基于潜类别回归对特异性过敏患者的分类研究  

Study on the classification of specific allergic patients based on latent class regression analysis

作　　者：王蕊[1] 张茂祥[1] 李萌[1] 刘妍妍[1] 于磊[1] 刘美娜[1] 

机构地区：[1]哈尔滨医科大学公共卫生学院卫生统计教研室,黑龙江哈尔滨150086

出　　处：《实用预防医学》2015年第5期517-519,共3页Practical Preventive Medicine

基　　金：国家重大专项课题(2011ZX08011-005)

摘　　要：目的应用潜类别回归分析特异性过敏患者的潜在分类,为过敏性疾病的早期预防提供依据。方法利用SAS软件的PROC LCA模块,对1 124名特异性过敏患者进行潜类别回归分析研究。结果分析获得特异性过敏患者的4个潜在类别及潜在类别的概率。潜在类别命名分别为:霉蟑吸入组,发生概率是0.3134;户外花粉组,发生概率是0.3646;户尘螨组,发生概率是0.2799;综合过敏组,概率是0.0422。性别为协变量时,潜在类别间差异有统计学意义(P<0.05)。结论应用潜类别回归对过敏患者进行分类,获得的潜类别为过敏原的聚类组合,过敏原间的关联通过所属潜类别产生,利用潜类别预防特异性过敏的发生更具实际意义。Objective To analyze the classification of specific allergic patients based on latent class regression analysis for early prevention of allergic diseases. Methods Using PROC LCA module of SAS software,latent class regression analysis was performed on 1,124 specific allergic patients. Results Four latent class memberships and their probability of occurring were obtained. The four latent class memberships were named as mold and cockroach inhalation group,outdoor pollen group,dermatophagoides pteronyssinus group and comprehensive allergens group,whose probability of occurring was 0. 3134,0. 3646,0. 2799 and0. 0422,respectively. Having sex as covariate,gender distribution was found to be significantly different among the four latent class memberships（ P 〈0. 05）. Conclusions The classification of specific allergic patients obtained through latent class regression analysis is the combination of allergens,and the association between allergens is produced by the latent class. Using latent class membership to prevent the occurrence of specific allergy is of more practical significance.

关 键 词：特异性过敏 潜类别 潜类别回归分析 协变量 

分 类 号：R593.1[医药卫生—内科学]