Skip to main content
. Author manuscript; available in PMC: 2011 Oct 13.
Published in final edited form as: J Pharmacokinet Pharmacodyn. 2009 Jun 27;36(4):317–339. doi: 10.1007/s10928-009-9124-x

Table III.

Relative Bias and root mean square error (RMSE) in % evaluated from 1000 simulated data sets with Mbase under H0 for the volume of distribution (V/F), its interindividual variance ( ωV/F2) and the residual error parameter (b) for the N=40/n=4, N=80/n=2, N=100/n=4,1 and N=200/n=4 designs and from 1000 simulated data sets with Mmult under H1 for V/F, β1, β2, ωV/F2 and b for the N=40/n=4, N=80/n=2 and N=100/n=4,1 designs, using estimates from SAEM and FOCE-I when available in (8)

Parameter N=40/n=4 N=80/n=2 N=100/n=4,1 N=200/n=4
SAEM FOCE-I SAEM SAEM SAEM FOCE-I
MbaseunderH0
Biais (%) V/F 0.23 2.9 0.04 0.62 0.08 1.4
ωV/F2
−2.8 −0.6 0.2 −4.2 −0.8 0.7
b −0.3 −1.9 −3.8 −0.9 0.008 −1.8
RMSE (%) V/F 8.6 9.5 8.5 11.8 3.8 11.1
ωV/F 28.1 28.9 27.8 38.5 13.4 13.3
b 8.8 10.3 15.8 12.4 4.0 4.8
MmultunderH1
Biais (%) V/F 4.1 6.7 3.9 5.1
β1 −1.0 −0.8 −2.5 −1.4
β2 −1.0 −1.0 −1.8 −1.3
ωV/F −7.5 −5.2 −1.3 −7.1
b −0.6 −2.2 −3.5 −0.02
RMSE (%) V/F 17.9 19.2 15.0 19.9
β1 19.9 20.0 15.3 18.1
β2 21.7 21.7 16.5 21.3
ωV/F 29.7 29.6 26.7 39.1
b 9.22 10.1 16.8 13.0

With FOCE-I, convergence was achieved and thus estimates were obtained from 969 and 950 data sets under H0 and H1 respectively for N=40/n=4 and 978 data sets under H0 for N=200/n=4