p_flexsurvmix: Transition probabilities from a flexsurvmix model

Description Usage Arguments Details Value

View source: R/summary.flexsurvmix.R


These quantities are variously known as transition probabilities, or state occupancy probabilities, or values of the "cumulative incidence" function, or values of the "subdistribution" function. They are the probabilities that an individual has experienced an event of a particular kind by time t.


p_flexsurvmix(x, newdata = NULL, startname = "start", t = 1, B = NULL)



Fitted model object returned from flexsurvmix.


Data frame or list of covariate values. If omitted for a model with covariates, a default is used, defined by all combinations of factors if the only covariates in the model are factors, or all covariate values of zero if there are any non-factor covariates in the model.


Name of the state where individuals start. This considers the model as a multi-state model where people start in this state, and may transition to one of the competing events.


Vector of times t to calculate the probabilities of transition by.


Number of simulations to use to compute 95% confidence intervals, based on the asymptotic multivariate normal distribution of the basic parameter estimates. If B=NULL then intervals are not computed.


Note that "cumulative incidence" is a misnomer, as "incidence" typically means a hazard, and the quantities computed here are not cumulative hazards, but probabilities.


A data frame with transition probabilities by time, covariate value and destination state.

flexsurv documentation built on Sept. 13, 2021, 5:07 p.m.