Risk-Based Quality Management in Device Manufacturing

Abstract

Medical device manufacturing quality management has traditionally relied on end-of-line inspection and batch release testing -- approaches that detect defects after they occur rather than preventing them. Risk-based quality management (RBQM) redirects resources from uniform inspection toward proactive identification and control of the process parameters most likely to affect patient safety, aligning with the ISO 13485:2016 emphasis on risk-proportionate quality processes and the FDA's shift toward process analytical technology. This study presents the Risk-Based Quality Management Assessment Framework (RBQMAF), evaluating five quality management strategies -- traditional inspection-centric QMS, statistical process control (SPC), process analytical technology (PAT), failure mode and effects analysis-driven control (FMEA-DC), and AI-augmented predictive quality -- across four manufacturing quality tasks: incoming material qualification, in-process defect prevention, final product conformance assurance, and CAPA effectiveness. Our Manufacturing Quality Performance Score (MQPS) integrates defect escape rate, process capability index, CAPA closure effectiveness, inspection efficiency, and regulatory audit readiness. AI-augmented predictive quality achieved the highest MQPS (0.932) through gradient-boosted models trained on 2.8 million process parameter records that predict non-conformances 4.2 hours before detection by traditional methods, enabling pre-emptive intervention that reduced scrap rates by 38%. FMEA-driven control achieved the highest regulatory audit readiness (0.960) through systematic risk linkage between process controls and patient safety outcomes.