1 / 68

The Scientific & Technical Foundations of SNOMED Clinical Terms

The Scientific & Technical Foundations of SNOMED Clinical Terms. Kent A. Spackman, MD PhD Scientific Director, SNOMED. October 2, 2006. The Scientific & Technical Foundations of SNOMED Clinical Terms Overview. Codes and classifications Basic principles of design for a clinical terminology

claude
Download Presentation

The Scientific & Technical Foundations of SNOMED Clinical Terms

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The Scientific & Technical Foundations of SNOMED Clinical Terms Kent A. Spackman, MD PhD Scientific Director, SNOMED October 2, 2006

  2. The Scientific & Technical Foundations of SNOMED Clinical TermsOverview • Codes and classifications • Basic principles of design for a clinical terminology • evolution & convergence • Clinical statements & information models • Language, meaning, and ontology • Description logic & terminology models • Curation & change management • Technical standardization for • distribution, interchange & reference

  3. The Scientific & Technical Foundations of SNOMED Clinical TermsOverview • Codes and classifications • Basic principles of design for a clinical terminology • evolution & convergence • Clinical statements & information models • Language, meaning, and ontology • Description logic & terminology models • Curation & change management • Technical standardization for • distribution, interchange & reference

  4. Codes • Most people have experience with coding systems where: • The primary purposes of coding are reimbursement and epidemiology • Coding is done using a pre-existing record • It follows rules designed to help you choose the one right code that best expresses the main diagnosis, or the main procedure done. • Double counting - for epidemiology or reimbursement - is a bad thing

  5. Codes for electronic health records • Coding is done without a pre-existing record, but rather to express statements about the patient, in order to create a coded record • Double counting is a good thing whenever the statement has fidelity to the clinical situation Roger Côté

  6. The Scientific & Technical Foundations of SNOMED Clinical TermsOverview • Codes and classifications • Basic principles of design for a clinical terminology • evolution & convergence • Clinical statements & information models • Language, meaning, and ontology • Description logic & terminology models • Curation & change management • Technical standardization for • distribution, interchange & reference

  7. SNOMED’s approach • Evolutionary design • Convergence • Authoritative sources should be integrated, not left scattered and incompatible • Clinicians determine meaning • URU understandable, reproducible, usable • Terminology model must interact with information models and knowledge bases

  8. Evolutionary design • Campbell, K, Cohn, S, Chute, C, Rennels, G, and Shortliffe, E. Gálapagos: Computer-based support for evolution of a convergent medical terminology. in AMIA Fall Symposium. 1996, pp. 269-273. • Campbell, K, Scalable methodologies for distributed development of logic-based convergent medical terminology. Methods of Information in Medicine, 37:426-439,1998. Keith Campbell

  9. Evolutionary Design (1) • Evolution without pre-ordained design • Accumulation of desirable features • Evidence-based assessment • All of the scientific and technical foundations are subject to revision based on best evidence • Graceful evolution • “graceful” implies the preservation of the value of encoded data • Heterogeneity of perspectives

  10. Evolutionary Design (2) • Dealing with Scale • Participatory consensus-based approach • Experts must be involved • Semantics-based concurrency control • Description logic underpinnings • Configuration management tools • Exemplified by the “Galapagos” tool set

  11. Convergence:Source of Concept Codes in SNOMED CT

  12. URU criteria for evolutionary changes • Definitions should be Understandable by average clinicians, given brief explanations • We assess understandability by examining Reproducibility • We can ignore distinctions for which there is no Use in health care

  13. `I don't know what you mean by "glory,"' Alice said. Humpty Dumpty smiled contemptuously. `Of course you don't -- till I tell you. I meant "there's a nice knock-down argument for you!"' `But "glory" doesn't mean "a nice knock-down argument,"' Alice objected.`When I use a word,' Humpty Dumpty said in rather a scornful tone, `it means just what I choose it to mean -- neither more nor less.' `The question is,' said Alice, `whether you can make words mean so many different things.' `The question is,' said Humpty Dumpty, `which is to be master - - that's all.' From Through the Looking Glass, Lewis Carrol

  14. Clinicians Determine Meaning • SNOMED is not the language police • We are not trying to tell people which words they should or should not use that’s up to the clinical professions • E.g. Should dermatologists still use the term “pyogenic granuloma” for the small blood vessel tumor that is neither pyogenic nor a granuloma? • We also are not trying to tell clinicians how to operationally apply meanings to specific cases again, that is up to the clinical professions • E.g. What systolic and diastolic limits should be used for determining whether a patient has hypertension 140/90 ?

