Interface of Inference Models with Concept and Medical Record Models

Medical information systems and standards are increasingly based on principled models of at least three distinct sorts of information – patient data, concepts (terminology), and guidelines (decision support). Well defined interfaces are required between the three types of model to allow development to proceed independently. Two of the major issues to be dealt with in the defining of such interfaces are the interaction between ontological and inferential ions – how general notions such as ‘abnormal cardiovascular finding’ are abstracted from concrete data – and the management of the meaning of information in guidelines in different contexts. This paper explores these two issues and their ramifications.

• Type: Conference
• Year: 2001
• Where: AIME
• Authors: Alan L. Rector, Peter D. Johnson, Samson W. Tu, Chris Wroe, Jeremy Rogers
• Keywords: AIME 2001, Artificial Intelligence, Distinct Sorts, Medical Information Systems, ‘abnormal Cardiovascular Finding