validation/test_code/04_MultiarmTestCode.R

In Sterniii3/drugdevelopR: Utility-Based Optimal Phase II/III Drug Development Planning

#' @editor Lukas D Sauer
#' @editDate 2022-12-23
test_that("04.01", {
  # Multiarm time-to-event endpoints
  
  res <- optimal_multiarm(alpha = 0.025,
                          beta = 0.1,
                          hr1 = 0.75, hr2 = 0.85,
                          ec = 0.6,
                          n2min = 10, n2max = 200, stepn2 = 1,
                          hrgomin = 0.71, hrgomax = 0.9, stephrgo = 0.01,
                          steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                          b1 = 1000, b2 = 2000, b3 = 3000,
                          num_cl = 12,
                          c02 = 100, c03 = 150,
                          c2 = 0.75, c3 = 1,
                          strategy = 1
               )
  
  expect_equal(res$u, 8.56, tolerance = 0.005)
  expect_equal(res$n2, 141)
  expect_equal(res$n3, 391)
  expect_equal(res$n, 532)
  expect_equal(res$HRgo, 0.82)
  expect_equal(res$pgo, 0.72)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-23
test_that("04.02", {
  # Multiarm time-to-event endpoints
  
  res <- optimal_multiarm(alpha = 0.025,
                          beta = 0.1,
                          hr1 = 0.75, hr2 = 0.85,
                          ec = 0.6,
                          n2min = 10, n2max = 200, stepn2 = 1,
                          hrgomin = 0.71, hrgomax = 0.9, stephrgo = 0.01,
                          steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                          b1 = 1000, b2 = 2000, b3 = 3000,
                          num_cl = 12,
                          c02 = 100, c03 = 150,
                          c2 = 0.75, c3 = 1,
                          strategy = 2
  )
  
  expect_equal(res$u, 4.36, tolerance = 0.005)
  expect_equal(res$n2, 79)
  expect_equal(res$n3, 426)
  expect_equal(res$n, 505)
  expect_equal(res$HRgo, 0.78)
  expect_equal(res$pgo, 0.65)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-23
test_that("04.03", {
  # Multiarm time-to-event endpoints
  
  res <- optimal_multiarm(alpha = 0.025,
                          beta = 0.1,
                          hr1 = 0.75, hr2 = 0.85,
                          ec = 0.6,
                          n2min = 10, n2max = 200, stepn2 = 1,
                          hrgomin = 0.71, hrgomax = 0.9, stephrgo = 0.01,
                          steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                          b1 = 1000, b2 = 2000, b3 = 3000,
                          num_cl = 12,
                          c02 = 100, c03 = 150,
                          c2 = 0.75, c3 = 1,
                          strategy = 3,
                          N = 520
  )
  
  # Strategy 1 = best promising
  expect_equal(res[1,]$u, 8.34, tolerance = 0.005)
  expect_equal(res[1,]$n2, 133)
  expect_equal(res[1,]$n3, 383)
  expect_equal(res[1,]$n, 516)
  expect_equal(res[1,]$HRgo, 0.82)
  # Strategy 2 = all promising -- shouldn't change compared to 04.02
  expect_equal(res[2,]$u, 4.36, tolerance = 0.005)
  expect_equal(res[2,]$n2, 79)
  expect_equal(res[2,]$n3, 426)
  expect_equal(res[2,]$n, 505)
  expect_equal(res[2,]$HRgo, 0.78)
  expect_equal(res[2,]$pgo, 0.65)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-23
test_that("04.04", {
  # Multiarm time-to-event endpoints
  start_time_3 = Sys.time()
  
