Get epidemic losses from a DAEDALUS model run — daedalus

Calculate the costs of an epidemic and any mitigation measures, either from a Daedalus model run, or from a timeseries of epidemic data and relevant demographic and economic parameters.

Usage

# S3 method for class 'daedalus_output'
get_costs(
  x,
  summarise_as = c("none", "total", "domain"),
  productivity_loss_infection = 1,
  ...
)

# S3 method for class 'data.frame'
get_costs(
  x,
  comp_non_working,
  comp_infected,
  comp_dead,
  workforce,
  daily_gva,
  productivity_loss_infection,
  vsl_by_age,
  life_expectancy,
  value_school_year,
  n_students,
  edu_effectiveness_remote,
  npi_data = NULL,
  summarise_as = c("none", "total", "domain"),
  ...
)

get_costs(x, ...)

Arguments

x

A <daedalus_output> object from a call to daedalus(), or a <data.frame> of an epidemic timeseries.

If x is a <data.frame>, it must have the columns "time", "compartment", "age_group", "econ_sector" and "value", giving the model time, the epidemiological compartment, and the number of individuals of each age group and economic sector in each compartment at each time. See Details for more on the expectations around how this dataset is organised.

summarise_as

A string from among "none", "total", or "domain", for how the costs should be returned. Select "none", the default, for the raw costs along with overall and domain-specific totals; "total" for the overall cost, and "domain" for the total costs per domain; the domains are 'economic', 'education', and 'life years'.

productivity_loss_infection

A single number in the range $[0, 1]$ giving the loss in productivity associated with symptomatic infection. Currently defaults to 1.0 for compatibility with earlier function versions.

...

<dynamic-dots> Other arguments to class methods.

comp_non_working

A character vector giving the names of compartments in which individuals are assumed to have a reduced ability to work. This may (and should) include compartments where individuals are infected and symptomatic, as well as hospitalised or dead.

comp_infected

A character vector giving the name of compartments in which individuals are assumed to be infected with reduced ability to work. This should not include compartments where individuals are assumed to have no ability to work (hospitalised or dead).

comp_dead

A string giving the name of the compartment holding the count of deaths. Note that only a single deaths compartment is supported, so models with multiple death states should pass the final state. Used to calculate the value of lives lost.

workforce

A numeric vector of the number of workers in each economic sector of the population. Must be absolute numbers (i.e., not scaled to thousands or millions). Must have the same length as daily_gva.

daily_gva

A numeric vector of the daily gross-value added (GVA) in million dollars, by each economic sector of the population (or territory or state) in which the epidemic occurs.

vsl_by_age

A numeric vector of the value of a statistical life (VSL) for each age age group in x.

life_expectancy

A numeric vector of the life-expectancy per age group. Expected to be the same length as vsl_by_age.

value_school_year

A single number giving the value of a school year. Users can use the helper function get_value_school_year() if a population's gross income per-capita is known to apply the internal Daedalus model for valuing a school year, or a value obtained in some other way.

n_students

A single number for the number of students in the population, whose education is expected to be impacted by the epidemic.

edu_effectiveness_remote

A single number for the effectiveness of remote education, in the range $[0, 1]$.

npi_data

A list with numeric elements that mirrors the elements of a <daedalus_npi>, with the following names:

npi_times_start: A numeric vector of the start times of any NPIs modelled.
npi_times_end: A numeric vector of the end times of modelled NPIs.
npi_durations: A numeric vector of the durations of NPIs.
npi_periods: A numeric vector giving integer-ish elements in the ranges specified by the start and end times.
openness: Either a numeric vector (if modelling a single NPI), or a list of numeric vectors (if modelling multiple NPIs), giving the openness coefficients of NPI regimes. Vector lengths must be the same as workforce.

Value

A list of different cost values, including the total cost. See Details for more information.

Details

Expectations for non-Daedalus model data

When a <data.frame> x is passed, it must have at least one row, at least some non-missing values, and:

Columns "time", "age_group", "compartment", "econ_sector", and "value", giving the model time (numeric), age group identifier (character), epidemiological compartment identifier (character), economic sector identifier (character), and the value of the number of individuals in each age, epi-compartment, and economic group at each time (numeric). Numeric values must be non-negative $> 0$, non-missing, and finite.
Compartments in comp_infected, comp_non_working, and comp_dead must be found in x$compartment.
All compartments of comp_infected and comp_dead must be in comp_non_working.

Output

The total cost in million dollars is returned as total_cost. This is comprised of the following costs.

Economic costs

A three element list of economic_cost_total, the total costs from pandemic impacts on economic sectors, including both costs of lost gross value added (GVA) due to pandemic-control restrictions or closures (economic_cost_closures), and pandemic-related absences due to illness and death (economic_cost_absences).

Educational costs

A three element list of education_cost_total, the total costs from pandemic impacts on education due to pandemic-control restrictions or closures (education_cost_closures), and pandemic-related absences due to illness and death (education_cost_absences).

Life-value lost

A four-element vector (for the number of age groups) giving the value of life-years lost per age group. This is calculated as the life-expectancy of each age group times the value of a statistical life, with all years assumed to have the same value.

Life-years lost

A four-element vector (for the number of age groups) giving the value of life-years lost per age group. This is calculated as the life-expectancy of each age group times the number of deaths in that age group. No quality adjustment is applied.

Examples

# for <daedalus_output> objects from `daedalus()`
o <- daedalus("CAN", "sars_cov_1", time_end = 100)
get_costs(o, "domain")
#>     economic    education   life_value   life_years 
#>   0.71370354   0.06463232  36.94926465 802.37273941 

# for a data.frame of epi compartment timeseries, as from any epi model
output <- daedalus("Canada", "influenza_1918", time_end = 100)

data <- get_data(output)
comp_non_working <- c(
  "infect_symp",
  "hospitalised_recov",
  "hospitalised_death",
  "dead"
)
comp_infected <- "infect_symp"
comp_dead <- "dead"

daily_gva <- output$country_parameters$gva
workforce <- output$country_parameters$workers
vsl_by_age <- output$country_parameters$vsl
life_expectancy <- output$country_parameters$life_expectancy

value_school_year <- 1e6 # 1 million dollars
n_students <- output$country_parameters$demography[3L]

edu_effectiveness_remote <- 0.33

productivity_loss_infection <- 1.0

get_costs(
  data,
  comp_non_working,
  comp_infected,
  comp_dead,
  daily_gva,
  workforce,
  vsl_by_age,
  life_expectancy,
  value_school_year,
  n_students,
  productivity_loss_infection = productivity_loss_infection,
  edu_effectiveness_remote = edu_effectiveness_remote,
  summarise_as = "domain"
)
#>     economic    education   life_value   life_years 
#> 190967154574  16840260109       709967     15417306