One-stage designs

OneStageDesign implements a one-stage design as special case of a two-stage design, i.e. as sub-class of TwoStageDesign. This is possible by defining n₂ = 0, c = c₁^f = c₁^e, c₂(x₁) = ifelse(x₁< c, Inf, -Inf). No integration pivots etc are required (set to NaN).

OneStageDesign(n, c)

# S4 method for OneStageDesign
TwoStageDesign(n1, order = 5L, eps = 0.01, ...)

# S4 method for OneStageDesign
plot(x, y, ...)

Arguments

n

sample size (stage-one sample size)

c

rejection boundary (c = c₁^f = c₁^e)

n1

OneStageDesign object to convert, overloaded from TwoStageDesign

order

integer >= 2, default is 5; order of Gaussian quadrature integration rule to use for new TwoStageDesign.

eps

numeric > 0, default = .01; the single critical value c must be split in a continuation interval [c1f, c1e]; this is given by c +/- eps.

...

further optional arguments

x

design to plot

y

not used

Details

Note that the default plot,TwoStageDesign-method method is not supported for OneStageDesign objects.

See also

TwoStageDesign, GroupSequentialDesign

Examples

design <- OneStageDesign(30, 1.96)
summary(design)
#> OneStageDesign: n1 =  30 
#> 
          futility | continue | efficacy
#> 
      x1:     1.96 |   NaN |  1.96
#> 
  c2(x1):     +Inf |    NA |  -Inf
#> 
  n2(x1):        0 |     0 |     0
#> 

design <- TwoStageDesign(design)
summary(design)
#> TwoStageDesign: n1 =  30 
#> 
          futility |            continue           | efficacy
#> 
      x1:     1.95 |  1.95  1.95  1.96  1.97  1.97 |  1.97
#> 
  c2(x1):     +Inf | +3.00 +3.00 +0.00 -3.00 -3.00 |  -Inf
#> 
  n2(x1):        0 |     0     0     0     0     0 |     0
#>