Bridging Epoch: Mapping Two Clinical Trial Ontologies 10th - - PowerPoint PPT Presentation

SLIDE 1

Bridging Epoch: Mapping Two Clinical Trial Ontologies

10th International Protégé Conference July 17, 2007

Samson W. Tu1, Douglas Fridsma2, Ravi D. Shankar1, Martin O'Connor1, Amar K. Das1, David B. Parrish3

1Stanford University, USA 2University of Pittsburgh, USA 3The Immune Tolerance Network, Pittsburgh, PA, USA

SLIDE 2

2

Problem: Ontologies and semantic interoperability

Ontology 1 Ontology 2

SLIDE 3

3

BRIDG: Biomedical Research Integration Group

Part of US NCI Cancer Biomedical

Informatics Grid (caBIG)

Stakeholders include US FDA, HL7,

CDISC

Create shared domain model for

protocol-driven clinical research

Comprehensive Consensus-based Abstract and context neutral

BRIDG

SLIDE 4

4

SLIDE 5

5

SLIDE 6

6

EPOCH: Immune Tolerance Network clinical trial ontologies

Immune Tolerance Network (ITN) International collaborative

research effort that sponsors clinical trials and mechanistic assays on immune tolerance

EPOCH clinical trial model Developed at Stanford Medical

Informatics

Designed to provide semantic

foundation for management of clinical trials

EPOCH

SLIDE 7

7

Management of clinical trials involves complex data and multiple groups

Protocol Group

Schedule

• f Events

CRO

CRF

Data Center

A ccessio n A ccessio n ID C

• r

eID (F K ) F ilen am e T r an sm issio n D t T r an sactio n C

• d

eID (F K ) R eco r d D t U ser ID (F K ) A r ch ive A r ch iveID C

• r

eID ( F K ) P r

• cessID

( F K ) N am e D ir ecto r y R eco r d D t U ser ID ( F K ) B ar co d e B ar co d eID S tu d yID (F K ) S u b jectID (F K ) V isitID ( F K ) B ar co d e C

• llectio

n D t R eco r d D t U ser ID ( F K ) C

• n

fig P ar am eter C

• n

fig P ar am eter ID P r

• cessID

(F K ) N am e V alu e R eco r d D t U ser ID (F K ) D escr ip tio n R eco r d D t U ser ID (F K ) C

• r

e_A ssays C

• r

eID (F K ) A ssayID (F K ) S p ecim en T yp eID ( F K ) R eco r d D t U ser ID (F K ) C

• r

e_C

• n

tacts C

• r

eID ( F K ) C

• n

tactT yp eID (F K ) U ser ID ( F K ) R eco r d D t C

• r

e_Issu es Issu eID C

• r

e S tu d y C

• r

esam p leID R eq u ested B y D escr ip tio n R eq u estD t R esp

• n

se S tatu s C

• r

eS am p leID (F K ) C

• r

eS am p le C

• r

eS am p leID S tu d yID ( F K ) A ccessio n ID (F K ) P r

• cessin

g C

• d

eID (F K ) S am p leT yp eID (F K ) R ecd D t A n alysisD t T ech n am e F ilen am e E xter n alID B ar co d eID 9 S p ecim en T yp eID (F K ) R eco r d D t U ser ID (F K ) C

• r

eS am p le_D etail C

• r

eS am p leID (F K ) B ar co d eID (F K ) QC S am p leID (F K ) V alid atio n C

• d

eID (F K ) V alid atio n D t R eco r d D t U ser ID (F K ) C

• r

eS am p le_D evC

• d

e C

• r

eS am p leID (F K ) D evC

• d

eID (F K ) N

• t

es R eco r d D t U ser ID (F K ) D evC

• d

e D evC

• d

eID C

• d

e D escr ip tio n GetC

• m

m en ts R eco r d D t U ser ID (F K ) am p leID (F K ) D (F K ) D(F K ) F K ) D(F K ) (F K ) m e eID (F K ) n t d D t D( F K ) F ield F ield ID N am e D escr ip tio n R eco r d D t U ser ID (F K ) F ileR eq u est R eq u estID R eq u estD t C

• m

p leted D t E m ailA d d r e ss S tatu s F ileN am e F ileT yp e M essag e A r ch iveID (F K ) U ser N am e F ileR eq u estL ist F ileR eq u estL istID S tatu s F ileN am e D ir ecto r y R eq u estID (F K ) M essag e K ) p cr R esu lt_Qu al C

• r

eS am p leid (F K ) T estID (F K ) R ep licateID B aseV isitID (F K ) Qu alifier A tL east S am p leQu ality C alib r ato r Qu ality R eco r d D t U ser ID (F K ) P r

