Patterns: a simple but expressive data modelling formalism

Abstract

The creation of a clinical application requires models that describe the structure of data in a way that can be displayed, exchanged and stored. A number of approaches for this have been proposed and are in widespread use. However, these are often complex and/or have shortcomings in the breadth of data that they are able to represent. The annotations facility provided by many computer languages could be used to include information shaping the development and run-time behaviour of a clinical application. If this were comprehensive, then annotations alone would be sufficient for conceptual modelling. A model for representing such annotations is presented and some examples shown and discussed. The paper concludes that such a formalism is simple to use while developing semantic concepts but is capable of representing information from many models simultaneously. It is well suited to the needs of clinical teams seeking consensus on the structure of records.