Description Usage Arguments Details Value Author(s) References See Also Examples

`curereg`

fits parametric regression models with cure fraction for survival
data. This function extends the `flexsurvreg`

by the inclusion of the cure fraction
in the formulation and adds the Marshall-Olkin extreme value distribution in the comprehensive
roll of parametric distributions avaliable.

1 2 |

`formula` |
an object of class " |

`cureformula` |
a formula defining the cure rate model. In the case of no covariate effect
on the cure rate, set |

`data` |
the data set of class |

`weights` |
optional prior weights for the data. |

`timedist` |
survival distribution for the non-cured individuals. This can be set as:
The exponential, Weibull, log-normal and log-logistic distributions have the same
parameterization defined in |

`ncausedist` |
distribution of the number of competing causes of the event. This can be
set as |

`subset` |
optional numeric vector specifying the subset observations from the full data set. |

`na.action` |
a function indicating what should happen when |

`inits` |
optional list with the initial values for the parameters. This list should be
set as |

`method` |
The optimisation method to be used. |

`...` |
optional arguments for the For situations where the default The argument |

Note that the arguments `ncausedist`

and `timedist`

set up the model to
be fitted. This means that if `ncausedist = "poisson"`

and `timedist = "genf"`

the fitted model is obtained considering the promotion-time model with generalised F
responses. The improper density function of this model is available in `dgenfpt`

.
If `ncausedist = "bernoulli"`

and `timedist = "genf"`

has been set then the fitted
model is calculated considering the standard mixture model with generalised F responses.
The improper density function of this model is available in `dgenfms`

.
`d____ms`

and `p____ms`

correspond to the improper density and probability
functions for standard mixture models, respectively, where `____`

can be any of the
distributions in `timedist`

. Similarly, `d____pt`

and `p____pt`

correspond
to the improper density and probability functions for promotion time models.

The relationship between the linear predictor in `formula`

and the time-to-event of
the non-cured elements is logarithmic as in the accelerated failure time models
(Lawless, 2003). However, for `cureformula`

, if `ncausedist = "bernoulli"`

the relationship is similar to a logistic regression model
(`family = binomial(link = " logit ")`

in `glm`

) and if
`ncausedist = "poisson"`

is similar to a Poisson model with logarithmic link
function (`family = poisson(link = "log")`

in `glm`

). See the references for
more details.

A list of class "curereg" containing information about the fitted model. Components of interest to users may include:

`call ` |
the matched call. |

`coefficients ` |
a named vector of coefficients obtained via Maximum Likelihood (see Details). |

`std.error ` |
a named vector of the estimated standard errors for the coefficients (see Details). |

`vcov ` |
A matrix of the estimated covariances between the coefficient estimates in the predictor of the model. |

`loglik ` |
log-likelihood. |

`AIC ` |
AIC the (generalized) Akaike Information Criterion for the fitted model. |

Rumenick Pereira da Silva [email protected]

Jackson, C.H. Christopher Jackson (2015). flexsurv: Flexible Parametric Survival and Multi-State Models. R package version 0.7. https://CRAN.R-project.org/package=flexsurv.

Maller, R. A., & Zhou, X. (1996). Survival analysis with long-term survivors. New York: Wiley.

Lawless, J. F. (2011). Statistical models and methods for lifetime data (Vol. 362). John Wiley & Sons.

Ortega, E. M., Cancho, V. G., & Paula, G. A. (2009). Generalized log-gamma regression models with cure fraction. Lifetime Data Analysis, 15(1), 79-106.

Peng, Y., Dear, K. B., & Denham, J. W. (1998). A generalized F mixture model for cure rate estimation. Statistics in medicine, 17(8), 813-830.

`confint.curereg`

for confidence intervals for the coefficients,
`curefraction`

for predict cure fractions from fitted `curereg`

model,
`plot.curereg`

and `lines.curereg`

to plot fitted survival, hazards
and cumulative hazards from models fitted by `curereg`

.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 | ```
## fit Marshall-Olkin extended extreme value standart mixture model
data(e1684)
fitmo <- curereg(Surv(FAILTIME, FAILCENS) ~ TRT + SEX + AGE, cureformula = ~ TRT + SEX + AGE,
data = e1684, timedist = "moeev", ncausedist = "bernoulli")
# Output of 'curereg' object
fitmo
# Extract Model Coefficients
coef(fitmo)
# Extract model coefficients:
# Terms: failure time distribution model
coef(fitmo, terms = "time")
# Terms: cure probability model
coef(fitmo, terms = "cure")
# Object summaries
summary(fitmo)
# Information criterion
AIC(fitmo) # Akaike information criterion
BIC(fitmo) # Bayesian information criterion
# Extract Log-Likelihood
logLik(fitmo)
# Calculate Variance-Covariance Matrix for a Fitted Model Object
vcov(fitmo)
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.