Skip to main content

Table 2 Prevalence of events assumed in each trial, according to the scenario and to the application data set, and weight and maximal rates of misclassifications assigned to each event to build the decision rule

From: A Bayesian non-inferiority approach using experts’ margin elicitation – application to the monitoring of safety events

   Event
   Death IVH [1] NEC [2] Retinopathy
Simulation study      
All scenarios Prevalence in FD arm 0.39 0.15 0.06 0.04
Scenario A Prevalence in HD arm 0.39 0.15 0.06 0.04
  Good decision [3] Acc Acc Acc Acc
Scenario B Prevalence in HD arm 0.26 0.10 0.04 0.03
  Good decision [3] Acc Acc Acc Acc
Scenario C Prevalence in HD arm 0.58 0.23 0.09 0.06
  Good decision [3] U U Acc Acc
Scenario D Prevalence in HD arm 0.78 0.30 0.12 0.08
  Good decision [3] U U U U
Scenario E Prevalence in HD arm 1.00 0.45 0.18 0.12
  Good decision [3] U U U U
Weight[4] 100 88 70 60
Maximal misclassifications rates     
Class a misclassifications [5] 0.10 0.10 0.10 0.10
Class b misclassifications [6] 0.10 0.16 0.25 0.30
Data set for application     
Prevalence in FD arm 0.39 0.15 0.06 0.04
Prevalence in HD arm 0.47 0.23 0.12 0.08
Good decision [3] U U U U
  1. HD arm: half-dose arm; FD arm: full-dose arm
  2. 1IVH: Intraventricular haemorrhage
  3. 2NEC: Necrotizing enterocolitis
  4. 3Good decision= What have been considered as good decision for each scenario and event: “Acc” if the difference of prevalence of events is Acceptable, “U” if the difference is Unacceptable
  5. 4Weight = Relative severity of the event according to the experts
  6. 5Class a misclassifications rate: Trials that conclude that the difference between arms is Unacceptable, among trials with acceptable difference
  7. 6Class b misclassifications rate: Trials that conclude that the difference between arms is Acceptable, among trials with unacceptable difference