ABSTRACT
With growing need in the field, the application of computers in nursing has been frequently studied to improve the quality of nursing care in Korea. But the development of useful clinical programs has not received adequate alternatives. The aim of this study is to compare of two Nursing Diagnosis Systems - Neutral Network and Expert System. The simulated output of each Nursing Diagnosis System and the Judgment from the researcher and two professors of nursing were comparatively examined. The misdiagnosis rate of Nursing Diagnosis System using the Neural Network was nine percent, while the Nursing Diagnosis System using the Expert System showed consistency with those three experts in every aspect. Accordingly, the result of this study demonstrated the possibility of application of a nursing diagnosis system as another nursing tool.
Subject(s)
Diagnostic Errors , Expert Systems , Judgment , Korea , Nursing Care , Nursing Diagnosis , NursingABSTRACT
Expert systems emerged during early 1970s have become one of the most important innovations in computer science. They actively work in various areas and achieve a big success especially in medical domain. It is to be difficult to obtain reliable recognitions about each diagnosis that has common characteristics and to manage data about over 140 nursing diagnoses by using existing Nursing Diagnosis System with artificial neural network. But a rule-based approach has been studied to automate diagnosis and is more effective that nursing diagnosis depends on objective data and has more common characteristics. In this paper, Nursing Diagnosis System was built with a rule-based approach due to the above reasons. Nursing Diagnosis System is implemented by using rules of production system, a kind of knowledge representation technique based on standardized nursing care plans. The ten nursing diagnoses, found most frequently from the patients on medical-surgical wards were determined to develop Nursing Diagnosis System. To verify the usefulness of Nursing Diagnosis System was tested with 80 patient's data. A diagnosis was judged correct if the system and the nursing experts agreed on diagnosing. The nursing experts evaluated 100 percents of the system's diagnosis was deemed correct. Judging from the results given above, the system seems highly reliable. In this work, I show that it can contribute to clinical application of nursing process as another nursing tool.