混合线性模型在婴幼儿生长发育研究中的应用

Application of Linear Mixed Model in Researches of Infant Growth

目的应用混合线性模型,分析婴幼儿生长发育轨迹的影响因素,建立生长发育轨迹的预测模型。方法采用SAS软件MIXED过程分析456例足月健康新生儿的生长发育资料,建立基于混合线性模型的0-12月龄婴儿体重生长轨迹预测模型。结果 0-12月龄婴儿体重发育轨迹主要受出生体重、性别、喂养方式、母亲文化程度、家庭经济状况的影响。拟合的混合线性模型不仅考虑了个体内和个体间变异,还可方便地加入个体效应与其他因素的交互项,更客观地解释各因素对婴儿生长发育的影响。结论混合线性模型是研究婴幼儿生长发育规律的一种有效、实用的方法。

Objective To investigate determinants and establish a prediction model of infant growth trajectory based on linear mixed model. Methods We used the M IXED procedure of SAS softw are to analyze the grow th data of 456 cases of fulltime healthy new borns,and to establish a prediction model of w eight grow th trajectory of infants from birth until 12 months of age. Results Birth w eight,gender,feeding mode,maternal education,and family economic status had a significant effect on w eight grow th trajectory of infants from birth until 12 months of age. The linear mixed model fitting not only considered the variations w ithin and betw een individuals,but also added the interactions betw een the individual effects and other factors easily,w hich could explain the impact of various factors on infant grow th more objectively. Conclusion The linear mixed model is an effective and practical method to study the rule of grow th and development of infants.