R: Non Proportional Hazard and Non Log-Linear effect

NPHNLL {flexrsurv}

R Documentation

Non Proportional Hazard and Non Log-Linear effect

Description

Generate the design matrix of spline basis for both non log-linear and non proportional effect.

Usage

NPHNLL(x, 
       timevar, 
       model = c("additive", "multiplicative"), 
       Spline = c("b-spline", "tp-spline", "tpi-spline"), 
       Knots = NULL, 
       Degree = 3, 
       Intercept = FALSE, 
       Boundary.knots = range(x), 
       Knots.t = NULL, 
       Degree.t = 3,
       Intercept.t = (model == "multiplicative"), 
       Boundary.knots.t = c(0, max(timevar)), 
       outer.ok = TRUE, 
       Keep.duplicates = TRUE, 
       xdimnames = ":XxXxXXxXxX ", 
       tdimnames = ":TtTtTTtTtT ")

Arguments

`x`	the predictor variable.
`timevar`	the time variable.
`model`	character string specifying the type of model for both non-proportionnal and non linear effects. The model `method=="additive"` assumes effects as explained in Remontet et al.(2007), the model `method=="multiplicative"` assumes effects as explained in Mahboubi et al. (2011).
`Spline`	a character string specifying the type of spline basis. "b-spline" for B-spline basis, "tp-spline" for truncated power basis and "tpi-spline" for monotone (increasing) truncated power basis.
`Knots`	the internal breakpoints that define the spline used to estimate the NLL part of effect. By default there are none.
`Degree`	degree of splines of variable which are considered.
`Intercept`	a logical value indicating whether intercept/first basis of spline should be considered.
`Boundary.knots`	range of variable which is analysed.
`Knots.t`	the internal breakpoints that define the spline used to estimate the NPH part of effect. By default there are none.
`Degree.t`	degree of splines of time variable which are considered.
`Intercept.t`	a logical value indicating whether intercept/first basis of spline should be considered.
`Boundary.knots.t`	range of time period which is analysed.
`Keep.duplicates`	Should duplicate interior knots be kept or removed. Defaults is `FALSE`, which removes duplicate knots with a warning if duplicate interior knots are found.
`outer.ok`	logical indicating how are managed `timevar` or `x` values outside the knots. If `FALSE`, return `NA`, if `TRUE`, return `0` for the corresponding `timevar` or `x` values.
`xdimnames`	string to build dimnames of `x` bases
`tdimnames`	string to build dimnames of `timevar` bases

Details

NPHNLL is based on package orthogonalsplinebasis

References

Mahboubi, A., M. Abrahamowicz, et al. (2011). "Flexible modeling of the effects of continuous prognostic factors in relative survival." Stat Med 30(12): 1351-1365. doi:10.1002/sim.4208