clef: clinical e-science framework
DESCRIPTION
The CLEF Chronicle: Transforming Patient Records into an E-Science Resource Jeremy Rogers, Colin Puleston, Alan Rector James Cunningham, Bill Wheeldin, Jay Kola Bio-Health Informatics Group Department of Computer Science University of Manchester. CLEF: Clinical E-Science Framework. - PowerPoint PPT PresentationTRANSCRIPT
The CLEF Chronicle: Transforming Patient Records into an E-Science Resource
Jeremy Rogers, Colin Puleston, Alan RectorJames Cunningham, Bill Wheeldin, Jay Kola
Bio-Health Informatics GroupDepartment of Computer Science
University of Manchester
CLEF: Clinical E-Science Framework
• Improving the storage and processing of Electronic Health Records to enhance general clinical care
• Supporting clinical research via the creation of a clinical research repository, known as the CLEF Chronicle
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2…located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years…whilst remaining in remission for the full extent of this period
THEN:
ALSO…
Chronicle Query
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2 …located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years…whilst remaining in remission for the full extent of this period
THEN:
ALSO…
Concepts from ExternalKnowledge Sources (EKS)
Properties from ExternalKnowledge Sources (EKS)
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2 …located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years…whilst remaining in remission for the full extent of this period
THEN:
ALSO…
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2 …located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years…whilst remaining in remission for the full extent of this period
THEN:
ALSO…
mastectomy is-a surgical-intervention
shin part-of lower-leg part-of leg
Implicit RelationshipsBetween EKS Concepts
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2 …located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years …whilst remaining in remission for the full extent of this period
THEN:
ALSO…
Temporal Information
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2 …located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years …whilst remaining in remission for the full extent of this period
THEN:
ALSO…
ARBITRARY TEMPORAL SEQUENCES
Temporal Abstractions
WHAT PERCENTAGE OF PATIENTS WHO…
Had cancer with stage of stage-2 …located somewhere in the leg…with primary tumour…that doubled in size within a 3 month period
FIRST:
Underwent surgical-intervention to remove all tumours
THEN:
Survived for at least ten years …whilst remaining in remission for the full extent of this period
THEN:
ALSO…
…whilst remaining in remission for the full extent of this period
…that doubled in size within a 3 month period
Chronicle System: Overview
(1) Chronicle Representation
Chronicle Representation
1
(2) Chronicle Repository + Query Engine
Chronicle Representation
Chronicle Repository
Query Engine
1
2
(3) ‘Chroniclisation’ Process
Chronicle Representation
Chronicle Repository
Query Engine
Chronicliser
EHR Repository(UCL)
Text Processor (Sheffield)
13
2
(4) Chronicle Simulator
Chronicle Representation
Chronicle Repository
Query Engine
Chronicle Simulator
Chronicliser
EHR Repository(UCL)
Text Processor (Sheffield)
13
24
(5) Browser + Query GUIs
Chronicle Representation
Chronicle Repository
Simple Browser +Query Formulator
Query Engine
Query Formulator(Open University)
Chronicle Simulator
Chronicliser
EHR Repository(UCL)
Text Processor (Sheffield)
13
24
5
ChronicleRepresentation
Temporal Representation
end point
start point
SPAN Event
occurrence point
SNAPEvent
Time
Temporal Representation
end point
start point
SPAN Event
occurrence point
SNAPEvent
Note: For the Patient Chronicle the atomic time-unit equals one-day…
Time
…hence, for example, Surgical-Operations and Consultations are SNAP Events
Temporal Representation
end point
start point
SPAN Event
occurrence point
SNAPEvent
Example: X-ray performed on specific day …with associated
set of results
Time
Temporal Representation
Time
end point
start point
SPAN Event
occurrence point
SNAPEvent
Example: Period of employment as Plumber, spanning specific time-period
Temporal Representation
end point
start point
Structured SPAN Event
Time
SNAP SNAPSNAPSNAP
Temporal Representation
end point
start point
Structured SPAN Event
Time
SNAP SNAPSNAPSNAP
Example: History of Tumour over specific time-period …
…with set of ‘snapshots’ representing same Tumour at specific time-points
Temporal Representation
end point
start point
Structured SPAN Event
Time
SNAP SNAPSNAPSNAP
Example cont.: Each SNAP has associated value for tumour-size attribute…
…whilst SPAN has set of ‘temporal-abstractions’ (e.g. max, min, etc.) summarising the tumour-size attribute
Clinical Model
Chronicle Representation
Generic Model
Clinical KnowledgeService
Chronicle Model
Java Object Model
ExternalKnowledge
Sources (EKS)Ontologies,
Databases, etc.
