easysurv (version 2.0.1)

fit_models: Fit Survival Models

Description

Fits survival models to the provided data using the specified engine and returns various outputs including model parameters, goodness of fit, and estimates of median survival.

Usage

fit_models(
  data,
  time,
  event,
  predict_by = NULL,
  covariates = NULL,
  dists = c("exp", "gamma", "gengamma", "gompertz", "llogis", "lnorm", "weibull"),
  engine = "flexsurv",
  k = c(1, 2, 3),
  scale = "hazard",
  add_time_0 = TRUE,
  ...
)

Value

A list containing information about the fit_models() call, the distributions attempted, goodness of fit, fit averages, and cure fractions (if applicable).

Arguments

data

A data frame containing the survival data.

time

The name of the column in data containing the time-to-event information.

event

The name of the column in data indicating whether the event of interest occurred.

predict_by

(Optional) The name of the column in data defining the prediction variable.

covariates

(Optional) A character vector specifying the names of covariates to be included in the model.

dists

(Optional) A character vector specifying the distribution(s) to be fitted.

When the engine parameter is set to "flexsurv", options are "exp", "exponential", "gamma", "genf", "genf.orig", "gengamma", "gengamma.orig", "gompertz", "llogis", "lnorm", "lognormal", "weibull", "weibullPH".

When the engine parameter is set to "flexsurvcure", options are "exp", "gamma", "gengamma", "gompertz", "llogis", "lnorm", "weibull".

When the engine parameter is set to "flexsurvspline", dists are ignored in favor of k and scale parameters.

When the engine parameter is set to "survival", options are "exponential", "extreme", "gaussian", "loggaussian" (same as lognormal), "logistic", "lognormal", "rayleigh", "weibull".

Default is c("exp", "gamma", "gengamma", "gompertz", "llogis", "lnorm", "weibull") which applies to flexsurv-related engines.

engine

(Optional) The survival analysis engine to be used. Options are "flexsurv", "flexsurvcure", "flexsurvspline", and "survival". Default is "flexsurv".

• "flexsurv" uses flexsurv::flexsurvreg().

• "flexsurvspline" uses flexsurv::flexsurvspline().

• "flexsurvcure" uses flexsurvcure::flexsurvcure().

• "survival" uses survival::survreg().

k

(Optional) A numeric vector specifying the number of knots for spline-based models. Default is c(1, 2, 3) to test different numbers.

scale

(Optional) A character vector specifying the scale parameter(s) for spline-based models. Options are "hazard", "odds", and "normal". Default is "hazard".

add_time_0

Optional. Uses survival::survfit0() to add a starting time of 0 to the KM survfit object. This may be useful for plotting the KM at a subsequent stage (in surv_plots). Default is TRUE.

...

Additional arguments just to catch them and avoid errors.

Examples

models <- fit_models(
  data = easysurv::easy_bc,
  time = "recyrs",
  event = "censrec",
  predict_by = "group",
  covariates = "group"
)

models