Workplace Incident Investigation: From Human Error to Root Cause

Why shouldn’t an investigation stop at “human error”?

Many incident reports reach the same conclusion: “the worker did not follow the procedure” or “employee carelessness”. It is a comfortable finding because it closes the file quickly, but it does not prevent the next incident. The next worker will be doing the job under exactly the same conditions.

The UK Health and Safety Executive (HSE) says investigations that conclude operator error was the sole cause are rarely acceptable, and that human error usually has underlying causes such as poor training, supervision or equipment design. ^[1] The US Occupational Safety and Health Administration (OSHA) says conclusions such as “the worker was careless” or “the employee did not follow safety procedures” do not get at the root causes, and asks the team, whenever a shortcoming is found, to establish why it existed and why it had not been addressed earlier. ^[2]

OSHA also warns that an investigation looking for someone to blame stops short of the root causes and makes workers and managers less willing to speak openly. ^[2]

What should be investigated, and at what level?

HSG245 distinguishes an accident, which results in injury or ill health, from a near miss, which caused no harm but could have, and an undesired circumstance, a condition with the potential to cause harm before any event has occurred. ^[1] OSHA encourages investigating near misses as well as injuries, and prefers the word “incident” to “accident” because most injuries are preventable. ^[2]

Not every event needs a full team. The test HSG245 sets is the worst outcome that could have occurred and the likelihood of recurrence, not the size of the actual injury. ^[1] A tool dropped from height that hit nobody may deserve a higher level of investigation than a minor cut to the hand.

Three layers of cause: immediate, underlying and root

HSG245 divides causes into three layers: ^[1]

The OSHA and US Environmental Protection Agency fact sheet defines a root cause as a fundamental, underlying, system-related reason why an incident occurred. ^[3] Fixing only the immediate cause may remove a symptom but not the problem, and an incident often has more than one root cause. ^[3]

Types of human error, and why the remedies differ

Not all “human error” is the same. HSG245 distinguishes: ^[1]

• Slips and lapses: a familiar action carried out wrongly or a step forgotten, which happens even to experienced workers.
• Mistakes: applying the wrong rule, or reaching the wrong conclusion in an unfamiliar situation.
• Violations: deliberately breaking the rules or cutting corners.

The remedy depends on the type: slips and lapses call for design measures such as colour coding, checklists and interlocks; mistakes call for training, safe working procedures and equipment design; and violations call for simple, workable rules backed by supervision and monitoring. ^[1] Even a violation needs a question: was the procedure workable, and did production pressure or incentives encourage the shortcut? ^[2]

The four steps of an investigation

1. Gather information. Once first aid is given and the area is safe, preserve the scene and record it with photographs, measurements and sketches. ^{[1, 2]} Start as soon as practicable: memory is best straight after the event, and evidence can be lost. ^[1] Collect the documents too: maintenance and training records, equipment manuals and earlier corrective action recommendations. ^[2]

In interviews, make clear that the purpose is fact-finding, not fault-finding, interview people in their own language with an interpreter if needed, ask permission before recording, and let the person describe what happened without interruption before asking what could have prevented it. ^[2] This matters especially on Gulf work sites, where the workforce speaks many languages.

2. Analyse information. Put the events in chronological order, then identify the immediate, underlying and root causes by asking “why?” repeatedly. ^[1]

3. Identify risk control measures. Identify the controls that failed or were missing, and rank them: eliminate the risk first, then tackle it at source, then minimise it through procedures and protective equipment. ^[1]

4. Action plan and implementation. Turn each control into an action with an owner and a deadline, as detailed in the action plan section below. ^[1]

Statutory reporting does not wait for the investigation

Reporting requirements differ from country to country. In the UAE, for example, employers must notify the Ministry of Human Resources and Emiratisation (MoHRE) of an injury or death at the workplace within 48 hours, and establishments with 50 or more workers must keep a record of injuries and have a mechanism for internal reporting and investigation. ^[5] Elsewhere in the GCC, check with the competent labour or human resources ministry, and do not hold back a notification until the investigation is complete.

How do you reach the root cause?

Ask four core questions: what happened, how did it happen, why did it happen, and what needs to be corrected? ^[3] Then keep asking “why?” until you reach failings in management, design, planning, organisation or operations. ^[2]

OSHA suggests practical questions: if a procedure was not followed, was it out of date, and was training adequate? If a guard or machine failed, why was it not repaired? Was there a staffing shortage on the day? ^[2] The OSHA fact sheet recommends combining tools: brainstorming and checklists may be enough for simple incidents, while complex ones need logic or event trees with a timeline. ^[3]

For incidents in chemical and process industries, the Center for Chemical Process Safety (CCPS) publishes specialist guidelines covering team formation, scene preservation, witness interviews, root cause identification and recommendations, and notes that its methods also apply to occupational health and safety incidents. ^[4]

Hypothetical example: a hand injury at a conveyor

A worker’s hand is injured while clearing a jam on a conveyor without stopping and isolating it. The first draft of the report says: “The worker did not follow the isolation procedure.” Had the investigation stopped there, the recommendation would have been to warn and retrain the worker.

Asking “why?” reveals more: the isolation point is far from where jams occur, jams happen several times a shift because maintenance is behind, and the supervisor tolerates quick clearing to keep production going. The root causes therefore lie in maintenance, design and supervision, and the recommendations become: fix the cause of the jams, add an isolation point nearby, fit a guard that prevents access while running, and review production expectations with supervisors. The details are hypothetical, to illustrate the method.

The action plan: how do you stop it happening again?

• SMART objectives: specific, measurable, agreed, realistic and timescaled. ^[1]
• Immediate action on high risks, or stop the work until controls are in place. ^[1]
• An owner and a deadline for every action, with one senior manager overseeing the whole plan. ^[1]
• Review the relevant risk assessments and safe working procedures after any incident. ^[1]
• Tell workers and their representatives about the findings and the progress of implementation. ^[1]

Some corrections take time and need senior management support, but OSHA stresses that fixing root causes at programme level is what prevents recurrence. ^[2]

Conclusion

Make “human error” the start of the investigation, not its end. Set the level by the potential outcome, gather evidence early, keep asking “why?” until you reach causes in the system, then turn the findings into a plan with owners, deadlines and follow-up.