Skip to contents

Cumulative Incidence with Competing Risks

Source: R/tbl_cumulative_incidence.R
tbl_cumulative_incidence.Rd

Aalen-Johansen estimate of the cumulative incidence of each cause at chosen times, optionally by group. Unlike `1 - Kaplan-Meier` with the other causes treated as censored, this does not overstate a cause's incidence when competing events remove subjects from risk.

Usage

tbl_cumulative_incidence(
  data,
  time_col,
  event_col,
  group_col = NULL,
  times = NULL,
  conf_level = 0.95
)

Arguments

data

A data frame, one row per subject.

time_col

Follow-up time.

event_col

Event type: 0 (or the first level of a factor) for censored, then one code per cause.

group_col

Optional grouping column.

times

Times at which to report. Default: the quartiles of the observed event times.

conf_level

Confidence level of the pointwise interval.

Value

A tibble with `group` (when `group_col` is used), `cause`, `time`, `cif`, `std.error`, `conf.low`, `conf.high`.

Scope

This describes cumulative incidence. It does not fit a Fine-Gray subdistribution hazard model; for that use `cmprsk::crr()` or `tidycmprsk`.

See also

[tbl_landmark_cutpoints()]

Examples

set.seed(1)
d <- data.frame(time = rexp(200, 0.1), type = sample(0:2, 200, TRUE, prob = c(0.3, 0.4, 0.3)),
                arm = rep(c("A", "B"), 100))
tbl_cumulative_incidence(d, "time", "type", group_col = "arm", times = c(5, 10))
#> # A tibble: 8 × 7
#>   group cause  time   cif std.error conf.low conf.high
#>   <chr> <chr> <dbl> <dbl>     <dbl>    <dbl>     <dbl>
#> 1 A     1         5 0.193    0.0410   0.128      0.293
#> 2 A     1        10 0.304    0.0493   0.221      0.418
#> 3 A     2         5 0.105    0.0313   0.0581     0.188
#> 4 A     2        10 0.141    0.0363   0.0848     0.233
#> 5 B     1         5 0.116    0.0330   0.0667     0.203
#> 6 B     1        10 0.268    0.0471   0.190      0.378
#> 7 B     2         5 0.105    0.0314   0.0582     0.188
#> 8 B     2        10 0.212    0.0434   0.142      0.316