Creates a log(-log(survival)) vs log(time) plot for visual assessment of the proportional hazards assumption. Parallel curves indicate the PH assumption is likely satisfied.
Usage
create_loglog_plot(
data,
time_var,
event_var,
trt_var,
title = "Log-Log Survival Plot",
xlab = "Log(Time)",
ylab = "Log(-Log(Survival))",
show_censor = TRUE,
colors = NULL,
base_size = 11,
conf_level = 0.95
)Arguments
- data
ADaMData object or data frame with time-to-event data
- time_var
Character. Name of the time variable
- event_var
Character. Name of the event variable (1=event, 0=censor). If "CNSR" (ADaM censoring flag), it will be inverted (0=event becomes 1).
- trt_var
Character. Name of the treatment variable
- title
Character. Plot title (default: "Log-Log Survival Plot")
- xlab
Character. X-axis label (default: "Log(Time)")
- ylab
Character. Y-axis label (default: "Log(-Log(Survival))")
- show_censor
Logical, show censoring marks as crosses (default: TRUE).
- colors
Named character vector of colors for each treatment group. If NULL, uses
getOption("pharmhand.palette")or the default palette.- base_size
Base font size for plot text elements (default: 11).
- conf_level
Confidence level for survival fit (default: 0.95)
Details
Under the Cox proportional hazards assumption, the log-log transformed survival curves should be approximately parallel. Crossing or diverging curves suggest the PH assumption may be violated.
This is a complementary visual diagnostic to the statistical test provided by test_ph_assumption(). Censor marks are supported; median lines, CI bands, risk tables, and landmarks are intentionally omitted to keep the diagnostic scale uncluttered.
Examples
if (FALSE) { # \dontrun{
# Log-log plot to assess PH assumption
plot <- create_loglog_plot(
data = adtte,
time_var = "AVAL",
event_var = "CNSR",
trt_var = "TRT01P",
title = "Log-Log Plot: PH Assumption Check"
)
print(plot)
} # }