The Usability of Applying Mining Technology to Support Decision Making in Nursing

• Main Authors: Pei-Hung Liao, 廖珮宏
• Other Authors: Woei-Chyn Chu
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/29745191561197012094

博士 === 國立陽明大學 === 醫學工程研究所 === 102 === In 2009, Taiwan’s Department of Health announced the establishment of electronic medicalrecords, which also expedited the adoption of electronic nursing records. The purpose of thisresearch is to use data-mining technologies and forecasting algorithms to help developingappropriate nursing terms and making appropriate nursing decisions, as well as to survey thesatisfaction of nursing staff who execute nursing decision, to review the consistency of nursing terms, and to evaluate the completeness of nursing records. This study was conductedin the educational hospitals in northern Taiwan by collecting and analyzing existing nursingrecords . We used data-mining technologies to create a databank of key nursing terms, tocompare different artificial intelligence algorithms, and to establish a supporting system formaking right nursing decisions. Our research tools included measuring scales of job satisfaction and questionnaires for reviewing the completeness of nursing records. We observed the changes in the time used by the nursing staff to do records, the completeness of their contents, and satisfaction of nursing decisions. This research used CKIP Chinese word segmentation system, IBM SPSS 20.0 Module for Windows, and Clementine 12.0 English version to mine and analyze our data. The databank of key nursing terms was created mainly by using CKIP Chinese word segmentation system and classification rules of decision tree.Next, the optimality forecasting model was established by comparing adaptive-network based fuzzy inference system (ANFIS) and back-propagation neural network (BPN) algorithm.Several statistic analytical tools, such as independent t-test, pair samples t-test, Chi-square,one-way ANOVA, and Pearson’s correlation were conducted to test the clinical data . Results of our research showed that the creation of a databank for key nursing terms and a forecasting model for supporting decision-making could significantly improve the consistency in nursingterms used and enhance work efficiency of nursing staff. The prediction rate of optimality fornursing decisions was as high as 80%. It is expected that the purpose model and its results can be use in clinical training and teaching.