[[alternative]]Applied cluster analysis and classification tree in school nursing --based on the physical fitness among vocational high school students

[[abstract]]中 文 摘 要邁向21世紀，體適能是競爭的來源。健康促進計畫的設計與管理為學校護理人員重要能力之一。然而，目前體適能研究，可分為兩大方向：其一為介入措施實施前後之比較，另一則以學生的體適能與體適能常模比較，以了解體適能狀況。由於傳統迴歸模型具有三種限制：一是無法考量多元自變項的關聯性；二是對於同一時間須同時檢定三個或更多的交互作用時，結果難以解釋;三是傳統迴歸模式須符合常態分佈之前提假設，且自變項與依變項應呈線性關係，若違反上述假設可能產生預測及分類誤差情形，且介入計畫發展也來自母群體的平均數結果，並未考慮次族群的需要，使得執行成效可能存在預期落差。因此將目標次族群進行具體的區隔與分類，將有助於降低健康不平等的議題。本研究以雲林縣某高中職93及94學年度新生體適能資料庫為樣本，運用群聚分析與分類樹技術。結果發現，群聚分析能將男生分為優等、中等、劣等3群，女生則可分為優等、劣等2群，且經過變異數分析均達統計顯著差異，顯示分群有意義，再利用生理測量值、學業成績及名次做校標相關檢定，男生大致顯示達統計顯著差異。分類樹方面，93學年度男生體適能歸屬於劣等的建議，第一優先介入族群為坐姿體前彎小於33.5公分且身體質量質數小於25.46，94學年度男生體適能歸屬於劣等的建議，第一優先介入族群為坐姿體前彎小於28.5公分且身體質量質數小於25.0267且跑走小於534.5秒。整體而言，群聚分析，可以在最短時間內，將資料予以分群，使管理者在執行管理時，可縮小管理範圍及節省成本;分類樹可以建構一簡易及精確的體適能評估系統，尋找介入的優先組合路徑及次族群，具體指出高危險族群，兩者合併使用將有助於學校護士個案管理之效率及健康促進計畫之計畫與執行之思考，進而提昇體適能，以達健康促進目的。關鍵字：群聚分析(cluster analysis)、分類樹技術(classification tree)、學校衛生護理(school nursing)、體適能(physical fitness)[[abstract]]AbstractWith the onset of the 21st century, physical fitnees is one of the sources for keen competition. And the design and management of the health promotion plan is kind of capibilities in measuring school nursing. Nowadays, the study filed of physical fitnesss can be divided into two subfields： One is a comparison before and after the intervention measures. The other is comparing students’ physical fitness with the normal model of physical fitness in order to get a best understanding of the state of phsyical fitness. There are three kinds of limitation to traditional regression models: first, the impossibility of taking into account the association among independent variables. Second, the impossibility of explaining three or more reciprocal interactions at the same time. Third, traditional regression models is supposed to conform to the presupposition of the normal distribution, and the relation between independent variables and dependent variables is expected to be linear. If not, the errors of prediction and classification may arise. In addition, the development of the intervention plan comes from the averages of mother’s groups without considering the needs of subgroups. Thus, the effect of execution might be affected. For this reason, the targeted subgroups are substantially classified and divided in an attempt to reduce the inequality issues concerning health. Based on the database of the newcomers’ physical fitness in a Yun-Lin county’s high school during the 93 and 94 school years, this study employs the techniques of cluster analysis and classification tree. The study results show that according to cluster analysis, the males students can be divided into the high level, the middle level, and the low level, while the female students can be classified into the superior level and the inferior level. Through the variance analysis, the remarkable differenec in statistics is shown, signifying that the group dividing is meaningul. And then, the physical measuring figure, academic scores,and positions in academic performance are used as dependent variables to make examinations; the results show that the statistics, on the part of the male students, is significantly different. As to the classification tree, in the 93 schoool year, the pysical fitness of male students belongs to the low level. In the first intervention group, 41 is the number of those whose sit-and-reach is smaller than 33.5 mins, and whose the prime of the bodily quality is bigger than 25.46. What’s more, in the 94 school year, as to those whose physical fitness belong to the low level, the first intervention group is the ones whose sit and bend forward is smaller than 28.5 mins, and whose prime of body mass index is bigger than 25.02, and whose 1600 m running time is smaller than 534.5 seconds. The number of this group is 36. Overall, cluster analysis can sort out, in the shortest time, the data so that the managers, in conducting affairs, can narrow down the scope and save costs. On the other hand, the classification tree can construct a simple and exact system of evaluating physical fitness, beneficial for seeking a better route for intervention and subgroups, concretely pointing out the high-risk groups. Combining these two methods will be helpful for the staff of school nursing in enhancing the efficiency of case management and in executing the health promotion plan. In this way, the physical fitness can be improved, and than health can be bettered. Key Words : cluster analysis, classification tree, school nursing, physical fitnes