fixed_cure: Builds survival models for diseases with cured fractions...

Description Usage Arguments Details Value

View source: R/fixedcure.R


Fits a cure model which assumes that if an individual has survived beyond a set time-point then they are considered cured and their mortality reverts to population levels. Please read the detailed description below for how to use this model.


  formula = NULL,
  data = NULL,
  cure_time = 10 * 365.25,
  daily_survival = NULL,
  population_covariates = NULL,
  dist = c("exponential", "weibull", "lognormal")



Formula specifying survival function, as used in prevalence with the surv_formula argument. Must be in days.


A data frame with the corresponding column names provided in form.


Time-limit at which a patient is considered cured. Note that if this is 0 or negative then survival will be based purely off the population rates (anything passed into formula and data will be ignored).


A data frame comprising population survival as a daily probability for as long as possible, ideally 100 years (36525 days). Defaults to using UK population survival from the UKmortality data set. It must contain columns 'age' and 'surv', providing the age (in days) and survival probability at that age respectively. It can also be stratified by other variables that are found in the survival formula for this model, such as sex.


A character vector containing fields to stratify population survival by in addition to age, as descripted in Details below. These must be the names of columns in both data and daily_survival. If not provided then defaults to the fields that are present in both data and daily_survival.


The distribution used by the default parametric survival model.


To model population survival, population mortality tables are required, as specified by the daily_survival argument. If not provided, then the default population mortality is that of the UK population, which goes up to 100 years of age. If a simulated individual has expected lifespan longer than the maximum age in the mortality table then they are estimated to have died at this age limit, which is why it is advantageous to provide as many accurate survival probabilities as possible.

Due to the linking with the registry data and the ability for user-specified mortality tables, there are stricter requirements on the survival models used in cure models than elsewhere. For example, the time-scale of the survival model specified in formula must be in days so that it matches up with the mortality tables. Likewise, age in years must be included as a covariate in the survival model


An object of class fixedcure that can be passed into prevalence.

rprev documentation built on May 4, 2021, 5:08 p.m.