EKS
EKSRelated
Inference
Clinical Model
Chronicle Representation
Generic Model
EKSRelated
Inference
Clinical KnowledgeService
EKS
Chronicle Representation is embedded within a generic Knowledge Driven Architecture
Clinical Model
Generic Model
Generic Model
Clinical KnowledgeService
EKS
Including… SNAP/SPAN temporal representation Temporal abstraction mechanisms EKS-concept handling
Generic modelling classes…
EKSRelated
Inference
Clinical Model
Clinical Model
Generic Model
Clinical KnowledgeService
EKS
Extends generic model with clinical-specific classes
Examples… SNAPS: ProblemSnapshot, SnapClinicalProcedure, etc. SPANS: ProblemHistory, ClinicalRegime, etc.EKS
RelatedInference
Clinical Model
External Knowledge Sources (EKS)
Generic Model
Clinical KnowledgeService
EKS
Detailed (time-neutral) clinical knowledge sources
Currently: Single OWL ontologyPossibly: Multiple ontologies, databases, etc.
EKSRelated
Inference
Clinical Model
External Knowledge Sources (EKS)
Generic Model
EKSRelated
Inference
Clinical KnowledgeService
EKS
Provide… Hierarchies of concepts Sets of inter-concept relationships Sets of instance-descriptor properties attached to concepts
Clinical Model
EKS-Related Inference
Generic Model
EKSRelated
Inference
Clinical KnowledgeService
EKS
Drive… Dynamic data creation Query formulation
Currently: Description-Logic based reasonerPossibly: Rule-bases, procedural code, etc.
Arbitrarily complex inference mechanisms…
Clinical Model
EKS-Related Inference
Generic Model
EKSRelated
Inference
Clinical KnowledgeService
EKS
Note: Full EKS-related inference is neither appropriate, nor required, for (time-critical) execution of queries over thousands of patient chronicles
Clinical Model
Clinical Knowledge Service
Generic Model
Clinical KnowledgeService
EKS
Provides transparent access to…External knowledge sourcesEKS-related inference
EKSRelated
Inference
Simple interface…Takes: Instance of concept X, including set of descriptor values
Returns: Updated descriptor-set for X (including updated constraints)
Problem-Types
ProblemHistory
snapshots[]
ProblemSnapshot
location type
Bodily-Locations
ProblemSnapshotProblem
Snapshot
Chronicle Representation:
ExampleRepresentation of the history of a specific clinical problem* as
displayed by a particular patient
* A ‘problem’ is either a pathology (e.g. cancer) or some
manifestation of a pathology (e.g. a specific tumour)
Chronicle Model
Objects
Problem-Types
ProblemHistory
snapshots[]
ProblemSnapshot
location type
Bodily-Locations
ProblemSnapshotProblem
Snapshot
Problem-Types
SPAN Event
SNAP Events
ProblemHistory
snapshots[]
ProblemSnapshot
location type
Bodily-Locations
ProblemSnapshotProblem
Snapshot
External Knowledge
Sources (EKS)
Problem-Types
ProblemHistory
snapshots[]
ProblemSnapshot
location type
Bodily-Locations
ProblemSnapshotProblem
Snapshot
‘type’ concept selected from
EKS
ProblemHistory
snapshots[]
ProblemSnapshot
location type
Tumour
ProblemSnapshotProblem
Snapshot
Bodily-Locations
IntegerHistory
ProblemHistory
snapshots[]
ProblemSnapshot
location type
IntegerSnapshot
tumour-size
IntegerSnapshotInteger
Snapshottumour-size
Tumour
ProblemSnapshotProblem
Snapshot
Bodily-Locations
‘descriptor’ variables derived
from ‘type’ concept
ProblemHistory
snapshots[]
ProblemSnapshot
location type
IntegerSnapshot
IntegerHistorytumour-size
