Bone densitometry, regardless of the specific technique, is not perfectly r
eproducible even when consistently performed in exact accordance with the m
anufacturer's recommendations. Precision must be quantified at each densito
metry facility in precision studies of the various skeletal sites used for
monitoring. The precision, as the root-mean-square standard deviation or ro
ot-mean-square coefficient of variation, is then used to determine the chan
ge in bone density that constitutes the least significant change and the mi
nimum interval between follow-up measurements. Until precision studies are
performed, the least significant change cannot be determined for any level
of statistical confidence, making the interpretation of serial studies impo
ssible.