Potential Severity vs Actual Outcome vs Barrier Failure: Which Investigation Depth Fits?

A worker steps away from a suspended load one second before the sling tears. No one is hurt, so the first report describes a near miss. In the same week, another site records a minor hand injury caused by a pinch point, even though the guard was removed during a maintenance task. The second event looks worse in the statistics, but the first may reveal the more urgent failure.

The investigation question is therefore not simply, "How severe was the injury?" The stronger question asks what this event tells us about the harm that was possible, the barriers that were supposed to prevent it, and the decisions required to keep the exposure from returning.

Evaluation criteria for choosing investigation depth

Investigation depth should match the decision the organization needs to make, not the amount of paperwork available. Three criteria make that decision more disciplined. Potential severity tests the credible harm that could have occurred. Actual outcome records the harm that did occur. Barrier failure tests the reliability of the controls that stood between exposure and consequence.

These criteria are related, although they are not interchangeable. A serious injury with strong controls may require a different investigation from a harmless event in which every critical barrier was unavailable. James Reason’s work on latent failures helps explain why the visible event is only one layer of the accident pathway. The investigation must therefore examine the conditions that allowed the pathway to remain open.

Potential severity asks what the event could credibly produce

Potential severity is not a license to invent a catastrophic story. It is a disciplined reconstruction of the credible harm that the operating conditions allowed. The investigator should identify the energy source, the person’s position, the time available to respond, the number of people exposed, and the control whose failure would have changed the outcome.

That distinction matters because a lucky escape can look harmless after the fact. A suspended load that misses a worker by a narrow margin does not become low risk because the worker moved. The near miss may show that the exclusion zone, lift plan, supervision, or equipment condition did not provide the separation the task required.

Frank Bird’s loss-control work and the Heinrich safety pyramid are useful here as reminders that precursor events deserve attention before the consequence becomes irreversible. They are not substitutes for evidence. They are a reason to ask whether the event exposed a repeatable pathway that could produce a different outcome under slightly different timing.

Actual outcome records the harm that did occur

Actual outcome remains essential because people need treatment, regulators may require notification, and the organization must describe what happened accurately. The problem begins when the outcome becomes the entire classification. A first-aid case can receive a shallow review even when the same failed barrier could expose many workers, while a serious injury can attract a large report that never changes the work.

The outcome lens also protects the investigation from exaggeration. A credible review distinguishes observed facts from possible consequences, which keeps the team from presenting a speculative worst case as if it had happened. This separation improves trust with workers and gives leaders a clearer basis for action.

For supervisors, the practical test is simple. Record the consequence precisely, then ask whether the consequence explains the full risk. If it does not, add the potential-severity and barrier-failure lenses before deciding that the event is routine.

Barrier failure reveals why the exposure remained possible

A barrier can fail even when a checklist was completed. It may be technically present but unavailable at the point of work, easy to bypass, misunderstood by the crew, or unable to tolerate the conditions in which the task is actually performed. A guard removed for maintenance, an alarm silenced during production, or a rescue plan that cannot reach the work area are all different forms of barrier failure.

Evidence preservation in the first hour matters because barrier performance is often reconstructed from memory after the physical condition has changed. Photographs, equipment status, permit records, control-room logs, witness accounts, and maintenance history should be connected to the question of how the barrier was expected to work and what happened instead.

Andreza Araujo’s *Safety Culture: From Theory to Practice* treats culture as something visible in decisions and routines, not as a slogan detached from operating conditions. The same principle applies to investigations. A barrier is credible only when the system makes its correct use possible under real production, maintenance, and staffing pressures.

Investigation depth should follow the strongest signal

The three criteria can point in different directions. A minor injury with a major barrier failure may require more organizational attention than a serious injury caused by a rare, well-controlled deviation. The right response is not to average the signals. It is to identify which signal demands the most authority, evidence, and change.

Supervisors need a fast triage question set

The first supervisor conversation should protect the scene and the people involved before it seeks a polished explanation. Ask what energy or exposure was present, who could have been harmed, what changed the outcome, which control was expected to work, and whether that control is reliable elsewhere. Each question should lead to evidence rather than opinion.

When the answer points to high potential or a material barrier failure, the supervisor should preserve the condition, notify the responsible operational leader, and prevent the same task from restarting under unchanged assumptions. That is not the same as stopping every operation indefinitely. It is a way to keep the organization from treating an unresolved control question as a closed event.

The investigation route should also be separated from the later barrier review and action-effectiveness audit, because a good cause analysis does not prove that the new control works in the field.

Decision matrix for EHS and operational leaders

EHS leaders can use the matrix below during the first review meeting. It keeps the discussion anchored to the decision that must be made, while the operational leader remains accountable for the work design and resources that determine whether exposure changes.

Recommendation by context

For routine, low-energy events with no credible escalation and no meaningful control weakness, a focused local review may be enough. For high-potential near misses, repeated exposures, or failures involving critical controls, use a formal investigation even when no one needed medical treatment. For serious outcomes, combine consequence management with a separate test of potential severity and barrier reliability.

Across Andreza Araujo’s work with multinational operations and more than 250 cultural-transformation projects, the practical lesson is consistent with the leadership focus in *Make The Difference: Be a Leader in Health & Safety*. The investigation becomes useful when it changes a decision that the operation can see, own, and verify. A longer report is not the same as a stronger control.

The most defensible classification is therefore three-dimensional. Actual outcome tells you what happened to people and assets. Potential severity tells you what the pathway could credibly produce. Barrier failure tells you why the system allowed the pathway to remain open. When those signals disagree, investigate to the strongest one rather than the smallest consequence.