Skip to contents

EVPPI wrapper function for total population YLLs

call_evppi.Rd

This function gets the ithim results into the correct format and calls the compute_evppi() script to calculate the EVPPIs for all input parameters and required outcomes and scenarios for the entire population considered in the model

Usage

call_evppi(
  parameter_samples,
  outcome_voi_list,
  outcome,
  cities,
  voi_add_sum,
  NSCEN,
  NSAMPLES,
  scenario_names
)

Arguments

parameter_samples

table containing all the input parameter variables for the different model runs for all cities

outcome_voi_list

vector detailing the outcomes to be considered in the VoI analysis

outcome

total yll outcome for all outcome age categories per city and scenario and disease combination, also combined city result (sum)

cities

vector of cities

voi_add_sum

if the sum of YLLs across all disease outcomes is to be considered

NSCEN

number of scenarios (not incl. baseline)

NSAMPLES

number of times the model was run for each city

scenario_names

gives the names of the scenarios (incl baseline)

Value

list containg the following elements:

evppi_df containing the evppi values for all input parameters and scenario and disease outcomes

evppi_outcome_names, a vector containing all the scenario outcome combinations

Details

The function performs the following steps:

  • create a vector containing the global parameters but not including the emission inventory parameters as they are not independent of each other

  • loop through the cities:

    • extract the city specific parameters (excluding the CO2 and PM emission inventory parameters)

    • extract the outcomes of interest for each scenario using the outcome_voi_list

    • if voi_add_sum == TRUE, calculate the total YLLs by summing across all diseases in the outcome_voi_list - this only makes sense if the diseases in the outcome_voi_list are independent of each other

    • call the compute_evppi() function to calculate the expected values of partially perfect information (EVPPI) for all parameters and diseases of interest

    • if NSAMPLES >= 1000 then also calculate the EVPPI values for the emission inventory parameters