Simulate serosurvey data based on various FOI models. — simulate

This function generates binned serosurvey data based on either a time-varying FOI model, an age-varying FOI model, or an age-and-time-varying FOI model. In all cases, it is possible to optionally include seroreversion. This function allows construction of serosurveys with binned age groups, and it generates uncertainty in the distribution of a sample size within an age bin through multinomial sampling.

Usage

simulate_serosurvey(model, foi, survey_features, seroreversion_rate = 0)

Arguments

model: A string specifying the model type which can be one of 'age', 'time', 'age-time'.
foi: A dataframe containing the force of infection (FOI) values. For time-varying models the columns should be 'year', 'foi'. For age-varying models the columns should be 'age', 'foi'. For age-and-time-varying models the columns should be 'age', 'time', 'foi'.
survey_features: A dataframe containing information about the binned age groups and sample sizes for each. It should contain columns: 'age_min', 'age_max', 'n_sample'.
seroreversion_rate: A non-negative value determining the rate of seroreversion (per year). Default is 0.

Value

A dataframe with simulated serosurvey data, including age group information, overall sample sizes, the number of seropositive individuals, and other survey features.

Examples

# time-varying model
foi_df <- data.frame(
  year = seq(1990, 2009, 1),
  foi = rnorm(20, 0.1, 0.01)
)
survey_features <- data.frame(
  age_min = c(1, 3, 15),
  age_max = c(2, 14, 20),
  n_sample = c(1000, 2000, 1500))
serosurvey <- simulate_serosurvey(
model = "time",
foi = foi_df,
survey_features = survey_features)

# age-varying model
foi_df <- data.frame(
  age = seq(1, 20, 1),
  foi = rnorm(20, 0.1, 0.01)
)
survey_features <- data.frame(
  age_min = c(1, 3, 15),
  age_max = c(2, 14, 20),
  n_sample = c(1000, 2000, 1500))
serosurvey <- simulate_serosurvey(
model = "age",
foi = foi_df,
survey_features = survey_features)

# age-and-time varying model
foi_df <- expand.grid(
  year = seq(1990, 2009, 1),
  age = seq(1, 20, 1)
)
foi_df$foi <- rnorm(20 * 20, 0.1, 0.01)
survey_features <- data.frame(
  age_min = c(1, 3, 15),
  age_max = c(2, 14, 20),
  n_sample = c(1000, 2000, 1500))
serosurvey <- simulate_serosurvey(
model = "age-time",
foi = foi_df,
survey_features = survey_features)