Estimation of fetal weight by fundal height measurement using regression and decision tree models in Ahwaz Imam Khomeini Hospital in winter 2014

Abstract

Introduction: To better control of pregnancy period and appropriate decision for termination of pregnancy, especially in high-risk cases, estimation of fetal weight is important. This study was performed with aim to evaluate fetal weight by fundal height measurement to access an easy, inexpensive and practical method in deprived areas. Methods: This analytical epidemiological study was performed on 220 pregnant women with special conditions who had referred Ahwaz Imam Khomeini Hospital for delivery in 2014. The uterine height was measured using a tape meter by a midwife and estimated fetal weight was obtained using Ben Johnson formula. The birth weight was measured immediately after birth with standard Seca scale and other factors affecting birth weight were studied. Then, data were studied by Regression and Decision Tree. Data were analyzed by SPSS software (version 17) and Chi-square and ANOVA regression tests. P<0.05 was considered significant. Results: Mean of real birth weight was 3244.52 gr and mean of estimated birth weight was 3304.18 gr. There was a direct and significant relationship between mean of birth weight and fundal height (P=0.01). The formula obtained from the regression was "W= 80.6 ×Fundal Height (cm) + 48.94 × number of delivery". Moreover, the well-estimated birth weight was obtained using the modeling with the decision tree based on the fundal height. Conclusion: Measuring the fundal height can be used as an appropriate criteria to predict the fetal weight. This feature can be used to predict some of the problems related to low birth weight and abnormal fetal growth before delivery.

Keywords

• decision tree
• estimated fetal weight
• fundal height
• regression