R T U New York State Center of

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Disease Ontology Workshop The Ontology of Diagnosis (to be quoted as work in progress) Baltimore, MD, USA. November 7, 2006 Werner CEUSTERS Center of Excellence in Bioinformatics and Life Sciences Department of Psychiatry, University at Buffalo, NY, USA National Center for Biomedical Ontology http: //www. org. buffalo. edu/RTU

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Outline 1. 2. 3. 4. Background notions on ontology. What is a diagnosis ? What is a disease ? Diagnostic decision making from an ontological and epistemological perspective. 5. The impact of change on making diagnoses. 6. (if enough time) Tracking changes in representations.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Part 1: Background Reality, ontologies, views, …

R T U New York State Center of Excellence in Bioinformatics & Life Sciences A realist view of the world • The world consists of – entities that are • Either particulars or universals; • Either occurrents or continuants; and, – relationships between these entities of the form • Either dependent or independent; • <particular , universal> • <particular , particular> • <universal , universal> e. g. is-instance-of, e. g. is-member-of e. g. isa (is-subtype-of) Smith B, Kusnierczyk W, Schober D, Ceusters W. Towards a Reference Terminology for Ontology Research and Development in the Biomedical Domain. Proceedings of KR-MED 2006, November 8, 2006, Baltimore MD, USA

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Three levels of reality 1. The world exists ‘as it is’ prior to a cognitive agent’s perception thereof; 2. Cognitive agents build up ‘in their minds’ cognitive representations of the world; 3. To make these representations publicly accessible in some enduring fashion, they create representational artifacts that are fixed in some medium. Smith B, Kusnierczyk W, Schober D, Ceusters W. Towards a Reference Terminology for Ontology Research and Development in the Biomedical Domain. Proceedings of KR-MED 2006, November 8, 2006, Baltimore MD, USA

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Representational artifacts • Ideally built out of representational units and relationships that mirror the entities and their relationships in reality. Primarily about particulars Primarily about universals and defined classes Non-Formalized Progress notes, discharge letters, medical summaries, maps, . . . Medical textbooks, scientific theories, . . . Inventories, referent tracking database, . . . Ontologies, terminologies, . . .

R T U New York State Center of Excellence in Bioinformatics & Life Sciences An example: Referent Tracking Now! That should clear up a few things around here ! • Purpose: – explicit reference to the concrete individual entities relevant to the accurate description of each patient’s condition, therapies, outcomes, . . . • Method: – Introduce an Instance Unique Identifier (IUI) for each relevant particular (individual) entity Ceusters W, Smith B. Strategies for Referent Tracking in Electronic Health Records. J Biomed Inform. 2006 Jun; 39(3): 362 -78.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences From non-formalized to formalized repositories through Referent Tracking ‘John Doe’s person inst-of at t 2 #10 ‘John Smith’s instance-of at t 1 liver inst-of at t 2 #20 liver instance-of at t 1 tumor inst-of at t 2 #30 tumor #1 instance-of at t 1 liver #2 tumor #3 was treated #4 instance-of at t 1 with RPCI’s irradiation device’ #5 treating clinic instance-of at t 1 #6 device inst-of at t 2 #5 inst-of at t 2 #6 #40 was treated with RPCI’s irradiation device’

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Repository – Ontology “collaboration”

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Ontology of Biomedical Reality (under development) pathological formation caused by disease pathological anatomical structure caused by isa tumor Liver cell adenoma isa Hepatocellular carcinoma hepatoblastoma isa aberrant DNA molecule

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Part 2: What is a diagnosis ?

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Making a diagnosis (naive view) instance-of at t caused #105 by

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Towards the full picture (1) • Level of biomedical reality: – Persons, diseases, pathological structures and formations, . . . do exist as particulars (p, d, ps, pf, . . . ) and universals (P, D, PS, PF, . . . ), and are related in specific ways prior to our perception; – Biomedical reality changes: • d’s, p’s, . . . come and go; D’s, P’s, . . . only come (? ) • Level of biomedical science and case perception: – Mirrors reality only partially – Evolves over time towards better understanding

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Towards the full picture (2) • Level of concretizations – Mirrors biomedical science and case perception only partially • Editing mistakes • Leaving out diseases or pathological behaviours for nonbiomedical reasons – Smoking • Adding non-pathological behaviour as a disease – homosexuality