  15. Integration of Terminologies • SNOMED is also not duplicating the work of established official consensus groups • Instead, it is providing an integrated resource where the various terminologies are available for electronic interoperability applications. Examples: • The International Society for Blood Transfusion ISBT provides a set of codes and names for red cell antigens and their antibodies • The WHO periodically revises numerous classifications of malignant neoplasms

  16. The Scientific & Technical Foundations of SNOMED Clinical TermsOverview • Codes and classifications • Basic principles of design for a clinical terminology • evolution & convergence • Clinical statements & information models • Language, meaning, and ontology • Description logic & terminology models • Curation & change management • Technical standardization for • distribution, interchange & reference

  17. Statements as the foundation of EPR • The electronic patient record is fundamentally about what is said (statements). • Rector AL, Nowlan WA, Kay S. Foundations for an Electronic Medical Record. Meth Inform Med 30:179-86, 1991. • Rector AL, Nowlan WA, Kay S, Goble CA, Howkins TJ. A framework for modelling the Electronic Medical Record. Meth Inform Med 32:109-19, 1993.

  18. Clinical Decision Support Model + Inference Rules Interface Interface Interface Information Model + Patient Data Structures Terminology Model + Compositional Expressions Diagram based on Figure 1 in Rector AL et al. “Interface of Inference Models with Concept and Medical Record Models” AIME 2001: 314-323

  19. Clinical Decision Support Model + Inference Rules IFTwoblood cultures, drawn through Antibiotic removal device, more than 30 minutes apart, grows no organism, THENdiscontinueantibiotic. Interface Interface Interface HL7 RIM SNOMED CT Information Model + Patient Data Structures Terminology Model + Expressions What test was performed? How many were done? At what time? What device was used? What was the result of the test? 30088009 blood culture 55512120 antibiotic removal device 264868006 No growth 281789004 antibiotic therapy 223438000 advice to discontinue a procedure Diagram based on Figure 1 in Rector AL et al. “Interface of Inference Models with Concept and Medical Record Models” AIME 2001: 314-323

  20. HL7 Clinical statement model extract ActRelationship Informant Observation Patient id Performer Substance Administration Author RelatedEntity DataEnterer Supply AssignedEntity Verifier Procedure Encounter RecordTarget Patient id Act RelatedEntity Organizer Subject Specimen ActReference

  21. The Scientific & Technical Foundations of SNOMED Clinical TermsOverview • Codes and classifications • Basic principles of design for a clinical terminology • evolution & convergence • Clinical statements & information models • Language, meaning, and ontology • Description logic & terminology models • Curation & change management • Technical standardization for • distribution, interchange & reference

  22. Ogden-Richards semiotic triangleas a foundation for understanding terminologies

  23. Cimino’s “desiderata” 1998 • Cimino, J, Desiderata for controlled medical vocabularies in the twenty-first century. Methods of Information in Medicine, 37:394-403,1998. • Requirements that are a consequence of assuming the terminology is multi-purpose Jim Cimino

  24. Codes, concepts and meanings • One meaning per code • Concepts themselves are in people’s heads. • And the things entities that the concepts reference via their meaning are in the real world. • No duplicates • No ambiguous meanings • But the meaning of a code may be at a very broad level of generality • In talking about the terminology, we sometimes are sloppy in saying “concept” when we mean “code” or “meaning of a code”

  25. Formal ontological analysis • Rector AL, Rogers J, “Patterns, Properties and Minimizing Commitment: Reconstruction of the GALEN Upper Ontology in OWL",  Core Ontologies Workshop, EKAW 2004. • Smith B, Ceusters W, Kumar A, Rosse C, “On Carcinomas and Other Pathological Entities”, Comparative and Functional Genomics 6 (7-8): 379 - 387, 2005. • Rosse, C, et al, “A Strategy for Improving and Integrating Biomedical Ontologies”, Proceedings of AMIA Symposium 2005, Washington DC, 639–643.