• cess

P r

• cessID

N am e R eco r d D t U ser ID ( F K ) P r

• cessin

g C

• d

e P r

• cessin

g C

• d

eID C

• d

e D escr ip tio n R eco r d D t U ser ID (F K ) S M S _A cc S tu d y_ID S ite_ID P ar ticip an V isit_N u m C

• llectio

n C

• llectio

n B ar co d e S p ecim en L ab S ite_D ev_ C

• r

e_D ev O_S tu d y_ O_S ite_ID O_P ar tici O_V isit_N O_C

• llec

O_C

• llec

O_S p ecim In valid S ta In valid C

• In

valid C

• In

valid S ID In valid P ID In valid E n R eco r d D t S M S _Issu ed _Q u er ies Issu eID (F K ) Qu er y_ID (F K ) R eco r d D t R eso lved S M S _Issu es Issu eId C

• r

eS am p leID ( F K ) T extM sg S tatu s R eco r d D t S M S _Qu Qu er y S tu d y Qu er y S ite_q Qu er y Qu er y Qu er y Qu er y D ate_ D ate_ D ate_ In itiat R eso l E xter n R eso l D ate_ Qu er y S

• u

r ce S

• u

r ceID C

• r

eID (F K ) V er sio n D escr ip tio n S tar tD t E n d D t R eco r d D t U ser ID (F K ) S

• u

r ce_F ield s S

• u

r ceF ield ID S

• u

r ceID ( F K ) F ield ID ( F K ) Or d in al R eco r d D t S kip F lag R eco r d D t U ser ID (F K ) S am p leT yp e S am p leT yp eID N am e R eco r d D t U ser ID (F K ) C

• d

e C

• n

vN u m IsL

• calL

ab IR B A p p r

• ved

lD t IT N A p p r

• ved

D t B u d g etA p p r

• ved

D t A ctivatio n D t R eco r d D t U ser ID (F K ) S u b ject S u b jectID S iteID (F K ) P ar ticip an tID S tu d yGr

• u

p ID (F K ) R eco r d D t U ser ID ( F K ) T r an sactio n C

• d

e T r an sactio n C

• d

eID C

• d

e D escr ip tio n R eco r d D t U ser ID ( F K ) V alid atio n C

• d

e V alid atio n C

• d

eID C

• d

e D escr ip tio n R eco r d D t U ser ID (F K ) r p tD ataE xtr actR eq u est R eq u estN u m b er D tOfR eq u est D tR eq u ir ed B y R eq u esto r N am e R eq u esto r P h

• n

e S tu d yN u m C

• r

e S tu d yD ata QC S am p les R esear ch D evelo p m en t S tar tD t E n d D t V isitsD etail D ataF

• r

m at In clu d eD eviatio n C

• d

es In clu d eT r tC

• h
• r

t In clu d eU n valid ated S am p les S p ecialIn str u ctio n s A ssig n ed T

• A

ssig n ed D t P lan S tatu s D tP lan A vailab le D tP lan F in al R eq u estC lo sed B y C lo sed D t S iteD etails In fo r m aticsN

• te

R ecU p d ated B y R ecU p d ated On D ataE xtr actT yp e D( F K ) r eID (F K ) F K ) D t (F K ) S cee g t E n r

• llm

en t D t S tu d yGr

• u

p ID ( F K ) D

• n
• r

ID R ecip ien tID R eco r d D t S cr een in g F ailu r eR esaso n ID S cr een in g F ailu r eC

• m

m en t U ser ID (F K ) E T L M ap p in g D etail M ap D tlID S

• u

r ceC

• lu

m n T ar g etT ab le T ar g etC

• lu

m n T yp e M ap H d r ID E T L M ap p in g H ead er M ap H d r ID M ap p in g N am e S

• u

r ce V er sio n Gen er icL

• ad

Gen L

• ad

ID F ileN am e F ileD ir R

• wN

u m b er C

• lu

m n N am e C ellV alu e S

• u

r ce V er sio n d t E n ter ed L

• ad

ID Gen er icL

• ad

B atch GL B atch ID F ileD ir R eco r d D t F ileN am e H L A B u sin essR u leD etail B R D etailID T ar g etT ab le T ar g etC

• lu

m n B R V alu e T ar g etC

• lu

m n T yp e B R H d r ID H L A B u sin essR u leH ead er B R H d r ID V er sio n B R N am e H L A D ataIm p

