Journal of Ultrasound in Medicine, Vol 7, Issue 5 245-249, Copyright © 1988 by American Institute of Ultrasound in Medicine
Interobserver variability. A source of error in obstetric ultrasound
I. E. Zador, R. J. Sokol and L. Chik
Department of Obstetrics and Gynecology, Wayne State University, Hutzel Hospital, Detroit, Michigan 48201.
Several sources of error encountered in obstetrical ultrasound examination
have been analyzed from a computerized ultrasound database. The variability
in measuring fetal landmarks and visualization of fetal structures were
found to be significantly different among three registered sonographers who
examined 1,410 consecutive second and third trimester patients. The rate of
successful measurement of biparietal diameter (BPD) and average abdominal
diameter (AAD), as well as successful visualization of fetal stomach and/or
kidneys were statistically different among the three sonographers. Only the
fetal femur length was consistently measured by all three sonographers. The
bias in terms of actual millimeters measured was significantly different
for both BPD and AAD. The bias related to the BPD was in the magnitude of 1
mm, while the bias related to the AAD was almost 3 mm. The clinical
significance of these findings and the value of the computer to perform
periodic checks to assure quality control in a busy ultrasound service are
discussed.
