Incident investigation and root cause on a construction site

The steps, in order

1. Make the scene safe and care for people. Medical help, stopping the work, controlling remaining hazards and securing the area come before any questions.
2. Preserve evidence. Photograph the scene, keep equipment and materials as they were where it is safe to do so, and note conditions: weather, lighting, noise, time of day.
3. Gather facts. Interview witnesses separately and soon, ask open questions, and collect documents such as the JHA, permits, inspection records, training records and maintenance logs.
4. Reconstruct the sequence. Build a timeline from what was planned to what happened.
5. Find causes. Look past the immediate cause to the conditions and decisions behind it.
6. Choose corrective actions. Match them to the causes, with owners and dates.
7. Report, share and follow up. Meet your recording and reporting duties, share lessons with other crews and confirm the fixes held.

Step 1 matters on exam day. If a question asks what to do first after an injury, it is almost never "start the report."

Immediate cause vs root cause

The immediate cause is what happened at the moment: a worker stepped onto an unsecured cover and fell. The root causes are the system conditions behind that moment: no one owned the cover inspection, the crew was rushed, the plan did not list the opening. Fix only the immediate cause and the same event returns with a different worker.

The 5 whys on an invented example

Suppose a laborer's hand is cut by a blade on a saw with a missing guard.

1. Why was the hand cut? The hand contacted the moving blade.
2. Why was the blade exposed? The guard was removed.
3. Why was the guard removed? It blocked a cut the crew needed to make.
4. Why was the saw used for that cut? The right tool was not on site.
5. Why was it not on site? The task was not planned, so no tool was ordered.

The root cause is in planning, not in the worker's hand. The strongest fix is to plan the task and supply the right tool, with a check that guards stay on. A reminder to "be careful" would not touch it. Notice that five is a guide, not a rule. Stop when you reach a cause the organization can actually change.

Other methods you should recognize

• Causal factor charting: a timeline with the conditions that contributed at each point.
• Fishbone diagram: groups causes into categories such as people, equipment, method, materials and environment.
• Fault tree: starts with the event and works backward through the combinations that led to it.

You do not need to master each. You need to recognize them and know they all ask the same question: what conditions made this possible?

Near misses count

A near miss is an event that could have hurt someone and did not. Treat it as free information. Programs that investigate near misses find problems before they injure anyone, and the exam rewards that reasoning in program questions.

Recording and reporting

After the investigation starts, separate duties come into play: recording injuries on the OSHA logs and reporting certain events to OSHA within set time limits. Those rules are in 29 CFR Part 1904, covered in the recordkeeping guide. Remember that Part 1904 is not Part 1926, a distinction the exam likes. After an event, you can also compute your rates with the incident rate calculator.

Wrong answers that sound right

• "Discipline the worker." An investigation that stops at blame teaches workers to hide incidents.
• "Retrain the crew." Training is an administrative control. It is fine as one action, rarely the whole fix.
• "Close the case once the injury has healed." Corrective actions need follow up.
• "Interview everyone together to save time." Group interviews let accounts drift toward one version.

Try it

Take the free 10-question check and see how you handle first-action and root cause stems. It needs no signup and every answer comes with an explanation.

FAQ

How soon after an incident should I interview witnesses?

As soon as it is safe and practical, because memory fades and accounts change when people talk to each other.

Is the 5 whys method required?

No. It is one common technique. What matters is that the method gets past the immediate cause.

Who should lead an investigation?

Someone trained in the method who is not defensive about the outcome, with input from the supervisor and workers who know the task.