Related Experiment Videos
The comparison of reliabilities in dental imaging methods
1Department of Medical Biometry, Epidemiology and Informatics, University Hospital Mainz, Germany. krummi@imbei.uni-mainz.de
Objectives:
Common practice in the statistical comparison of imaging instruments with limited reproducibility consists in the separate estimation of the instrument's reliabilities. However, as soon as one of the imaging methods is subject to item-specific bias (which has to be expected in many dentomaxillofacial imaging procedures), this approach will end in severe errors in reliability computation and in corresponding erroneous clinical conclusions. This paper seeks to point out these effects and to illustrate a more appropriate model for the comparison of instrumental reliabilities.
Methods:
A standard reliability model was adjusted for item-specific bias and illustrated by the comparison of twice repeated planimetric cephalometry versus twice repeated noninvasive orthodontic video imaging (based on the Digigraph 100 device) in 50 children; the anterior cranial base length was used for illustration.
Results:
The proposed model revealed pronounced inferiority of the video-based imaging system concerning its reliability compared with the X-ray based standard. Analysis using separate estimation of the two reliabilities would result in the reverse conclusion and thus falsely establish video imaging, which is in fact less reliable, as a superior diagnostic method.
Conclusion:
The reliabilities of dentomaxillofacial imaging methods have to be adjusted for potential item-specific bias to avoid the erroneous conclusion of the superiority of a diagnostic innovation.