Showing Results from Cox Proportional Hazard Models in R with simPH

September 22, 2013

By Ahra Wu

[Ed. note: This post was contributed by Christopher Gandrud, a postdoctoral researcher at the Hertie School of Governance, and originally featured on his blog.]

Effectively showing estimates and uncertainty from Cox Proportional Hazard (PH) models, especially for interactive and non-linear effects, can be challenging with currently available software. So, researchers often just simply display a results table. These are pretty useless for Cox PH models. It is difficult to decipher a simple linear variable’s estimated effect and basically impossible to understand time interactions, interactions between variables, and nonlinear effects without the reader further calculating quantities of interest for a variety of fitted values.

So, I’ve been putting together the simPH R package to hopefully make it easier to show results from Cox PH models. The package uses plots of post-estimation simulations (the same idea behind the plotting facilities in the Zelig package) to show Cox PH estimates and the uncertainty surrounding them.

Here I want to give an overview of how to use simPH. First the general process, then a few examples. Have a look at this paper for details about the math behind the graphs.

General Steps

There are three steps to use simPH:

  1. Estimate a Cox PH model in the usual way with the coxph command in the survival package.
  2. Simulate quantities of interest–hazard ratiosfirst differencesmarginal effectrelative hazards, or hazard rates–with the appropriate simPH simulation command.
  3. Plot the simulations with the simGG method.

A Few Examples

Here are some basic examples that illustrate the process and key syntax. Before getting started you can install simPH in the normal R way from CRAN.

Linear effects

Let’s look at a simple example with a linear non-interactive effect. The data set I’m using is from Carpenter (2002). It is included with simPH. See the package documentation for details.

First, let’s estimate a Cox PH model where the event of interest is a drug receiving FDA approval.

# Load packages
library(survival)
library(simPH)

# Load Carpenter (2002) data
data("CarpenterFdaData")

# Estimate survival model
 M1 

Now say we want to examine the effect that the number of FDA staff reviewing a drug application has on it being accepted. This variable is called stafcder in the model we just estimated. To do this let’s use simPH to simulate hazard ratios. We will simulate hazard ratios for units j and l, i.e. \frac{h_{j}(t)}{h_{l}(t)} using simPH’s coxsimLinear command, because we estimate the effect of the number of staff as linear. In the following code we use the x_{j} argument to set the j values. We could use x_{l} also, but as we don’t coxsimLinear assumes all x_{l} are 0.

Notice that the argument spin = TRUE. This finds the shortest 95% probability interval of the simulations using the SPIn method. SPIn can be especially useful for showing simulated quantities of interest generated from Cox PH models, because then can often be crowded close to a lower boundary (0 in the case of hazard rates). We should be more interested in the area with the highest probability–most simulated values–rather than an arbitrary central interval. That being said, if spin = FALSE, then we will simply find the central 95% interval of the simulations.

# Simulate and plot Hazard Ratios for stafcder variable
Sim1 

Notice in the plot that each simulation is plotted as an individual point. These are all of the simulations in the shortest 95% probability interval. Each point has a bit of transparency (they are 90% transparent by default). So the plot is visually weighted; the darker areas of the graph have a higher concentration of the simulations. This gives us a very clear picture of the simulation distribution, i.e. the estimated effect and our uncertainty about it.

If you don’t want to plot every point, you can simply use ribbons showing the constricted 95% and the middle 50% of this interval. To do this simply use the ribbons = TRUE argument.

simGG(Sim1, ribbons = TRUE, alpha = 0.5)

Notice the alpha = 0.5 argument. This increased the transparency of the widest ribbon to 50%.

The syntax we’ve used here is very similar to what we use when we are working with nonlinear effects estimated with polynomials and splines. Post-estimation simulations can be run with the coxsimPoly and coxsimSpline commands. See the simPH documentation for more examples.

Interactions

The syntax for two-way interactions simulated with the coxsimInteract command is a little different. Using the same data, let’s look at how to show results for interactions. In the following model we are interacting two variables lethal and prevgenx. We can think of these variables as X1 and X2, respectively. For interactions it can be useful to examine the marginal effect. To find the marginal effect of a one unit increase in the lethal variable given various values of prevgenx let’s use the following code:

# Estimate the model
M2 

The order of the X1 and X2 variables in the interactions matters. The marginal effect is always calculated for the X1 variable over a range of X2 values.

Notice also that we set the x and y axis labels with the xlab and ylab arguments.

Time-varying Effects

Finally, let’s look at how to use with the coxsimtvc command to show results from effects that we estimate to vary over time. Here we are going to use another data set that is also included with simPH. The event of interest in the following model is when deliberation is stopped on a European Union directive (the model is from Licht (2011)). We will create hazard ratios for the effect that the number of backlogged items (backlog) has on deliberation time. We will estimate the effect as a log-time interaction.

# Load Golub & Steunenberg (2007) data. The data is included with simPH.
data("GolubEUPData")

# Create natural log-time interactions
Golubtvc 

Much of the first half of the code is dedicated to creating the log-time interactions with the tvc command.

Now we simply create the simulations for a range of values of backlog and plot them. Note in the following code that we tell coxsimtvc the name of both the base backlog variable and its log-time interaction term Lbacklog using the btvc argument. We also need to tell coxsimtvc that it is a log-time interaction with tfun = "log".

# Create simtvc object for relative hazard
Sim2 
simGG Plots are ggplot2

Finally, because almost every plot created by simGG is a ggplot2 plot, you can use almost the full range of customisation options that that package allows. See the ggplot2 documentation for many examples.

See also: Post