  26. Formal ontological analysis • Benefit: • Better definitions • More reproducibility • Possible danger: • Introduction of non-useful distinctions • Proliferation of codes

  27. Some ontological distinctionsof value for refining SNOMED CT • Continuant vs occurrent • Discrete vs mass • Individual vs collection/group • others

  28. The Scientific & Technical Foundations of SNOMED Clinical TermsOverview • Codes and classifications • Basic principles of design for a clinical terminology • evolution & convergence • Clinical statements & information models • Language, meaning, and ontology • Description logic & terminology models • Curation & change management • Technical standardization for • distribution, interchange & reference

  29. Logic-based definitionsand the terminology model • Conceptual graphs • K. E. Campbell, A. K. Das, & M. A. Musen. A Logical Foundation for Representation of Clinical Data. Journal of the American Medical Informatics Association 13:218-232, 1994. • Description logic • Spackman KA, Campbell KE, and Côté RA, SNOMED-RT: A reference Terminology for Health Care. Journal of the American Medical Informatics Association Symposium special issue: pp. 640-644, 1997. • Read code semantic definitions • Price C, O’Neil M, Bentley TE, Brown PJB. Exploring the ontology of surgical procedures in the Read Thesaurus. Meth Inform Med 37:420-425, 1998. • Brown PJB, O’Neil M, Price C. Semantic definition of disorders in Version 3 of the Read Codes. Meth Inform Med 37:415-419, 1998.

  30. Description Logic FoundationIn a Nutshell: • SNOMED CT’s formal concept definitions are based on a description logic with the following features: • Conjunction • Existential restrictions • Role hierarchies • Plus “role groups” • Plus role composition “transfers-thru” or “right identity” or “restricted role-value maps” • Currently: • direct-substance o has-active-ingredient

  31. DL Basics • Concepts are given formal definitions • e.g. “Red car” is a kind of “Car”, and has color “Red” • e.g. “Lung disorder” is a kind of disorder, and has site “Lung” • Definitions are expressed in description logic • conjunction logical “and” п • existentially quantified role restrictions R.C • red_car = car п  color.red • lung_disorder = disorder п  site.lung

  32. Concept & role forming operators& terminological axioms = not used in SNOMED CT

  33. Evolution of SNOMED’s DL Notation follows Donini in Ch.3 Description Logic Handbook Except here omeans restricted role-value maps

  34. Need for Role Groups • When a single concept may have more than one value for a particular attribute • for example, “bone fusion with tendon transfer” • method = fusion, site = bone, and • method = transfer, site = tendon • And, one attribute-value pair needs to be associated with another. • How can we specify that the fusion is done to the bone and not to the tendon? and that the transfer is done to the tendon and not to the bone?

  35. Role Grouping Logical Form: A Nested Existential Restriction • 9RG:(9R1:C1⊓ 9R2:C2) ⊓ 9RG:(9R3:C3) • Distributed as rows in relationships table: CR3C3 0 CR1C1 1 CR2C2 1

  36. Right Identity Other names: “restricted role-value maps” “propagates via” “transfers-thru” “role composition” R S⊑ R xRy ⋀ySz →xRz • femurFracture site femur • headOfFemurFracture site headOfFemur • headOfFemur part-of Femur • Allows the automated inference that: • headOfFemurFracture is-a FemurFracture

  37. Description logic for anatomy Avoiding right identity by using SEP triples Liver Structure XM0Ps Liver structure T-62000 Liver is-a is-a part-of Liver Part Entire liver T-D0535 Liver part 7N330 Liver is-a Schulz, Romacker and Hahn: “Part-whole reasoning in medical ontologies revisited: Introducing SEP triplets into classification-based description logics”, Proc. AMIA Fall 1998. Lobe of liver

  38. SNOMED RT T-62000 Liver T-D0535 Liver part CTV3 XM0Ps Liver structure 7N330 Liver Example: Liver • How is T-62000 used in pathology? • The site of tumors and other morphologies • Does it necessarily mean the entire liver? • No • In what sense is 7N330 Liver a subtype of XM0Ps Liver structure? • XM0Ps means “the entire liver OR some part of the liver” • 7N330 means “entire liver”..

