Representation and execution of temporal criteria for guideline-based medical decision support at the Intensive Care Unit

This paper presents the development of a framework for representation and execution of temporal criteria in guideline-based Medical Decision Support Systems (MDSSs). An MDSS is a system that provides medical advice based on patient information and on knowledge of medical best practices as represented in ontologies or other forms of knowledge representation. The proposed framework consists of a temporal ontology that defines time-related concepts and their interrelationships, a temporal data model, a mechanism for executing temporal criteria, and the interfaces with data sources. In order to extract requirements for our framework, we studied practice guidelines used at the ICU and we contacted health-care professionals. We implemented our framework in the Gaston MDSS and connected it to the database of the ICU department of the Catharina Hospital in Eindhoven, the Netherlands. We tested it and found it to provide a solid base for temporal functionality in a MDSS. We will further evaluate it in a clinical environment.