical Trials/Research
Patients
Clinical Practice
Some interesting developments …
•
•
•
Payors are performing analyses to enable
— Employers to better identify health issues and optimize offerings
— Employees/members to make better medical decisions
— For cost/utilization optimization and claim adjudication.
Providers are performing clinical studies and reviews:
— To evaluate the quality and consistency of clinical care
— To perform clinical research and evaluate clinical protocols
Pharmaceuticals are performing:
— Clinical Trials
— Evaluating secondary uses of healthcare data, e.g., use of EMRs
for clinical research
HCLS Ecosystem: Goal State
NIH
(Research)
Universities,
AMCs
FDA
CDC
Pharmaceutical
Companies
Patients, Public
CROs
Hospitals Doctors
Payors
Clinical Observations Interoperability will be a
Critical Enabler to realize this Vision!
From FDA, CDC
Functional Requirements
• X identifies the Use Cases, Systems and Functional Requirement under consideration of the
COI Task Force
• Based on the Functional Requirements Specification developed by EHRVA/HIMSS
Need for a bi-directional EMR – CTMS Link:
Shareable Open Source Models of Clinical Data
Healthcare Provider 1
Healthcare Provider 2
Clinical
…
Healthcare Provider N
…..
Clinical Trial 1
Clinical Trial 2
Clinical
Observations
…
Observations
Open
Source
Clinical
Models
- DCM
- SDTM
- BRIDG
- Snomed
- MedDRA
- NCIT
Clinical Trial M
Use Case: Patient Screening
Research Coordinator
selects protocol for patient
screening:
-
Clinical Research Protocol
Eligibility Criteria:
- Inclusion
- Exclusion
EMR DATA
Meds
Research
Coordinator
views list of
patients and
selects which
ones to approach
in person for
evaluation and
recruitment.
Clinical
Evaluation and
Recruitment
Diagnoses
Procedures
Demographics
Patient MR #
Potentially
Eligible for
Protocol
# Criteria
Met / Total
Criteria in
Protocol
Criteria #1
(Pass/Fail/
Researcher
Needs to
Evaluate)
No Criteria #2
(Pass/Fail/
Researcher
Needs to
Evaluate)
Criteria #3
(Pass/Fail/
Researcher
Needs to
Evaluate)
…
0011111
Yes
6/8 criteria
met
Pass
Pass
Pass
…
0022222
No
3/8 criteria
met
Pass
Fail
Pass
…
0033333
Yes
5/8 criteria
met
Pass
Pass
Fail
…
…
…
…
…
…
…
…
* Thanks to Rachel Richesson
COI Demo – Clinical Trial Eligibility Criteria
Use Case Step-Through
1.
2.
3.
4.
5.
(Textual) specification of the eligibility criteria for a given clinical trial
Ontology-based translation of the eligibility criteria into SPARQL queries
Translation of the SPARQL queries into database-specific queries
Execution of the queries at the databases –
results contain all eligible patients
Return of a list of eligible patients to clinical trial administrator
COI Demo – Selecting Inclusion Criteria
Inclusion in SDTM
based ontology
SDTM based
clinical trial
ontology
COI Demo – Drug Ontology Inference
Drug ontology
Exclusion in Drug
ontology
COI Demo – Selecting Mapping Rules
#check all drugs that "may_treat obese"
{?A rdfs:subClassOf ?B; rdfs:label ?D.
?B a owl:Restriction;
owl:onProperty :may_treat;
owl:someValuesFrom :C0028754}
=>
{?D a :WeightLoseDrug}.
Medication
:M0271 a sdtm:Medication;
spl:classCode 6809 ; #metformin
sdtm:subject :P0006;
sdtm:dosePerAdministration [
sdtm:hasValue 500;
sdtm:hasUnit "mg„ ];
sdtm:startDateTime
"20070101T00:00:00"^^xsd:dateTime ;
sdtm:endDateTime
"2008-0101T00:00:00"^^xsd:dateTime .
Criteria in SPARQL
?medication1 sdtm:subject ?patient ;
spl:activeIngredient ?ingredient1 .
metformin
?ingredient1 spl:classCode 6809 .
OPTIONAL {
?medication2 sdtm:subject ?patient ;
spl:activeIngredient ?ingredient2 .
?ingredient2 spl:classCode 11289 .
} FILTER (!BOUND(?medication2))
anticoagulant
Exclusion Criteria
SDTM to HL7 Transformation
Clinical Trial Ontology
sdtm:Medication
{
sdtm:dosePerAdministration
hl7:SubstanceAdministration
hl7:doseQuantity
?x a sdtm:Medication ;
sdtm:dosePerAdministration ?y
} => {
?x hl7:SubstanceAdministration ;
hl7:doseQuantity ?y
}
Clinical Practice Ontology
HL7 to EMR Database Transformation
SPARQL in Clinical Practice Ontology
hl7:SubstanceAdministration
hl7:doseQuantity
Item_Medication:EntryName
?takes .
