线性回归预测胎儿出生体质量的应用价值探讨

Application value of linear regression in predicting fetal birth weight

目的根据胎儿各超声测量值,应用线性回归构建方程预测胎儿出生体质量,并分析各超声测量参数与胎儿出生体质量的相关性。方法回顾性收集159名新生儿出生体质量及分娩前3 d内胎儿双顶径(biparietal diameter,BPD)、头围(head circumference,HC)、腹围(abdominal circumference,AC)、股骨长(femur length,FL)测量数据,使用SPSS 26.0统计学软件依据胎儿超声测量值和胎儿实际超声体质量推导出预测体质量公式,应用线性方程公式计算各胎儿出生体质量预测值,分析对比各胎儿体质量实际值与预测值,分析BPD、HC、AC、FL与胎儿出生体质量的相关性。结果根据各超声测量值推导出线性方程为y(weight,g)=-5527.11+56.81x(HC,cm)+135.38x(AC,cm)+321.03x(FL,cm),F=134.61,P<0.001,方程具有统计学意义,R^(2)=0.837,方程拟合良好,使用公式计算的胎儿体质量预测值与实际值相关性显著(r=0.851,P<0.05),97.5%的散点在2个标准学生化残差内,胎儿AC、HC、FL与胎儿体质量相关,标准化影响系数从大到小为AC>FL>HC。结论应用线性回归预测胎儿出生体质量有效可靠,有较好的临床使用价值。

Objective To predict fetal birth weight by using linear regression equation based on fetal ultrasound measurements and to analyze the correlation between ultrasound parameters and fetal birth weight.Methods The birth weight of 159 newborns and the measurement data of fetal biparietal diameter(BPD),head circumference(HC),abdominal circumference(AC)and femur length(FL)3 days before delivery were collected retrospectively.The predictive weight formula was derived by using SPSS 26.0 software according to the fetal ultrasonic measurement and the actual fetal ultrasonic weight.The predictive value of each fetal birth weight was calculated by using the linear equation formula,and the actual and predictive values of each fetal weight were analyzed and compared.The correlations between biparietal diameter(BPD),head circumference(HC),abdominal circumference(AC),femur length(FL)and fetal birth weight were analyzed.Results The linear equation was deduced according to the measured values:y(weight,g)=-5527.11+56.81x(HC,cm)+135.38x(AC,cm)+321.03x(FL,cm),F=134.61,P<0.001,the equation had statistical significance,R^(2)=0.837,the equation fitted well,the predicted value calculated by the formula was significantly correlated with the actual value(r=0.851,P<0.05),97.5%scatter points were within the two standardized residuals,fetal AC,HC,FL were correlated with fetal weight,and the standardized influence coefficient from large to small was AC>FL>HC.Conclusion The data of fetal birth weight predicted by linear regression is effective and reliable,and has good clinical value.