山西省运城市居民年住院次数影响因素分析——基于零膨胀及其相关回归模型

Analysis of influencing factors of annual hospitalization frequency of residents in Yuncheng City,Shanxi Province—based on zero-inflated models and its correlation regression models

[目的]通过拟合居民年住院次数的最优模型,筛选出影响山西省运城市居民年住院次数的主要影响因素。[方法]分别构建负二项回归、Poisson回归、零膨胀负二项回归和零膨胀Poisson回归4种模型,经比较得出最优模型,并据此开展居民年住院次数的影响因素分析。[结果]近一年内0次住院共计4870例(88.18%)。模型分析显示:性别、年龄、文化程度、居住区域、患有慢性病者、家庭年人均收入及常住人口数是居民年住院次数的影响因素(P<0.05);年龄、职业、家庭常住人口数、患有慢性病情况、吸烟、居住区域及健康知识获取途径是居民住院次数出现零膨胀现象的影响因素(P<0.05)。[结论]山西省运城市居民的年住院次数零膨胀负二项回归模型拟合效果最优。家庭年人均收入处于中等水平、患有慢性病人群就医行为较多;年龄偏大、家庭常住人口数为4~6人、城区居住、患有慢性病及经常吸烟的人群的住院潜在需求更大。

Objective By fitting the optimal model of the annual hospitalization frequency of residents,the main influencing factors of the annual hospitalization frequency of residents in Yuncheng City,Shanxi Province are screened.Methods Four models of negative binomial regression,Poisson regression,zero-inflated negative binomial regression and zero-inflated Poisson regression were constructed respectively,and the optimal model was obtained through comparison.The influencing factors of residents annual hospitalization times were analyzed based on this.Results A total of 4870 cases(88.18%)were hospitalized for zero times in the past year.Model analysis showed that gender,age,educational background,residential areas,patients with chronic diseases,annual per capita income of the family and the number of permanent residents were the factors influencing annual hospitalizations(P<0.05).Age,occupation,permanent resident population,chronic diseases,smoking,residential areas and access to health knowledge were the influencing factors of zero inflation of hospitalizations(P<0.05).Conclusions The fitting effect of the zero-inflated negative binomial regression model of annual hospitalizations of residents in Yuncheng City,Shanxi Province is the best.The annual per capita income of the family is in the middle level,and the people with chronic diseases tend to seek medical treatment.The potential demand for hospitalization is greater among people who are senior,live in urban areas,and smoke regularly,with chronic diseases,and whose family population is 4-6.