I was trying to reproduce model from reference Snelder et al (2014) British Journal of Pharmacology 171 5076–5092.
It is indirect response model for heart rate changes in dog including (i) feedback mechanisms and (ii) alterations in heart rate changes due to manual handling.
Equations are :
dHR/dt =Kin (1+ CR).(1-FB.MAP).(1+HD)-Kout.HR
CR is circadian rhythm = amp.cos(2p (t+hor)/24)
FB= feedback component
MAP= maximum arterial pressure
Kin and kout are usual constants
HD= handling effect=P. exp[-kHD.(t-thd) when t>thd
Authors had used NONMEM to work out this model.
When I tried to use it in Phoenix NLME, it failed to fit the data and I think it is due to 'discontinuity' introduced by 'time event t-thd' in handling effect.
Can anyone suggest how to overcome this using sequence statement or alternative PML syntax. I want to use following 2 handling effects to model QTc change in indirect resposne model:
Handling effects: (repeat of handling effect every 24 hour modelled using 'tsld')
CR = Amp1 * cos (2 * 3.1416 * (t - Phi1)/24)
Indirect response model for QTc change
deriv(QTc = ((kin)*(1+CR)*(1=HD1)(1+HD2+Edrug))-kout*QTc)