Skip to contents

Fit a semi-parametric model for interval-censored data

Source: R/ic_sp2.R
ic_sp2.Rd

This function fits a semi-parametric model for interval-censored data using either the proportional hazards or proportional odds model. Aliases ic_sp_ph() and ic_sp_po() are provided for convenience.

Usage

ic_sp2(
  formula,
  data,
  weights,
  subset,
  na.action,
  B = c(0, 1),
  control = ic_sp_control(...),
  model = c("ph", "po"),
  profile_ci = NULL,
  ...
)

ic_sp_ph(
  formula,
  data,
  weights,
  subset,
  na.action,
  B = c(0, 1),
  control = ic_sp_control(...),
  model = c("ph", "po"),
  profile_ci = NULL,
  ...
)

ic_sp_po(
  formula,
  data,
  weights,
  subset,
  na.action,
  B = c(0, 1),
  control = ic_sp_control(...),
  model = c("ph", "po"),
  profile_ci = NULL,
  ...
)

Arguments

formula

A model formula with Surv(l, u, type = 'interval2') response and covariates on the right-hand side. May also contain strata() terms.

data

A data frame containing the variables in the formula, including strata terms.

weights

Optional vector of weights for each observation, or the name of a variable in data containing the weights.

subset

Optional expression indicating a subset of the rows of data to be used in the fit.

na.action

Optional function to handle missing data. Default is na.omit.

B

A vector of length 2 giving the lower and upper bounds for the observation times. Default is c(0, 1).

control

A list of control settings, with defaults created by ic_sp_control().

model

Type of model to fit. Choices are "ph" for proportional hazards and "po" for proportional odds. Default is "ph". This is normally determined by the function aliases ic_sp_ph and ic_sp_po.

profile_ci

Confidence level for profile likelihood confidence intervals based on the likelihood ratio. Set to NULL to skip this profile likelihood confidence interval calculations.

...

Additional arguments passed to control.

Value

An object of class "ic_sp2", a list containing the fitted model information, including coefficients, variance-covariance matrix, and other details.

Examples

fit_ph <- ic_sp_ph(Surv(l, u, type = "interval2") ~ grp, data = miceData, profile_ci = 0.95)
summary(fit_ph)
#> Call:
#> ic_sp_ph(formula = Surv(l, u, type = "interval2") ~ grp, data = miceData, 
#>     profile_ci = 0.95)
#> 
#> Coefficients:
#>       Estimate Std.Error z.value p.value
#> grpge   0.6785    0.3537  1.9184   0.055
#> 
#> Log-likelihood: -76.56894 
#> Iterations: 28 

# The CI for the regression parameter(s) based on the likelihood ratio:
fit_ph$profile_ci
#>            lower    upper
#> grpge -0.1590267 1.366948

confint(fit_ph, level = 0.95)
#>              2.5%    97.5%
#> grpge -0.01470717 1.371635