Exploiting decision theory concepts within clinical guideline systems: Toward a general approach

Supporting therapy selection is a fundamental task for a system for computerized management of clinical guidelines (GL). To this end, decision theory concepts could provide significant advances. In this article, we propose a systematic analysis of the main GL representation primitives and of how they could be related to decision theory concepts. The knowledge representation contribution we provide can be seen as a basis for implementing a decision support tool within any of the systems described in the literature: As a matter of fact, at a sufficiently abstract level, the GL primitives we treat are shared by all of the systems. Such a tool could be adopted when executing a GL on a single patient (in clinical practice) and for simulation purposes. In particular, a decision theory tool based on this analysis is being implemented in the GLARE system.