bin.fi.walsh: Original fragility index calculation, for binary data
In brb225/FragilityTools: Toolbox For Patient Count Fragility Measures

Description Usage Arguments Value Examples

This function implements the original algorithm to calculate a fragility index, as it was proposed by Walsh et al. (2014). The algorithm determines a single group in which to make outcome modifications. We also implement a variant which considers outcome modifications in both directions (i.e. increasing and decreasing event counts), which is a hybrid of the algorithm due to Walsh et al. (2014) and the algorithm commonly used for the "reverse" fragility index.

1	bin.fi.walsh(crosstab, get.p, alpha = 0.05, dir = "both", group = "event")

`crosstab`	a 2x2 contingency table with groups on the rows
`get.p`	a function which accepts a 2x2 matrix and outputs a p value
`alpha`	a numeric for the significance cutoff, 0.05 by default
`dir`	a character, either "left", "right", or "both". The default is "left". Walsh originally used 'left' which increases the outcome count in the left column (which is typically events)
`group`	a character specifying how to choose the single group in which to make modifications. The options are: 'event' for the fewest events (default), 'nonevent' for the fewest nonevents, or 'both' for the fewest overall. We assume that events are in the first column.