Cannot specify confidence interval width of simulation summary invpc_tte

[atredennick]

The confidence interval calculated over the Kaplan-Meier curves estimated from simulated data is currently fixed at 90% (0.05 and 0.95 quantiles of KM curves from simulated data). From the vpc_tte function code starting on line 211:

sim_km <- all_dat %>%
      dplyr::group_by(strat, bin) %>%
      dplyr::summarise (bin_mid = head(bin_mid,1),
                 bin_min = head(bin_min,1),
                 bin_max = head(bin_max,1),
                 qmin = quantile(surv, 0.05),
                 qmax = quantile(surv, 0.95),
                 qmed = median(surv),
#                         lower_med = median(lower, 0.05),
#                         upper_med = median(upper, 0.05),
                 step = 0)

This can also lead to confusion where a user might think that specifying ci = c(0.025, 0.975) in the arguments would lead to a different CI, but if in fact does not. In fact, the ci argument in this function applies to the Kaplan-Meier estimate from the observed data, which one only observes when show = list(obs_ci = TRUE) is specified.

Would it be possible to pass a new argument like ci_sim as a vector to the summarizing function of the sim_km object such that the user can specify any CI of choice? Thanks for considering!