bootstrap

[MikeJSeo]

I realize there are two ways of doing bootstrap now..
What we initially had was using boot function and defining statistic (bootstrap_HR)

HR_bootstraps <- boot(
data = ipd_centered,
statistic = bootstrap_HR,
centered_colnames = centered_colnames,
pseudo_ipd = pseudo_ipd,
model = Surv(TIME, EVENT == 1) ~ ARM,
ref_treat = "B",
R = 1000
)

Other way is storing all weights from bootstrapped data and calculating this manually without using boot function, which I think Greg is implementing in the wrapper.

cluster_boot_out <- boot(pair_ids, cluster_boot_fun, R = 999)
cluster_boot_out
boot.ci(cluster_boot_out, type = "perc")
boot.ci(cluster_boot_out, type = "bca")