Appraising surgeons learning sonography: measuring measurement variability

Sonography is increasingly used by various clinicians in several task-specific situations. We present a method of assessing 1 aspect of a nonspecialist's ability in performing sonography, measurement variability.

An experienced radiologist and a surgeon who had received 4 weeks of sonography training each made triplicate measurements of gallbladder dimensions in 19 adult patients undergoing abdominal sonography. The observers examined the subjects in immediate succession using the same ultrasound equipment and were blinded to each other's measurements until the end of the study. Intraobserver measurement variability rates were calculated for gallbladder length, height, and width by using analysis of variance techniques and were expressed as a within-subject standard deviation and a repeatability coefficient.

In terms of intraobserver measurements, the estimated within-subject standard deviations were comparable for the radiologist and the surgeon (length, 0.22 versus 0.17 cm; height, 0.14 versus 0.14 cm; and width, 0.12 versus 0.14 cm, respectively), as were the repeatability coefficients. The interobserver variability showed good agreement as well; the estimated within-subject standard deviations (and 95% limits of agreement) obtained were length, 0.47 cm (-1.08 to 1.48 cm); height, 0.34 cm (-1.08 to 0.49 cm); and width, 0.28 cm (-0.80 to 0.81 cm).

The ability to measure organ dimensions accurately is an essential sonographic skill that can be readily appraised. The findings from this comparison study illustrate that this 1 small aspect of sonography can be learned by various clinicians and that their ability in this task can be appraised. In a clinical setting, the small degree of variability in measurements made by the experienced radiologist and the sonography-trainee surgeon is likely to be satisfactory.