add_cov: Adding covariate to multistate data

In niklasmaltzahn/multistate: simulation and estimation of multistate models

Description

The function merges a multistate dataset and a dataset of a time dependent covariate.

Usage

add_cov(data, cov_data)

Arguments

data	Multistate date of the type from sim_fun.
cov_data	covariate data with column one with columns (id, time, value).

Examples

jump_types <- matrix(c(NA, "exp", "exp", NA, "exp",
                       "exp", NA ,NA,"exp", "exp",
                       "exp", NA, NA, "exp", "exp",
                       NA, "exp", "exp", NA, "exp",
                       rep(NA, times = 5)), 5, 5, byrow = T)
tmat <- tmat_function(jump_types)

lambda <- matrix(c(NA, 0.1, 0.15, NA, 0.08,
                   0.12, NA ,NA,0.15, 0.05,
                   0.1, NA, NA, 0.1, 0.08,
                   NA, 0.12, 0.1, NA, 0.05,
                   rep(NA, times = 5)), 5, 5, byrow = T)

par_list2 <- list(lambda = lambda)

data_sim <- sim_fun(init_probs = c(0.5,0,0.5,rep(0, times = 2)), jump_types, pars = par_list2,
                   cens_time = 1000, nbr = 1000)


new_states <- list(state1 = 1:4, state2 = 5)
    new_tmat <- matrix(NA, 2, 2)
    new_tmat[1,2] <- 1

data_margial <- merge_states(data_sim, new_states, new_tmat)

A_data <- cbind(data_sim[,c("id", "Tstop")], as.numeric(data_sim$from %in% c(2, 4)))
names(A_data) <- c("id", "Tstop", "A")

data_margial <- add_cov(data_margial, cov_data = A_data)
data_margial$L <- as.numeric(data_margial$from %in% c(3, 4))

niklasmaltzahn/multistate documentation built on Nov. 4, 2019, 10:06 p.m.