住院跌倒患者的数据挖掘与跌倒防范对策分析

Data mining and preventive measures for 239 cases of falls in hospitalized adults

目的找出住院跌倒患者数据中有价值的关联规则,为制订院内跌倒预防措施提供参考。方法基于数据挖掘的原理和方法,采集浙江省某三级甲等医院2011年—2016年上报的239例跌倒患者的7 170项数据,采用Apriori算法进行数据挖掘,再使用卡方检验分别对得到的关联规则进行有效性评定。结果通过条件设定,得出245条关联规则,再经卡方检验筛选出94条规则,最后结合专业知识分析获取强关联规则18条。包括年龄≥65岁、夜间发生、有高血压史、日常生活活动能力三级或四级、有糖尿病史、步态不稳、跌倒前活动内容与排泄相关、跌倒风险认知不足等。结论护理人员应该加强夜间时段的防范、预防再次跌倒的发生、重视排泄相关性跌倒的管理、提高老年患者对跌倒防范的认知水平和关注多个疾病诊断的患者。

Objective To find out valuable association patterns hidden in the events of falls among hospitalized patients,and to provide scientific references for prevention of falls. Methods According to basic principles of data mining,totally 7 170 records of 239 cases of falls in a tertiary hospital in Zhejiang Province from 2011 to 2016 were collected. Apriori algorithm was conducted to mine association patterns,and chi-square test was used to test effectiveness for strong association patterns. Results Through the condition setting,245 association patterns were obtained,and 94 patterns were selected by chi-square test. Finally,18 strong patterns were obtained via analysis ,,4 with professional knowledge. Conclusion Through data mining of falls in hospitalized patients,figuring out strong association patterns for factors of falls ,can help to discover system＇s weaknesses ,and to provide scientific and accu- rate references for further development of targeted preventive measures.