Coordinating taxonomies: Key to re-usable concept representations

The means and consequences of this strategy as implemented in the GALEN project are explored, discussing both the strengths — providing greater detail, greater computer support, and avoiding many arguments which are endemic in discussions of classification systems — and the limitations intrinsic in such a formal approach.

Abstract

A unified controlled medical vocabulary has been cited as one of the grand challenges facing Medical Informatics. We would restate this challenge as 'achieving a reusable and application-independent representation of medical concepts.' Achieving a reusable representation of medical concepts is a pre-requisite for meeting two key strategic goals of the next decade of the development in medical informatics: interoperability and cumulative development. A key strategy for achieving re-usability is to separate concepts into their component parts, organise those parts in nearly pure hierarchies, and then recombine into composite representations which can be classified flexibly and automatically. This paper explores the means and consequences of this strategy as implemented in the GALEN project. It discusses both the strengths — providing greater detail, greater computer support, and avoiding many arguments which are endemic in discussions of classification systems — and the limitations intrinsic in such a formal approach.