基于贝叶斯网络模型的HIV感染者无保护性行为影响因素分析  

The influencing factors of unprotected sexual behavior among HIV-infected patients:a study based on Bayesian network model

作　　者：王建萍 武媛婷 姚添 蓝光华[4] 鲁鸿燕[4] 向绍密 李淑珍 冯永亮 王素萍[1,3] WANG Jian-ping;WU Yuan-ting;YAO Tian;LAN Guang-hua;LU Hong-yan;XIANG Shao-mi;LI Shu-zhen;FENG Yong-liang;WANG Su-ping(Department of Epidemiology,School of Public Health,Shanxi Medical University,Taiyuan 030001,China;Department of Ultrasound,Tongmei General Hospital of Sinopharm,Datong 037000,China;Center of Clinical Epidemiology and Evidence Based Medicine,Shanxi Medical University,Taiyuan 030001,China;Institute for HIV/AIDS Prevention and Control,Guangxi Zhuang Autonomous Regional Center for Disease Control and Prevention,Nanning 530028,China;Ningming County Center for Disease Control and Prevention of Guangxi Zhuang Autonomous Regional,Ningming 532500,China)

机构地区：[1]山西医科大学公共卫生学院流行病学教研室,太原030001 [2]国药同煤总医院超声科,大同037000 [3]山西医科大学临床流行病学与循证医学中心,太原030001 [4]广西壮族自治区疾病预防控制中心艾滋病防制所,南宁530028 [5]广西壮族自治区宁明县疾病预防控制中心,宁明532500

出　　处：《中华疾病控制杂志》2022年第8期961-966,共6页Chinese Journal of Disease Control & Prevention

基　　金：国家科技重大专项(2018ZX10721202,2012ZX10002001)。

摘　　要：目的了解HIV感染者无保护性行为发生情况,探讨其无保护性行为的影响因素及因素间的网络关系。方法采用横断面研究方法,于2015年3月―2015年7月,对广西壮族自治区两所疾病预防控制中心的HIV感染者进行面对面问卷调查,收集其一般人口学特征、HIV治疗情况、近三月的性行为、健康状况和吸毒情况等。采用单因素、多因素非条件logistic回归分析模型进行变量的初步筛选,采用“bnlearn”贝叶斯网络软件包进行模型构建,Netica软件进行模型推理。结果HIV感染者无保护性行为发生率为47.47%(441/929)。多因素logistic回归分析模型分析结果显示,年龄大、多性伴、文化程度低、家庭人均月收入低和未接受抗反转录病毒治疗(antiretroviral therapy,ART)是无保护性行为的危险因素。贝叶斯网络模型结果显示,多性伴、接受ART、文化程度、家庭人均月收入及年龄与无保护性行为直接相关,目前已婚或同居通过影响多性伴与无保护性行为间接相关,职业通过影响家庭人均月收入与无保护性行为间接相关。结论HIV感染者无保护性行为发生率较高,对多性伴、未接受ART、文化程度较低、家庭人均月收入较低、年龄较大、单身及无业者应加强健康教育,减少其无保护性行为的发生。Objective To investigate the prevalence of unprotected sexual behavior among HIV-infected patients,and to explore the influencing factors of unprotected sexual behavior and their network relationships.Methods HIV-infected patients managed by two centers for Disease Control and Prevention in Guangxi were selected from March to July 2015 in a cross-sectional study.Data of general demographic characteristics,HIV treatment status,sexual behavior in the last three months,health status and drug use status were collected through face-to-face interview.Single-factor analysis and multi-factor logistic regression were used to initially screen the variables,“bnlearn”Bayesian network software package was used for model construction,and Netica software was used for model inference.Results The incidence of unprotected sexual behavior among HIV-infected patients was 47.47%(441/929).The results of multi-factor logistic regression analysis showed that age,multiple partners,education level,per capita monthly income of family,and receiving ART were influencing factors of unprotected sexual behavior.The results of Bayesian network model showed that multiple sexual partners,receiving ART therapy,education level,per capita monthly income of family,and age were directly related to unprotected sexual behavior,currently being married or cohabiting was indirectly related to unprotected sexual behavior through affecting multiple sexual partners,and occupation was indirectly related to unprotected sexual behavior through affecting per capita monthly income of family.Conclusions HIV-infected patients have a high incidence of unprotected sexual behavior.Health education should be strengthened for those having multiple sexual partners,not receiving ART,with low level of education,having low per capita monthly income of family,being older,being currently single and being unemployed,so as to reduce the occurrence of unprotected sexual behavior.

关 键 词：贝叶斯网络模型 HIV感染者 无保护性行为 影响因素 

分 类 号：R183.7[医药卫生—流行病学]