What it is
Failure Modes and Effects Analysis (FMEA) is a structured, proactive risk tool that identifies how a device, process, or software could fail, evaluates effects and causes, and prioritizes actions to reduce risk before harm occurs (IEC 60812; ISO 14971:2019). Used across design, manufacturing, and post-market control to improve reliability and patient safety.
Regulatory framework
- Risk management: ISO 14971:2019 (device risk processes; links FMEA outputs to risk controls and benefit-risk).
- Method standard: IEC 60812 (principles, FMEA types, criticality analysis).
- EU: Risk reduction and state-of-the-art controls required (MDR Annex I Ch. I §1–9; IVDR Annex I); integrate with PMS/PMCF (MDR Annex XIV, Arts. 83–86).
- US: FDA QMSR (21 CFR 820 aligned with ISO 13485) expects documented risk control; use FMEA as part of design/production controls; link to complaints, MDR (21 CFR 803), and corrections/removals (21 CFR 806).
- QMS: ISO 13485:2016 (design/production controls, CAPA, feedback) uses risk-based methods such as FMEA.
Key elements
- Scope: design FMEA (dFMEA), process FMEA (pFMEA), use-related FMEA (uFMEA), and software-focused variants (tie to IEC 62304).
- Failure mode, effect, cause, and existing controls captured with traceable IDs (to specs, drawings, software items).
- Severity, occurrence, and detection ratings; prioritize via RPN or risk matrix consistent with ISO 14971 criteria.
- Action plan: additional controls, owners, due dates, and expected risk reduction.
- Living file: revisions on change, field feedback, CAPA, or new hazards.
Process — how it works
- Plan: Define scope, interfaces, and rating scales aligned with your risk policy; assemble a cross-functional team.
- Map: Break the design/process into functions and steps; list credible failure modes and effects.
- Rate: Assign severity, occurrence, detection; calculate priority (RPN or matrix) and document rationale.
- Act: Select risk controls (design changes, process controls, alarms, information for safety); update files and verification plans.
- Verify: Test that controls work; update ratings and residual risk; link to DHF/DMR and labeling.
- Monitor: Trend complaints, nonconformities, and PMS; feed CAPA and refresh the FMEA on changes.
Common pitfalls
- Treating FMEA as a one-time checklist instead of a controlled, revisable record.
- Using RPN alone without severity-based escalation or clinical context.
- Poor linkage to hazards/hazardous situations, usability, and software risk files.
- Inconsistent scales across projects, making priorities unreliable.
- Actions logged without verification of effectiveness or DHF/DMR updates.
Quick checks
- Are high-severity items escalated regardless of RPN?
- Do actions map to ISO 14971 controls and verification tests?
- Is the FMEA versioned and updated after changes, complaints, or CAPA?
- Are user errors and use scenarios included (uFMEA) and tied to usability engineering?
FAQ
Is FMEA required by regulators?
Regulations require effective risk management (ISO 14971; MDR Annex I; FDA QMSR). FMEA is not mandated by name, yet it is an accepted, often expected method to meet those requirements.
What’s the difference between dFMEA and pFMEA?
dFMEA evaluates design failures and effects on the patient/user; pFMEA evaluates manufacturing/service steps to prevent nonconforming product from reaching users.
Should we use RPN or a risk matrix?
Either may be used. Define rules in your risk procedure, emphasize severity-driven escalation, and avoid masking high-severity issues with low detection scores.
How does FMEA connect to CAPA?
Trends and nonconformities feed CAPA; selected actions and effectiveness checks must update the FMEA, risk files, and controlled documents.
When should we update the FMEA?
At design changes, process changes, new complaints or hazards, field actions, or when PMS/PMCF data indicate new risks or higher occurrence.
::contentReference[oaicite:0]{index=0}