Change axis ranges for a ggplot based on a Predict object (rms package)

Question

I am using the rms package to perform Cox regression with age as restricted cubic splines with 4 knots, see reproducible code below from the rms package documentation. My problem is: How do I change the y axis interval for the ggplot-output? I tried adding scale_y_continuous(limits = c(0,20)) but it didn't alter the axis

library(ggplot)
library(rms)

# NOT RUN {
# Simulate data from a population model in which the log hazard
# function is linear in age and there is no age x sex interaction

n <- 1000
set.seed(731)
age <- 50 + 12*rnorm(n)
label(age) <- "Age"
sex <- factor(sample(c('Male','Female'), n, 
                     rep=TRUE, prob=c(.6, .4)))
cens <- 15*runif(n)
h <- .02*exp(.04*(age-50)+.8*(sex=='Female'))
dt <- -log(runif(n))/h
label(dt) <- 'Follow-up Time'
e <- ifelse(dt <= cens,1,0)
dt <- pmin(dt, cens)
units(dt) <- "Year"
dd <- datadist(age, sex)
options(datadist='dd')
S <- Surv(dt,e)

f <- cph(S ~ rcs(age,4) + sex, x=TRUE, y=TRUE)
cox.zph(f, "rank")             # tests of PH
anova(f)
ggplot(Predict(f, age, sex, fun=exp)) + # plot age effect, 2 curves for 2 sexes
  scale_y_continuous(limits = c(0,20))

Allan Cameron · Accepted Answer · 2022-04-20T10:00:46.503

Unusually, the Predict class of rms has its own S3 method for ggplot, which automatically adds position scales, coordinates and geom layers. This makes it easier to plot Predict objects, but limits extensibility. In particular, it already sets the y limits via a CoordCartesian, which over-rides any y axis scales you add.

The simplest way round this is to add a new coord_cartesian which will over-ride the existing Coord object (though also generate a message).

ggplot(Predict(f, age, sex, fun=exp)) + # plot age effect, 2 curves for 2 sexes
  coord_cartesian(ylim = c(0, 20))
#> Coordinate system already present. Adding new coordinate system, 
#> which will replace the existing one.

The alternative is to store the plot and change its coord limits, which doesn't generate a message but is a bit "hacky"

p <- ggplot(Predict(f, age, sex, fun=exp))
p$coordinates$limits$y <- c(0, 20)
p

You can also change the y axis limits to NULL in the above code, which will allow your axis limits to be set using scale_y_continuous instead.

This solved the problem! Is there any specific reason for scale_y_continuous() not to work? Just to understand the underlying functionality. — Pontus Hedberg, Apr 20 '22 at 09:48

score 2 · Answer 2 · answered Apr 21 '22 at 10:26

Although a bit laborious, you could always create the plot from first principles by converting the object to a data.frame first. The benefit with that way is you have full control over what you want to add.

age_data <- as.data.frame(Predict(f, age, sex, fun = exp))
ggplot(age_data, aes(x = age, y = yhat, colour = sex)) + 
  geom_line() +
  geom_ribbon(aes(ymin = lower, ymax = upper), alpha = 0.2) +
  scale_y_continuous(limits = c(0, 20))

@Allans approach is much simpler for your particular request though!

Change axis ranges for a ggplot based on a Predict object (rms package)

2 Answers2