CompetingRiskRegression

Fine-Gray regression family for associate$regression()

Description

CompetingRiskRegression implements Fine-Gray subdistribution-hazards models through the suggested {cmprsk} package.

The class prepares a competing-risk analysis frame, converts the event and competing indicators to the status coding required by {cmprsk}, fits the Fine-Gray model, extracts predictor coefficient rows on the log-subdistribution-hazard scale, and derives analysis-level cumulative-incidence artifacts for downstream plotting.

In the user-facing associate$regression() call, outcomes is the biological or clinical endpoint being modeled and is supplied as the event indicator for the event of interest. time is the observed age or follow-up time at endpoint, competing event, or censoring. competing is the separate indicator for an event that prevents later observation of the endpoint. A dementia-onset analysis with death before dementia therefore uses dementia onset as outcomes, age at onset/death/last observation as time, and non-demented death as competing.

Summary rows follow the crr association schema. Competing-risk artifacts include observed.curves, predicted.curves, prediction.grid, and group.summary.

Super class

PolyGenius::RegressionFamily -> CompetingRiskRegression

Methods

Public methods

Inherited methods

Method new()

Register the canonical family key used by the engine and summary rows.

Usage 
CompetingRiskRegression$new()
Returns

The initialized family object with family.name = “crr”.

Method validate.frame()

Validate competing-risk-specific inputs and package availability.

Usage 
CompetingRiskRegression$validate.frame(frame, cell, fit.specs, conf.level)
Arguments
frame

Resolved analysis frame restricted to the current fit cell.

cell

Named list describing the current fit-grid cell.

fit.specs

Additional family-specific fitting arguments.

conf.level

Confidence level for intervals.

Returns

Invisibly TRUE when the Fine-Gray inputs are valid.

Method prepare.frame()

Build the complete-case competing-risk payload and design matrix for one fit.

Usage 
CompetingRiskRegression$prepare.frame(frame, cell, fit.specs, conf.level)
Arguments
frame

Resolved analysis frame restricted to the current fit cell.

cell

Named list describing the current fit-grid cell.

fit.specs

Additional family-specific fitting arguments.

conf.level

Confidence level for intervals.

Returns

A compact list with raw data, design matrix, and retained row ids.

Method build.formula()

Build the symbolic Fine-Gray formula for one fit payload.

Usage 
CompetingRiskRegression$build.formula(frame)
Arguments
frame

Fit payload returned by prepare.frame().

Returns

A model formula object.

Method fit.model()

Fit the Fine-Gray regression using the prepared payload.

Usage 
CompetingRiskRegression$fit.model(formula, frame, ...)
Arguments
formula

Model formula returned by build.formula().

frame

Family-specific payload returned by prepare.frame().

…

Additional family-specific fitting arguments.

Returns

A fitted cmprsk::crr object.

Method build.summary.rows()

Convert fitted Fine-Gray coefficients to standardized regression summary rows.

Usage 
CompetingRiskRegression$build.summary.rows(
  fit,
  fit.data,
  cell,
  fit.specs,
  conf.level,
  formula
)
Arguments
fit

Fitted cmprsk::crr object.

fit.data

Family-specific payload returned by prepare.frame().

cell

Named list describing the current fit-grid cell.

fit.specs

Additional family-specific fitting arguments.

conf.level

Confidence level for intervals.

formula

Display formula string.

Returns

A data.table of regression summary rows for the focal predictor.

Method build.artifacts()

Return observed and adjusted cumulative-incidence artifacts.

Usage 
CompetingRiskRegression$build.artifacts(
  fit,
  fit.data,
  cell,
  fit.specs,
  conf.level,
  formula
)
Arguments
fit

Fitted cmprsk::crr object.

fit.data

Family-specific payload returned by prepare.frame().

cell

Named list describing the current fit-grid cell.

fit.specs

Additional family-specific fitting arguments.

conf.level

Confidence level for intervals.

formula

Display formula string.

Returns

A named list of observed curves, predicted curves, prediction profiles, and group summary.

Method clone()

The objects of this class are cloneable with this method.

Usage 
CompetingRiskRegression$clone(deep = FALSE)
Arguments
deep

Whether to make a deep clone.