Hi (again - apologies for the inundation of comments!):

Is there a way in the cohort modelling side of hesim to specify that in one strategy, the cohort starts in one state and in a second strategy the cohort starts in the other? I am comparing a cohort of patients starting on dialysis and a cohort starting on liver transplants.

I see that you can specify which state the model starts in directly by setting start_stateprobs = c(...) in create_CohortDtstm - but this sets the starting state to be the same for all strategies simulated.

I've managed to kind of do this with a hack (see below), but it isn't ideal and I thought either there may be a neater solution or else it could be something useful for future development if you are so inclined.

A second question on the below code: you'll see I've adapted your time-dependent drugs cost code to try and crowbar in a one-off cost of transplant. But when I look at final costs - the drug cost is much higher than the upper bound estimate of my drug cost - and since the cost of transplant should only be added once per simulation I can't figure out why I appear to be adding in more than one transplant cost - even though it should only execute when time == 0?

`tparams_def <- define_tparams({

ntx_switch <- ifelse(strategy_name == "Transplant only" & time < 1e-2, 0.85, 0) # 0.85 chosen as highest value to shift people across from dialysis to transplant (ntx) without sending state occupancy negative

list( tpmatrix = tpmatrix( C, ntx_switch, 0, p_rec_noecu, p_dial_death, # transitions from dialysis (never had recurrence) p_norec_dial, C, p_rec_noecu, 0, p_ntx_death, # transitions from ntx + no-recurrence 0, 0, C, p_rec_dial, p_ntx_death, # transitions from ntx + recurrence 0, 0, 0, C, p_dial_death, # transitions from dialysis post ntx with graft loss with recurrence or else direct recurrence from dialysis 0, 0, 0, 0, 1 # death category - don't edit ), utility = u, costs = list(

add a transplant cost in first cycle

  drug = ifelse(strategy_name == "Transplant only" & time == 0,
                c_transplant, 0),
  # community_medical = c_cmed,
  direct_medical = c_dmed
)

) }, times = c(0,4*49))`

hesim-dev / hesim

Starting in different states of the model to begin with #10

add a transplant cost in first cycle