plantation: Plantation (regular forest) plot
In tobyjohnson/gtx: Genetics ToolboX

Description Usage Arguments Details Author(s) Examples

Provides a simple forest plot. Allows an arbitrary functional relationship between the plotting scale and the numerical values displayed, useful hazard ratios or odds ratios where plotting on a log scale is preferred.

plantation(ntext, beta, ci.lo, ci.hi, se, alpha = 0.05,
                       FUN = I, pvals = TRUE, meta = TRUE,
                       xzero = 0, xlim, xticks, xlab = "Effect",
                       digits = 2,
                       groups = list("Fixed effect meta-analysis" = 1:length(beta)))

`ntext`	Names for each item
`beta`	Effect sizes on the plotting scale
`ci.lo`	Lower CI limits
`ci.hi`	Upper CI limits
`se`	Standard errors
`alpha`	Alpha level for CIs
`FUN`	Function to transform from plotting scale to printing scale
`pvals`	Whether to write P-values
`meta`	Calculate meta-analysis
`xzero`	Value of X at which to
`xlim`	X-axis limits on plotting scale
`xticks`	X-axis ticks on plotting scale
`xlab`	X-axis label
`digits`	digits
`groups`	List of groups for meta-analyses

The user can supply either standard errors (SEs) or upper and lower limits of confidence intervals (CIs). Whichever is missing is calculated from the other, assuming the CIs are symmetric and at the level alpha.

A key concept is the separation of the printing scale (which is the scale that effect sizes are conventionally reported on), and the plotting scale (which should typically be the same as the analysis scale; the scale on which normal theory for the estimator best holds). The most common examples are the printing scale being hazard ratio (HR) or odds ratio (OR), and the plotting scale being (natural) log HR or log OR. A key feature of plantation is that it allows an arbitrary function FUN to map the plotting scale onto the printing scale.

See the examples and vignette.

Toby Johnson Toby.x.Johnson@gsk.com

plantation(paste("Study ", 1:4, " (drug ", c("A", "A", "B", "B"), ")", sep = ""),
           c(0.41, 0.06, 0.39, 0.50),
           se = c(0.12, 0.17, 0.18, 0.25),
           FUN = exp, digits = 2,
           xlim = log(c(0.8, 2.5)), xticks = log(c(0.8, 1, 1.5, 2, 2.5)),
           xlab = "HR", 
           groups = list("Fixed effect all drug A" = 1:2,
             "Fixed effect all drug B" = 3:4,
             "Fixed effect all studies" = 1:4))