Perform rapid unweighted or weighted cooccurrence analysis

Usage

cooccurrence_analysis(
  data,
  covariates,
  possible_exposures = ms::pheinfox[, phecode],
  weight_var = NULL,
  n_cores = 1,
  parallel = FALSE,
  min_case_count = 20,
  min_overlap_count = 5,
  evalue = TRUE,
  verbose = FALSE,
  logistf_pl = TRUE,
  detect_separation = TRUE
)

Arguments

data: dataset containing outcome, exposures, covariates, and, if weighted, weight variable
covariates: vector of covariates names
possible_exposures: vector of exposure names
weight_var: weight variable name
n_cores: number of cores over which to parallelize (default = 1)
parallel: logical; parallelize analyses (default = FALSE)
min_case_count: minimum number of cases to consider an exposure (default = 20)
min_overlap_count: minimum number of cases with exposure to consider an exposure (default = 5)
evalue: logical; calculate evalues (default = TRUE)
verbose: logical; print notes (default = FALSE)
logistf_pl: logical; use profile likelihood (TRUE) for confidence intervals and tests or Wald (FALSE) in logistf; (default = TRUE)
detect_separation: logical; use detect_separation (TRUE) or skip and use logistf regardless (FALSE); (default = TRUE)

Value

data.table of cooccurrence analysis results

Examples

if (FALSE) { # \dontrun{
# TBD
} # }