🔥 8/10
12 hours ago
Building Clinical Software For A World That Changes Faster Than Code
In environments where software influences real clinical decisions, rigor is not optional. It is the condition that makes progress safe.
- The U.S. FDA has approved over 500 AI/ML-based medical devices as of 2023, yet only 12% include continuous learning algorithms, highlighting the regulatory gap.
- Predetermined change control plans (PCCPs), proposed by the FDA in 2025, aim to pre-approve algorithm updates for AI clinical software, reducing re-certification delays.
- A 2022 study in JAMA found that two widely used commercial sepsis prediction algorithms had an area under the curve (AUC) below 0.65 across diverse hospital systems, underscoring drift risks.
By Boris Berat, Forbes Councils Member
·
3 min read
·
Forbes
Read story →
