RE: [NMusers] Coding a triple absorption model

[Mark Sale - Next Level Solutions]

Title: Coding a triple absorption model

Mahesh,   Consider modeling F2 as a fraction of the part of bioavailablilty left over after F1:    LF1 = THETA(1)*EXP(ETA(1))  F1= LF1/(1+LF1)                        LF2 = THETA(2)*EXP(ETA(2))   F2 = (1-F1)*(LF2/(1+LF2))     ;LF2 DESCRIBES FRACTION OF THE STILL AVAILABLE DOSE (1-F1)  F3 = 1-F1-F2   Mark Sale MD Next Level Solutions, LLC www.NextLevelSolns.com 919-846-9185 -------- Original Message -------- Subject: [NMusers] Coding a triple absorption model From: "Samtani, Mahesh [PRDUS]" <msamt...@its.jnj.com> Date: Thu, June 04, 2009 4:34 pm To: <nmusers@globomaxnm.com> Dear NMusers, I am using a dual absorption model as follows (The dose is split into 2 fractions and the splitting fraction, THETA(.), is an estimated parameter in the model): TPH = LOG(THETA(.)/(1-THETA(.))) TH  = EXP(TPH+ETA(.))/(1+EXP(TPH+ETA(.))) F2  = TH F1  = 1-TH However, upon prolonged sampling we can observe a 3rd shallow phase. I would like to use a 3-depot absorption process with F1, F2 and F3=1-F1-F2. Both F1 and F2 need etas using some type of transformation such that neither one of them exceed 1 nor does their sum exceed 1. Can some one kindly provide the NONMEM code. Thanking you in advance, Mahesh