R T U New York State Center of Excellence in Bioinformatics & Life Sciences My working definition of ‘diagnosis’ Any configuration of representational units which is believed to mirror the portion of reality consisting of an organism’s disease and the relationships this disease enjoys with the entities that caused the disease or influence its course, whereby some part of this configuration of representational units refers to the universal of which that disease is believed to be an instance, or the defined class of which it is believed to be a member.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences BTW: universal versus defined class instance-of at t “lung inflammation”

R T U New York State Center of Excellence in Bioinformatics & Life Sciences A well-formed diagnosis of ‘pneumococal pneumonia’ • A configuration of Disease representational units; isa • Believed to mirror the person’s disease; Pneumococcal pneumonia • Believed to mirror the disease’s cause; Instance-of at t 1 • Refers to the universal of which the disease is #78 #56 caused John’s portion John’s believed to be an by of pneumococs Pneumonia instance.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some motivations and consequences (1) • No use of debatable or ambiguous notions such as proposition, statement, assertion, fact, . . . • The same diagnosis can be expressed in various forms. Disease isa Pneumococcal pneumonia Instance-of at t 1 #78 caused by #56 Portion of pneumococs caused by isa Pneumonia Instance-of at t 1 #56 caused by #78 at t 1

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some motivations and consequences (2) • A diagnosis can be of level 2 or level 3, i. e. either in the mind of a cognitive agent, or in some physical form. • Allows for a clean interpretation of assertions of the sort ‘these patients have the same diagnosis’: The configuration of representational units is such that the parts which do not refer to the particulars related to the respective patients, refer to the same portion of reality.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Distinct but similar diagnoses Pneumococcal pneumonia Instance-of at t 1 #78 John’s portion of pneumococs caused by Instance-of at t 2 #56 #956 John’s Pneumonia Bob’s pneumonia caused by #2087 Bob’s portion of pneumococs

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some motivations and consequences (3) • Allows evenly clean interpretations for the wealth of ‘modified’ diagnoses: – With respect to the author of the representation: • ‘nursing diagnosis’, ‘referral diagnosis’ – When created: • ‘post-operative diagnosis’, ‘admitting diagnosis’, ‘final diagnosis’ – Degree of the belief: • ‘uncertain diagnosis’, ‘preliminary diagnosis’

R T U New York State Center of Excellence in Bioinformatics & Life Sciences But the definition requires working out: • At the level of biomedical reality: – What is a disease ? – What are the entities that cause a specific disease to exist or influence its course ? – What are the relationships between these entities and the disease ? • At the level of representational artifacts: – How do they relate to reality ? – How keeping track of changes in reality ?

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Part 3: What is a disease ?

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Latest WHO definition • A disease is: – an interconnected set of one or more dysfunctions in one or more body systems including: • a pattern of signs, symptoms and findings (symptomatology manifestations) • a pattern or patterns of development over time (course and outcome) • a common underlying causal mechanism (etiology) – linking to underling genetic factors (genotypes, phenotypes and endophenotypes) and to interacting environmental factors – and possibly: to a pattern or patterns of response to interventions (treatment response).

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Latest WHO definition • A disease is: – an interconnected set of one or more dysfunctions in one or more body systems including: • a pattern of signs, symptoms and findings (symptomatology manifestations) • a pattern or patterns of development over time (course and outcome) • a common underlying causal mechanism (etiology) – linking to underling genetic factors (genotypes, phenotypes and endophenotypes) and to interacting environmental factors – and possibly: to a pattern or patterns of response to. . . I assume interventions (treatment response). ‘of one single person’.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Latest WHO definition • A disease is: – an interconnected set of one or more dysfunctions in one or more body systems including: • a pattern of signs, symptoms and findings (symptomatology manifestations) • a pattern or patterns of development over time (course and outcome) • a common underlying causal mechanism (etiology) What isgenetic referred by ‘dysfunction’ – linking to underling factorsto(genotypes, phenotypes ? and endophenotypes) and to interacting environmental - A non-canonical functioningfactors ? – and possibly: to a pattern or patterns of response to - A functioning which may lead to interventions (treatment response). harming the diseased person ?

