| Type: | Package |
| Title: | Structural Additive Cumulative Intensity Models with IV |
| Version: | 2.1.0 |
| Date: | 2022-01-01 |
| Author: | Andrew Ying |
| Maintainer: | Andrew Ying <aying9339@gmail.com> |
| Description: | An instrumental variable estimator under structural cumulative additive intensity model is fitted, that leverages initial randomization as the IV. The estimator can be used to fit an additive hazards model under time to event data which handles treatment switching (treatment crossover) correctly. We also provide a consistent variance estimate. |
| License: | GPL-2 | GPL-3 [expanded from: GPL (≥ 2)] |
| Imports: | Rcpp |
| LinkingTo: | Rcpp, RcppArmadillo |
| Depends: | R (≥ 4.0) |
| RoxygenNote: | 7.1.2 |
| Encoding: | UTF-8 |
| NeedsCompilation: | yes |
| Packaged: | 2022-01-30 01:08:41 UTC; andrewying |
| Repository: | CRAN |
| Date/Publication: | 2022-01-30 03:10:02 UTC |
This function is to compute the centralized instrument
Description
This function is to compute the centralized instrument
Usage
IV_center(Z, weights)
This is the main function to compute the estimator in C++
Description
This is the main function to compute the estimator in C++
Usage
invalidivsacim_est(
time,
event,
stime,
Z,
Z_c,
D_status,
D_status_c,
Z_model_mat,
eps_1,
D_model_mat,
eps_2,
weights
)
Arguments
time |
censored event time |
event |
event indicator |
stime |
unique sorted noncensored event time |
Z |
the instrument |
Z_c |
centered IV |
D_status |
treatment process at each noncensored event time |
D_status_c |
centered treatment process at each noncensored event time |
Z_model_mat |
bread matrix from modeling IV |
eps_1 |
influence function from modeling IV |
D_model_mat |
bread matrix from modeling treatment |
eps_2 |
influence function from modeling treatment |
weights |
optional weights |
Fitting a Cumulative Intensity Model for Exposure Effects with Instrumental Variables
Description
ivsacim is used to fit cumulative intensity models for exposure effects with instrumental variables.
Usage
ivsacim(
time,
event,
instrument,
IV_valid = TRUE,
treatment_init,
treatment_shift_time = NULL,
max_time = NULL,
max_time_bet = NULL,
n_sim = 0,
weights = NULL
)
Arguments
time |
the censored event time |
event |
event indicator |
instrument |
the instrumental variable |
IV_valid |
whether assuming IV satisfies the exclusion restriction |
treatment_init |
the initial treatment assignment |
treatment_shift_time |
the shift time of each subject, if no shift for a subject, set as 0 |
max_time |
the max time that we threshold for nonconstant effect |
max_time_bet |
the max time that we threshold for constant effect |
n_sim |
the number of resampling, set as 0 if no resampling is needed |
weights |
optional weights used in the estimating equation |
Value
ivsacim returns an object of class "ivsacim". An object of class "ivsacim" is a list containing the following components:
stime |
an estimate of the baseline hazards function |
dB_D |
an estimate of the increment of the treatment effect |
B_D |
an estimate of the treatment effect |
beta_D |
an estimate of the constant treatment effect |
B_D_se |
an estimate of the variance covariance matrix of B_D |
beta_D_se |
an estimate of the constant treatment effect |
by_prod |
a byproduct, that will used by other functions |
Examples
n = 400
event = rbinom(n, 1, 0.8)
IV = rbinom(n, 1, 0.5)
trt_init = IV
trt_shift = rep(0, n)
time = rexp(n)/(0.5 + trt_init * 0.2)
max_t = 3
max_t_bet = 3
n_sim = 0
fit <- ivsacim(time, event, IV, TRUE, trt_init,
trt_shift, max_t, max_t_bet, n_sim)
This is the main function to compute the estimator in C++
Description
This is the main function to compute the estimator in C++
Usage
ivsacim_est(time, event, stime, Zc, D_status, eps_1, Zc_dot, weights)
Arguments
time |
censored event time |
event |
event indicator |
stime |
unique sorted noncensored event time |
Zc |
centered IV |
D_status |
treatment process at each noncensored event time |
eps_1 |
influence function from modeling IV |
Zc_dot |
bread matrix from modeling IV |
weights |
optional weights |
Plotting Estimated Cumulative Intensity function with Pointwise Confidence Intervals
Description
The function will plot the estimated cumulative intensity function of the treatment after fitting. Corresponding pointwise confidence intervals at level alpha are also included.
Usage
## S3 method for class 'ivsacim'
plot(x, gof = FALSE, ...)
Arguments
x |
the fitting object after fitting IVSACIM model |
gof |
whether to draw the goodness-of-fit plot |
... |
the other arguments you want to put in the built-in plot function |
Value
No return value, called for side effects
Examples
n = 400
event = rbinom(n, 1, 0.8)
IV = rbinom(n, 1, 0.5)
trt_init = IV
trt_shift = rep(0, n)
time = rexp(n)/(0.5 + trt_init * 0.2)
max_t = 3
max_t_bet = 3
n_sim = 100
fit <- ivsacim(time, event, IV, IV_valid = TRUE, trt_init, trt_shift, max_t, max_t_bet, n_sim)
plot(fit, main = "", xlab = "Time", ylab = "Cumulative Intensity Function")
plot(fit, gof = TRUE, xlab = "Time", ylab = "")
Summarizing Cumulative Intensity Function of Treatment with Instrumental Variables Estimation Using Structural Additive Cumulative Intensity Models
Description
summary method for class "ivsacim".
Usage
## S3 method for class 'ivsacim'
summary(object, ...)
## S3 method for class 'summary.ivsacim'
print(x, ...)
Arguments
object |
an object of class "ivsacim", usually, a result of a call to ivsacim. |
... |
further arguments passed to or from other methods. |
x |
an object of class "summary.ivsacim", usually, a result of a call to summary.ivsacim. |
Details
print.summary.ivsacim tries to be smart about formatting coefficients, an estimated variance covariance matrix of the coeffieients, Z-values and the corresponding P-values.
Value
The function summary.ivsacim computes and returns a list of summary statistics of the fitted model given in object.
Examples
n = 400
event = rbinom(n, 1, 0.8)
IV = rbinom(n, 1, 0.5)
trt_init = IV
trt_shift = rep(0, n)
time = rexp(n)/(0.5 + trt_init * 0.2)
max_t = 3
max_t_bet = 3
n_sim = 0
fit <- ivsacim(time, event, IV, IV_valid = TRUE, trt_init, trt_shift, max_t, max_t_bet, n_sim)
summary(fit)
This function is to generate a grid of treatment status
Description
This function is to generate a grid of treatment status
Usage
treatment_status(N, K, stime, treatment_init, treatment_shift_time, max_time)
This function is to compute the centralized treatment process given the covariates Z, L at each sorted unique event time
Description
This function is to compute the centralized treatment process given the covariates Z, L at each sorted unique event time
Usage
trt_center(D_status, Z, weights)