Skip to contents

Simulate one paired replicate for post hoc design evaluation

single_paired_rep.Rd

Simulates a single trial and records everything needed to evaluate a family of designs afterwards, without baking any design into the run: interim statistics and the predictive probability (PP) at the futility look, and the test statistics at the efficacy look and the final analysis of the real trial (which is never stopped here).

Usage

single_paired_rep(
  i,
  seed,
  n_c,
  n_t,
  total_events,
  futility_IF,
  efficacy_IF = 0.75,
  design,
  control_model,
  effect_model,
  recruitment_model,
  truth = NULL,
  force_state = NULL,
  analysis_model_LRT,
  mw_t_stars = c(2, 6),
  update_priors_sims,
  PP_sims,
  compute_PP = TRUE
)

Arguments

i

Replicate index.

seed

Base seed. Replicate i uses set.seed(seed * 1e5 + i), which must be below .Machine$integer.max; the JAGS seed is drawn from that stream, so the replicate is fully reproducible.

n_c, n_t

Planned number of patients in the control / treatment arm.

total_events

Maximum planned number of events.

futility_IF

Information fraction of the futility (PP) look.

efficacy_IF

Information fraction of the interim efficacy look (default 0.75).

design

An rpact group-sequential design containing the efficacy look at efficacy_IF and the final analysis (and optionally the futility look, with zero alpha spent).

control_model, effect_model

The analysis prior, used for the interim posterior update (see update_priors). control_model$parameter_mode must be "Distribution".

recruitment_model

Recruitment specification. Must be method = "power" with power = 1 (see PP_func).

truth

Either a list with the true lambda_c, delay_time, post_delay_HR and optionally gamma_c (exponential control arm if NULL) and state, or NULL, in which case the truth is drawn from control_model and effect_model via simulate_trial_with_recruitment(), optionally with the latent state fixed by force_state.

force_state

NULL, or 1 (no separation), 2 (immediate separation) or 3 (delayed separation). Only used when truth = NULL.

analysis_model_LRT

Analysis specification for the log-rank test (method = "LRT", alpha, alternative_hypothesis), used for the PP calculation and the *_LRT statistics.

mw_t_stars

Values of \(t^*\) for the modestly weighted log-rank statistics (default c(2, 6)).

update_priors_sims

Number of posterior samples per chain (required when compute_PP = TRUE).

PP_sims

Number of posterior-predictive draws (required when compute_PP = TRUE).

compute_PP

If FALSE, skip the interim posterior update and the PP calculation: PP_val, converged, rhat_max, n_nonfinite_rhat, P_Z1, P_Z2 and P_Z3 are NA, and all other columns are identical to those from compute_PP = TRUE with the same seed. Such output supports rules D1, D2, D4 and D5 but not D3. Default TRUE.

Value

A one-row data frame with columns: rep_id, seed_used, state, true_lambda_c, true_gamma_c (1 for an exponential control arm), true_delay, true_HR; interim t_int, n_int, frac_recruited_int (enrolled / planned), fu_gt3_int, fu_gt6_int (fraction of enrolled treated patients with follow-up t_int - rec_time above 3 and 6), Z_int_LRT, Z_int_MW_t<t>; PP_val, converged, rhat_max, n_nonfinite_rhat, P_Z1, P_Z2, P_Z3; t_eff, n_eff, Z_eff_LRT, Z_eff_MW_t<t>, t_fin, n_fin, Z_fin_LRT, Z_fin_MW_t<t>; the D2 outcome continuation_success, continuation_stop_time, continuation_sample_size; aliases t_interim and sample_size_interim (for summarize_grid_by_kappa); and error (NA, or the error message if the replicate failed, in which case all other outcome columns are NA).

Details

All designs are then summaries of the returned columns. With critical values \(c_{eff}\), \(c_{fin}\) from design and a final-only critical value \(c_1\):

  • D1 (fixed design): success if Z_fin_LRT > c_1.

  • D2 (efficacy look only): success if Z_eff_LRT > c_eff, otherwise Z_fin_LRT > c_fin. This is continuation_success.

  • D3 (D2 plus PP futility): stop for futility if PP_val < kappa; see summarize_grid_by_kappa.

  • D4, D5 (D2 plus Z futility): stop for futility if Z_int_LRT (or Z_int_MW_t*) is below a cutoff.

See also

run_paired_scenario