example-pkpd-model.md

Example: PKPD model

Below is an example of a simple indirect-response model. With PK-PD models, the inital state of the PD system often depends on specific model parts. We can define the state of the ODE system statically using the A_init= argument, but this will not any (individual) parameters into account. However, we can also define initial states dynamically using the state_init argument, so that for each inidividual the desired initial state will be calculated:

p_pkpd <- list(CL = 5, V  = 50,
               KIN = .02, KOUT=.5, EFF = 0.2)

pkpd <- new_ode_model(
  code = "
    dAdt[1] = -(CL/V) * A[1];
    conc = A[1]/V;
    dAdt[2] = KIN * 1/(1+EFF*conc) - KOUT*A[2];
  ",
  state_init = "A[2] = KIN/KOUT;"
)

dat <- sim_ode (ode = "pkpd",
                n_ind = 25,
                omega = cv_to_omega(par_cv = list("CL"=0.1, "V"=0.1, "KIN" = .05, "KOUT"= 0.1), p),
                par = p_pkpd,
                regimen = r1,
                verbose = FALSE)

ggplot(dat, aes(x=t, y=y, colour=factor(id))) +
  geom_line() +
  scale_colour_discrete(guide=FALSE) +
  facet_wrap(~comp, scales="free")

Combine PK and PD models

As shown above, a PK-PD model can be written as a single set of differential equations. However, we often develop PK and PD models separately and e.g. want to plug various PK models into existing PD models. This can be done easily in PKPDsim by specifying the model parts in a list to the code argument:

pkpd <- new_ode_model(code = list(pk = "
                                    dAdt[1] = -(CL/V) * A[1];
                                    conc = A[1]/V;
                                  ",
                                  pd = "
                                    dAdt[1] = KIN * 1/(1+EFF*conc) - KOUT*A[1];
                                  "),
                      state_init = list(pd = "A[1] = KIN/KOUT;"))