SAFETY FIRSTPEOPLE FIRSTPASQUALE MAGLIULO

MY WORK

Root cause analysis

From an anomaly to the conditions that made it possible: understand the problem before selecting an action.

A cause analysis starts with a factual problem statement: what happened, where, when and how it differed from the expected result. It can address a nonconformity, a recurring error or an operational difficulty. The aim is to identify conditions that can be improved and verify the result.

The 5 Whys help develop questions. They do not require stopping at the fifth answer or finding a single cause. Each link needs evidence. ‘Human error’ opens further questions about information, tools, competence and work organisation; it does not close the analysis.

A five-stage approach

  1. 1. Define the problem

    Separate facts, effects and hypotheses. Identify the expected result and the observed deviation.

  2. 2. Gather evidence

    Examine records, document versions and operating conditions, and listen to the people involved.

  3. 3. Explore the whys

    Develop several branches where needed and verify the proposed links between conditions and effects.

  4. 4. Select actions

    Connect each action to an identified condition, with an owner and a completion date.

  5. 5. Check effectiveness

    Agree how the result will be assessed and check whether the problem recurs.

Illustrative example · Outdated information

Three issued documents contain outdated data. The following chain is a set of hypotheses to test, not a report of an actual case.

  1. Why was the data outdated? An earlier template version was used.
  2. Why was it available? The superseded copy remained accessible without a clear status label.
  3. Why had it not been withdrawn? The release workflow did not include checking operational copies.
  4. Why was that check missing? No owner had been assigned to withdrawal and verification.
  5. Why had the weakness not emerged? Process reviews tracked deadlines without measuring version errors.

Compare this explanation with documents, dates and interviews before accepting it. Explore alternative branches, including interfaces, workload, access and training. A plausible chain is not yet a demonstrated cause.

Sources and further reading

Explore also: Incident investigations ↗

Latest insights

Human error as a starting point

Human error as a starting point

Describing an error does not explain why it was likely. Available information, interfaces, workload and conditions help explain how the decision arose. Knowing that someone made…

Read the insight ↗
Five whys: stay with the evidence

Five whys: stay with the evidence

Repeating why can open an investigation, but each step needs support. An elegant chain without evidence can move away from how the process actually works. How do you connect…

Read the insight ↗

HSE retrospective 2025

Illustrative image of the work environment · Woodworking incidents: look beyond the final movement

Root cause analysis and incident investigations

Woodworking incidents: look beyond the final movement

A hand approaching a dangerous part is often the final visible step in a much longer sequence. The InforMO publication released by INAIL on 16 December 2025 examines serious and fatal incidents in the wood industry. Its prevention value lies in connecting the event to equipment, working methods and …

Read the article ↗
Explore the 2025 articles ↗