Pool estimates using the Rubin Rules

Usage

pooling_rubin(estimate, std.error, df = NA_real_, term = NULL)

Arguments

estimate: List. Each element of the list is a vector of pseudo estimates of length \(m\).
std.error: List. The standard error of the estimates.
df: Numeric. Indicates the degrees of freedom of the test.
term: Character. Indicates the names of the estimates.

Value

A data frame with the pooled results.

Details

This function is adapted from the mice::pool() function.

References

Rubin, D.B. (1987). Multiple Imputation for Nonresponse in Surveys. New York: John Wiley and Sons.

Van Buuren, S., Groothuis-Oudshoorn, K., & Robitzsch, A. (2019). Package ‘mice’: multivariate imputation by chained equations. CRAN Repos.

Examples

if (FALSE) {
library(survival)
data(tvtdata2)

list_data_surv <-
replicate(n = 3,
    {tvt_matching_date(tvtdata2,date_statrisk = stos,treatna,eos,eos01)},
    simplify = F)

list_data_surv <- lapply(list_data_surv, \(di){
    di%>%
      bind_cols(tvtdata2%>%
      select(eos01,fct_2lvl,fct_4lvl,xcov)%>%
      slice(tvt_new$id_row))})

list_coxph <- lapply(list_data_surv, \(di){
    coxph(Surv(time_eos_tv,state_eos_tv)~treat_tv+xcov,data= di)})

estim <- lapply(list_coxph,coef)
var <- lapply(list_coxph,\(x){sqrt(diag(x$var))})
df <- list_coxph[[1]]$n-length(lestim[[1]])
term <- names(lestim[[1]])
pooling_rubin(estim,var,df,term)
}