Filter the Noise, Illuminate the Signal
Filter the Noise, Illuminate the Signal
EMA qualified PROCOVA™ (Prognostic Covariate Adjustment); with this you can increase confidence in clinical results and potentially use smaller sample size for Phase II and III randomised controlled trials for regulatory decision making.
Is this a game changer for medicinal development? What can it do and can not do? Let us answer some common questions for the medicinal developers.
The landscape of AI in drug development has shifted from "Can we do this?" to "How do we establish credibility?"
In this video, we break down the FDA’s framework for Risk-Based Credibility. It ultimately comes down to a simple mental equation:
Model Influence + Decision Consequence = Model Risk
We explore this through two contrasting scenarios:
1️⃣ High Risk: AI acting as the "Sole Determinant" for patient safety stratification.
2️⃣ Medium Risk: AI acting as a "Support Tool" for manufacturing fill-volume checks.
Yes - but, how do you prove the algorithm is safe and effective to use as a diagnostic?
Read the model credibility assessment, applying FDA 7-step risk-based approach outlined in the 2025 draft guidance.