  res <- optimal_multiarm(alpha = 0.025,
                          beta = 0.1,
                          hr1 = 0.75, hr2 = 0.85,
                          ec = 0.6,
                          n2min = 10, n2max = 200, stepn2 = 1,
                          hrgomin = 0.71, hrgomax = 0.9, stephrgo = 0.01,
                          steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                          b1 = 1000, b2 = 2000, b3 = 3000,
                          num_cl = 12,
                          c02 = 100, c03 = 150,
                          c2 = 0.75, c3 = 1,
                          strategy = 2
  )
  end_time_3 = Sys.time()
  time_elapsed_num_cl_3 = end_time_3 - start_time_3
  start_time_1 = Sys.time()
  res <- optimal_multiarm(alpha = 0.025,
                          beta = 0.1,
                          hr1 = 0.75, hr2 = 0.85,
                          ec = 0.6,
                          n2min = 10, n2max = 200, stepn2 = 1,
                          hrgomin = 0.71, hrgomax = 0.9, stephrgo = 0.01,
                          steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                          b1 = 1000, b2 = 2000, b3 = 3000,
                          num_cl = 6,
                          c02 = 100, c03 = 150,
                          c2 = 0.75, c3 = 1,
                          strategy = 2
  )
  
  end_time_1 = Sys.time()
  time_elapsed_num_cl_1 = end_time_1 - start_time_1
  expect_true(time_elapsed_num_cl_1 > time_elapsed_num_cl_3)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-23
test_that("04.05", {
  # Multiarm binary endpoints
  
  res <- optimal_multiarm_binary(alpha = 0.025,
                          beta = 0.1,
                          p0 = 0.5, p11 = 0.3, p12 = 0.4,
                          n2min = 200, n2max = 400, stepn2 = 2,
                          rrgomin = 0.7, rrgomax = 0.9, steprrgo = 0.01,
                          steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                          b1 = 1000, b2 = 2000, b3 = 3000,
                          num_cl = 12,
                          c02 = 100, c03 = 150,
                          c2 = 0.75, c3 = 1,
                          strategy = 1
  )
  
  expect_equal(res$u, 1264.64, tolerance = 0.005)
  expect_equal(res$n2, 386)
  expect_equal(res$n3, 344)
  expect_equal(res$n, 730)
  expect_equal(res$RRgo, 0.86)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-29
test_that("04.06", {
  # Multiarm binary endpoints with strategy 2
  
  res <- optimal_multiarm_binary(alpha = 0.025,
                                 beta = 0.1,
                                 p0 = 0.5, p11 = 0.3, p12 = 0.4,
                                 n2min = 200, n2max = 400, stepn2 = 2,
                                 rrgomin = 0.7, rrgomax = 0.9, steprrgo = 0.01,
                                 steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                                 b1 = 1000, b2 = 2000, b3 = 3000,
                                 num_cl = 12,
                                 c02 = 100, c03 = 150,
                                 c2 = 0.75, c3 = 1,
                                 strategy = 2
  )
  
  expect_equal(res$u, 1281.74, tolerance = 0.005)
  expect_equal(res$n2, 312)
  expect_equal(res$n3, 561)
  expect_equal(res$n, 873)
  expect_equal(res$RRgo, 0.77)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-29
test_that("04.07", {
  # Multiarm binary endpoints with strategy 3
  
  res <- optimal_multiarm_binary(alpha = 0.025,
                                 beta = 0.1,
                                 p0 = 0.5, p11 = 0.3, p12 = 0.4,
                                 n2min = 200, n2max = 400, stepn2 = 2,
                                 rrgomin = 0.7, rrgomax = 0.9, steprrgo = 0.01,
                                 steps1 = 1, stepm1 = 0.95, stepl1 = 0.85,
                                 b1 = 1000, b2 = 2000, b3 = 3000,
                                 num_cl = 12,
                                 c02 = 100, c03 = 150,
                                 c2 = 0.75, c3 = 1,
                                 strategy = 3,
                                 S = 0.85
  )
  
