Generate simulated contingency tables with the option of incorporating adverse event correlation within clusters.

Usage

generate_contin_table_with_clustered_AE(
  row_marginal,
  column_marginal,
  signal_mat,
  contin_table = NULL,
  AE_idx = NULL,
  n_rep = 1,
  rho = NULL,
  seed = NULL
)

Arguments

row_marginal: Marginal sums for the rows of the contingency table.
column_marginal: Marginal sums for the columns of the contingency table.
signal_mat: A data matrix of the same dimension as the contingency table with entries representing the signal strength. The values should be greater or equal to 1, where 1 indicates no signal, and values greater than 1 indicate signal.
contin_table: A data matrix of an \(I\) x \(J\) contingency table with row (adverse event) and column (drug or vaccine) names, of which the row and column marginals are used to generate the simulated data. Please first check the input contingency table using the function check_and_fix_contin_table(). Default is NULL.
AE_idx: A data frame or list. In case of data frame it must contain two variables idx and AE, where idx indicates the cluster index (a number), and AE lists the adverse event names. See the statin49_AE_idx for statin49 data as an example. In case of a list, make sure the cluster indices are aligned with the corresponding row marginal.
n_rep: Number of contingency tables to be generated.
rho: A numeric value, matrix, or NULL indicating the correlation structure.If a numeric value (float or int) is provided, it represents the correlation value rho to be used between all elements within each cluster specified by AE_idx. If a matrix is provided, it must be a square matrix with dimensions equal to the number of rows in contin_table. In this case, rho defines the correlation structure directly, and AE_idx is not used. If rho is NULL, a covariance matrix is generated based on the correlation coefficients of contin_table.
seed: An optional integer to set the seed for reproducibility. If NULL, no seed is set.

Value

A list of n_rep simulated contingency tables.

Examples

# using statin49 as an example
data(statin49)
data(statin49_AE_idx)

# Prepare a matrix of signal strength with the same dimension as
# statin49, where 1 indicates no signal and values > 1 indicate
# signal
lambda_matrix <- matrix(1, nrow = nrow(statin49), ncol = ncol(statin49))

# Assign the cell (45,1) with signal strength 4
lambda_matrix[45, 1] <- 4

# Generate 5 simulated tables
set.seed(123)
simulated_tables <- generate_contin_table_with_clustered_AE(
  contin_table = statin49,
  n_rep = 5,
  AE_idx = statin49_AE_idx,
  rho = 0.5,
  signal_mat = lambda_matrix
)