AN EXPERT SYSTEM IN EMERGENCY MEDICINE

Abstract

The current emergency room setting in National University Hospital with a patient turn-around of three hours leads to both unnecessary hospital admission and potentially harmful premature discharge. Accordingly, an intermediate supervisory facility has been designed: the Extended Diagnostic and Treatment Unit (EDTU). This is designed to be a protocol-driven, nurse-run ward that treats patients suffering from acute diseases such as asthma and chest pain. The newly setup EDTU will have a maximum capacity of eight patients at one time. In order to facilitate the protocol-driven nature of the work, demand arose for the implementation of a digital expert system, capable of integrating digital data from bed-side monitoring instruments and patients' particulars. To that end, a rule-based expert system is developed. In this system, the rules agreed upon by experienced doctors are described by flow charts which can be easily programmed using if-then structure. Users can determine a strategy to diagnose and treat the patient through the interactive question-and-answer sequence. All the functional requirements of this system are fulfilled by the idea of object-oriented programming on the platform of visual basic.net. The software product has a user-friendly interface that is concise and straightforward. The expert system for EDTU hopes to provide new solutions to practical problems, which shows that quality improvements and cost effectiveness can occur simultaneously in the real world. Moreover, it can be extended to case-based decision modeling by gathering knowledge from multiple on-line sources.