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Latest WHO definition • A disease is: – an interconnected set of one or more dysfunctions in one or more body systems including: • a pattern of signs, symptoms and findings (symptomatology manifestations) • a pattern or patterns of development over time (course and outcome) • a common underlying causal mechanism (etiology) Is not an ontological entity but a – linking to underling genetic factors (genotypes, phenotypes and representational entity. endophenotypes) and to interacting environmental factors – and possibly: to a pattern or patterns of response to more. . . Better: a disease involves one or interventions (treatment But then it is response). left open what a disease IS !

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Latest WHO definition • A disease is: – an interconnected set of one or more dysfunctions in one or more body systems including: • a pattern of signs, symptoms and findings (symptomatology manifestations) • a pattern or patterns of development over time (course and outcome) • a common underlying causal mechanism (etiology) – linking to underling genetic factors (genotypes, phenotypes and endophenotypes) and to interacting environmental factors Or: the processual entity composed of the – and possibly: to a pattern or patterns of response to interventions (treatment response). dysfunctionings. . .

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Latest WHO definition • A disease is: – an interconnected set of one or more dysfunctions in one or more body systems including: • a pattern of signs, symptoms and findings (symptomatology manifestations) • a pattern or patterns of development over time (course and outcome) • a common underlying causal mechanism (etiology) – linking to underling genetic factors (genotypes, phenotypes and principle factors of the a endophenotypes) and to. Violates interactingthe environmental – and possibly: to a pattern or patterns of response to would priori reality: diseases interventions (treatment response). only exist if we look for them.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences ‘Ontology inspired’ definition • An organism (or part of an organism) is diseased if and only if – it includes among its parts pathological anatomical structures which compromise the organism’s physiological processes to the degree that they give rise to symptoms and signs. • An anatomical structure is pathological whenever: – it has come into being as a result of changes in some pre-existing canonical anatomical structure – through processes other than the expression of the normal complement of genes of an organism of the given type, and – is predisposed to have health-related consequences for the organism in question manifested by symptoms and signs. Smith B, Kumar A, Ceusters W, Rosse C. On carcinomas and other pathological entities. Comparative and Functional Genomics, Volume 6, Issue 7 -8 (October - December 2005), p 379 -387.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some origines of pathological structures • At molecular level – Defective piece of DNA (“gene”) – Transcription mistakes or mutations • Random event • Environmental effect (radiation, chemical, virusses. . . ) Results in cellular dysfunction or structural damage • At higher levels of granularity – Traumatic Results in cellular, organ or system damage

R T U New York State Center of Excellence in Bioinformatics & Life Sciences ‘Ontology inspired’ definition • An organism (or part of an organism) is diseased if and only if – it includes among its parts pathological anatomical structures which compromise the organism’s physiological processes to the degree that they give rise to symptoms and signs. • An anatomical structure is pathological whenever: – it has come into being as a result of changes in some pre-existing canonical anatomical structure – through processes other than the expression of the normal complement of genes of an organism of the given A clever tricktype, forandnot having us to – is predisposed to have health-related consequences for the organism in commit at that question manifested by symptoms andtime signs. what a disease IS. Smith B, Kumar A, Ceusters W, Rosse C. On carcinomas and other pathological entities. Comparative and Functional Genomics, Volume 6, Issue 7 -8 (October - December 2005), p 379 -387.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences 1 st tentative definition of disease a disposition which, when realized, has health-related consequences for the wellbeing of the organism or any of its parts.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Somewhat problematic • Notion of “well-being” easily leads into circularity when associated with ‘non-diseased’. • ‘having health-related consequences’ may also be positive. • It says very little about the nature of the disposition.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences 2 nd tentative definition of disease A disease is a disposition acquired by a relatively isolated causal system (RICS) to have its external membrane or covering damaged through the coming into existence of a pathological structure.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences An entity is a relatively isolated causal system iff (1) • The entity has an external boundary which is established via a physical covering or membrane which extends continuously across all or almost all of its surface. • The events transpiring within the entity have characteristic magnitudes which either fall or fall not within a certain spectrum of allowed values. The latter are distinguished by the fact that they will, in cumulation, lead to the entity’s ceasing to exist. Barry Smith and Berit Brogaard. Sixteen Days. Journal of Medicine and Philosophy 2003, Vol. 28, No. 1, pp. 45– 78