  # Return results for both strategies
  expect_equal(res[1,]$u, -9999)

  expect_equal(res[2,]$u, 1281.28, tolerance = 0.005)
  expect_equal(res[2,]$n2, 326)
  expect_equal(res[2,]$n3, 565)
  expect_equal(res[2,]$n, 891)
  expect_equal(res[2,]$RRgo, 0.77)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-29
test_that("04.08", {
  # Multiarm normal endpoints with strategy 1
  
  res <- optimal_multiarm_normal(alpha = 0.05,
                                 beta = 0.1,
                                 Delta1 = 0.175, Delta2 = 0.225,
                                 n2min = 10, n2max = 200, stepn2 = 2,
                                 kappamin = 0.02, kappamax = 0.3, stepkappa = 0.02,
                                 steps1 = 0, stepm1 = 0.5, stepl1 = 0.8,
                                 b1 = 1000, b2 = 3000, b3 = 5000,
                                 num_cl = 12,
                                 c02 = 15, c03 = 20,
                                 c2 = 0.675, c3 = 0.72,
                                 strategy = 1
  )
  
  expect_equal(res$u, 109.9, tolerance = 0.005)
  expect_equal(res$n2, 56)
  expect_equal(res$n3, 205)
  expect_equal(res$n, 261)
  expect_equal(res$Kappa, 0.16)
  expect_equal(res$sProg, 0.32)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-29
test_that("04.09", {
  # Multiarm normal endpoints with strategy 2
  
  res <- optimal_multiarm_normal(alpha = 0.05,
                                 beta = 0.1,
                                 Delta1 = 0.175, Delta2 = 0.225,
                                 n2min = 10, n2max = 200, stepn2 = 2,
                                 kappamin = 0.02, kappamax = 0.3, stepkappa = 0.02,
                                 steps1 = 0, stepm1 = 0.5, stepl1 = 0.8,
                                 b1 = 1000, b2 = 3000, b3 = 5000,
                                 num_cl = 12,
                                 c02 = 15, c03 = 20,
                                 c2 = 0.675, c3 = 0.72,
                                 strategy = 2
  )
  
  expect_equal(res$u, 107.09, tolerance = 0.005)
  expect_equal(res$n2, 30)
  expect_equal(res$n3, 247)
  expect_equal(res$n, 277)
  expect_equal(res$Kappa, 0.20)
  expect_equal(res$sProg, 0.33)
})
#' @editor Lukas D Sauer
#' @editDate 2022-12-29
test_that("04.10", {
  # Multiarm normal endpoints with strategy 3 (i.e. strategies 1 and 2)
  
  res <- optimal_multiarm_normal(alpha = 0.05,
                                 beta = 0.1,
                                 Delta1 = 0.175, Delta2 = 0.225,
                                 n2min = 10, n2max = 200, stepn2 = 2,
                                 kappamin = 0.02, kappamax = 0.3, stepkappa = 0.02,
                                 steps1 = 0, stepm1 = 0.5, stepl1 = 0.8,
                                 b1 = 1000, b2 = 3000, b3 = 5000,
                                 num_cl = 12,
                                 c02 = 15, c03 = 20,
                                 c2 = 0.675, c3 = 0.72,
                                 strategy = 3,
                                 K = 200
  )
  
  expect_equal(res[1,]$u, 109.68, tolerance = 0.005)
  expect_equal(res[1,]$n2, 46)
  expect_equal(res[1,]$n3, 190)
  expect_equal(res[1,]$n, 236)
  expect_equal(res[1,]$K2, 46)
  expect_equal(res[1,]$K3, 151)
  expect_equal(res[1,]$K, 200)
  expect_equal(res[2,]$u, 107.06, tolerance = 0.005)
  expect_equal(res[2,]$n2, 28)
  expect_equal(res[2,]$n3, 208)
  expect_equal(res[2,]$n, 236)
  expect_equal(res[2,]$K2, 34)
  expect_equal(res[2,]$K3, 163)
  expect_equal(res[2,]$K, 200)
})