KNOWLEDGE-BASED TEMPORAL ABSTRACTION IN CLINICAL DOMAINS

We have defined a knowledge-based framework for the creation of abstra ct, interval-based concepts from time-stamped clinical data, the knowl edge-based temporal-abstraction (KBTA) method. The KBTA method decompo ses its task into five subtasks; for each subtask we propose a formal solving mechanism. Our framework emphasizes explicit representation of knowledge required for abstraction of time-oriented clinical data, an d facilitates its acquisition, maintenance, reuse and sharing. The RES UME system implements the KBTA method. We tested RESUME in several cli nical-monitoring domains, including the domain of monitoring patients who have insulin-dependent diabetes. We acquired from a diabetes-thera py expert diabetes-therapy temporal-abstraction knowledge. Two diabete s-therapy experts (including the first one) created temporal abstracti ons from about 800 points of diabetic-patients' data. RESUME generated about 80% of the abstractions agreed by both experts; about 97% of th e generated abstractions were valid. We discuss the advantages and lim itations of the current architecture.