摘要
随着医院信息化的发展,将医疗大数据应用于疾病风险预测领域中的研究逐渐增多。住院病案首页是一类质量较高的医疗大数据,包含患者的基本信息、疾病诊断、手术和操作、住院管理等数据。本研究以北京市两家医院近三年脑卒中患者住院病案首页为研究对象,提取性别、年龄、入院途径、脑卒中类型、慢性合并症等信息作为输入特征,构建基于K近邻分类算法的死亡风险预测模型,分析不同因素对于死亡风险预测的影响。
As the development of hospital information architecture, more and more researchers focusd on the prediction of mortality risk using medical records. The home page of inpatient medical record is an important kind of medical data with high quality, including inpatients’ basic information, disease diagnoses, procedural and examination, and hospitalization index. In this paper, we collected home pages of inpatient medical records from two hospitals in the past three years, and extracted patients’ age, sex, stroke type, admission path, and chronic comorbidities and survival status. They were used as predictors to build a K-nearest neighbors based predictive model for in-hospital mortality risk of stroke patients.
作者
游佳
李爱阳
王森林
魏岚
费晓璐
陈卉
YOU Jia;LI Aiyang;WANG Senlin;WEI Lan;FEI Xiaolu;CHEN Hui(School of Biomedical Engineering,Capital Medical University,Beijing 100069,China;Department of Gynecology;Department of Internal Medicine,Beijing Miyun Hospital,Beijing 101500,China;Information Centre,Xuanwu Hospital,Capital Medical University,Beijing 100053,China)
出处
《中国医疗设备》
2018年第10期60-63,共4页
China Medical Devices
关键词
住院病案首页
脑卒中
K近邻算法
脑血管病
home page of inpatient medical records
stroke
K nearest neighbor
cerebrovascular disease