Medication:ItemID
?indicItem;
{
hl7:substanceAdministration
[
a
hl7:SubstanceAdministration ;
hl7:consumable [
hl7:displayName
?takes ;
spl:activeIngredient [
spl:classCode ?ingred
]
] ;} => {
{
?indicItem
Item_Medication:PatientID
?person;
Item_Medication:PerformedDTTM
?indicDate ;
Item_Medication:EntryName
?takes .
.}
SQL to EMR Database
Pushing Query to Database
•
•
SPARQL in SDTM ontology to SPARQL in HL7 ontology
SPARQL in HL7 ontology to SQL in EMR database
List of eligible patients
SPARQL
EMR
HL7 DCM/RIM
CT Eligibility
SPARQL
SQL
SPARQL in SDTM
PREFIX sdtm: <http://www.sdtm.org/vocabulary#>
PREFIX spl: <http://www.hl7.org/v3ballot/xml/infrastructure/vocabulary/vocabulary#>
SELECT ?patient ?dob ?sex ?takes ?indicDate?contra
WHERE {
?patient a sdtm:Patient ;
sdtm:middleName ?middleName ;
sdtm:dateTimeOfBirth ?dob ;
sdtm:sex ?sex .
[
sdtm:subject ?patient ;
sdtm:standardizedMedicationName ?takes ;
spl:activeIngredient [ spl:classCode ?code ] ;
sdtm:startDateTimeOfMedication ?indicDate
].
OPTIONAL {
[
sdtm:subject ?patient ;
sdtm:standardizedMedicationName ?contra ;
spl:activeIngredient [ spl:classCode 11289 ] ;
sdtm:effectiveTime [
sdtm:startDateTimeOfMedication ?contraDate
].
} FILTER (!BOUND(?contra) && ?code = 6809)
}
SDTM-HL7 Mapping Rules
CONSTRUCT {
?patient
a
sdtm:Patient ;
sdtm:middleName
?middleName ;
sdtm:dateTimeOfBirth ?dob ;
sdtm:sex
?sex .
[
a
sdtm:ConcomitantMedication ;
sdtm:subject
?patient ;
sdtm:standardizedMedicationName
?takes ;
spl:activeIngredient [ spl:classCode
?ingred ] ;
sdtm:startDateTimeOfMedication ?start
] .} WHERE {
?patient
a
hl7:Person ;
hl7:entityName
?middleName ;
hl7:livingSubjectBirthTime
?dob ;
hl7:administrativeGenderCodePrintName ?sex ;
hl7:substanceAdministration
[
a
hl7:SubstanceAdministration ;
hl7:consumable [
hl7:displayName
?takes ;
spl:activeIngredient [ spl:classCode
];
hl7:effectiveTime [ hl7:start ?start ]
].
}
?ingred ]
SPARQL in HL7 Via SWtranformer
PREFIX hl7: <http://www.hl7.org/v3ballot/xml/infrastructure/vocabulary/vocabulary#>
SELECT ?patient ?dob ?sex ?takes ?indicDate
WHERE
{
?patient
hl7:entityName
?middleName .
?patient
hl7:livingSubjectBirthTime
?dob .
?patient
hl7:administrativeGenderCodePrintName ?sex .
?patient
a
hl7:Person .
?patient
hl7:substanceAdministration ?b0035D918_gen0 .
?b0035D918_gen0 hl7:consumable
?b0035C798_gen1 .
?b0035D918_gen0 a
hl7:SubstanceAdministration> .
?b0035D918_gen0 hl7:effectiveTime ?b0035C5E8_gen3 .
?b0035C798_gen1 hl7:displayName
?takes .
?b0035C798_gen1 hl7:activeIngredient ?b0035C848_gen2 .
?b0035C848_gen2 hl7:classCode
?code .
?b0035C5E8_gen3 hl7:start
?indicDate .
FILTER ( ?code = 6809 )
}
HL – Database Mapping Rules: Tables
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX Person: <http://hospital.example/DB/Person#>
PREFIX Sex_DE: <http://hospital.example/DB/Sex_DE#>
PREFIX Item_Medication: <http://hospital.example/DB/Item_Medication#>
PREFIX Medication: <http://hospital.example/DB/Medication#>
PREFIX Medication_DE: <http://hospital.example/DB/Medication_DE#>
PREFIX NDCcodes: <http://hospital.example/DB/NDCcodes#>
HL – Database Mapping Rules: Schema
CONSTRUCT {
?person
a
hl7:Person ;
hl7:entityName
?middleName ;
hl7:livingSubjectBirthTime
?dob ;
hl7:administrativeGenderCodePrintName ?sex ;
hl7:substanceAdministration
[ a
hl7:SubstanceAdministration ;
hl7:consumable
[
hl7:displayName
?takes ;
spl:activeIngredient [ spl:classCode ?ingred]
];
hl7:effectiveTime [ hl7:start ?indicDate ]
].
