A comparison of three methods of assessing inter‐observer variation applied to ultrasonic fetal measurement in the third trimester

The inter‐observer variation of fetal biparietal diameter, head circumference, abdominal circumference, and femoral length measured in the third trimester by ultrasound was studied by three different methods—coefficient of variation, correlation coefficient, and the 95% limits of agreement. The coefficients of variation were 1·6 to 3·7%, the correlation coefficients were 0·89 to 0·98 with P values of <0·001, yet the limits of agreement when applied to centile charts were found to be too wide to separate reliably small fetuses from those that were not small. We conclude that the limits of agreement is the preferred method of assessing inter‐observer variation, and we suggest that future research into growth retardation move away from ultrasonic measurements of the fetus.