Clinical versus sonographic estimation of foetal weight in southwest Nigeria

Abstract

A prospective study was conducted at Obafemi Awolowo University Teaching Hospital Complex, Ile-Ife, Nigeria, between 3 January and 31May 2004, to compare the accuracy of clinical and ultrasonographic estimation of foetal weight at term. One hundred pregnant women who fulfilled the inclusion criteria had their foetal weight estimated independently using clinical and ultrasonographic methods. Accuracy was determined by percentage error, absolute percentage error, and proportion of estimates within 10% of actual birth-weight (birth-weight of +10%). Statistical analysis was done using the paired t-test, the Wilcoxon signed-rank test, and the chi-square test. The study sample had an actual average birth-weight of 3,255+622 (range 2,150-4,950) g. Overall, the clinical method overestimated birth-weight, while ultrasound underestimated it. The mean absolute percentage error of the clinical method was smaller than that of the sonographic method, and the number of estimates within 10% of actual birth-weight for the clinical method (70%) was greater than for the sonographic method (68%); the difference was not statistically significant. In the low birth-weight (<2,500 g) group, the mean errors of sonographic estimates were significantly smaller, and significantly more sonographic estimates (66.7%) were within 10% of actual birth-weight than those of the clinical method (41.7%). No statistically significant difference was observed in all the measures of accuracy for the normal birth-weight range of 2,500-<4,000 g and in the macrosonic group (> or =4,000 g), except that, while the ultrasonographic method underestimated birth-weight, the clinical method overestimated it. Clinical estimation of birth-weight is as accurate as routine ultrasonographic estimation, except in low-birth-weight babies. Therefore, when the clinical method suggests weight smaller than 2,500 g, subsequent sonographic estimation is recommended to yield a better prediction and to further evaluate foetal well-being.

Fig. 1.

Scatter diagrams of actual birthweight by estimated foetal weight

Fig. 2.

Overall distribution of percentage error terms

Fig. 3.

Overall distribution of absolute percentage error terms