Measurement criteria for group echocardiographic studies

This study reports the reproducibility achieved both within an observer and between observers measuring left heart echocardiographic dimensions using standard measurement criteria in school age children. These criteria define acceptable interfaces as: 1) a single dominant line; 2) continuity at the point of measurement of at least 5 mm in length; and 3) display of specific motion characteristics for the cardiac structure to be measured. The within observer intraclass correlation coefficient for interventricular septum, diastolic dimension, systolic dimension, left ventricular posterior wall, left atrium and aorta averaged 0.94 (range 0.87--0.98). The coefficient of variation averaged 3.6% (range 2--7.5%). The least reproducible dimension was left ventricular posterior wall. The day-to-day variability within a single subject of echocardiographic determined cardiac dimensions was small for all measurements. The standard deviation of error for all variables was no more than 0.3 mm. A system using screening criteria and standardized measurement techniques can provide the degree of precision necessary to begin investigating echocardiographic differences in group studies.