R T U New York State Center of Excellence in Bioinformatics & Life Sciences An entity is a relatively isolated causal system iff (2) • The external membrane or covering serves as a shield to protect the entity from those causal influences deriving from its exterior which are likely to give rise to events which are outside its spectrum of allowed values. • The entity contains within itself its own mechanisms which are able to maintain (or, in cases of disturbance, to re-establish) sequences of events falling within the spectrum of allowed values. The entity also contains within itself mechanisms for reconstituting or replacing its external membrane or covering in case of damage. Barry Smith and Berit Brogaard. Sixteen Days. Journal of Medicine and Philosophy 2003, Vol. 28, No. 1, pp. 45– 78

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Entities qualifying as relatively isolated causal system • Organisms – Human beings – Unicellular organisms –. . . • Certain parts of organisms – Cells – Organ systems – Some organs – Maximal portions of tissues Opens the possibility for an entity to be a disease of a body part without being a disease of the organism of which that body part is a part, yet being a disease in that organism.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences But the definition requires working out: • At the level of biomedical reality: – What is a disease ? – What are the entities that cause a specific disease to exist or influence its course ? – What are the relationships between these entities and the disease ? • At the level of representational artifacts: – How do they relate to reality ? – How keeping track of changes in reality ?

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Part 4: Diagnostic decision making from an ontological and epistemological perspective How to come to a diagnosis ?

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some common notions given an ontological flavor • Clinical picture: – A spatiotemporal entity – Comparable to a theatre play • Which has temporal parts which form the course of the disease • Which has spatial parts which are the pathological formations and pathological anatomical structures. • Pathological entity: – Any part of a clinical picture. • Pathological process: – an organismal process that contributes to the dysfunctioning of the organism or any of its parts.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Disease and course of disease processual entity disposition Joe’s disease ? actualized in caused by Joe’s disease’s course part-of Creations of pathological formations and of pathological anatomical structures in Joe

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some counterintuitive aspects (1) • If in a RICS a disposition is realized in such a way that it can’t get that type of disease anymore (because the RICS will certainly die or becomes immune), then the “disease” would have come to an end, but it’s course not. • We tend to say “the disposition for the disease”, while the disposition IS the disease. Thus perhaps better (? ): ‘Disease’ ‘diseasedisposition’

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some counterintuitive aspects (2) • The creations of pathological formations and of pathological anatomical structures in Joe lead to new dispositions but are these dispositions distinct diseases ? – A virus enters one cell in Joe 1 disease ? – The virus enters a second cell 2 nd disease ? • … But ! 1) The absence of counterintuitive notions associated with colloquial use of the word ‘disease’ is an illusion. They are there, but unnoticed. 2) Word usage is such that language treats disease as continuant.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Beliefs about diseases and clinical pictures (1) • A particular person may exhibit particular pathological entities. • A particular clinical picture starts to exist at the time the first pathological process that is a realization of a particular disease starts to exist. • A particular clinical picture ceases to exist when the last pathological process that is a realization of a particular disease comes to an end, and when there are no more pathological formations or pathological anatomical structures that were formed by the course of the disease.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Beliefs about the relevant portion of reality (2) • A particular disease exists in a particular person before a particular clinical picture is present in that person. • There can be no clinical picture without a disease. • A person may exhibit different particular diseases during his lifetime, some or all of them of the same disease type, and some or all of them at the same time or during overlapping time spans.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Beliefs about diseases and clinical pictures (3) • A clinical picture can continue to exist although its disease ceased to exist: – Remaining scars – Healthy carriers of virusses • Instances of the same disease may lead to nonsimilar clinical pictures, even in the same person.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Beliefs about the relevant portion of reality (4) • The course of some particular disease may lead to other particular diseases of the same or different disease types in a particular person. – Varicella (? ) as a child, zoster as an adult • A particular disease of type A in one person may lead to a particular clinical picture of type B, while a particular disease of that same type A in another person may lead to a clinical picture of a totally different type. • …

R T U New York State Center of Excellence in Bioinformatics & Life Sciences On ‘findings’ and ‘observations’ (1) • • • A particular pathological entity may at a certain time be undetectable by any observation method or technique available to an observer, including the person exhibiting the pathological entity itself. A particular observation may produce false results and thus simulate the existence of a pathological entity. An observer may observe or fail to observe a detectable particular pathological entity.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences On ‘findings’ and ‘observations’ (2) • When an observer perceives a particular pathological entity, he might judge it – (1) to be an instance of the universal of which it is indeed an instance in reality, – (2) to be an instance of another universal (and thus be in error), or – (3) he might be not able to make an association with any universal at all. • Distinct manifestations of ‘the same type’ may be pathological or not: – Singing naked under the shower versus in front of The White House • . . .

