Title: Ontology and the future of Evidence-Based Medicine Dagstuhl May 23th, 2006
1Ontology and the future ofEvidence-Based
Medicine Dagstuhl May 23th, 2006
- Werner Ceusters, MD
- Ontology Research Group
- Center of Excellence in Bioinformatics Life
Sciences - SUNY at Buffalo, NY
2Evidence Based Medicine
- the integration of best research evidence with
clinical expertise and patient values. - best research evidence clinically relevant
patient centered research into the accuracy and
precision of diagnostic tests, the power of
prognostic markers, and the efficacy and safety
of care regimens. - clinical expertise the ability to use
clinical skills and past experience to rapidly
identify each patient's unique health state. - patient values the unique preferences,
concerns and expectations each patient brings to
a clinical encounter and which must be integrated
into clinical decisions if they are to serve him.
3Application of Evidence Based Medicine
- Now
- Decisions based on (motivated/justified by) the
outcomes of (reproducable) results of
well-designed studies - Guidelines and protocols
- Evidence is hard to get, takes time to
accumulate. - Future
- Each discovered fact or expressed belief should
instantly become available as contributing to
evidence, wherever its description is generated.
4Future scenarios
- Data entered about a successful treatment of a
case in X generates a suggestion for a similar
case in Y - Submission of a new paper to Pubmed on some ADR
triggers an alert in EHR systems worldwide for
those patients that might be at risk -
- ? From reactive care to proactive care
5Some problem areas
- Pharmaceutical Industry
- Optimise drug discovery
- Make messy databases more useful for everybody
- Consumer health
- Opposing forces
- Quality of information
- Make them consume
- Malpractice suits
- Public sector health
- Cost containment
- Cost effectiviness of treatment, prevention
- Bio-informatics world
- How to find out that a discovery is a new
discovery ?
6An action plan for a European eHealth Area.
- By the start of 2005
- MS and EC should agree on an overall approach to
benchmarking in order to assess the quantitative,
including economic, and qualitative impacts of
e-Health. - By end 2006
- in order to achieve a seamless exchange of health
information across Europe through common
structures and ontologies, MS, in collaboration
with the EC, should identify and outline
interoperability standards for health data
messages and electronic health records, taking
into account best practices and relevant
standardisation efforts. - By end 2008
- the majority of all European health organisations
and health regions (communities, counties,
districts) should be able to provide online
services such as teleconsultation (second medical
opinion), e-prescription, e-referral,
telemonitoring and telecare.
Failed
7One key issue Semantic Interoperability
- Working definition
- Two information systems are semantically
interoperable if and only if each can carry out
the tasks for which it was designed using data
and information taken from the other as
seemlessly as using its own data and information.
system Any organized assembly of resources and
procedures united and regulated by interaction or
interdependence to accomplish a set of specific
functions.
information system a system, whether automated
or manual, that comprises people, machines,
and/or methods organized to collect, process,
transmit, and disseminate data that represent
user information.
8Essential components
Communication Interpretation
- People physicians, nurses, patients, healthcare
administrators, ... - Machines
- to make humans interact with the EHR,
- to transmit data from one EHR to another
- to enter data (lab analysers, EMR monitors, ...)
- to interprete data (alerts, quality assessment,
protocol selection, ...) - Data and information (data in context)
- Procedures
9Understanding content (1)
John Doe has a pyogenic granuloma of the left
thumb
John Doe has a pyogenic granuloma of the left
thumb
10Understanding content (2)
ltrecordgt ltpatientgtJohn Doelt/patientgt ltdiagnosisgtpy
ogenic granuloma of the left thumblt/diagnosisgt lt/r
ecordgt
ltrecordgt ltsubjectgt John Doe lt/subjectgt ltdiagnosisgt
pyogenic granuloma of the left thumb
lt/diagnosisgt lt/recordgt
11Understanding content (3)
lt129465004gt lt116154003gtJohn Doelt/116154003gt lt
8319008 gt 17372009 ltfinding sitegt 76505004
ltlateralitygt7771000lt/lateralitygt lt/finding
sitegt lt/ 8319008 gt lt/129465004gt
12Ontology based on Unqualified Realism
- Accepts the existence of
- a real world outside mind and language
- a structure in that world prior to mind and
language (universals / particulars) - Rejects nominalism, conceptualism, ontology as a
matter of agreement on conceptualizations - Uses reality as a benchmark for testing the
quality of ontologies as artifacts by building
appropriate logics with referential semantics
(rather than model-theoretic)
13Relevance for EHR Semantic Interoperability
14Relevance for EHR Semantic Interoperability
The realist approach
R E A L I T Y
L O G O L K A I S N S G
Ontology
EHR
15Terminology
- A theory concerned with those aspects of the
nature and the functions of language which permit
the efficient representation and transmission of
items of knowledge (J. Sager) - Precise and appropriate terminologies provide
important facilities for human communication (J.
