Sources of variability in carotid duplex examination: a prospective study.

Interobserver and intraobserver variability of ultrasonic, duplex carotid artery examinations was studied in a prospective, randomized, and blinded clinical trial. Forty-eight patients were examined by two technologists, yielding 96 carotid artery examinations. The kappa statistic was calculated to determine the degree of agreement corrected for chance. The kappa value between examinations by different technologists was 0.476. Variability occurred at both steps in the examination procedure: (1) obtaining the velocity waveforms (kappa = 0.536); and (2) using these waveforms to classify the extent of carotid disease (kappa = 0.609 for interobserver variability in reading waveforms). Minimal to moderate disease categories accounted for most of the variability. There was little disagreement in categorizing lesions as greater than or less than 50% diameter reduction. Intraobserver variability in rereading spectral waveforms was minimal (kappa = 0.842 and 0.894). Recognition of disturbed flow patterns in normal carotid bulbs may reduce variability.