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Part 5: Making diagnoses in an ever changing world

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 Reality p 3 IUI-#3 Belief O-#0 O-#2 O-#1 = “refers to” = what constitutes the meaning of representational units …. Therefore: O-#0 is meaningless

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 Total ignorance

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 False belief in the existence of a type, e. g. unicorns Note: happens also at the level of particulars: e. g. the planet ‘Vulcan’

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 The coming into existence of a new universal remains unnoticed: ‘AIDS’ existed before being discovered.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 The coming into being of a new particular remains unnoticed: e. g. John Doe’s colonic polyp, which from that time on, is an instance of U 1.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 An advance in science: the existence of U 1 is acknowledged.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 The existence of John Doe’s benign colonic polyp is discovered, however, without being recognized as such. Rather, it is believed to be an instance of what in reality is a fantasy.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 Another advance in science: the ‘concept’ O-#0 is rightfully abandoned, necessitating therefor to reconsider of what p 3 must be believed to be an instance of.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 It is, rightfully believed that p 3 is an instance of U 1. It raises, amongst other things, the question to what point in history this belief can be extended.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 p 3 changes from a benign into a malignant tumor, at a time that science did not discover malignancy yet. p 3 is now wrongly believed to be an instance of U 1.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 Advance in science: “malignancy” is discovered. However, that it applies to John Doe’s polyp has not yet been noticed.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 John Doe’s polyp becomes recognized as an instance of a malignant tumor.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 John Doe’s polyp was irradiated and believed to have vanished, while in reality, it isn’t.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 John Doe is lucky: his tumor indeed disappeared. His physicians who believed it was already gone, are lucky also: they escape a law suite.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Reality versus beliefs, both in evolution t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 For ‘utilitarian’ reasons, the “pragmatic engineers” remove malignant tumors from their ontology: if it is not believed to exist, you can’t get law suites for failures in recognizing instances.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some other possible situations t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 A particular is believed to exist longer than it really does. e. g. “Elvis is not dead”, or the innumerous EHRs that state the patient taking some drug while he stopped.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Some other possible situations t U 1 U 2 R p 3 IUI-#3 B O-#0 Artifacts on radiographic images O-#2 O-#1

R T U New York State Center of Excellence in Bioinformatics & Life Sciences And in this, I thus far ignored … t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1 Relationships amongst universals (R) or beliefs therein (B)

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Mistakes, discoveries, being lucky, having bad luck Mistakes t U 1 U 2 R p 3 IUI-#3 B O-#0 O-#2 O-#1

R T U New York State Center of Excellence in Bioinformatics & Life Sciences A crucial question … • Do we have some means to assess how good we are doing in our understanding of reality ? • Some might argue: no, because every representational unit will always rest on a belief ! • My belief is: yes, if one keeps track of the reasons, in function of the three levels discussed earlier, for which one’s beliefs have changed.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Part 6: tracking changes in representations Ceusters W, Smith B. Towards A Realism-Based Metric for Quality Assurance in Ontology Matching. Forthcoming in Proceedings of FOIS-2006, Baltimore, Maryland, November 9 -11, 2006. Ceusters W, Smith B. A Realism-Based Approach to the Evolution of Biomedical Ontologies. Forthcoming in Proceedings of AMIA 2006, Washington DC, November 11 -15, 2006.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Remember our 3 fundamentally different levels 1. the reality on the side of the patient; 2. the cognitive representations of this reality embodied in observations and interpretations on the part of clinicians and others; 3. the publicly accessible concretizations of such cognitive representations in representational artifacts of various sorts, of which ontologies and terminologies are examples.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Example: a person (in this room) ’s phenotypic gender • Reality: – Male – Female Other types of phenotypic gender ? • Cognitive representation – [male] – [female] • In the EHR: – “male” – “female” – “unknown”

