Medical Device Vigilance Manual

The Methodology of Causality Assessment

Chapter 3 — Root Cause Analysis, Human Factors, and Determination Frameworks

1 Device Causality Categories

Click each category to view its defining criteria. The scale represents strength of causal association.

Certain
▶
  • Failure is reproducible in laboratory/bench testing of the suspect device
  • No plausible alternative cause identified after thorough investigation
  • Positive “rechallenge” in stress testing — re-exposure to operating conditions replicates the adverse event mechanism
Probable
▶
  • Reasonable temporal link between device use and the adverse event
  • Device failure or malfunction confirmed through investigation
  • Alternative causes considered but judged unlikely on the balance of evidence
Possible
▶
  • Temporal link exists between device use and the event
  • Underlying disease progression, user error, or concomitant device/drug factors are also plausible explanations
  • Insufficient evidence to confirm or exclude device contribution
Unlikely
▶
  • Improbable that the device contributed to the adverse event
  • Alternative factors (patient condition, concomitant treatments, environment) provide a substantially better explanation
  • Device investigation reveals no defect or performance deviation
Not Related
▶
  • Comprehensive device investigation is conclusive — no defect, malfunction, or performance issue found
  • The event is fully explained by non-device factors
  • Temporal association alone is insufficient to maintain a causal link

2 Root Cause Analysis & Human Factors Engineering

The paradigm shift from blaming the operator to evaluating the system.

Traditional View
“User Error”

Single-cause attribution that ends the investigation prematurely and fails to capture systemic design deficiencies.

⟶
Modern HFE Approach

Multi-Factor Root Cause Determination

Affordances (Physical / Visual Cues) Cognitive Load Training Adequacy UI Design Environmental Factors
▼ ▼ ▼ ▼ ▼
Root Cause Determination
IMDRF Annex D — D0102
Device Difficult to Operate — covers design-related use difficulties including poor affordance design, confusing labeling, and inadequate feedback mechanisms.
IMDRF Annex D — D0107
Cybersecurity — addresses vulnerabilities in connected devices, unauthorized access, data integrity failures, and software-related hazards.
Example — Pen-Injector Needle-Stick Injury: Rather than labeling this as “user error,” an HFE analysis examines the cap removal affordance, whether the safety mechanism provided adequate visual/tactile feedback, whether the grip design inadvertently guided the user’s fingers toward the needle path, and whether environmental factors (e.g., low lighting, rushed clinical setting) contributed to the injury.

3 Algorithmic vs. Global Introspection

Two complementary frameworks for reaching a causality determination.

Global Introspection

A holistic approach relying on engineering judgment and clinical expertise. The assessor considers the totality of evidence — device history, patient factors, clinical context, and field data — to form an integrated opinion.

Strength Captures unique, context-dependent device-patient interactions that checklists may miss.

Weakness Subjective by nature; assessments vary between experts, yielding poor inter-rater reproducibility.

Algorithmic (Naranjo Adaptation)

A questionnaire-based scoring method adapted from the Naranjo drug ADR scale for medical devices. Dechallenge (removal of device) and rechallenge (re-application or stress testing) questions are highly relevant for reusable devices.

Strength Structured, reproducible, auditable — well suited for regulatory submissions.

Weakness May oversimplify complex multi-factor device incidents.

≥9 Definite
5–8 Probable
1–4 Possible
≤0 Doubtful

4 Investigation Methods

IMDRF Annex B — standardized approaches to device incident investigation.

🔍
B01

Testing of Actual Device

Physical and functional examination of the specific device involved in the incident, including visual inspection, electrical safety testing, and performance verification against specifications.

⚖
B02

Testing of Similar / Equivalent Device

When the actual device is unavailable or destroyed, a device from the same batch, model, or equivalent design is tested to identify potential systematic defects or failure modes.

⚗
B03

Laboratory Evaluation

Controlled bench testing under simulated clinical conditions, including accelerated aging, biocompatibility analysis, materials characterization, and failure mode replication.

📄
B04

Review of Design Documentation

Examination of the device master record, risk management file (ISO 14971), design verification/validation reports, and manufacturing records to identify design-stage root causes.

Key Insight

In modern device causality, a “rechallenge” is often performed in a laboratory setting where the suspect device is subjected to stress testing to replicate the failure, rather than re-exposing patients. This distinguishes device vigilance from pharmaceutical causality assessment, where a clinical rechallenge may sometimes be ethically permissible.