bin.fi.samplesize: Sample size calculator for 2x2 tables, taking into account...
In brb225/FragilityTools: Toolbox For Patient Count Fragility Measures

This function is a wrapper around general.fi.samplesize, see the documentation there.

bin.fi.samplesize(
  min.fi = 10,
  min.power = 0.8,
  alpha = 0.05,
  tau = 0.5,
  test = "fisher",
  algorithm = "original",
  theta1 = 0.3,
  theta2 = 0.7,
  row.prop = 1/2,
  niters = 5,
  cl = NULL
)

`min.fi`	the smallest acceptable QUANTILE fragility index. When NULL, the FI calculation is skipped and sample_size_init_fi is taken to produce the desired FI.
`min.power`	the smallest acceptable power. When NULL, the power calculation is skipped and sample_size_init_power is taken to produce the desired power.
`alpha`	a numeric for the size of the p value based test
`tau`	the quantile of FI to bound, default 1/2
`test`	a string specifying which p value based test to use. By default 'fisher' for Fisher's exact test. Can also specify `chisq.prop` for Pearson's chi square test.
`algorithm`	A string specifying the algorithm to use to calculate fragility indices. The default is "walsh" (or 'original'). Alternatives include "greedy" for using greedy.fi
`theta1`	a numeric for the event rate in the first group, for simulating the alternative model
`theta2`	a numeric for the event rate in the second group, for simulating the alternative model
`row.prop`	a numeric for the proportion of patients in the first group, for simulating the alternative model
`niters`	the number of iterations to run the algorithm. The final output is the median across iterations.