超长住院日影响因素的Logistic回归分析

Analysis on risk factor associated with long staying in hospital by using Logistic regression model

探索内科临床部超长住院病人特征和超长住院病人的影响因素,为缩短病人住院日提供有效措施。病人信息来自HIS统计集成与管理决策支持系统的ODS资源库,由于病人住院日服从偏态分布,以几何均数表示平均水平。将住院天数大于lg-1(χ-lg±Slg)定义为超长住院日,利用非条件Logistic回归模型分析超长住院日的影响因素。Logistic回归分析显示病人住院次数、身份、年龄、科室、治疗结果和是否有院内感染对病人住院日超长的影响因素有统计学意义(P<0.01)。结果表明:(1)超长住院日主要受到病人自身健康状况、病情、病种和住院期间是否出现院内感染的影响;(2)合理安排病人检查、控制院内感染是缩短住院日的有效措施;(3)根据科室病种特点制定平均住院日控制值,可使管理更加科学有效。

To analyse the Days of Patient Stay in Hospital （DPSH） in medicine departments and its risk factors associated with long staying. The data was extracted from ODS of HIS Statistical Integration and Decision-making Management Support System （HSIDMSS） . Geometric mean was used to stand for the average level of DPSH since the distribution of it was positive skew. The threshold of long staying in hospital was defined as more than. An unconditional logistic regression model was performed for risk factors analysis, which the variables were selected by forward procedure. The result of logistic regression indicated that times of being in hospital, identity of cases, age department where case being, outcome of treatment and nosocomial infected were statistically significant （P〈0.01） . The conclusion shown; （1） Long DPSH was associated with health status of case, state of serious, characteristic of disease, and nosocomial infectious; （2） Arrange the plan of test and treatment of case reasonable and control nosocomial infected in hospital could be effective measure for shortening DPSH; （3） It will improve pertinence that map out the value for controlling the indicator of average DPSH according the characteristic of diseases in different departments.