calibrate_thresholds.md
In ppseq: Design Clinical Trials using Sequential Predictive Probability Monitoring

one-sample simulate data

Code
  sim_dat1(p = 0.1, n = c(5, 10))
Output
  # A tibble: 2 x 2
       n1    y1
    <dbl> <int>
  1     5     0
  2    10     1

two-sample simulate data

Code
  sim_dat1(p = c(0.1, 0.3), n = cbind(c(5, 10), c(5, 10)))
Output
  # A tibble: 2 x 4
       n0    n1    y0    y1
    <dbl> <dbl> <int> <int>
  1     5     5     0     2
  2    10    10     0     5

two-sample vector argument to n

Code
  sim_dat1(p = c(0.1, 0.3), n = c(5, 5))
Output
  # A tibble: 1 x 4
       n0    n1    y0    y1
    <dbl> <dbl> <int> <int>
  1     5     5     0     2

evaluate threshold one-sample case

Code
  eval_thresh(dat1, 0.95, 0.3, p0 = 0.1, delta = NULL, S = 500, N = 25)
Output
  # A tibble: 1 x 6
       n1    y1 pp_threshold ppp_threshold   ppp positive
    <dbl> <int>        <dbl>         <dbl> <dbl> <lgl>   
  1     5     0         0.95           0.3 0.094 FALSE

evaluate threshold two-sample case

Code
  eval_thresh(dat2, 0.95, 0.3, p0 = NULL, delta = 0, S = 500, N = c(25, 25))
Output
  # A tibble: 1 x 8
       n0    n1    y0    y1 pp_threshold ppp_threshold   ppp positive
    <dbl> <dbl> <int> <int>        <dbl>         <dbl> <dbl> <lgl>   
  1    10    10     0     5         0.95           0.3 0.988 TRUE

one-sample calibrate thresholds

  pp_threshold ppp_threshold mean_n1_null prop_pos_null prop_stopped_null
1          0.9          0.05           25          0.05                 0
  mean_n1_alt prop_pos_alt prop_stopped_alt
1          25         0.91                0

two-sample calibrate thresholds

  pp_threshold ppp_threshold mean_n0_null mean_n1_null prop_pos_null
1          0.9           0.2        15.25        15.25          0.07
  prop_stopped_null mean_n0_alt mean_n1_alt prop_pos_alt prop_stopped_alt
1              0.65       23.95       23.95         0.88             0.07