Gamper)
16Ontology
- An ontology is a representation of some
pre-existing domain of reality which - (1) reflects the properties of the objects within
its domain in such a way that there obtains
a systematic correlation between reality and the
representation itself, - (2) is intelligible to a domain expert
- (3) is formalized in a way that allows it to
support automatic information processing
17A division of labour
- Terminology
- Communication amongst humans
- Communication between human and machine
- Ontology
- Representation of reality inside a machine
- Communication amongst machines
- Interpretation by machines
18Todays biggest problem a confusion between
terminology and ontology
- The conditions to be agreed upon when to use a
certain term to denote an entity, are often
different than the conditions which make an
entity what it is. - Trees would still be different from rabbits if
there were no humans to agree on how these things
should be called. - ontos means being. The link with reality
tends to be forgotten one concentrates on the
models instead of on the reality.
19What to do about it ? (1)
- Research
- Revision of the appropriatness of concept-based
terminology for our purposes - Relationship between models and that part of
reality that the models want to represent - Adequacy of current tools and languages for
representation - Boundaries between terminology and ontology and
the place of each in semantic interoperability in
healthcare
20What to do about it ? (2)
- Training and awareness
- Make people more critical wrt terminology and
ontology promisses - What is needed must be based on needs, not on the
popularity of a new concept - But in a system, its not just your own needs, it
is each components needs ! - Towards an ontology of ontologies
- First description
- Then quality criteria
21Ultimate goal
Ontology
continuant
disorder
person
CAG repeat
EHR
Juvenile HD
IUI-1 affects IUI-2 IUI-3 affects
IUI-2 IUI-1 causes IUI-3 ...
Referent Tracking Database
223 fundamentally different in levels
- the reality on the side of the patient
- the cognitive representations of this reality
embodied in observations and interpretations on
the part of clinicians and others - the publicly accessible concretizations of such
cognitive representations in representational
artifacts of various sorts, of which ontologies
and terminologies are examples.
23Example a person (in this room) s phenotypic
gender
- Reality
- Male
- Female
- Cognitive representation
- male
- female
- In the EHR
- male
- female
- unknown
244 fundamental reasons for making changes
- changes in the underlying reality
- does the appearance of an entry (in a new version
of an ontology or in an EHR) relate to the
appearance of an entity or a relationship among
entities in reality ? - changes in our (scientific) understanding
- reassessments of what is considered to be
relevant for inclusion (notion of purpose), or - encoding mistakes introduced during data entry or
ontology development.
25Key requirement
- Any change in an ontology or data repository
should be associated with the reason for that
change !
26Example a person (in this room) s gender in the
EHR
- In John Smiths EHR
- At t1 male at t2 female
- What are the possibilities ?
- Change in reality transgender surgery
- Change in understanding it was female from the
very beginning but interpreted wrongly - ( No change in relevance )
- Correction of data entry mistake
- (was understood as male, but wrongly transcribed)
27Possible combinations
OE objective existence ORV objective
relevance BE belief in existence BRV
belief in relevance Int. intended encoding
Ref. manner in which the expression refers
G typology which results when the factor of
external reality is ignored. E number of
errors when measured against the benchmark of
reality. P/A presence/absence of term.
28Possible evolutions
29Towards an implementation
- A client-server application in which the server
is composed of four layers - the Web Server Layer (WSL) provides the interface
to clients via web services - the RT Core application programming interface
(API) encapsulates the data services related to
storage and retrieval. Its Security Module
validates the access rights before any data
service - the database layer stores all the RT data, and
- the reasoner layer (RL) performs inferences upon
specific requests, based on the information
available in the database and, if available, the
ontologies that have been used for the
descriptions of the portions of reality.
Shahid Manzoor
30Schematic representation
31Simple Graph Representation
32Complete structure
33UML-diagram for the entities in the RDF-graph
34Querying the RTDB using ontologies (SPARQL)
Retrieve particulars that are related to the
universal face
1 PREFIX rts lthttp//ecorgt 2 PREFIX fma
lthttp//FMAgt 3 SELECT ?p ?u ?v 4 WHERE ?p
rtsrelation ?u . 5 ?u a rtsPtoU . 6 ?u rtsu
?t . 7 ?t a fmaFace . 8
35Querying the RTDB using ontologies (SPARQL)
1 WHERE ?p rtsrelation ?rf . 2 ?rf a rtsPtoP
. 3 ?rf rtsp ?f . 4 ?f a fmaHead . 5 ?f
rtsrelation ?rd . 6 ?rd a rtsPtoP . 7 ?rd p
?d . 8 ?d a disDISEASES AND INJURIES .
Retrieve patients with diseases in the head
36Test interface