# score_cr()


Bundled competing-risks evaluation: concordance, Brier, and integrated Brier.


Usage

``` python
score_cr(
    surv,
    incidence_prob,
    *,
    data=None,
    cause,
    times,
    tau=None,
    conf_level=0.95,
    format=None,
)
```


Computes the main prediction-performance metrics for a cause-specific cumulative incidence model in one call, returning a tidy summary frame.


## Parameters


`surv: Surv | Outcome | str`  
A multistate [Surv](Surv.md#greenwood.Surv) response. An [Outcome](Outcome.md#greenwood.Outcome) or a formula response such as `'Surv(time, status == 2)'` is also accepted, with its columns read from `data`.

`incidence_prob: Any`  
Predicted CIF for the cause of interest, shape `(n_subjects, n_times)`.

`cause: int`  
Integer cause code.

`times: Any`  
Evaluation times (1-D array-like), matching the columns of `incidence_prob`.

`tau: float | None = None`  
Truncation time for the concordance index. Defaults to the last value in `times`.

`conf_level: float = ``0.95`  
Confidence level for the concordance CI.

`format: str | None = None`  
DataFrame backend for the returned table.

`data: Any = None`  
A data frame (pandas, Polars, PyArrow, DuckDB, a lazy frame, …) holding the columns named by the response and `incidence_prob`. When `surv` is an [Outcome](Outcome.md#greenwood.Outcome) or a formula, rows with a missing value in any column used are dropped first, along with the matching rows of any arrays passed alongside.


## Returns


`DataFrame`  
One row per metric with columns `metric`, `estimate`, and (for the concordance) `se`, `conf_low`, `conf_high`.


## Examples


``` python
import greenwood as gw
import numpy as np

mg = gw.load_dataset("mgus2", backend="polars")
first = gw.Outcome.first_event(
    endpoints={"pcm": ("ptime", "pstat"), "death": ("futime", "death")}
)

fg = gw.FineGray(cause="pcm").fit(first, covariates=["age", "sex"], data=mg)
eval_times = np.array([120, 240, 360])
cif_pred = fg.predict_cumulative_incidence(
    mg, times=eval_times, format="pandas"
).drop(columns="time").values.T

gw.score_cr(
    first, incidence_prob=cif_pred, cause=1, times=eval_times, data=mg, format="polars"
)
```


shape: (5, 5)

| metric             | estimate | se       | conf_low | conf_high |
|--------------------|----------|----------|----------|-----------|
| str                | f64      | f64      | f64      | f64       |
| "concordance"      | 0.664422 | 0.044569 | 0.57231  | 0.745517  |
| "brier(120)"       | 0.087738 | NaN      | NaN      | NaN       |
| "brier(240)"       | 0.147587 | NaN      | NaN      | NaN       |
| "brier(360)"       | 0.199861 | NaN      | NaN      | NaN       |
| "integrated_brier" | 0.145693 | NaN      | NaN      | NaN       |