Tables of the MRCT analytic layer, over cells and sets

Tidy data frames from the existing analytic functions, without quadrature of their own: ahr_truths_table() loops over ahr_limit() and ahr_design_table() over ahr_design().

Usage

ahr_truths_table(
  x,
  sets = c("all", "nonregion", "region", "region_Gstar"),
  b = NULL,
  h = NULL,
  analysis_time = NULL,
  max_entry = NULL,
  ...
)

ahr_design_table(x, b, h, n, analysis_time = NULL, max_entry = NULL, ...)

Arguments

x: The cell or cells: an ahr_population() result, a DGM from inject_mrct_structure(), or a named list of either. The list's names label the cells.
sets: The sets, as ahr_limit() takes them: a character vector of set names ("all", "nonregion", "region", "region_Gstar", "Hstar", "Hstar_region", "Hstar_nonregion"), or a list of those and logical vectors of length pop$N. A name on an element labels the set, and a logical vector needs one.
b, h: Hazard ratios outside and inside $H^*$: one number, or a vector named by cell. In ahr_truths_table(), both NULL (the default) use the DGM's own per-patient log hazard ratios.
analysis_time, max_entry: Passed to ahr_population() when x holds DGMs, and not used for an ahr_population() result, which carries its own.
...: Passed to ahr_limit() (for instance horizon, b_R, h_R) or ahr_design() (b_R, h_R).
n: The trial size.

Value

ahr_truths_table(): a plain data frame with one row per cell and set: cell, set, horizon (the analysis time), marghr (the marginal hazard ratio at the horizon), marghr_inf (with no censoring), ahr, cde.

ahr_design_table(): a plain data frame with one row per cell: cell, n, b, h; the five probabilities of ahr_design() under their names there (p_itt_ub_lt1, p_region_itt_gt1, p_region_Gstar_lt1, p_trigger, p_trigger_and_Gstar_lt1); p_Gstar_lt1_given_trigger, their ratio; and the seven rows above, r1_itt_ub_lt1, r2_region_itt_gt1, r3_consistency, r4_non_consistency, r5_oracle_consistency, r6_oracle_non_consistency, r7_consistency_or_oracle.

Details

ahr_design_table()'s r1_ to r7_ columns are the seven design probabilities of a regional-consistency analysis. With $A$ = {ITT ub < 1.0}, $B$ = {region ITT > 1.0} and $C^*$ = {region($G^*$) < 1.0}: (1) $A$, (2) $B$, (3) consistency $A \setminus B$, (4) non-consistency (the trigger) $A \cap B$, (5) oracle consistency $A \cap B \cap C^*$, (6) oracle non-consistency, and (7) consistency or oracle consistency. The function stops unless (3) + (4) = (1), (5) + (6) = (4) and (3) + (5) = (7) at 1e-12.

Examples

set.seed(1)
n <- 400
d <- data.frame(age = rnorm(n, 50, 10), lab = rnorm(n, 300, 100),
                f1 = rbinom(n, 1, 0.4), treat = rbinom(n, 1, 0.5))
d$tte <- rexp(n, 0.05 * exp(-0.5 * d$treat))
d$event <- as.integer(d$tte < 40)
d$tte <- pmin(d$tte, 40)
dgm <- suppressWarnings(inject_mrct_structure(
  d, outcome_var = "tte", event_var = "event", treatment_var = "treat",
  continuous_vars = c("age", "lab"), factor_vars = "f1", x_pred = "lab",
  spline_spec = list(knot = 300, zeta = 800, log_hrs = rep(log(0.55), 3)),
  region = list(prevalence = 0.25, or_pred = 1, x1_vars = "f1", or_x1 = 3,
                loghr = -log(3), band = list(cut = 230, or = 1)),
  x3 = list(vars = "lab", cuts = list(lab = 230), loghr = log(3 / 0.55)),
  n_super = 2000, expand = "bootstrap", seed = 7L,
  continuous_vars_cens = c("age", "lab"), factor_vars_cens = "f1"))
pop <- ahr_population(dgm, analysis_time = 30, max_entry = 6)
ahr_truths_table(pop, sets = c("all", "region", "region_Gstar"), b = 0.55, h = 3)
#>   cell          set horizon    marghr marghr_inf       ahr      cde
#> 1 cell          all      30 0.9027294  0.8078874 0.9079770 1.340451
#> 2 cell       region      30 1.1084587  0.7639976 0.9296125 1.388062
#> 3 cell region_Gstar      30 0.5503904  0.5518979 0.5500000 0.550000
ahr_design_table(pop, b = 0.55, h = 3, n = 800)
#>   cell   n    b h p_itt_ub_lt1 p_region_itt_gt1 p_region_Gstar_lt1  p_trigger
#> 1 cell 800 0.55 3    0.2003976         0.669416          0.9610761 0.08177303
#>   p_trigger_and_Gstar_lt1 p_Gstar_lt1_given_trigger r1_itt_ub_lt1
#> 1              0.07989874                 0.9770793     0.2003976
#>   r2_region_itt_gt1 r3_consistency r4_non_consistency r5_oracle_consistency
#> 1          0.669416      0.1186246         0.08177303            0.07989874
#>   r6_oracle_non_consistency r7_consistency_or_oracle
#> 1               0.001874295                0.1985233