潜变量增长混合模型在学龄儿童体质指数变化轨迹分析中的应用

Application of Latent Growth Mixture Modeling in the Trajectory of Body Mass Index of Primary Students

目的探讨潜变量增长混合模型(latent growth mixture modeling,GMM)和潜类增长模型(latent class growth model,LCGM)在识别儿童体重增长变化潜在类别上的应用。方法以大连市932名6~12岁学龄儿童的体检纵向数据为例。运用Mplus8.3软件构建不同性别儿童体质指数(body mass index,BMI)变化的GMM和LCGM模型。结果LCGM模型对男女学龄儿童的生长轨迹均识别出3个增长趋势不同的亚组:“稳定组”、“肥胖组”、“偏瘦组”;GMM模型对男性学龄儿童的生长轨迹识别出2个增长趋势不同的亚组:“稳定增长组”和“肥胖增长组”。结论GMM和LCGM模型可以识别学龄儿童BMI发展轨迹的异质性,拓展了描述儿童体重动态变化的方法研究。

Objective To explore the application of latent growth mixture modeling(GMM)and latent class growth model(LCGM)to identify potential classes of children′s weight gain.Methods This study included 932 school-age children aged 6~12 in Dalian.GMM and LCGM of body mass index of children in different genders were constructed by mplus 8.3 software.Results LCGM identified three subgroups with different trends in the growth trajectory for males and females,which were“stable group”,“obese group”and“lean group”.GMM identified two subgroups with different growth trends(“stable growth group”and“obesity growth group”).Conclusion GMM and LCGM can identify the heterogeneity of the development of school-age children′s body mass index,and expand the method on dynamic change of weight.