Hi,
I am trying to simultaneously model PK/PD data from drug input that is similar to short-term infusion with the new Phoenix PML language but keep getting error (no error message but %CV was not calculated for parameter estimate and the shape of the curve is totally off).
When I use WNL ASCII, this code works for me.
T=X
tau = 24
IF T <= tau THEN
ti = T
ELSE
ti = tau
ENDIF
Ceff = Css*(1-exp(-Ke*ti))*exp(-Ke*(T-ti))
d(z1) = Kin * (1 - Emax * Ceff/ (Ceff + EC50)) - Kout * z(1)
However because I have data from several doses, the new Phoenix model would be more convenient. So similarly, I tried this code (which didn't work):
double(ti)
Ceff = Css*(1-exp(-Ke*ti))*exp(-Ke*(t-ti))
dosepoint(Ceff, idosevar = Css)
Tau=24
sequence{
ti=t;
sleep(Tau);
ti=Tau;
}
deriv(E = Kin * (1 - Emax * Ceff/ (Ceff + EC50)) - Kout * E)
sequence{E = 100}
error(EEps = 0.1)
observe(EObs = E + EEps)
There was no error message but %CV was not calculated for parameter estimates and the shape of the estimated profile was totally wrong, i.e. flat at 100% (baseline) for the first 24 h (Tau) before drug effect starting after Tau.
I suspected that the problem was with the sequence statement. What would be the correct way to model this IF statement with Phoenix modeling?
Thank you!
Pam
Edited by pharaveej, 26 October 2017 - 04:38 PM.