This is interesting study from 1988 evaluated the capability of four hazard identification techniques.
Using a collection of prior incidents and accidents, three groups were formed to evaluate which of the contributing factors to the incidents may have been identified if one of these methods had been used prospectively.
Of the 157 contributing factors identified in advance, results were:
• HAZOP, the most successful, found 36% of contributors
• AEA (action error analysis), 21%
• FMEA (failure mode and effect analysis), 17%
• MORT (management oversight and risk tree), the least successful, 6%.
When pooled together (accounting for all of the common contributors), all four techniques identified only around 55% of total factors – leaving a significant proportion unaccounted for.
The authors found that the techniques predominately addressed the technical subsystems, but not as capable of addressing human, information and management subsystems.
Some of their recommendations included:
• using multiple HAZID techniques
• a need for more systematic questioning of design factors, and
• a need to develop tools that better account for people and management factors.
I think the last point is critical and well-aligned with more modern research on systems thinking and human-centred techniques.
Link to the review in LinkedIn: https://www.linkedin.com/pulse/repost-evaluation-validity-four-hazard-identification-ben-hutchinson/?trackingId=QsPSVriJcuc4xrGpdjHgKw%3D%3D
SafetyInsights.org 