Re: [COI] Clinical Observations Interoperability Telcon January 15 2008

[Samson Tu]

Kashyap, Vipul wrote:   From: Samson Tu [mailto:[EMAIL PROTECTED]] In the SAGE project, the approach we took was that (1) patient data are represented by information model entities derived from HL7 RIM, (2) legal patient data are constrained by DCM, (3) standard terminologies such as SNOMED CT provides codes used in instances of information model entities, (4) criteria are logical expressions composed from HL7 data types, variables, functions, and queries on information structures.We had a collection of templates for logical expressions that correspond to stereotypical GELLO expressions.   [VK] Is there a write up or document discussing the above approach, hopefully with some illustrative examples?  Also, would it be possible for you to present these ideas on an upcoming COI call?  Tu SW, Campbell JR, Glasgow J, Nyman MA, McClure R, McClay J, Parker C, Hrabak KM, Berg D, Weida T, Mansfield JG, Musen MA, Abarbanel RM, The sage guideline model: Achievements and overview. J Am Med Inform Assoc 2007. 14(5): 589-598. See the "Standard Terminology, Patient Information Model, and Clinical Expressions" section. More details can be found in http://sage.wherever.org/references/docs/SAGEGuidelineModelSpec1.65.pdf [VK]  [Open-world assumption] is an interesting issue and maybe we could evolve some best practices on "open world issues in eligibility criteria". The current approach I have seen is proving that no model exists for the negation of Exclusion criteria. Not sure whether this is generalizable.  The problem is that sometimes you want to make OWA sometimes not. In my past work on eligibility criteria, I wrote specialized reasoners to deal with the cases. In terms of best practices, I think what you  want is to show that (appropriately formulated)  Exclusion criteria not true.   Also, the terminology - information model distinction appears to be an artificial one, at least from the OWL/DL perspective. There is no reason why both of them cannot be represented in an OWL reasoner. Representation in OWL is not the issue. The relationship between terminology & information model is. Alan Rector's paper talks about the relationship. Samson -- Samson Tu email: [EMAIL PROTECTED] Senior Research Scientist web: www.stanford.edu/~swt/ Center for Biomedical Informatics Research phone: 1-650-725-3391 Stanford University fax: 1-650-725-7944