  39. Role Hierarchies • Equipment • Direct-device • Indirect-device • Using • Access-instrument

  40. What DL classifiers can we use ? • 1996: Kaiser Convergent Medical Terminology Project together with College of American Pathologists CAP • Kaiser Colorado used K-REP description logic classifier from IBM • 1999-present: UK NHS and CAP • K-REP developers moved on to Apelon Corp., developed Ontylog classifier • Recent successful experiments with other classifiers • 2005: Cel+ • 2006: Fact++ • ?others

  41. SNOMED’s Concept Model • Top-level categories & FSN suffixes • Attributes

  42. Suffixes Top Level Suffixes administrative concept body structure situation environment / location event finding linkage concept observable entity organism physical force physical object procedure product qualifier value record artifact social concept special concept specimen staging scale substance

  43. Suffix subtypes Parent SuffixChild Suffix body structure cell structure body structure cell body structure morphologic abnormality cell structure cell environment / location environment environment / location geographic location finding disorder linkage concept attribute linkage concept link assertion morphologic abnormality cell person occupation physical object product procedure regime/therapy qualifier value administrative concept social concept ethnic group social concept life style social concept occupation social concept person social concept racial group social concept religion/philosophy special concept inactive concept special concept namespace concept special concept navigational concept staging scale assessment scale staging scale tumor staging Suffix subtypes

  44. Number of attributes relationship types in the SNOMED concept model

  45. Clinical finding attributes Acquired body structure, Anatomical concepts Finding site Associated morphology Morphologically abnormal structure Clinical finding, Substance, Physical object, Physical force, Events, Organisms, Pharmacological / Biological product, Procedure Associated with Clinical finding After Clinical finding, Procedure, event Due to Clinical finding, Event Organism, Substance, Physical object, Physical force Causative agent Has interpretation Findings values, Result comments Laboratory procedure, Observable entity, Patient evaluation procedure Interprets

  46. Clinical finding attributes cont. First episode, New episode, Ongoing episode Episodicity Pathological process Pathological process Sudden, Gradual Onset Course Courses Clinical finding Has definitional manifestation Clinical finding Occurrence Periods of life Mild, Moderate, Severe Severity Procedure Finding method Performer of method, Subject of record Provider of history other than subject, Subject of record or other provider of history Finding informer

  47. Examples of procedure attributes • Procedure site • Specifies the site affected by a procedure • Relates “Procedure” and “Body structure” Aortic aneurysm repair procedure Procedure-site: Aortic structure body structure

  48. Examples of procedure attributes • Method • Specifies the action being performed to accomplish the procedure Incision of ureter procedure Method: Incision - action qualifier value

  49. Procedure attributes Procedure site Acquired body structure, Anatomical concepts Procedure site - direct Acquired body structure, Anatomical concepts Procedure site - indirect Procedure Morphology Morphologically abnormal structure Indirect morphology Direct morphology Procedure device Device, Physical force Procedure Direct device Device Indirect device Device Using Device, Physical force Access instrument Endoscope and subtypes Method Action Direct substance Substance, Pharmaceutical / biologic product

  50. Procedure attributes proposed 2007 Procedure site Acquired body structure, Anatomical concepts Procedure site - direct Acquired body structure, Anatomical concepts Procedure site - indirect Procedure Morphology Morphologically abnormal structure Indirect morphology Direct morphology Procedure device Procedure Device Direct device Indirect device Using device Endoscope and subtypes Using access device Method Action Direct substance Substance, Pharmaceutical / biologic product

More Related