mmcif_pd_cond: Computes Marginal Measures Using Two Observations

In mmcif: Mixed Multivariate Cumulative Incidence Functions

Computes Marginal Measures Using Two Observations

Description

Computes bivariate figures such as conditional CIFs, survival probabilities, and hazards or bivariate CIFs, densities, and survival probabilities.

Usage

mmcif_pd_cond(
  par,
  object,
  newdata,
  cause,
  time,
  left_trunc = NULL,
  ghq_data = object$ghq_data,
  strata = NULL,
  which_cond,
  type_cond = "derivative",
  type_obs = "cumulative"
)

mmcif_pd_bivariate(
  par,
  object,
  newdata,
  cause,
  time,
  left_trunc = NULL,
  ghq_data = object$ghq_data,
  strata = NULL,
  use_log = FALSE,
  type = c("cumulative", "cumulative")
)

Arguments

par	numeric vector with the model parameters.
object	an object from mmcif_data.
newdata	a data.frame with data for the observations. It needs to have two rows.
cause	an integer vector with the cause of each outcome. If there are n_causes of outcome, then the vector should have values in 1:(n_causes + 1) with n_causes + 1 indicating censoring.
time	a numeric vector with the observed times.
left_trunc	numeric vector with left-truncation times. NULL implies that there are not any individuals with left-truncation.
ghq_data	the Gauss-Hermite quadrature nodes and weights to use. It should be a list with two elements called "node" and "weight". A default is provided if NULL is passed.
strata	an integer vector or a factor vector with the strata of each individual. NULL implies that there are no strata.
which_cond	an integer with value one or two for the index of the individual that is being conditioned on.
type_cond	a character for the type of outcome that is being conditioned on. "derivative" for the derivative of a CIF or "cumulative" for a CIF or the survival probability.
type_obs	a character the type of conditional measure. It can be "derivative" for the derivative of a CIF, "cumulative" for a CIF or the survival probability, and "hazard" for the hazard.
use_log	a logical for whether the returned output should be on the log scale.
type	a 2D character vector for the type of measures for each observation. The elements can be "derivative" for the derivative of a CIF or "cumulative" for a CIF or the survival probability.

Value

A numeric scalar with the requested quantity.

See Also

mmcif_pd_univariate and mmcif_fit.

Examples

if(require(mets)){
  data(prt)

  # truncate the time
  max_time <- 90
  prt <- within(prt, {
    status[time >= max_time] <- 0
    time <- pmin(time, max_time)
  })

  # select the DZ twins and re-code the status
  prt_use <- subset(prt, zyg == "DZ") |>
    transform(status = ifelse(status == 0, 3L, status))

  # Gauss Hermite quadrature nodes and weights from fastGHQuad::gaussHermiteData
  ghq_data <- list(
    node = c(-3.43615911883774, -2.53273167423279, -1.75668364929988, -1.03661082978951,
             -0.342901327223705, 0.342901327223705, 1.03661082978951, 1.75668364929988,
             2.53273167423279, 3.43615911883774),
    weight = c(7.6404328552326e-06, 0.00134364574678124, 0.0338743944554811, 0.240138611082314,
               0.610862633735326,0.610862633735326, 0.240138611082315, 0.033874394455481,
               0.00134364574678124, 7.64043285523265e-06))

  # setup the object for the computation
  mmcif_obj <- mmcif_data(
    ~ country - 1, prt_use, status, time, id, max_time,
    2L, strata = country, ghq_data = ghq_data)

  # previous estimates
  par <- c(0.727279974859164, 0.640534073288067, 0.429437766165371, 0.434367104339573,
           -2.4737847536253, -1.49576564624673, -1.89966050143904, -1.58881346649412,
           -5.5431198001029, -3.5328359024178, -5.82305147022587, -3.4531896212114,
           -5.29132887832377, -3.36106297109548, -6.03690322125729, -3.49516746825624,
           2.55000711185704, 2.71995985605891, 2.61971498736444, 3.05976391058032,
           -5.97173564860957, -3.37912051983482, -5.14324860374941, -3.36396780694965,
           -6.02337246348561, -3.03754644968859, -5.51267338700737, -3.01148582224673,
           2.69665543753264, 2.59359057553995, 2.7938341786374, 2.70689750644755,
           -0.362056555418564, 0.24088005091276, 0.124070380635372, -0.246152029808377,
           -0.0445628476462479, -0.911485513197845, -0.27911988106887, -0.359648419277058,
           -0.242711959678559, -6.84897302527358)

  # the test data we will use
  test_dat <- data.frame(
    country = factor(c("Norway", "Norway"), levels(prt_use$country)),
    status = c(1L, 2L), time = c(60, 75))

  # probability that both experience the event prior to the two times
  mmcif_pd_bivariate(
    par = par, object = mmcif_obj, newdata = test_dat, cause = status,
    strata = country, ghq_data = ghq_data, time = time, type =
      c("cumulative", "cumulative")) |>
    print()

  # density that one experiences an event at the point and the other
  # experiences an event prior to the point
  mmcif_pd_bivariate(
    par = par, object = mmcif_obj, newdata = test_dat, cause = status,
    strata = country, ghq_data = ghq_data, time = time, type =
      c("derivative", "cumulative")) |>
    print()

  # probability that both survive up to the passed points
  mmcif_pd_bivariate(
    par = par, object = mmcif_obj, newdata = test_dat, cause = c(3L, 3L),
    strata = country, ghq_data = ghq_data, time = time, type =
      c("cumulative", "cumulative")) |>
    print()

  # conditional hazard given that the other experiences an event prior to time
  mmcif_pd_cond(
    par = par, object = mmcif_obj, newdata = test_dat, cause = status,
    strata = country, ghq_data = ghq_data, time = time, which_cond = 1L,
    type_cond = "cumulative", type_obs = "hazard") |>
    print()

  # conditional CIF given that the other experiences an event prior to the
  # time
  mmcif_pd_cond(
    par = par, object = mmcif_obj, newdata = test_dat, cause = c(2L, 2L),
    strata = country, ghq_data = ghq_data, time = time, which_cond = 1L,
    type_cond = "cumulative", type_obs = "cumulative") |>
    print()

  # same but given that the other experiences the event at the point
  mmcif_pd_cond(
    par = par, object = mmcif_obj, newdata = test_dat, cause = c(2L, 2L),
    strata = country, ghq_data = ghq_data, time = time, which_cond = 1L,
    type_cond = "derivative", type_obs = "cumulative") |>
    print()
}

mmcif documentation built on July 18, 2022, 1:06 a.m.