Provide projections from calibrated simulations by changing RO, contact matrices or bed availability.

This extends previous projections as you can pass in lists of each argument that then get passed to each simulation replicate.

projections(
  r,
  time_period = NULL,
  R0 = NULL,
  R0_change = NULL,
  tt_R0 = 0,
  contact_matrix_set = NULL,
  contact_matrix_set_change = NULL,
  tt_contact_matrix = 0,
  hosp_bed_capacity = NULL,
  hosp_bed_capacity_change = NULL,
  tt_hosp_beds = 0,
  ICU_bed_capacity = NULL,
  ICU_bed_capacity_change = NULL,
  tt_ICU_beds = 0,
  to_be_run = TRUE,
  model_user_args = NULL
)

Arguments

r	Calibrated `{squire_simulation}` object.
time_period	How many days is the projection. Deafult = NULL, which will carry the projection forward from t = 0 in the calibration (i.e. the number of days set in calibrate using forecast)
R0	Numeric vector for R0 from t = 0 in the calibration. E.g. `R0 = c(2, 1)`. Default = NULL, which will use `R0_change` to alter R0 if provided.
R0_change	Numeric vector for relative changes in R0 relative to the final R0 used in the calibration (i.e. at t = 0 in the calibration) E.g. `R0 = c(0.8, 0.5)`. Default = NULL, which will use `R0` to parameterise changes in R0 if provided.
tt_R0	Change time points for R0
contact_matrix_set	Contact matrices used in simulation. Default = NULL, which will use `contact_matrix_set_change` to alter the contact matrix if provided.
contact_matrix_set_change	Numeric vector for relative changes in the contact matrix realtive to the final contact matrix used in the calibration (i.e. at t = 0 in the calibration). E.g. `contact_matrix_set_change = c(0.8, 0.5)`. Default = NULL, which will use `contact_matrix_set` to parameterise changes in contact matrices if if provided.
tt_contact_matrix	Time change points for matrix change. Default = 0
hosp_bed_capacity	Numeric vector for hospital bed capacity from t = 0 in the calibration. Default = NULL, which will use `hosp_bed_capacity_change` to alter hosp_bed_capacity if provided.
hosp_bed_capacity_change	Numeric vector for relative changes in hospital bed capacity relative to the final hospital bed capacity used in the calibration (i.e. at t = 0 in the calibration). E.g. `hosp_bed_capacity_change = c(0.8, 0.5)`. Default = NULL, which will use `hosp_bed_capacity` to parameterise changes in hospital bed capacity if provided.
tt_hosp_beds	Change time points for hosp_bed_capacity
ICU_bed_capacity	Numeric vector for ICU bed capacity from t = 0 in the calibration. Default = NULL, which will use `ICU_bed_capacity_change` to alter ICU_bed_capacity if provided.
ICU_bed_capacity_change	Numeric vector for relative changes in ICU bed capacity relative to the final ICU bed capacity used in the calibration (i.e. at t = 0 in the calibration). E.g. `ICU_bed_capacity_change = c(0.8, 0.5)`. Default = NULL, which will use `ICU_bed_capacity` to parameterise changes in ICU bed capacity if provided.
tt_ICU_beds	Change time points for ICU_bed_capacity
to_be_run	List of logicals for whether each replicate should be run. Default = TRUE, which causes all replictes to be run.
model_user_args	List of other parameters to be passed to the model to be run. Default = NULL. An example would be: `list( list( "prob_severe" = runif(17), "tt_dur_get_ox_survive" = c(0, 10), "gamma_get_ox_survive" = 0.2), list( "prob_severe" = runif(17), "tt_dur_get_mv_survive" = c(0, 5), "gamma_get_mv_survive" = 0.1) )` The list should be the same length as the number of replicates in the simulations. Each list element should then be a list with elements named to match the arguments expected by the odin model with `r`. Above would be suitable to set the model parameters for a simulation with 2 replicates. You do not have to have the same arguments in each list.

Details

The user can specify changes to R0, contact matrices and bed provision, which will come into effect from the current day in the calibration. These changes can either set these to be specific values or change them relative to their values in the original simulation. If no change is requested, the simulation will use parameters chosen for the calibration run. This extends previous versions of projections as you can now pass in lists of each argument that then get passed to each simulation replicate.