Objectives: 1. To assess the accuracy of the gestation-adjusted projection method in estimating birth weight in Thai pregnant women. 2. To determine the efficiency of the gestation-adjusted projection method for the detection of low actual birth weight, normal birth weight, and large fetus groups.
Study design: Diagnostic clinical trial.
Material and Method: The present study was conducted on 328 uncomplicated pregnancies that were monitored at the Department of Obstetrics and Gynecology, Charoenkrung Pracharuk Hospital during the period of August 1 to November 30, 2006. The fetal biometry was measured by ultrasound at 34 weeks’ gestation or after. Fetal weight was calculated according to Hadlock’s formula. The extrapolation technique is based on the gestation-adjusted projection (GAP) method. The accuracy of the method was assessed by analyzing the weight predictions in relation to the actual birth weight (ABW). Main outcome measurements were simple error, absolute error, absolute percentage error and accuracy within 10% of ABW.
Results: The accuracy within 10% of ABW was 76.5% (95% CI 71.9, 81.1). The estimation tended to be underestimated (-134.5 + 235.2 grams). The mean of absolute error and of absolute percentage error were 226.2 + 148.8 grams and 7.2 + 4.5% respectively. The smallest observed mean difference was obtained in the large fetus group (birth weight > 4,000 grams) and the largest one was obtained in the normal birth weight group (birth weight 2,500-4,000 grams). The accuracy amongst possible contributing factors were compared and analyzed. The sensitivity and specificity for prediction of birth weight (BW) lower than 2,500 grams (g); 2,500-4,000 g and more than 4,000 g were 60% and 93.8%; 92.5% and 56.5%; 33.3% and 98.8%, respectively. The positive predictive value (PPV) and negative predictive value (NPV) of each BW group were 38.7% and 97.3%; 96.6% and 36.1%; 20% and 99.4%, respectively. The post-test likelihood when the test is negative of each BW group was 2.7%, 63.9%, and 0.6%, respectively. The likelihood ratio of a positive result (LR+) and negative result (LR-) of each BW group were 10 and 0.4; 2.1 and 0.1; 33 and 0.8, respectively. The test efficiency (TE) of each BW group was 91.7%, 89.9%, and 98.2%, respectively.
Conclusion: The GAP of estimated fetal weight (EFW) is able to accomplish the prediction of BW in Thai pregnant women with good accuracy. The diagnostic performance of this method for detection of low birth weight (LBW) and large fetus group are acceptable because of its high specificity, high NPV, high LR+, and low post-test likelihood ratio when the test is negative.

Keywords: Gestation-adjusted projection method, Sonographic estimation of fetal weight, Birth weight