Predict overall survival and cumulative incidence functions

Usage

predict_comprisks2ts(models = list(), newdata, u, s, ds = NULL)

Arguments

models: A list with cause-specific hazard models of class 'haz2ts', fitted via fit2ts(). The models have to be fitted on the same grid of points over the time axes.
newdata: A dataframe with columns cointaing the values of the variable u and the time scale s for which predictions are to be obtained.
u: The name of the variable in newdata containing values for the variable u.
s: The name of the variable in newdata containing values for the variable s.
ds: (optional) The distance between two consecutive points on the s axis. If not provided, an optimal minimum value will be chosen automatically and a warning is returned.

Value

A dataframe cointaing the values of u and s in newdata, the predicted survival probability and the values of the cumulative incidence functions (cif), for each combination of u and s. There will be as many values of the cif as the number of cause-specific hazard models.

Details

Predictions of cumulated quantities can be provided only within intervals of values on the s time scale.

Examples

id <- 1:20
u <- c(5.43, 3.25, 8.15, 5.53, 7.28, 6.61, 5.91, 4.94, 4.25, 3.86, 4.05, 6.86,
       4.94, 4.46, 2.14, 7.56, 5.55, 7.60, 6.46, 4.96)
s <- c(0.44, 4.89, 0.92, 1.81, 2.02, 1.55, 3.16, 6.36, 0.66, 2.02, 1.22, 3.96,
       7.07, 2.91, 3.38, 2.36, 1.74, 0.06, 5.76, 3.00)
ev <- c(1, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1)#'
fakedata <- as.data.frame(cbind(id, u, s, ev))
# cause 1
fakedata2ts1 <- prepare_data(u = fakedata$u,
                            s_out = fakedata$s,
                            ev = fakedata$ev,
                            ds = .5)
#> `s_in = NULL`. I will use `s_in = 0` for all observations.
#> `s_in = NULL`. I will use `s_in = 0` for all observations.
# Fit a fake model - not optimal smoothing for cause type 1
fakemod1 <- fit2ts(fakedata2ts1,
                  optim_method = "grid_search",
                  lrho = list(seq(1 ,1.5 ,.5),
                              seq(1 ,1.5 ,.5)))
# cause 2
fakedata2ts2 <- prepare_data(u = fakedata$u,
                              s_out = fakedata$s,
                              ev = 1-(fakedata$ev),
                              ds = .5)
#> `s_in = NULL`. I will use `s_in = 0` for all observations.
#> `s_in = NULL`. I will use `s_in = 0` for all observations.
# Fit a fake model - not optimal smoothing for cause 2
fakemod2 <- fit2ts(fakedata2ts2,
                    optim_method = "grid_search",
                    lrho = list(seq(1 ,1.5 ,.5),
                                seq(1 ,1.5 ,.5)))

newdata <- as.data.frame(cbind("u" =  c(2.5,3.4,6), "s" = c(.2,.5,1.3)))

predict_comprisks2ts(models = list(fakemod1, fakemod2),
                     newdata,
                     u = "u", s = "s")
#> chosen interval: ds = 0.2
#> Warning: Right boundary adjusted to max(x) = 7.5
#> Warning: Right boundary adjusted to max(x) = 7.5
#> Warning: Right boundary adjusted to max(x) = 7.5
#>      u   s  survival      cif_1       cif_2
#> 16 2.5 0.2 0.9314345 0.06385976 0.003495532
#> 32 3.4 0.5 0.8958052 0.09253255 0.009763526
#> 93 6.0 1.3 0.7107880 0.14481779 0.137412289