Inspection, Audit, and Incident Investigation
Core Concepts
Distinguish the three activities clearly. An inspection checks conditions and behaviours against requirements at a point in time. An audit evaluates whether the management system conforms to a defined standard and is functioning as intended. An investigation determines what happened in an incident and why, so recurrence is prevented.
Inspections need defined scope, frequency, competent inspectors, checklists, and a corrective action system with owners and due dates. Findings without tracked closure are worthless. Audits use criteria (regulation, internal standard, consensus standard), evidence sampling, and findings graded by significance.
Incident investigation should follow a consistent method: secure the scene and care for people, gather evidence (physical, documentary, witness), establish the sequence of events, identify root and contributing causes, and recommend controls. Avoid blame-focused conclusions; human error is usually a symptom of system conditions such as inadequate training, poor design, or production pressure. Tools include the 5 Whys, cause-and-effect (fishbone) analysis, and the loss causation model separating immediate causes (unsafe acts/conditions) from basic causes (management system failures).
Traps
- Calling an inspection an audit, or vice versa.
- Stopping at "worker was careless" rather than identifying why the error was likely.
- Writing recommendations without assigning responsibility and deadlines.
- Failing to distinguish near-misses from reportable incidents.
Rule of Thumb
Investigate to the level of a cause you can actually fix. If the corrective action is "retrain the worker," the investigation is probably incomplete.