joineRML (version 0.4.6)

plot.dynSurv: Plot a dynSurv object

Description

Plots the conditional time-to-event distribution for a new subject calculated using the dynSurv function.

Usage

# S3 method for dynSurv
plot(
  x,
  main = NULL,
  xlab = NULL,
  ylab1 = NULL,
  ylab2 = NULL,
  grid = TRUE,
  estimator,
  smooth = FALSE,
  ...
)

Value

A dynamic prediction plot.

Arguments

x

an object of class dynSurv calculated by the dynSurv function.

main

an overall title for the plot: see title.

xlab

a title for the x [time] axis: see title.

ylab1

a character vector of the titles for the K longitudinal outcomes y-axes: see title.

ylab2

a title for the event-time outcome axis: see title.

grid

adds a rectangular grid to an existing plot: see grid.

estimator

a character string that can take values mean or median to specify what prediction statistic is plotted from an objecting inheritting of class dynSurv. Default is estimator='median'. This argument is ignored for non-simulated dynSurv objects, i.e. those of type='first-order', as in that case a mode-based prediction is plotted.

smooth

logical: whether to overlay a smooth survival curve (see Details). Defaults to FALSE.

...

additional plotting arguments; currently limited to lwd and cex. See par for details.

Author

Graeme L. Hickey ()

Details

The joineRML package is based on a semi-parametric model, such that the baseline hazards function is left unspecified. For prediction, it might be preferable to have a smooth survival curve. Rather than changing modelling framework a prior, a constrained B-splines non-parametric median quantile curve is estimated using cobs, with a penalty function of \(\lambda=1\), and subject to constraints of monotonicity and \(S(t)=1\).

References

Ng P, Maechler M. A fast and efficient implementation of qualitatively constrained quantile smoothing splines. Statistical Modelling. 2007; 7(4): 315-328.

Rizopoulos D. Dynamic predictions and prospective accuracy in joint models for longitudinal and time-to-event data. Biometrics. 2011; 67: 819–829.

See Also

dynSurv

Examples

if (FALSE) {
# Fit a joint model with bivariate longitudinal outcomes

data(heart.valve)
hvd <- heart.valve[!is.na(heart.valve$log.grad) & !is.na(heart.valve$log.lvmi), ]

fit2 <- mjoint(
    formLongFixed = list("grad" = log.grad ~ time + sex + hs,
                         "lvmi" = log.lvmi ~ time + sex),
    formLongRandom = list("grad" = ~ 1 | num,
                          "lvmi" = ~ time | num),
    formSurv = Surv(fuyrs, status) ~ age,
    data = list(hvd, hvd),
    inits = list("gamma" = c(0.11, 1.51, 0.80)),
    timeVar = "time",
    verbose = TRUE)

hvd2 <- droplevels(hvd[hvd$num == 1, ])
out1 <- dynSurv(fit2, hvd2)
plot(out1, main = "Patient 1")
}

if (FALSE) {
# Monte Carlo simulation with 95% confidence intervals on plot

out2 <- dynSurv(fit2, hvd2, type = "simulated", M = 200)
plot(out2, main = "Patient 1")
}