应用分类树模型分析女性性功能障碍的影响因素  

Application of regression tree model in the analysis of the influencing factors of female sexual dysfunction

作　　者：魏颖 毛宝宏[2] 王剑[2] 王燕侠[2] 李致远[3] 刘青[2] WEI Ying;MAO Baohong;WANG Jian;WANG Yanxia;LI Zhiyuan;LIU Qin(2017 Professional Master Graduate Student Majoring in Gynaecology and Obstetrics,Gansu University of Chinese Medicine,Lanzhou 730000,Gansu,China;Scientific Research Center,Gansu Provincial Maternal and Child Care Hospital,Lanzhou 730050,Gansu,China;Frist Department of Gynecology,Gansu Provincial Maternal and Child Care Hospital,Lanzhou 730050,Gansu,China)

机构地区：[1]甘肃中医药大学2017级妇产科学专业硕士研究生,兰州730000 [2]甘肃省妇幼保健院科研中心,兰州730050 [3]甘肃省妇幼保健院妇一科,兰州730050

出　　处：《中国性科学》2019年第12期83-89,共7页Chinese Journal of Human Sexuality

基　　金：甘肃省重点研发计划项目(17YF1FA109);甘肃省自然科学基金(18JR3RA032)

摘　　要：目的采用分类树模型分析女性性功能障碍(female sexual disorders,FSD)的影响因素,探讨盆底功能障碍性疾病对女性性功能障碍的影响,为临床医师有效识别女性性功能障碍提供科学依据,改善患者的生活质量。方法选择甘肃省6个地区20岁及其以上的5 073名女性为研究对象。采用问卷调查方法,收集其人口学特征等信息;采用中文版女性性功能量表(CV-FSFI)对研究对象进行FSD评估,总分小于23.45分表明存在FSD。根据是否诊断为女性性功能障碍,将5 073名女性分为病例组和对照组。采用病例对照研究,应用分类树模型的卡方自动交互检测方法,探讨女性性功能障碍的影响因素。结果分类树模型共有3层、24个节点、5个终末节点,共筛选出年龄、分娩次数、职业、文化程度、高血压、UI类型6个解释变量,发现高龄、低文化程度、高血压疾病、压力性尿失禁/急迫性尿失禁、多次分娩和不同职业与FSD发生相关。分类树模型Risk统计量0.358,拟合效果尚可。结论分类树模型不仅可以拟合女性性功能障碍发病风险预测模型,针对不同特点人群明确不同影响因素。临床医师应当主动评估高龄、文化程度低、高血压和患有盆底功能障碍性疾病患者的性功能。Objective To analyze the influencing factors of female sexual dysfunction(FSD) using regression tree model and explore the impact of pelvic floor dysfunctions on the FSD, in order to provide a scientific basis for clinicians to identify FSD patient efficiently and improve the life quality of patients. Methods A total of 5 073(age≥20) women located at 6 cities/countries of Gansu Province participated in the study. The general characteristics of the research subjects was collected by questionnaire. The Chinese version of female sexual function scale(CV-FSFI) was used to evaluate the sexual function, and the total score less than 23.45 indicated the presence of FSD. According to the diagnosis of female sexual dysfunction, two groups were defined as the FSD and the non-FSD controls. The chi-square automatic interaction detection(CHAID) of regression trees model was used to estimate the association between the factors and FSD. Results Six explanatory variables(i.e, age, parity, occupation, level of education, hypertension, stress/urge urinary incontinence) were screened out in the prediction model. The model revealed that advanced age, low level of education, hypertension, stress urinary incontinence/urge urinary incontinence, more parities and some kinds of occupation were associated with FSD. The risk statistic of misclassification probability of the model was 0.385, which suggested that the model was intermediate level for predicting FSD. Conclusions Besides predicting the risk of FSD, regression tree model can also reveal the focused factors between women of different ages. Clinicians should evaluate the sexual function of female with elder age, low educational level, hypertension and pelvic floor dysfunction diseases.

关 键 词：女性性功能障碍 女性性功能量表 分类树模型 影响因素 

分 类 号：R714[医药卫生—妇产科学]