Authoritative Definitions (mapped to ISO 14971 & MDR)

Hazard vs. Harm — At a Glance

Mapping Nuance: Many-to-Many Relationships

In practice, mappings are rarely one-to-one. A single hazard can lead to several hazardous situations; one situation can produce multiple harms; and one harm may arise from multiple hazards via different sequences of events. Your risk file should allow many-to-many links and maintain unique IDs for each node and sequence step. ISO 14971:2019, Clauses 5–7 require traceability across these relationships.

How Hazards & Harms Fit into ISO 14971:2019 Risk Management

1. Hazard identification — enumerate energy sources, chemical/biological agents, software & data risks (incl. AI/ML), usability errors, cybersecurity threats. ISO 14971:2019, 5.4; IEC 62366-1 (usability); IEC 62304 (software).
2. Foreseeable sequence of events — credible misuses, failures, environmental conditions that lead to exposure. ISO 14971:2019, 5.4; Annex A (examples).
3. Hazardous situation — who/what is exposed and under what conditions. ISO 14971:2019, Clause 3 (definition).
4. Harm & severity — define plausible harms and rate clinical impact using a documented severity scale aligned with CER/clinical evaluation. ISO 14971:2019, 6; MDR Annex I (GSPR).
5. Probability of harm — estimate likelihood that the sequence reaches harm; consider detectability, protection layers, field data, and diagnostics. ISO 14971:2019, 5–6.
6. Risk evaluation — compare to pre-defined risk acceptability criteria (policy per ISO 14971:2019, 4.2 “state of the art”).
7. Risk control — apply 1→2→3 hierarchy; verify implementation & effectiveness (tests, validation, summative usability, cybersecurity verification). ISO 14971:2019, 7; MDR Annex I risk reduction “as far as possible”.
8. Residual risk & benefit-risk — re-estimate; if still high, add controls or justify with benefit-risk + PMCF/PMS signals; disclose residual risks in IFU if needed. ISO 14971:2019, 7.4; MDR Annex I §23 (information supplied).
9. Production & post-production — integrate PMS (complaints, trend reports, FSCA, vigilance, literature, registries) and feed back to the risk file. ISO 14971:2019, 10; MDR Chapter VII (Arts. 83–86).

Risk Estimation Beyond “Severity × Probability”

• Detectability — incorporate detection before harm (alarms, self-tests, UI affordances). If detection interrupts the sequence, the probability of harm may be lower than the probability of a fault. ISO 14971:2019, 5–6; IEC 62366-1; IEC 62304.
• Effectiveness of controls — verify implemented effectiveness (bench tests, EMC/60601-1, verification logs, summative usability). ISO 14971:2019, 7.3–7.4; IEC 60601-1 (electrical limits).
• Benefit-risk & state of the art — decisions must reference clinical benefit and current best practice; document rationale and alternatives considered. ISO 14971:2019, 4.2 & 7.4; MDR Annex I.

Concrete Examples (with Control Hierarchy & Verification)

Clinical Language: Use MedDRA for Harms

For consistency and audit defense, describe harms using standardized clinical terminology (e.g., MedDRA preferred terms). Align severity definitions with clinical categories and your CER/clinical evaluation. Map each harm in the risk file to a standardized term and code; use the same terms in PMS trending and vigilance reports. ISO 14971:2019 requires consistent definitions and traceability; MDR Chapter VII requires consistent PMS/vigilance reporting.

Copy-Paste Risk Statement Templates

Hazard: [energy/biological/software/usability/cybersecurity]
Sequence of Events: [misuse/failure/environmental]
Hazardous Situation: [who/what is exposed]
Harm (MedDRA PT): [clinical injury/health impact]
Severity: [S1..S5 with definition] · Probability of Harm: [P1..P5 with basis]
Initial Risk: [matrix cell] → Controls: [1 design / 2 protection / 3 info]
Verification of Effectiveness: [test/protocol/report id]
Residual Risk: [matrix cell] · IFU Residual Risk Disclosure: [yes/no + section]

Align with ISO 14971:2019 Clauses 5–7; ensure risk acceptability policy per 4.2 is referenced.

Traceability — From Hazard to Labeling

PRRC Oversight & Review Gates

Article 15 MDR requires the manufacturer to have a Person Responsible for Regulatory Compliance (PRRC) with specific qualifications and protected authority. Embed PRRC checkpoints in your risk lifecycle:

• Gate A — Policy & Criteria: Approve risk acceptability policy (ISO 14971:2019, 4.2 “state of the art”).
• Gate B — Analysis Complete: Confirm hazard inventory completeness (incl. usability, software/AI, cyber); challenge sequences & misuse coverage.
• Gate C — Control Design Freeze: Sign off that the 1→2→3 hierarchy has been applied and verification plans exist.
• Gate D — Residual Risk & IFU: Approve benefit-risk justifications; ensure residual risks are disclosed in IFU where required (MDR Annex I §23).
• Gate E — PMS Loop: Confirm PMS plan routes harm signals to the risk file (MDR Arts. 83–86; ISO 14971:2019, 10).