IntegerSnapshotInteger
Snapshottumour-size
Tumour
value:
time-point:
7
4/3/98
ProblemSnapshotProblem
Snapshot
Bodily-Locations
Values allocated to snapshot ‘descriptors’
ProblemHistory
snapshots[]
ProblemSnapshot
location type
IntegerSnapshot
IntegerHistorytumour-size
IntegerSnapshotInteger
Snapshottumour-size
Tumour
start-value:
end-value:
minimum:
maximum:
range:
increase-rate:
end-point:
Temporal Abstractions
start-point: 4/3/98
7
7/2/02
43
82
7
75
0.051
ProblemSnapshotProblem
Snapshot
Bodily-Locations
History ‘descriptor’ values derived automatically
Breast
‘location’ concept selected from EKS
ProblemHistory
snapshots[]
ProblemSnapshot
location type
IntegerSnapshot
IntegerHistorytumour-size
IntegerSnapshotInteger
Snapshottumour-size
Tumour
ProblemSnapshotProblem
Snapshot
her2-receptor
her2-receptor
Breast
ProblemHistory
snapshots[]
ProblemSnapshot
location type
IntegerSnapshot
IntegerHistorytumour-size
IntegerSnapshotInteger
Snapshottumour-size
Tumour
ProblemSnapshotProblem
SnapshotBoolean
SnapshotBooleanSnapshotBoolean
Snapshot
BooleanHistory
Additional ‘descriptor’ variables inferred via
EKS-related reasoning
her2-receptor
her2-receptor
Breast
ProblemHistory
snapshots[]
ProblemSnapshot
location type
IntegerSnapshot
IntegerHistorytumour-size
IntegerSnapshotInteger
Snapshottumour-size
Tumour
ProblemSnapshotProblem
SnapshotBoolean
SnapshotBooleanSnapshotBoolean
Snapshot
BooleanHistory
start-value:
end-value:
always-true:
always-false:
percent-true:
percent-false:
end-point:
start-point: 4/3/98
false
7/2/02
true
false
false
63.72
36.28
value:
time-point:
false
4/3/98
Values allocated/derived for new ‘descriptors’
Chronicle Repositoryand
Query Engine
Chronicle Query Engine: Requirements
• Querying over Large Numbers of patient chronicles
• Basic RDF/RDFS-Style Reasoning, involving:– Hierarchical relationships (is-a)– Property relationships (part-of, has-location, etc.)– Transitivity
• Temporal Reasoning, including:– Reasoning about temporal sequences– On-the-fly temporal abstraction
Chronicle Repository
• An RDF/RDFS-based repository (currently using Sesame RDF-store)
• RDF/RDFS representation to facilitate:– Querying over Large Numbers of patient
chronicles– Basic RDF/RDFS Reasoning (must incorporate
transitivity)• Additional Temporal Reasoning mechanisms
will be required (including on-the-fly temporal abstraction)
ChroniclisationProcess
Electronic Health Records (EHR)
• Document based:– One document per clinical procedure
• Minimally structured:– No inter-concept references– No inter-document references
• Mainly free-form text:– For human consumption– Incomplete information– Many implicit assumptions
Chroniclisation
• Complex heuristic process:– Input: Largely unstructured EHR data– Output: Highly structured chronicle data
• Process will involve:– Text processing– Co-reference resolution– Temporal reference resolution – Inference of implicit information
CLEF Chronicle: Summary
• Chronicle Representation:– Temporal Representation– External Knowledge Sources (OWL, etc.)– Complex EKS-related reasoning (DL, etc.)
• Chronicle Repository + Query Engine:– Querying large numbers of patient records– Simple EKS-related reasoning (RDF/RDFS)– Temporal Reasoning
• Chroniclisation Process:– Input: Largely unstructured EHR data– Output: Highly structured Chronicle data