|
Considerable variability found between 25-D assays. |
 |
|
EurekAlert: The researchers performed statistical analysis (linear regression and bias analysis) of the results. In 40 percent of the Architect-tested specimens and in 48 percent of the Centaur2-tested specimens, the results were at least 25 percent too high or 25 percent too low, compared with the third technique, Holmes reported. The maximum recommended allowable error for 25-hydroxyvitamin D tests is plus or minus 25 percent, and he stated that the sizes of the observed errors ranged from ? to +80 percent.
|