DecisionCurve (version 1.4)

plot_decision_curve: Plot the net benefit curves from a decision_curve object or many decision_curve objects

Description

Plot the net benefit curves from a decision_curve object or many decision_curve objects

Usage

plot_decision_curve(x, curve.names, cost.benefit.axis = TRUE,
  n.cost.benefits = 6, cost.benefits, standardize = TRUE,
  confidence.intervals, col, lty, lwd = 2, xlim, ylim, xlab, ylab,
  cost.benefit.xlab, legend.position = c("topright", "right", "bottomright",
  "bottom", "bottomleft", "left", "topleft", "top", "none"), ...)

Arguments

x

'decision_curve' object to plot or a list of 'decision_curve' objects. Assumes output from function 'decision_curve'

curve.names

vector of names to use when plotting legends.

cost.benefit.axis

logical (default TRUE) indicating whether to print an additional x-axis showing relative cost:benefit ratios in addition to risk thresholds.

n.cost.benefits

number of cost:benefit ratios to print if cost.benefit.axis = TRUE (default n.cost.benefit = 6).

cost.benefits

Character vector of the form c("c1:b1", "c2:b2", ..., "cn:bn") with integers ci, bi corresponding to specific cost:benefit ratios to print. Default allows the function to calculate these automatically.

standardize

logical (default TRUE) indicating whether to use the standardized net benefit (NB/disease prevalence) or not.

confidence.intervals

logical indicating whether to plot confidence intervals.

col

vector of color names to be used in plotting corresponding to the 'predictors' given. Default colors will be chosen from rainbow(..., v = .8). See details for more information on plot parameters.

lty

vector of linetypes.

lwd

vector of linewidths.

xlim

vector giving c(min, max) of x-axis. Defaults to c(min(thresholds), max(thresholds)).

ylim

vector giving c(min, max) of y-axis.

xlab

label of main x-axis.

ylab

label of y-axis.

cost.benefit.xlab

label of cost:benefit ratio axis.

legend.position

character vector giving position of legend. Options are "topright" (default), "right", "bottomright", "bottom", "bottomleft", "left", "topleft", "top", or "none".

...

other options directly send to plot()

Details

When k decision_curve objects are input, the first k elements of col, lty, lwd ... correspond to the curves provided. The next two elements (..., k+1, k+2) correspond to the attributes of the 'all' and 'none' curves. See below for an example.

Examples

# NOT RUN {
data(dcaData)
set.seed(123)
baseline.model <- decision_curve(Cancer~Age + Female + Smokes,
                                data = dcaData,
                                thresholds = seq(0, .4, by = .005),
                                bootstraps = 10)

#plot using the defaults
plot_decision_curve(baseline.model,  curve.names = "baseline model")

set.seed(123)
full.model <- decision_curve(Cancer~Age + Female + Smokes + Marker1 + Marker2,
                            data = dcaData,
                            thresholds = seq(0, .4, by = .005),
                            bootstraps = 10)

# for lwd, the first two positions correspond to the decision curves, then 'all' and 'none'
plot_decision_curve( list(baseline.model, full.model),
                    curve.names = c("Baseline model", "Full model"),
                    col = c("blue", "red"),
                    lty = c(1,2),
                    lwd = c(3,2, 2, 1),
                    legend.position = "bottomright")


plot_decision_curve( list(baseline.model, full.model),
                    curve.names = c("Baseline model", "Full model"),
                    col = c("blue", "red"),
                    confidence.intervals = FALSE,  #remove confidence intervals
                    cost.benefit.axis = FALSE, #remove cost benefit axis
                    legend.position = "none") #remove the legend

#Set specific cost:benefit ratios.

plot_decision_curve( list(baseline.model, full.model),
                    curve.names = c("Baseline model", "Full model"),
                    col = c("blue", "red"),
                    cost.benefits = c("1:1000", "1:4", "1:9", "2:3", "1:3"),
                    legend.position = "bottomright")

#Plot net benefit instead of standardize net benefit.

plot_decision_curve( list(baseline.model, full.model),
                    curve.names = c("Baseline model", "Full model"),
                    col = c("blue", "red"),
                    ylim = c(-0.05, 0.15), #set ylim
                    lty = c(2,1),
                    standardize = FALSE, #plot Net benefit instead of standardized net benefit
                   legend.position = "topright")


# }