Recording Associated Disorders Using SNOMED CT Ronald Cornet , - PowerPoint PPT Presentation

Recording Associated Disorders Using SNOMED CT Ronald Cornet , Nicolette de Keizer Dept. of Medical Informatics Academic Medical Center – University of Amsterdam r.cornet@amc.uva.nl Oral presentation at August 31, 2011 – Oslo

What it has to do with the theme of the conference: User Centered Networked Health Care MIE2011 - Ronald Cornet - 2

Patient Centered Health Care u Patient history § Chronic condition: Type II diabetes mellitus § Complication: neuropathic arthropathy MIE2011 - Ronald Cornet - 3

Patient Centered Health Care Physician Centered Health Care MIE2011 - Ronald Cornet - 4

Physician Centered Health Care u Patient history § Chronic condition: Type II diabetes mellitus § Complication: neuropathic arthropathy u Diabetologist § Type II diabetes mellitus with neuropathic arthropathy u Neurologist § diabetic neuropathic arthropathy u Rheumatologist § arthropathy associated with a neurological disorder MIE2011 - Ronald Cornet - 5

4 views on one situation u Not a problem per se u But … How to deal with this in “User Centered Networked Health Care” ? MIE2011 - Ronald Cornet - 6

Consistency of Views u Any combination of these findings can occur P. Type II diabetes mellitus + neuropathic arthropathy D. Type II diabetes mellitus with neuropathic arthropathy N. Diabetic neuropathic arthropathy R. Arthropathy associated with a neurological disorder P 1 P 2 P 3 P 4 P 5 P D N D R N R MIE2011 - Ronald Cornet - 7

Inconsistency of inference SNOMED CT enables inference u Type II diabetes mellitus + neuropathic arthropathy = Diabetes mellitus type 2 + Arthropathy u Type II diabetes mellitus with neuropathic arthropathy = Diabetes mellitus type 2 u Diabetic neuropathic arthropathy = Arthropathy + Diabetic complication u Arthropathy associated with a neurological disorder = Arthropathy MIE2011 - Ronald Cornet - 8

Diabetic neuropathic arthropathy MIE2011 - Ronald Cornet - 9

MIE2011 - Ronald Cornet - 10

Problem u 1 situation; 4 views u 4 inferences u None of these explicitly captures presence of “neurological disorder” § Only “neuropathic” or “associated with a neurological disorder” MIE2011 - Ronald Cornet - 11

Magnitude of the problem Lexical Logical disorder concepts formally defined disorder concepts described by with associated with , due to, means of ‘associated with’, and/or after relations to disorders. ‘due to’ and/or ‘after’ 2,804 2,981 780 Some 27% overlap MIE2011 - Ronald Cornet - 12

What should be inferred u The patient has diabetes mellitus type II associated associated with with u The patient has arthropathy associated with u The patient has disorder of nervous system MIE2011 - Ronald Cornet - 13

Challenge: finding solutions 1. Leave it to reasoning process § E.g., if patient P has Disease D 1 , and D 1 associated with Disease D 2 , then P has D 2 2. Store concept references and their interrelations 3. Store instance references and their interrelations MIE2011 - Ronald Cornet - 14

Solutions Pro Con Reasoner Simple modeling Complex retrieval and inference Simpler retrieval Need to tackle Concepts and inference terminology binding Instances Simplest retrieval Need to tackle and inference terminology binding + referent tracking MIE2011 - Ronald Cornet - 15

Conclusion u Disorders which are associated to other disorders can not be consistently recorded with different “views” u This poses a problem for retrieval/inference u It can be solved by explicit recording of constituting concepts or instances u This requires terminology binding (and referent tracking ) MIE2011 - Ronald Cornet - 16