Extensions to the Time-Oriented Database Model to Support Temporal Reasoning in Medical Expert Systems

• DOI: 10.1055/s-0038-1634816
• Publication Date: 2018-03-20T13:05:19Z
• Authors: M. G. Kahn; L. M. Fagan; S. Tu
• Keywords: Temporal database; Relevance; Logical data model; Data model (GIS)

Physicians faced with diagnostic and therapeutic decisions must reason about clinical features that change over time. Database-management systems (DBMS) can increase access to patient data, but most are limited in their ability store retrieve complex temporal information. The Time-Oriented Databank (TOD) model, the widely used data model for medical database systems, associates a single time stamp each observation. proper analysis of requires accounting multiple concurrent events may alter interpretation raw data. Most DBMSs cannot indexed by events. We describe two logical extensions TOD-based databases solve set reasoning problems we encountered constructing expert systems. A key feature both is stored partitioned into groupings, such as sequential visits, exacerbations, or other abstract have decision-making relevance. network (TNET) an object-oriented extends capabilities ONCOCIN, system provides chemotherapy advice. TNET uses persistent objects associate observations intervals during which “an event interest” occurred. second system, called extended (ETNET), extension simplification TNET. Like TNET, ETNET represent relevant intervals; unlike first however, contains methods (rules) be executed when “begins”, withdrawn “concludes”. capture relationships among recorded information not represented databases. Although they do all found decision making, these new structures enable encode relationships, based on contexts and, ETNET, modify available onset conclusion specific