A Symbolic Processing Approach to Measurement Interpretation in the Intensive Care Unit

摘要

This paper reports work in progress on a computer program to interpret physiological measurements from the intensive care unit. The Ventilator Manager (VM) program processes data from a patient monitoring system. VM can recognize unexpected events and make suggestions about achieving the therapeutic goals of mechanical ventilation. A form of production rules, drawn from the area of Artificial Intelligence research, is used to represent and manipulate knowledge about this dynamic clinical setting. An expectation-driven approach to measurement interpretation is discussed. The structure of the knowledge base of rules and sample output of the system are presented.