Human Factors in Accident Investigation

Introduction ‘To say accidents are due to human failing is like saying falls are due to gravity. It is true but it does not help us prevent them’ Trevor Kletz Aim today is to present methods that are known to help identify human failure in accident investigation and prevent reoccurrence Not a black art, a pragmatic and robust process

What we expect Methodical process for gathering information, analysing what went wrong (and right), and learning lessons in order to: Manage risk Prevent reoccurrence Retrospective tool, but can be powerful in promoting change

Accident reports What happened Who to When How it happened But not why Technical myopia Failure to consider human factors

Significance of human factors Up to 90% of accidents attributable to some degree to human failures Proportion and significance increasing as technical safety measures improve

Human failure taxonomy Human failures Intended actions Unintended actions Violation - Intended consequences Errors - Unintended consequences Mistakes Lapses Slips When the person decided to act without complying with a known rule or procedure When the person does what they meant to, but should have done something else When the person forgets to do something When the person does something, but not what they meant to do

Slip, lapse or mistake? Involuntary or non-intentional action No Was there intention in the action? Was there prior intention to act? No Spontaneous or subsidiary action Yes Yes Did the actions proceed as planned? Unintentional action (slip or lapse) No Yes Did the actions achieve their desired end? Intentional but mistaken action No Yes Successful action

How to apply Create timeline Identify significant behaviours Analyse behaviours Identify effective measures to prevent reoccurrence Record

Errors Slip When a person does something, but not what they meant to do Lapse When a person forgets to do something Both are unintended actions with unintended consequences

Mistakes When a person does something they intended to do, but should have done something else Rule based – choosing a standard solution for a known problem – the maintenance worker who selects the wrong isolation procedures Knowledge based

Mistakes Because the action is intended, mistakes are much harder to detect at the individual level People believe what they are doing is right and often dismiss evidence to the contrary Bias Tunnel vision

Violations Violation When a person decides to act without complying with a known rule or procedure Note that, in this context, there must be an known rule or procedure This is not a moral or ethical judgement

Violations

Violations Note that we all integrate rule violation into our day to day lives so the identification of a violation should not be regarded as a precursor to discipline Indeed, we tend to like those who break the rules

Violations

Violations Types of violations Routine Exceptional Acts of sabotage The key to the effective analysis of violations is to understand why What antecedents were present? What behaviour was observed? What consequences resulted?

Performance Influencing Factors Defined as ‘the characteristics of the job, the individual and the organization that influence behaviour’ Considered during behavioural analysis, often at the end of the process Very broad topic including a range of factors e.g. fatigue, group effects, design of equipment, mental wellbeing, task knowledge/complexity Often have a critical role in error causation but equally often overlooked.

Common issues Failure to correctly specify behaviour The individual involved The task they were engaged in at the time What they did (or did not do) What the outcome was Making early decisions and sticking to them As information becomes available, a mistake can become a violation Failure to identify the multiple behaviours contributing to an accident or incident Timeline critical

Why bother with any of it? Each failure type has a different set of solutions designed to prevent their reoccurrence. For example (not exhaustive): Slip/Lapse NOT training Hardware solutions Cross checks PIFs Error Training e.g. scenarios Group support Challenge Violations Behaviour modification Culture improvement

What to remember Human behaviour can be predicted with reasonable accuracy Correctly integrating HF into your accident investigation process will reap rewards – just look at the contemporary causation figures Separating error, mistake and violation represents a highly valuable first step

A final thought The most powerful influence on human behaviour is outcome Therefore managing human failure requires a high degree of corporate honesty: What behaviour is really rewarded? Are we willing to look at organizational factors, especially when we see rule breaking? Are we willing to make the investments that are likely to prevent reoccurrence? Are we willing to strive for objectivity and pragmatism?

Thanks "Don't fix the blame, fix the problem!"