Documentation tip: Maintain a PRRC Decision Log with date, device/version, gate, decision, evidence links, and required actions.

Post-Market Surveillance & Harm Signal Handling

1. Source harm signals: complaints, service records, trend reports, FSCA, vigilance, literature, registries, usability studies.
2. Classify & triage: map each signal to hazard → hazardous situation → harm (MedDRA term); re-rate severity/probability with clinical input.
3. Update risk file: revise estimates, add/adjust controls, document benefit-risk; escalate to CAPA as needed.
4. Update labeling/IFU: if residual risk understanding changes, align warnings/cautions (MDR Annex I §23).
5. Management review: trend residual risk; verify PMS effectiveness and PRRC sign-offs are current.

Refs: ISO 14971:2019, Clause 10; MDR Chapter VII (Arts. 83–86).

Common Pitfalls (and How to Avoid Them)

• Mixing terms: writing “harm” where you mean “hazard” confuses controls — train teams on definitions (ISO 14971:2019, Clause 3).
• Skipping sequences: jumping from hazard to harm without a credible sequence; include misuse & environmental factors (5.4).
• Evidence gaps: listing controls without verification of effectiveness (7.3–7.4); tie each control to test/report IDs.
• Labeling drift: warnings not reflecting residual risks; sync IFU with the risk file (MDR Annex I §23).
• Clinical language drift: harms described inconsistently; mandate MedDRA PTs and coding in risk file and PMS.

Audit-Ready Checklist

• Approved definitions for hazard, harm, hazardous situation, risk, sequence of events (ISO 14971:2019, Clause 3).
• Complete hazard inventory covering HW, SW (incl. AI/ML), usability, cybersecurity; many-to-many traceability to situations & harms (5–7).
• Risk acceptability policy approved before analysis (4.2), citing “state of the art”.
• 1→2→3 risk control hierarchy evidenced; verification of implementation & effectiveness logged (7.3–7.4).
• Residual risk decisions justified (7.4); IFU disclosures aligned (MDR Annex I §23).
• PRRC sign-offs captured at gates A–E; unresolved critical risks block release (MDR Art. 15).
• PMS feeds risk file; trending reviewed by management; CAPA integration proven (ISO 14971:2019, 10; MDR Arts. 83–86).
• Harms consistently coded with MedDRA; used in vigilance and PMS trending.

Practical, Audit-Ready Documentation Fields

Policy ID: RM-POL-001  |  Rev: 1.0
Scope: All devices & configurations
Basis: ISO 14971:2019 (4.2), MDR Annex I, state of the art
Criteria: R ≤ ALARP* only if hierarchy 1→2→3 exhausted; 
  benefits outweigh residual risk; IFU discloses residuals if needed.
Governance: Approved by QA Head & PRRC; reviewed annually.
*Use ALARP or company-specific matrix aligned to MDR “as far as possible”.

Date | Device/Version | Gate (A–E) | Decision | Evidence links 
| Residual risk statement | IFU impact | PMS trigger | Owner

Control ID | Control Type (1/2/3) | Verification Method | 
Protocol/Report ID | Result | Acceptance Criteria | Approver

Signal ID | Source | Hazard→Situation→Harm (MedDRA PT) | 
S/P re-rating | CAPA? | IFU update? | PRRC sign-off | Closed date

Trends Impacting Risk & Compliance (SaMD / AI/ML)

• Software lifecycle rigor — classify software items, maintain SOUP inventory, and prove risk control effectiveness per IEC 62304; map software hazards through to harms in the RMF.
• Use-related risk — formative/summative usability evidence (IEC 62366-1) directly supports detectability and control effectiveness.
• Transparency & updates — for data-driven models, document dataset risks, drift monitoring, and update controls; ensure PMS captures algorithm-related harms.

Standards cross-reference: ISO 14971:2019; IEC 62304; IEC 62366-1; IEC/TR 80002-1.

Bottom Line

Hazard is the potential source; harm is the outcome. Robust risk management connects them via credible sequences, prioritized controls, and verified effectiveness—under PRRC oversight with MedDRA-coded harms and PMS feedback. That’s how you stay compliant and protect patients.

Need an audit-ready Risk Management File?

We build ISO 14971 files aligned to MDR GSPR, set up PRRC review gates, and operationalize PMS feedback loops. SaMD and Class I/II devices can start this week.