Benchmark-Calibrated Mixture Prior

Prior Construction

Standard Trial Series ▼
Phase Include α (1-sided) Power*
Phase 2a
Phase 2b
Phase 3-1
Phase 3-2
*Power at effect = TPP (2.00)
❌
Status: Not calibrated
Click button to calibrate to benchmark

Beliefs about treatment effect based on the benchmark alone

Assumptions behind the benchmark-calibrated prior

  1. Parametric class: two-component normal mixture centred at the null effect and the TPP
  2. Key input: a separate estimate of the historical efficacy success rate of similar programs across Phase 2–3 development
  3. Calibration: the mixture weight is chosen so that the Prior Predictive PoS (PPoS) through a standard trial series matches this historical estimate