Objective: To compare the relative accuracy of predicting birth weight among registered diagnostic medical sonographers versus maternal-fetal medicine specialists.
Study design: Over 7 months all patients who delivered within 2 weeks and had sonographic measurements of femur length and head and abdominal circumferences by sonographers and physicians were included in the analysis. The exclusion criteria were multiple gestation and anomalous fetuses. Receiver operating-characteristic curves (ROC) were constructed to determine the ability to detect intrauterine growth restriction (IUGR; birth weight < 2,500 g) and macrosomia (birth weight > or = 4,000 g) among term (gestational age > or = 37 weeks) parturients. A level of p < 0.05 was considered significant.
Results: Among 365 patients recruited, the mean gestational age was 37.3 +/- 2.4 weeks with a mean birth weight of 3,083 +/- 72.5 g. Among term patients the prevalence of IUGR was 7.5% (18/238) and of macrosomia 12% (29/238). A significantly higher percentage of predictions were within 10% of the birth weight when obtained by sonographers (70%) than physicians (54%; p < 0.0001). Registered sonographers were significantly more likely to detect IUGR than the specialists (area under the ROC curves 0.97 +/- 0.02 vs. 0.92 +/- 0.02, respectively; p = 0.02). Both groups had similar accuracy in detecting macrosomic fetuses (area under the ROC curves 0.92 +/- 0.02 for sonographers and 0.90 +/- 0.02 for physicians; p = 0.40).
Conclusions: Prediction of birth weight is significantly more accurate when sonographers rather than maternal-fetal medicine specialists perform the ultrasonographic examination.