Knowledge base
How to Properly Investigate an Incident Under ISO 45001
Here is a simple-language version: ISO 45001 requires organizations to investigate incidents to identify their root or underlying causes and check whether the same or similar incidents could happen in other areas of the organization. This approach provides a practical way to meet that requirement and helps organizations take appropriate corrective actions to prevent incidents from happening again.
Neha Dvivedi · 16 August 2026
ISO 45001 requires organisations to investigate incidents to understand what happened, why it happened, and whether the same thing could happen again. A good investigation should focus on finding the real causes—not simply blaming an individual.
What Is an Incident?
An incident is any event related to work that causes or could cause injury or ill health.
This includes:
• Injuries
• Work-related ill health
• Near misses
• Incidents involving contractors
• Incidents involving visitors
If an organisation investigates only incidents that cause injuries, it may be missing important warning signs that could help prevent more serious incidents in the future.
Step 1: Make the Area Safe and Collect Information
First, make sure the area is safe and prevent further harm.
Then collect evidence that may disappear or change, such as:
• Photos of the scene
• Position of equipment and materials
• Condition of machinery
• Statements from people who witnessed the incident
Speak to people separately and as soon as possible. Memories can change quickly, especially after people discuss the incident with each other.
Ask “What happened?” rather than “Who is responsible?”. This encourages people to provide more honest and useful information.
Step 2: Establish What Happened
Before deciding why the incident happened, first understand exactly what happened and in what order.
Create a timeline covering:
• Events
• Working conditions
• Actions
• Decisions
Avoid deciding on a cause too early. Otherwise, investigators may look only for evidence that supports their first assumption.
Step 3: Identify All the Causes
An incident usually has more than one cause.
Look at three levels:
1. Immediate Causes
What directly caused the incident?
For example:
• An unsafe action
• Faulty equipment
• An unsafe workplace condition
2. Contributing Causes
What made the immediate cause more likely?
For example:
• Time pressure
• Poor lighting
• Equipment condition
• Lack of staff
• Fatigue
• Poor workplace layout
3. Underlying Causes
Why did those conditions exist in the first place?
This could involve:
• Poor planning
• Inadequate resources
• Maintenance decisions
• Procurement decisions
• Insufficient supervision
• Poor work design
A useful question is: “Why did this happen?” Keep asking why until you reach a management or system-level decision.
If the investigation ends only with “the worker made a mistake,” it may not have gone deep enough.
Step 4: Don't Immediately Blame Human Error
If the investigation identifies human error, ask:
1. Was the correct action realistically possible under the circumstances?
2. Could another trained and competent person have made the same mistake?
3. Did the way the work was designed make the mistake more likely?
If a reasonable person could have made the same error, the mistake may be a symptom of a system weakness, rather than the root cause.
Step 5: Choose Corrective Actions Using the Hierarchy of Controls
Corrective actions should follow the hierarchy of controls.
Before choosing training, procedures, or instructions, consider whether the hazard can be:
• Eliminated
• Replaced with something safer
• Controlled through engineering measures
Training and procedures are useful, but they are generally lower-level controls.
Therefore, if retraining is the only corrective action, ask whether stronger controls could address the hazard more effectively.
Step 6: Check Whether the Same Problem Exists Elsewhere
One of the most important parts of an investigation is checking whether a similar risk exists in other areas.
For example:
• If one machine has a missing safety guard, check other machines.
• If a permit was not issued for one job, review the permit process for similar jobs.
• If one department has a particular risk, check whether other departments face the same risk.
Think of every incident as a warning or sample that may reveal a wider problem.
Step 7: Complete the Actions and Communicate the Results
Do not simply assign corrective actions. Check that they were:
• Actually implemented
• Effective
• Preventing the problem from happening again
Then communicate the findings and changes to the workforce.
When employees see that reported incidents lead to real improvements, they are more likely to continue reporting incidents and near misses.
Common Weaknesses in Incident Investigations
Avoid these common mistakes:
1. Stopping the investigation at the immediate cause.
2. Automatically blaming operator error without checking whether it was foreseeable.
3. Making retraining the only corrective action.
4. Failing to check whether the same risk exists elsewhere.
5. Assigning actions without checking whether they were effective.
6. Recording near misses but not investigating them.
7. Failing to communicate the investigation results to employees.
The Key Message
A good ISO 45001 incident investigation should not simply answer “Who made the mistake?”
It should answer:
What happened? → Why did it happen? → What allowed it to happen? → Could it happen elsewhere? → What can we change to prevent it from happening again?
What this covers
See how this looks as a working system
Reading about a requirement and seeing the documentation that satisfies it are different things. In a short demo we open the actual manual, procedures and records set for ISO 45001, show you how each clause is answered and where your existing way of working already fits. You will know what implementation involves before you commit to it.
More reading
- ISO 45001:2018: Documentation and Compliance Requirements
A clause-by-clause guide to everything ISO 45001:2018 requires you to document, including what an auditor may ask to review for each requirement. Written as clear requirements, not as a checklist.
12 September 2026
- Australia's first mining industrial manslaughter conviction
Mastermyne was convicted in March 2026 and fined $7 million in May — a record WHS penalty. What changed, and what boards should read into it.
12 September 2026
- New Zealand Changes Workplace Safety Rules from April 2027
The Health and Safety at Work Amendment Act received Royal Assent in July 2026 and will come into effect in April 2027. The changes focus more closely on serious workplace risks that could result in fatalities.
12 September 2026
