survrep: Survival Analysis Results for Reports

In biostatUZH: Misc Tools of the Department of Biostatistics, EBPI, University of Zurich

Description

The analysis of survival data most often requires the production of Kaplan-Meier plots, estimated median survival time, and univariate estimates of hazard ratios (with associated confidence intervals and p-values). Often, this is performed separately for several possible risk factors. This function simplifies these tasks by providing a unified interface to simultaneously produce graphics, estimate median survival time, and compute hazard ratios with associated statistics from the Cox proportional hazards model.

Usage

survrep(ftime, fstatus, fgroup = NA,
        stmt.pl = c("bottomleft", "bottomright", "topright", "subtitle", "none")[1],
        legend.pl = c("bottomleft", "bottomright", "topright", "none")[3],
        output = c("plain", "text")[2],
        maintitle = "",
        ylbl = "Survival",
        xlbl = "Time",
        lbls = levels(fgroup),
        dig = 2, 
        conf.level = 0.95)

Arguments

ftime	Event times, censored or observed.
fstatus	Censoring indicator, 1 for event, 0 for censored.
fgroup	Kaplan-Meier plots, median survival and hazard rates will be computed by fgroup. If fgroup is missing, Kaplan-Meier plots and median survival will be produced for the entire dataset.
stmt.pl	Placement of the statement giving hazard ratios, confidence intervals and p-values in the plot. Ignored if fgroup = NA.
legend.pl	Placement of the legend in the plot. Ignored if fgroup = NA. A warning will be produced if stmt.pl = legend.pl.
output	Type of output, “plain” gives plaintext output for median survival and hazard ratios suitable for the creation of tables etc., while “text” gives output suitable for the text of reports. “text”-type output is always printed in the plots unless stmt.pl="none".
maintitle	Main title of the plot. Corresponds to the main option of plot.survfit.
ylbl	Label for the y-axis of the plot. Corresponds to the ylab option of plot.survfit.
xlbl	Label for the x-axis of the plot. Corresponds to the xlab option of plot.survfit.
lbls	Labels used to produce the legend, as well as the contrasts printed in the hazard ratio statements.
dig	Number of significant digits used in rounding for “text”-type output.
conf.level	Significance level for confidence interval for the median survival and hazard ratio estimates.

Value

The Kaplan-Meier plot is printed, as well as a list containing:

med	Matrix of median survival, and corresponding confidence intervals. If output = "text", this will be a vector of text giving median survival estimates and confidence intervals.
events	Matrix of hazard ratios, corresponding confidence intervals, and p-values from the Wald test. If output = "text", this will be a vector of text.

Author(s)

Sarah R. Haile

See Also

survfit, coxph, quantileKM

Examples

## use Acute Myelogenous Leukemia survival data contained in package 'survival'
time <- leukemia[, 1]; status <- leukemia[, 2]; x <- leukemia[, 3]
survrep(time, status, NA)
survrep(time, status, x)
survrep(time, status, x, stmt.pl = "bottomright", out = "plain", main = "Acute Myeloid Leukemia")
survrep(time, status, x, stmt.pl = "bottomright", out = "text", main = "AML")
survrep(time, status, x, stmt.pl = "subtitle", out = "text", main = "Acute Myeloid Leukemia")

## using Larynx data contained in package 'survival'
## to show behavior with risk factors with >2 levels
data(larynx)
larynx$stage <- factor(larynx$stage, 1:4,  c("I", "II", "III", "IV"))
with(larynx, survrep(time, death, stage, out = "t",
                     stmt.pl = "bottomright", main = "Larynx"))
with(larynx, survrep(time, death, stage, out = "t",
                     stmt.pl = "subtitle", main = ""))
with(larynx, survrep(time, death, stage, out = "p",
                     stmt.pl = "topright", legend.pl = "bottomright"))

biostatUZH documentation built on May 2, 2019, 6:06 p.m.