PwePred (version 1.0.0)

plot_event: Plot Cumulative Event Curve

Description

Plot cumulative event curve with right censoring data.

Usage

# S3 method for default
plot_event(time, event, abs_time=TRUE, additional_event=0,
         add=FALSE, plot=TRUE, xyswitch=FALSE, ...)
# S3 method for predict.pwexpm
plot_event(time, abs_time=TRUE, add=TRUE, plot=TRUE,
         xyswitch=FALSE, eval_at=NULL, ...)
# S3 method for predict.boot.pwexpm
plot_event(time, abs_time=TRUE, alpha=0.1, type='confidence',
         add=TRUE, plot=TRUE, xyswitch=FALSE, eval_at=NULL,
         show_CI=TRUE, CI_par=NULL, ...)

Value

When xyswith=FALSE, the function returns a data frame containing these columns:

time

sorted time at eval_at

n_event

predicted cumulative number of events

(alpha/2) n_event

(for predict.boot.pwexpm) lower alpha level predicted cumulative number of events

1-(alpha/2) n_event

(for predict.boot.pwexpm) upper alpha level predicted cumulative number of events

When xyswith=TRUE, the function returns a data frame containing these columns:

n_event

sorted cumulative number of events at eval_at

time

predicted required time

(alpha/2) time

(for predict.boot.pwexpm) lower alpha level predicted required time

1-(alpha/2) time

(for predict.boot.pwexpm) upper alpha level predicted required time

Arguments

time

observed/follow-up time from individual randomization time (abs_time=FALSE) or from the first subject randomization time (abs_time=TRUE); or a predicted object from predict.pwexpm, or a predicted object with bootstrapping from predict.boot.pwexpm.

abs_time

logical; if TRUE, time is the time from first randomization of the trial. if FALSE, time is the time from the randomization of each subject.

event

the status indicator, 0=censor, 1=event. Other choices are TRUE/FALSE (TRUE = event).

additional_event

adding the cumulative number of events by a constant number from the beginning.

add

logical; if TRUE add lines to current plot.

plot

logical; if FALSE, do not plot any lines, but return the line data

xyswitch

logical; if TRUE, x-axis will be cumulative number of events and y will be the time.

eval_at

a vector of the time (when xyswitch=FALSE) or the number of events (when xyswitch=TRUE) that you want to make prediction on.

alpha

the significance level of the confidence interval.

type

the type of prediction required. The default confidence returns the confidence interval without random error; the alternative predictive returns the predictive interval.

show_CI

logical; if TRUE add confidence interval of the estimated event curve.

CI_par

a list of parameters to control the apperance of lines of confidence intervals. The values pass to lines.

...

other arguments (e.g., lwd, etc.) are passed over to plot.

Author

Tianchen Xu

Details

A convenient function to calculate and plot the cumulative number of events.

Parameters in ... are passed to plot function to control the appearance of the event curve; parameters in CI_par are passed to lines function to control the appearance of confidence intervals. See examples for usage.

By default, plot_event plots a data frame in a new figure; and plots a predicted model in existing figure.

See Also

plot_survival

Examples

set.seed(1818)
event_dist <- function(n)rpwexpm(n, rate = c(0.1, 0.2), breakpoint = 14)
dat <- simdata(rand_rate = 20, drop_rate = 0.03, total_sample = 500,
               advanced_dist = list(event_dist=event_dist),
               add_column = c('censor_reason','event','followT','followT_abs'))
cut <- quantile(dat$randT, 0.8)
train <- cut_dat(var_randT = 'randT', cut = cut, data = dat,
                 var_followT = 'followT', var_followT_abs = 'followT_abs',
                 var_event = 'event', var_censor_reason = 'censor_reason')

fit_res3 <- pwexpm(Surv(followT, event), data=train, nbreak = 1)
fit_res_boot <- boot.pwexpm(fit_res3, nsim = 8) # here nsim=8 is for demo purpose,
# pls increase it in practice

drop_indicator <- ifelse(train$censor_reason=='drop_out' & !is.na(train$censor_reason),1,0)
fit_res_censor <- pwexpm_fit(train$followT, drop_indicator, nbreak = 0)
fit_res_censor_boot <- boot.pwexpm(fit_res_censor, nsim = 8)

cut_indicator <- train$censor_reason=='cut'
cut_indicator[is.na(cut_indicator)] <- 0

predicted_boot <- predict(fit_res_boot, cut_indicator = cut_indicator,
                          analysis_time = cut,  censor_model=fit_res_censor_boot,
                          future_rand=list(rand_rate=20, total_sample=NROW(dat)-NROW(train)))

plot_event(train$followT_abs, train$event, xlim=c(0,69), ylim=c(0,500))
plot_event(predicted_boot, eval_at = seq(40,90,5), CI_par = list(lty=3, lwd=2))

plot_event(train$followT_abs, train$event, xyswitch = TRUE, ylim=c(0,69), xlim=c(0,400))
plot_event(predicted_boot, xyswitch = TRUE,  eval_at = seq(250,400,50))