• r

tL

• g

Im p

• r

tL

• g

ID F ileN am e S tu d yN u m C

• d

e V isitN u m S p ecim en T yp e P ar ticip an tID B ar co d e D escr ip tio n R eco r d D t Op er ato r ID A ccessio n ID H L A E ver tL

• g

H L A E ven tL

• g

ID E ven tD esc R eco r d D t Op er ato r ID GL B atch ID S cr een in g F ailu r eR e aso n S cr een in g F ailu r eR easo n ID C

• d

e D escr ip tio n E n ab led S tu d yID T r an sp lan t T r an sp lan tId R ecip ien t_S u b jectId D

• n
• r

_S u b jectId Or g an R eco r d D t Op er ato r S tu d yId

Clinical Trial Data

Cim arron

I m m unoTrak

Core Labs

Assay Results

Tubes Manufacturer

Kit Report

Assay Group Operations Group

Specim en Table Tube Table

Queries Reports

Queries Reports

SLIDE 8

8

Three goals of EPOCH

• ntologies

Design tools to help

acquire and maintain knowledge about protocol and assay designs

Use this knowledge to

drive data collection during a trial

Implement querying

methods to support trial management, and ad hoc data analysis

SLIDE 9

9

EPOCH ontologies created in Protégé OWL

SLIDE 10

10

ITN wants to use BRIDG- compliant applications

ITN protocols encoded as EPOCH

knowledge bases to drive caBIG applications (e.g., Patient Study Calendar)

Challenge: Develop methods to Harmonize common subset of BRIDG &

EPOCH: shared semantics

Overcome representational mismatch

• Representation languages
• Representation choices
• (Terminological mismatch not consider here)

SLIDE 11

11

Approach taken

Semantic alignment Overcoming representation language

mismatch

Overcoming representation choice

mismatches

SLIDE 12

12

Approach taken

Semantic alignment

Use Excel spreadsheet to systematically

review and document possible mappings

Define necessary preconditions for

mapping

Overcoming representation language

mismatch

Overcoming representation choice

mismatches

SLIDE 13

13

Semantic Alignment: Excel spreadsheet

SLIDE 14

14

Semantic Alignment: Restrictions on EPOCH

Mapping from EPOCH to BRDG => Place

restrictions on EPOCH

Only one schedule of activities Period has no subperiods Limited temporal annotations … Define necessary conditions for mapping Formulate as DL definition of

“BRIDGClinicalTrial” subclass of epoch:ClinicalTrial

SLIDE 15

15

BRIDGClinicalTrial in EPOCH

Need trial-specific closure axioms to do

automated classification of EPOCH trials that can be mapped to BRIDG

SLIDE 16

16

Approach taken

Semantic alignment Overcoming representation language

mismatch

Overcoming representation choice

mismatches

SLIDE 17

17

Overcoming representation language mismatch: BRIDG-in-OWL

Scope: BRIDG Study Planned View + BRIDG Complex Data Types

SLIDE 18

18

Example: PlannedStudy

SLIDE 19

19

Modifications to BRIDG

Driven by Patient Study Calendar

application requirements

Added several associational

relationships

Modified some subsumption relations

SLIDE 20

20

Approach taken

Semantic alignment Overcoming representation language

mismatch

Overcoming representation choice

mismatches

SLIDE 21

21

Overcoming representation choice mismatch: Epoch example

Period Screening0 Arm1Cycle1 Arm1Cycle2

PeriodType=Screening PeriodType=Intervention PeriodType=Intervention

Arm2Cycle1 Arm2Cycle2

PeriodType=Intervention PeriodType=Intervention

Intervention epoch Screening epoch EPOCH BRIDG

SLIDE 22

22

SWRL rule to map epochs

EPOCH:periodTypes

• f periods

correspond to BRIDG:epochs

EPOCH

:periodType.label corresponds to BRIDG:epoch.code. displayName

SLIDE 23

23

Successfully used an EPOCH clinical trial to configure BRIDG Patient Study Calendar application

Herold protocol in EPOCH Herold protocol in BRIDG Herold protocol in PSC XML SWRL rules SWRL rules Patient Study Calendar Automated mappings except for one

relationship

Because of OWL/SWRL’s open-world

assumption, First epoch cannot be derived as an epoch that has no predecessor

SLIDE 24

24

Conclusions

Semantic interoperability requires Harmonization of subsets of

• ntologies/models

Overcoming mismatches in representation

languages and representation choices

OWL restrictions and SWRL rules help to

• vercome semantic and syntactic

mismatches

Possible future work Continued harmonization of BRIDG/EPOCH Scalability and (semi-)automation of method

SLIDE 25

25

Thank you!

Questions? Comments? Suggestions?