病例交叉研究中条件Logistic回归在Python中的实现  

Implementation of conditional Logistic regression in case-crossover study in Python

作　　者：张清煜 李培政 罗晨曦 李湘莹 马露[1] Qing-Yu ZHANG;Pei-Zheng LI;Chen-Xi LUO;Xiang-Ying LI;Lu MA(School of Public Health,Wuhan University,Wuhan 430071,China)

机构地区：[1]武汉大学公共卫生学院,武汉430071

出　　处：《数理医药学杂志》2023年第5期321-325,共5页Journal of Mathematical Medicine

基　　金：湖北省卫生健康委2021—2022年度科研项目(WJ2021F103)。

摘　　要：目的 探索病例交叉研究中条件Logistic回归在Python软件中的实现。方法以研究某地空气污染物NO2暴露与因肺部感染住院的关系作为实例,利用Python构建条件Logistic回归模型,比较其与常用统计软件R和SAS的建模过程以及统计分析结果的异同。结果 Python、R和SAS三种软件建模逻辑相似,Python的建模语言与其他统计软件相比稍显繁琐,与SAS在参数检验方法上也略有差异,但三种软件的参数估计结果完全相同。结论 Python软件可实现条件Logistic回归分析,进一步拓展了Python在统计分析的应用场景。Objective To explore the implementation of conditional Logistic regression in Python software for case-crossover study.Methods The relationship between exposure to air pollutant nitrogen dioxide and hospitalization due to pulmonary infection was studied as an example.The conditional Logistic regression model was constructed by using Python to compare the modeling process and statistical analysis results with common statistical software R and SAS.Results The modeling logic of Python,R and SAS is similar.Compared with the statistical softwares,the modeling language of Python is a little more complicated,and it is also slightly different from SAS in parameter test methods,but the parameter estimation results of the three softwares are identical.Conclusion Python software could realize conditional Logistic regression analysis,further expanding the application scenarios of Python in statistical analysis.

关 键 词：PYTHON 条件逻辑回归 病例交叉研究 

分 类 号：R195.1[医药卫生—卫生统计学] O212.1[医药卫生—卫生事业管理] TP312.1[医药卫生—公共卫生与预防医学]