運用資料探勘中分類技術於預測住院老人跌倒之研究

A Study of Predicting the Likelihood of Falls in Hospitalized Elderly Patients Using Data Mining Classification Techniques

陳榆臻(Yu-Chen Chen)

指導教授 ： 詹前隆

元智大學/資訊學院/資訊管理學系/碩士(2009年)

https://doi.org/10.6838/YZU.2009.00125

摘要

減少病患因跌倒所造成傷害風險為2007年美國評鑑機構聯合會之子目標，預防住院老人跌倒為世界各國重大公共衛生議題之一。本研究整理過去文獻之跌倒風險因子，並收集國內北部一間教學型醫院之602筆跌倒與非跌倒基本資料(年齡、入院方式等)；入院時的生理狀況變項(疾病診斷、肌力等)；入院、住院當中跌倒危險因子審核記錄；日常生活能力(巴氏量表)。其研究目的為運用資料探勘中的分類技術，建構一個住院跌倒之風險預測模式，找出影響有無跌倒之重要預測因子；除此之外，依不同科別檢視影響該科住院別病患之重要危險因子；以提供未來醫院在審視病患時，能在病患入院之初更有效的篩選最需醫療照護之病患；又病患從入院到出院這段過程中，生理狀況會因治療或其他因素而有所不同，因此本研究針對跌倒危險因子審核記錄，選取兩個時間點做探討，一為入院時，二為住院當中。研究結果發現，針對全部科別，過去一年是否有跌倒歷史為影響有無跌倒最重要之變數；但對外科而言，跌倒（出院）前是否步態不穩或需使用輔具以及入院時是否有意識欠清、無定向感、躁動不安是重要之預測因子；另外，外科病患若入院方式為急診也和跌倒有顯著相關。運動反應是本研究發現的新議題，運動反應評估為6分(能服從指令動作)以下，則此病患有高度跌倒風險，最後，本研究利用四種分類方法做分析，其平均訓練資料準確度都能從五成達到七成以上，而測試資料也達到七成以上，因此達到準確度之外，一般性也佳。

關鍵字

高齡者 ； 跌倒 ； 分類 ； 決策樹 ； 資料探勘

並列摘要

Reduce the risk of patient harm resulting from falls is a sub goal of the Joint Commission on Accreditation of Healthcare Organizations of 2007. So prevent the elderly from falling is one of the important public health issue of countries all over the world. The sub-project coordinated fall risk factors of past literature and collected 602 fall and non-fall admission data includes age, admission type, diagnosis, muscle power, admission and fall(discharge) fall risk assessment, Barthel indexs from researching hospital of north Taiwan. The research proposed to use data mining to construct inpatient fall risk forecast model and find significant patterns and prediction factors of fall and non-fall. In addition, the research according to different surgery first, and then look over that influences its surgery's important dangerous factor. It can provide hospital to examine patients more efficiency. Finally, the accuracy rate of training data were up to 70% ,and validation date were also achieved 70%. Therefore, the model can achieve not just accuracy, but general.

並列關鍵字

Geriatrics ； fall ； classification ； decision tree ； Data mining

參考文獻

[2].Barrett-Connor, E., Weiss, T. W., McHorney, C. A., Miller, P. D., A Siris, E. S. (2009). Predictors of falls among postmenopausal women: Results from the national osteoporosis risk assessment (NORA). Osteoporosis International, 20(5), 715-722.

[5].Brassington, G.S., King, A.C., Bliwise, D.L. (2000). Sleep problems as a risk factor for falls in a sample of community-dwelling adults aged 64-99 years. Journal of the American Geriatrics Society 48 (10), 1234-1240.

[7].Fayyad, U., Piatetsky-Shapiro, G., A Smyth, P. (1996). From data mining to knowledge discovery in databases. AI Magazine, 17(3), 37-53.

[8].Fischer, I. D., Krauss, M. J., Dunagan, W. C., Birge, S., Hitcho, E., Johnson, S., et al. (2005). Patterns and predictors of inpatient falls and fall-related injuries in a large academic hospital. Infection Control and Hospital Epidemiology, 26(10), 822-827.

[10].Härlein, J., Dassen, T., Halfens, R. J. G., A Heinze, C. (2009). Fall risk factors in older people with dementia or cognitive impairment: A systematic review. Journal of Advanced Nursing, 65(5), 922-933.