Intraobserver repeatability and interobserver reproducibility in musculoskeletal ultrasound imaging measurements.

OBJECTIVE To assess the repeatability and reproducibility of ultrasonographic measurements at the anterior surface of the femoral neck and iliofemoral ligament and on a human tissue-mimicking phantom. METHODS Two independent investigators studied 22 consecutive hips. One investigator had previous experience in musculoskeletal ultrasonography (US). The other investigator had undergone a short course in hip sonography (only 3 hours). Both investigators were blinded to their own and each other's results. On the phantom both observers had taken 10 vertical measurements at 6 cm deep where two objects were placed at 2 cm from each other. Calculation of measurement errors, percent errors and the Bland-Altman graphic technique were used for analysis of data. RESULTS After 132 examinations the first investigator's within-subject standard deviation was 0.4 mm. The intraobserver error was 4.75%. The second investigator's within-subject standard deviation was 0.6 mm and his intraobserver error was 7.00%. The interobserver error was 10.91%. After 20 phantom examinations the first investigators's intraobserver error was 1.11% and the second investigator's intraobserver error was 1.47%. CONCLUSION An inexperienced musculoskeletal sonographer can achieve an acceptable performance if given appropriate training.