msmplot: Combined plot of the multistate model and the cumulative...
In aallignol/mvna: Nelson-Aalen estimator of the cumulative hazard in multistate models

Description Usage Arguments Details Value Note Author(s) See Also Examples

This function uses the package Rgraphviz to display the multistate model. Next GrisBase is used to put the cumulative hazard plots near the arrows representing transitions.

## S3 method for class 'mvna'
msmplot(object, tr.choice, graph, layout = "dot",
        recip.edges = "combined", order = "LR", unit = "npc",
        width, height, just = "center",
        mgp = c(1.3, 0.4, 0), cex = 0.6,
        mtext = TRUE, label.plots, side = 3, lcex = 0.9,
        conf.int = TRUE, level = 0.05, xlab = "Time", ylab = "",
        ...)

`object`	An object of class `mvna`.
`tr.choice`	A character vector of the form ("from to", "from to") describing the transitions to plot. By default, the estimated cumulative hazards are displayed for all transitions.
`graph`	Optional. An object of class `Ragraph`. People familiar with the Rgraphviz and/or willing to have a better control on how the multistate model is displayed can provide here their own `Ragraph` object. If not provided, the multistate plot is done internally.
`layout`	If `graph` is not provided, permits to specify the layout of the multistate plot.
`recip.edges`	If `graph` is not provided, permits to specify how to handle reciprocated edges, defaults is 'combined'.
`order`	Order for the display of the nodes. Default is 'LR', that is from left to right.
`unit`	A string specifying the units used for creating the viewports in which the cumulative hazards will be plotted. Default is "npc". See the grid vignette for the possible values.
`width`	A vector giving the width of the viewport for plotting the cumulative hazards.
`height`	A vector giving the height of the viewport.
`just`	A string or numeric vector indicating the justification of the viewport.
`mgp`	Margin line for the axis title, axis labels and axis line in the cumulative hazards plots. Default is c(1.3, 0.4, 0), smaller that the default value in `par` due to the fact the plots will usually be small.
`cex`	A numerical value giving the amount by which plotting text and symbols should be magnified relative to the default.
`mtext`	A logical indicating whether to display a label above the plot.
`label.plots`	If `mtext = TRUE`, specifies the labels. Default is the transition numbers.
`side`	Gives the side where to put the label (1=bottom, 2=left, 3=top, 4=right). Default is 3.
`lcex`	Character expansion factor. This is an absolute value. Therefore, this value will be evaluate with respect with `cex` above.
`conf.int`	A logical indicating whether plot confidence intervals. Default is TRUE
`level`	Level for the confidence interval. Default is 0.05
`xlab`	Label for the x-axis.
`ylab`	Label for the y-axis.
`...`	Further arguments for the plot method.

This functions permits to display the cumulative hazards estimates in multistate models in an original manner. A drawing of the multistate model is provided along with the display of the estimated cumulative hazards that are put along the arrows that describes the possible transitions.

The multistate model is displayed using the Rgraphviz package. By default, the Ragraph object needed for the display is computed internally, letting few options for the user to be modified. For a finer control, the user can provide his own Ragraph object using the agopen function.

The cumulative hazards plots are positioned by the user using the grid.locator function. One has just to click where he wants the plot to be. The order of the transitions is displayed on the prompt. The width, height, angle and just arguments permit to adjust the way the plots are displayed.

object with the invisible flag set

The Rgraphviz package requires the graphviz software to be installed.

Rgraphviz binaries for windows are not available on CRAN, but are on bioconductor.

Arthur Allignol, arthur.allignol@uni-ulm.de

mvna

data(sir.cont)

sir.cont <- sir.cont[order(sir.cont$id, sir.cont$time), ]
for (i in 2:nrow(sir.cont)) {
  if (sir.cont$id[i]==sir.cont$id[i-1]) {
    if (sir.cont$time[i]==sir.cont$time[i-1]) {
      sir.cont$time[i-1] <- sir.cont$time[i-1] - 0.5
    }
  }
}

### Computation of the transition probabilities
# Possible transitions
tra <- matrix(ncol=3,nrow=3,FALSE)
tra[1, 2:3] <- TRUE
tra[2, c(1, 3)] <- TRUE

# mvna
temp <- mvna(sir.cont, c("0", "1", "2"), tra, "cens")

## Not run: 
msmplot(temp, width = c(0.25, 0.25, 0.2, 0.25),
        height = c(0.2, 0.2, 0.18, 0.2), lcex=0.8)

## End(Not run)