Appendix D. Calculation of the goals for bias and imprecision, as percent coefficient of variation [%CV], from the biological variation of T4 and T3

back

The suggested goal for maximum bias in diagnostic testing is:

(CV2[within-subject] + CV2[between-subject])1/2
__________________________________________
4

where CV is the mean coefficient of variation in the group.

 

 

The suggested goal for maximum bias in monitoring an individual is:

CV [within-subject]
___________________
4

 

The suggested goal for maximum imprecision in diagnostic testing is:

(CV2[within-subject] + CV2[between-subject])1/2
____________________________________________
2

 

and for monitoring an individual is:

(CV[within-subject])
__________________
2

 

  


 

The known %CVs are as follows (45-47):
___________%CV____________
  Free Total Free Total
  T4 T4 T3 T3

Mean within-subject CV (CVI)
9.5
5.1
7.9
10.4
Mean between-subject CV (CVG) 12.1 10.2 22.5 20.4
Mean composite biological CV (CVC)a 15.4 11.4 23.8 22.9

For diagnostic applications the relevant
biological variation is the composite
(total) biological variationa
 
Goals for maximum bias
(1/4 the relevant CV)
3.8 2.9 6.0 5.7
Goals for maximum imprecision
(1/2 the relevant CV)
7.7
5.7
11.9
11.5
For therapeutic applications the relevant
biological variation is the within-
subject biological variation		        

Goals for maximum bias
(1/4 the relevant CV)		 2.4 1.3 2.0 2.6

Goals for maximum imprecision
(1/2 the relevant CV)		 4.8 2.6 4.0 5.2


a Calculated as (CVI2 = CVG2)1/2