Two of the most important issues in l!leveloping data-driven medical decision support systems are the choice of knowledge representation and the methods and technique for integration with the patient database. The main purpose of the Arden Syntax for Medical Logic Modules, which is the method of choice in this thesis, is to create a standard representation for data-driven decision support systems and to facilitate sharing of medical knowledge among different institutions and clinics. The main feature of the syntax is that it is easily readable and writeable, thereby allowing maintenance and transparency. The knowledge within a domain is braken down inta a number of independent medical logic modules (MLMs), where each module together with its logic also holds a description of its evocation criteria in terms of events and conditions related to patient data.

This thesis addresses same of the problems with integration of a decision support system (OSS) inside a specific software engineering environment, together with issues in relation to database query handling. An architecture of a OSS based on Arden Syntax and its integration in an object-oriented environment is presented. The basic components of the OSS are the knowledge base composed of MLMs and the patient database with developed couplings between the knowledge base and the patient database. Since the MLM format is neutral to the local environment such as patient databases and computer platforms, these methods and tools are needed when realizing an efficient run-time OSS.