Skip to contents

Fine-Gray Subdistribution Hazard Regression for One Predictor

Source: R/tbl_fine_gray_single.R
tbl_fine_gray_single.Rd

Fits the Fine-Gray (1999) proportional subdistribution hazards model of one cause with the other causes treated as competing events, for one predictor at a time. It is the regression counterpart of [tbl_cumulative_incidence()], as [cox_single()] is for [km_single()]. Uses `cmprsk::crr()`.

Usage

tbl_fine_gray_single(
  data,
  time_col,
  event_col,
  feature,
  failcode = 1,
  cencode = 0,
  conf_level = 0.95,
  exponentiate = TRUE,
  ...
)

Arguments

data

A data frame, one row per subject.

time_col

Follow-up time.

event_col

Event type, as in [tbl_cumulative_incidence()]: a censoring code, then one code per cause.

feature

Name of the predictor column.

failcode

Code of the cause of interest (default 1).

cencode

Code for censoring (default 0).

conf_level

Confidence level.

exponentiate

Report the subdistribution hazard ratio (default); `FALSE` gives the coefficient on the log scale.

...

Passed to `cmprsk::crr()` (for example `maxiter`).

Value

A tibble with `feature`, `term`, `estimate`, `std.error` (of the log coefficient), `statistic`, `p.value`, `conf.low`, `conf.high`, `n` and `n_event` (events of the cause of interest among the rows used).

Interpretation

The estimate is a subdistribution hazard ratio (sHR). It describes the effect on the cumulative incidence of the cause of interest, not on the rate among those still event-free, so it is not comparable to a cause-specific hazard ratio from [cox_single()] with the competing events censored. Subjects with a competing event stay in the risk set, which is what `crr()` does. Rows with a missing time, event code or predictor are dropped. A factor or character predictor gets one row per non-reference level (the first level is the reference). Terms are named as `broom::tidy()` names them: a logical predictor gives `<feature>TRUE`, a factor `<feature><level>`.

See also

[tbl_cumulative_incidence()], [plot_cumulative_incidence()], [cox_single()]

Examples

if (requireNamespace("cmprsk", quietly = TRUE)) {
  set.seed(1)
  d <- data.frame(time = rexp(150, 0.1), type = sample(0:2, 150, TRUE, prob = c(0.3, 0.4, 0.3)),
                  age = rnorm(150, 60, 8), arm = rep(c("A", "B"), 75))
  tbl_fine_gray_single(d, "time", "type", "age")
  tbl_fine_gray_single(d, "time", "type", "arm", failcode = 2)
}
#> # A tibble: 1 × 10
#>   feature term  estimate std.error statistic p.value conf.low conf.high     n
#>   <chr>   <chr>    <dbl>     <dbl>     <dbl>   <dbl>    <dbl>     <dbl> <int>
#> 1 arm     armB      1.50     0.328      1.24   0.216    0.789      2.86   150
#> # ℹ 1 more variable: n_event <int>