} WHERE {
?person
Person:MiddleName
?middleName ;
Person:DateOfBirth
?dob ;
Person:SexDE
?sexEntry .
OPTIONAL { ?indicItem Item_Medication:PatientID
?person ;
Item_Medication:PerformedDTTM ?indicDate ;
Item_Medication:EntryName
?takes .
?indicMed
Medication:ItemID
?indicItem ;
Medication:DaysToTake
?indicDuration ;
Medication:MedDictDE
?indicDE .
?indicDE Medication_DE:NDC
?indicNDC .
}
}
Drug Class Information in CT #8
•
•
•
•
•
•
•
•
monotherapy with metformin, insulin secretagogue, or alpha-glucosidase inhibitors
and a low dose combination of all
Long term insulin therapy
Therapy with rosiglitazone (Avandia) or pioglitazone (Actos), or extendin-4
(Byetta), alone or in combination
corticosteroids
weightloss drugs e.g., Xenical (orlistat), Meridia (sibutramine), Acutrim
(phenylpropanol-amine), or similar medications
nonsteroidal anti-inflammatory drugs
Use of warfarin (Coumadin), clopidogrel (Plavix) or other anticoagulants
Use of probenecid (Benemid, Probalan), sulfinpyrazone (Anturane) or other
uricosuric agents
Prescription Information in Patient Database
•
"132139","131933","98630 ","GlipiZIDE-Metformin HCl 2.5-250 MG Tablet","54868079500
",98630,"2.5-250
","TABS","","MG
"," ","15","GlipiZIDE-Metformin HCl
","","GlipiZIDEMetformin HCl 2.5-250 MG Tablet“
•
"132152","131946","98629 ","GlipiZIDE-Metformin HCl 2.5-500 MG Tablet","54868518802
",98629,"2.5-500
","TABS","","MG
"," ","15","GlipiZIDE-Metformin HCl
","","GlipiZIDEMetformin HCl 2.5-500 MG Tablet“
•
"132407","132201","98628
500
","TABS","","MG
5-500 MG Tablet“
•
"132642","132436","C98630 ","GlipiZIDE-Metformin HCl TABS","54868079500 ",98630,"","TABS","","
"," ","15","GlipiZIDE-Metformin HCl
","","GlipiZIDE-Metformin HCl TABS"
","GlipiZIDE-Metformin HCl 5-500 MG Tablet","54868546702 ",98628,"5"," ","15","GlipiZIDE-Metformin HCl
","","GlipiZIDE-Metformin HCl
NDC Code
Drug Ontology By Stanford
from drug ontology documentation
Mapping Between CT and Patient Record
Drug Ontology
CT
MechanismOfAction
C1299007
metformin,
insulin secretagogue
GeneralDrugType
nonsteroidal anti-inflammatory
alpha-glucosidase inhibitors
C0050393
C0066535
C0025598
drugBank: DB00331
RxNORM: 6809
anticoagulants
uricosuric agents
NDC:54868079500:
GlipiZIDE-Metformin HCl 2.5-250 MG Tablet
NDC: 54868518802: GlipiZIDE-Metformin HCl 5-500 MG Tablet
NDC:54868079500:GlipiZIDE-Metformin HCl TABS
Advantages of Semantic Web Technologies
•
Plug and play use of multiple ontologies and information models based on
industry standards (e.g., CDISC, HL7).
•
Ability to access multiple points of view through declarative specification of
mappings.
— Mappings across CDISC/SDTM and HL7 based information models
— Mappings across terminologies such as NDC, RxNorm and Stanford’s Drug
Ontology
•
Ability to map across terminologies via compositional definition of concepts,
e.g., Obesity drugs
•
Late binding of coding systems and database schema
•
Transform SPARQL to SQL in real time, reflecting real time discovery and
integration needs
Next Steps
•
Solicit Feedback and Participation from the broader Biomedical
Informatics communities
http://esw.w3.org/topic/HCLS/ClinicalObservationsInteroperability
http://hcls.deri.org/coi/demo
•
Develop proof of concepts for a wider variety of use cases in
collaboration with various participants in the HCLS Ecosystem
— Adverse Drug Event Reporting and Resolution
— Clinical Trials Data Collection
— Pharmaco-vigilance