One row per arm, endpoint and modelled nominal time, giving the probability that a generated subject in that arm has an observation there. It is the fraction of the arm's patients who did, so attendance is drawn per visit rather than a real patient's visit set being reused.
Arguments
- x
A dataset from
synpmx_pca(), or its trial summary.
Value
A data frame with arm, endpoint, time, probability and
patients, the last being how many patients across the study hold that
cell at all.
Examples
data <- pmx_simulated_fixture(60)
roles <- pmx_roles(
id = "ID", time = "TIME", nominal_time = "NTIME", dv = "DV", amt = "AMT",
evid = "EVID", cmt = "CMT", dvid = "DVID", mdv = "MDV"
)
head(pca_visits(synpmx_pca(data, roles, seed = 1)))
#> arm endpoint time probability patients
#> 1 all cp 0.25 1 60
#> 2 all cp 1.00 1 60
#> 3 all cp 2.00 1 60
#> 4 all cp 6.00 1 60
#> 5 all cp 12.25 1 60
#> 6 all cp 13.00 1 60