Generating Macro-Temporality in Timed Transition Diagrams

Abstract

Decision support systems in medicine are designed to aid healthcare professionals on making clinical decisions. Clinical Algorithms derived from Clinical Practice Guidelines (CPGs) make explicit the knowledge necessary to assist physicians in order to make appropriate decisions. Decision support systems for healthcare procedures are supposed to answer questions about what to do and with what time restrictions. Unfortunately, so far we are not able to answer the second question, as clinical algorithms do not contain temporal constraints. Here, our objective is to produce explicit knowledge on temporal restrictions for healthcare procedures. This is reached by generating temporal models from hospital databases. First, we have identified macro-temporality as a constraint on the time required to evolve one step in a clinical algorithm. We have decided to use Timed Transition Diagrams (TTDs) as a structure to represent clinical algorithms, extended with macro-temporality constraints. Then we have identified three different data levels in hospital databases and we have proposed an algorithm to generate macro-temporality in TTDs for each data level.