simulate_cases: Simulate Cases by Date of Infection, Onset and Report

In epiforecasts/EpiNow: Estimate Realtime Case Counts and Time-varying Epidemiological Parameters

Description

Simulate Cases by Date of Infection, Onset and Report

Usage

simulate_cases(
  rts,
  initial_cases,
  initial_date,
  generation_interval,
  rdist = rpois,
  delay_def,
  incubation_def,
  reporting_effect,
  reporting_model,
  truncate_future = TRUE,
  type = "sample"
)

Arguments

rts	A dataframe of containing two variables rt and date with rt being numeric and date being a date.
initial_cases	Integer, initial number of cases.
initial_date	Date, (i.e as.Date("2020-02-01")). Starting date of the simulation.
generation_interval	Numeric vector describing the generation interval probability density
rdist	A function to be used to sample the number of cases. Must take two arguments with the first specfying the number of samples and the second the mean. Defaults to rpois if not supplied
delay_def	A single row data.table that defines the delay distribution (model, parameters and maximum delay for each model). See lognorm_dist_def for an example of the structure.
incubation_def	A single row data.table that defines the incubation distribution (model, parameters and maximum delay for each model). See lognorm_dist_def for an example of the structure.
reporting_effect	A numeric vector of length 7 that allows the scaling of reported cases by the day on which they report (1 = Monday, 7 = Sunday). By default no scaling occurs.
reporting_model	A function that takes a single numeric vector as an argument and returns a single numeric vector. Can be used to apply stochastic reporting effects. See the examples for details.
truncate_future	Logical, should cases be truncted if they occur after the first date reported in the data. Defaults to TRUE.
type	Character string indicating the method to use to transfrom counts. Supports either "sample" which approximates sampling or "median" would shift by the median of the distribution.

Value

A dataframe containing three variables: date, cases and reference.

Examples

## Define an initial rt vector 
rts <- c(rep(2, 20), (2 - 1:15 * 0.1), rep(0.5, 10))
rts
## Use the mean default generation interval for covid
generation_interval <- rowMeans(EpiNow::covid_generation_times)

## Sample a report delay as a lognormal
delay_def <- EpiNow::lognorm_dist_def(mean = 5, mean_sd = 1,
                                      sd = 3, sd_sd = 1, max_value = 30,
                                      samples = 1, to_log = TRUE)
                                      

## Sample a incubation period (again using the default for covid)
incubation_def <- EpiNow::lognorm_dist_def(mean = EpiNow::covid_incubation_period[1, ]$mean,
                                          mean_sd = EpiNow::covid_incubation_period[1, ]$mean_sd,
                                          sd = EpiNow::covid_incubation_period[1, ]$sd,
                                          sd_sd = EpiNow::covid_incubation_period[1, ]$sd_sd,
                                          max_value = 30, samples = 1)

## Simulate cases with a decrease in reporting at weekends and an increase on Monday                                     
simulated_cases <- simulate_cases(rts, initial_cases = 100 , initial_date = as.Date("2020-03-01"),
                    generation_interval = generation_interval, delay_def = delay_def, 
                   incubation_def = incubation_def, 
                   reporting_effect = c(1.1, rep(1, 4), 0.95, 0.95))
                   
print(simulated_cases)



## Simulate cases with a weekly reporting effect and stochastic noise in reporting (beyond the delay)                                  
simulated_cases <- simulate_cases(rts, initial_cases = 100 , initial_date = as.Date("2020-03-01"),
                    generation_interval = generation_interval, delay_def = delay_def, 
                   incubation_def = incubation_def, 
                   reporting_effect = c(1.1, rep(1, 4), 0.95, 0.95),
                   reporting_model = function(n) {
                      out <- suppressWarnings(rnbinom(length(n), as.integer(n), 0.5))
                      out <- ifelse(is.na(out), 0, out)
                      })
                   
print(simulated_cases)

epiforecasts/EpiNow documentation built on Oct. 26, 2020, 2:38 p.m.