基于病案首页数据挖掘的老年糖尿病患者住院日影响因素分析

目的：回顾性分析老年糖尿病患者住院日的影响因素，探讨其平均住院日的控制方法。方法收集主要诊断为糖尿病或糖尿病相关并发症的老年患者病案首页数据，将可能影响其住院日的因素分为社会学因素、疾病因素和临床因素。用Mann－Whitney U检验和Kruskal－Wallis H检验，逐个筛选出影响住院日的因素，对筛选出的影响因素进行logistic 回归，建立老年糖尿患者住院日分布的预测概率方程。结果（1）年龄（ P＜0.001）、职业（P＝0.001）、付费方式（P＜0.001）、住院次数（P＜0.001）、伴随疾病（P＜0.001）、慢性并发症（P＜0.001）、医院感染（P＜0.001）、手术（P＜0.001）、3 d确诊（P＜0.001）、转科（P＝0.007）、疗效（P＝0.003）等11个因素对住院日有影响，差异有统计学意义。（2）经logistic回归分析，共筛选出8个对糖尿病患者住院日有显著影响的因素，按各因素优势比大小依次为：医院感染、手术、伴随疾病、慢性并发症、住院次数、年龄、付费方式、疗效。结论老年糖尿病患者住院日受多种因素综合影响，医院感染、手术、合并伴随疾病或慢性并发症、多次住院、年龄增加、社会基本医疗保险付费、追求疗效等均可明显影响老年糖尿病患者住院日。

Factor analysis of the length of stay for diabetes according to the information on first page of medical records

Objective A retrospective study was conducted to investigate the length of hospitalization for the dia-betes patients and its control methods .Methods The data of diabetic patients were collected .Sociological factors , dis-ease factors , and clinical factors were included for analysis .Mann-Whitney U test and Kruskal-Wallis H test were used to screen the influencing factors of the length of stay .The correlations between the length of hospitalization and all relevant factors were conducted by logistic regression analysis .Results There were significant correlations between length of stay and age (P<0.001), occupation (P=0.001), payment (P<0.001), number of hospitalizations (P<0.001), ac-companying diseases ( P<0.001 ) , chronic complications ( P<0.001 ) , hospitalinfection ( P<0.001 ) , surgery ( P<0.001),diagnosis within three days (P<0.001), department transfer (P=0.007),or treatment efficacy (P=0.003), according to univariate analysis .According to the logistic regression analysis , eight factors were screened out as signifi-cantly affect factors on the length of stay of diabetic patients , including hospital infection , surgery, accompanying disea-ses, chronic complications, number of hospitalizations, age, payment method and treatment efficacy (ordered by odd ratio). Conclusion The length of hospitalization for the diabetes patients is influencedby many factors .Among them, hospitalin-fection, surgery, accompanying diseases , chronic complications , number of hospitalizations , age, payment method and treatment efficacy are the independent factors .