R T U New York State Center of Excellence in Bioinformatics & Life Sciences The specification bias in ontology evolution (1) • Ontology versions exhibit differences of the following sorts: – – – Add a subtree Delete a subtree Move a subtree to a different location Move a set of sibling classes to a different location Create a new abstraction and move a set of siblings down in a class hierarchy, creating a new superclass. – Delete a class, moving its subclasses to become subclasses of its superclass. – Split a class Noy, N. F. , Kunnatur, S. , Klein, M. , and Musen, M. A. Tracking changes during ontology evolution. In Proceeding – Merge classes of the 3 rd International Semantic Web Conference (ISWC 2004), Hiroshima, Japan, November 2004.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences The specification bias in ontology evolution (2) Yaozhong David Liang. Enabling Active Ontology Change Management within Semantic Webbased Applications Ph. D thesis October 2, 2006

R T U New York State Center of Excellence in Bioinformatics & Life Sciences They don’t care about the reasons for the changes • • changes in the underlying reality (does the appearance or disappearance of an entry in a new version of an ontology relate to the appearance or disappearance of entities or of relationships among entities? ); changes in our scientific understanding; reassessments of what is relevant for inclusion in an ontology; encoding mistakes introduced during ontology curation (for example through erroneous introduction of duplicate entries reflecting lack of attention to differences in spelling).

R T U New York State Center of Excellence in Bioinformatics & Life Sciences An “optimal” ontology (1) • Because ontologies, as conceived on realist terms, – are artifacts created for some purpose (e. g. to serve as controlled vocabulary, or to provide domain knowledge to a software application), – are at the same time intended to mirror reality, – should allow reasoning which is efficient from a computational point of view, • we argue that an optimal ontology should constitute a representation of all and only those portions of reality that are relevant for its purpose.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences An “optimal” ontology (2) • Each term in such an ontology would designate – (1) a single portion of reality (POR), which is – (2) relevant to the purposes of the ontology and such that – (3) the authors of the ontology intended to use this term to designate this POR, and – (4) there would be no PORs objectively relevant to these purposes that are not referred to in the ontology.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences But things may go wrong … • assertion errors: ontology developers may be in error as to what is the case in their target domain; • relevance errors: they may be in error as to what is objectively relevant to a given purpose; • encoding errors: they may not successfully encode their underlying cognitive representations, so that particular representational units fail to point to the intended PORs.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Key requirement for versioning Any change in an ontology or data repository should be associated with the reason for that change to be able to assess later what kind of mistake has been made !

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Example: a person (in this room) ’s gender in the EHR • In John Smith’s EHR: – At t 1: “male” at t 2: “female” • What are the possibilities ? • Change in reality: • transgender surgery • change in legal self-identification • 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)

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Ways representational units do or do not refer 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.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Ways representational units do or do not refer OE/BE value pairs Y/Y: correct assertion of the existence of a POR; Y/N: lack of awareness of a POR, reflecting an assertion error; N/N: correct assertion that some putative POR does not exist ; N/Y: the false belief that some putative POR exists.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Ways representational units do or do not refer Ref. : manner in which the expression refers R+: the encoding of the belief is correct R: the encoding is incorrect because it does not refer R-: it does refer, but to a POR other than the one which was intended.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Possible evolutions through versions

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Possible evolutions through versions An entity ceases to exist, but the representation is not updated:

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Updating is an active process • authors assume in good faith that all included representational units are of the P+1 type, and all they are aware of, but not included, of A+1 or A+2. • If they become aware of a mistake, they make a change under the assumption that their changes are also towards the P+1, A+1, or A+2 cases. • Thus at that time, they know of what type the previous entry must of have been under the belief what the current one is, and the reason for the change.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences This leads to a calculus … • NOT: – to demonstrate how good an individual version of an ontology is, • But rather – to measure how much it is believed to have been improved (hopefully) as compared to its predecessors.

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Chains of changes Belief at t reality at t+1: unintended encoding corrected If assumed that this is correct, then the unit at t must have been of type P-4, rather than P+1 a gain of +1 is assumed at t+2: the unit is assumed to refer to nothing, thus is removed from the representation assumption at t+2 then at t+1 it was then at t it was

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Additional goodie • Measures also the quality/skills of the ontology / repository authors

R T U New York State Center of Excellence in Bioinformatics & Life Sciences Can this be implemented ? • Manual burden: documenting the reason for a change clicking one radio button. – Note: additions require no extra effort • The change of belief revisions is computable • The proof of the pudding is in the eating: – Would SNOMED be a candidate (new grant) – For PHRs (one proposal written)