| Title: | Optimal Sample Size and Progression Criteria for Three-Outcome Trials |
| Version: | 1.0.3 |
| Description: | Find the optimal decision rules (AKA progression criteria) and sample size for clinical trials with three (stop/pause/go) outcomes. Both binary and continuous endpoints can be accommodated, as can cases where an adjustment is planned following a pause outcome. For more details see Wilson et al. (2024) <doi:10.1186/s12874-024-02351-x>. |
| License: | MIT + file LICENSE |
| Suggests: | knitr, rmarkdown, testthat (≥ 3.0.0), vdiffr |
| VignetteBuilder: | knitr |
| Config/testthat/edition: | 3 |
| Encoding: | UTF-8 |
| RoxygenNote: | 7.3.3 |
| URL: | https://github.com/DTWilson/tout, https://dtwilson.github.io/tout/ |
| BugReports: | https://github.com/DTWilson/tout/issues |
| NeedsCompilation: | no |
| Packaged: | 2025-11-05 13:16:03 UTC; meddwilb |
| Author: | Duncan Wilson 
[aut, cre, cph] |
| Maintainer: | Duncan Wilson <d.t.wilson@leeds.ac.uk> |
| Repository: | CRAN |
| Date/Publication: | 2025-11-07 14:00:02 UTC |
Plot sampling distributions of three-outcome designs
Description
Takes an object of class tout and plots sampling distributions under the
null and alternative hypotheses, highlighting which portions correspond
to stop, pause, and go outcomes.
Usage
## S3 method for class 'tout'
plot(x, ...)
Arguments
x |
object of class |
... |
further arguments passed to or from other methods. |
Value
no return value, called for side effects.
Print a tout object
Description
The default print method for a tout object.
Usage
## S3 method for class 'tout'
print(x, ...)
Arguments
x |
object of class |
... |
further arguments passed to or from other methods. |
Value
no return value, called for side effects.
Find optimal sample size and progression criteria
Description
Given a null and alternative hypothesis, this function finds the
lowest sample size such that a design with optimal progression criteria (as
determined by the function opt_pc) satisfies upper constraints on three
operating characteristics.
Usage
tout_design(
rho_0,
rho_1,
alpha_nom,
beta_nom,
gamma_nom = 1,
eta_0 = 0.5,
eta_1 = eta_0,
tau = c(0, 0),
max_n = NULL,
n = NULL,
x = NULL,
sigma = NULL
)
Arguments
rho_0 |
null hypothesis. |
rho_1 |
alternative hypothesis. |
alpha_nom |
nominal upper constraint on alpha. |
beta_nom |
nominal upper constraint on beta. |
gamma_nom |
nominal upper constraint on gamma. Defaults to 1. |
eta_0 |
probability of an incorrect decision under the null hypothesis after an intermediate result. Defaults to 0.5. |
eta_1 |
probability of an incorrect decision under the alternative hypothesis after an intermediate result. Defaults to eta_0. |
tau |
two element vector denoting lower and upper limits of the effect of adjustment. |
max_n |
optional upper limit to use in search over sample sizes. |
n |
optional sample size (optimised if left unspecified). |
x |
optional vector of decision thresholds (optimised if left unspecified). |
sigma |
standard deviation of outcome. If left unspecified, a binary outcome is assumed. |
Value
An object of class tout, which is a list containing the following components:
valid |
boolean indicating if the nominal constraints are met. |
n |
sample size. |
thesholds |
numeric vector of the two decision thresholds. |
alpha |
attained value of operating characteristic alpha. |
beta |
attained value of operating characteristic beta. |
gamma |
attained value of operating characteristic gamma. |
Examples
rho_0 <- 0.5
rho_1 <- 0.7
alpha_nom <- 0.05
beta_nom <- 0.2
tout_design(rho_0, rho_1, alpha_nom, beta_nom)
# Allowing for adjustment effects:
tout_design(rho_0, rho_1, alpha_nom, beta_nom, tau = c(0.08, 0.12))
# Allowing for different error probabilities following a pause decision
tout_design(rho_0, rho_1, alpha_nom, beta_nom, eta_0 = 0.3)
# Designs for continuous outcomes:
tout_design(rho_0 = 0, rho_1 = 0.4, alpha_nom, beta_nom, sigma = 1)