Saying what you mean, meaning what you say, sharing and re-using what has been said

Abstract

For describing patient data and medical knowledge there are in principle two alternatives to say what is meant – as free text on the one hand or in a more structured way on the other hand. In the case of free text, natural language processing systems try to determine what is meant when something is said. In the case of structured data entry or computer based knowledge processing, the objective is to control what is meant when something is said. The latter is especially important for a computer-based processing and interpretation of language and even more for a multiple usability of patient data and knowledge. Therefore those applications should be linked explicitly to an independent terminological component. Controlled taxonomic vocabularies like SNOMED, ICD and MeSH should be used in that context. But each of them has been developed for special purposes and primarily for manual use. For advanced medical data processing, a compositional approach like the one in the GALEN-project has been proposed. This is one answer to the question How to share and re-use what has been said?, at least from the semantic point of view. There are further – especially ontological – issues that should be addressed.