Population PKPD Mode.. - Pharsight Corporation
Population PKPD Mode.. - Pharsight Corporation
Population PKPD Mode.. - Pharsight Corporation
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
What to use in diagnostics ?<br />
●<br />
⎡Ω11<br />
Ω=<br />
⎢<br />
⎢<br />
Ω Ω<br />
⎢⎣<br />
0 0<br />
Simulation Based Diagnostics<br />
• Posterior Predictive Checks<br />
1<br />
Θ=<br />
⎢<br />
θ<br />
⎥<br />
⎢<br />
2 ⎥<br />
⎢θ<br />
⎥<br />
3<br />
12 22<br />
⎡σ<br />
♦<br />
♦<br />
♦<br />
♦<br />
♦<br />
Ω<br />
33<br />
⎤<br />
⎥<br />
⎥<br />
⎥⎦<br />
⎤<br />
11<br />
Σ= ⎢<br />
0 σ ⎥<br />
22<br />
⎣<br />
⎡θ<br />
⎤<br />
⎣<br />
⎦<br />
⎦<br />
Select a statistic of interest that a good model should not miss (e.g. AUC,<br />
Cmax, etc) and that can be derived from the raw data<br />
Simulate many data sets from the model (with uncertainty, TS2.2)<br />
Calculate the statistic from the N simulated data and J Replicates<br />
Calculate the statistic from the observed data<br />
Compare the statistic from raw and simulated data<br />
⎡Var<br />
⎤<br />
⎢<br />
Cov Var<br />
⎥<br />
VARCOV = ⎢<br />
⎥<br />
⎢ Cov Cov Var ⎥<br />
⎢<br />
⎥<br />
⎣Cov Cov Cov Var⎦<br />
Draw N Parameters form the<br />
final estimates and Variance<br />
Covariance Matrix<br />
Simulate Data according<br />
to the design of the study<br />
and compute the statistic<br />
Compare the statistic<br />
from raw and simulated<br />
data<br />
If we ignore parameter uncertainty this reduces to predictive check<br />
slide 47<br />
© <strong>Pharsight</strong> <strong>Corporation</strong> All Rights Reserved