Forklift load shift: near miss or reportable event?
A distribution warehouse handles about 60 vehicle movements a day. On Tuesday a forklift carrying a 900 kg pallet of canned goods rounded a blind corner and the load slid off the forks, landing 1.5 m from an agency picker. No one was hurt. The forklift driver, an agency worker on his fourth shift, said the load 'felt light' and he had not seen the picker. The supervisor recorded 'no accident - no action' in the shift log. Two similar load-shift near misses were logged in the previous five months, both on the same corner. The warehouse has a 5 mph limit, but there are no mirrors at the corner and pallets are banded only on two sides. The agency supplies drivers with a one-hour induction; the site's own induction is three hours. The safety manager is on leave.
1. Is the event reportable under RIDDOR? What should be recorded, and why does the distinction matter?
Read the worked answer
Reportability turns on injury and the specified list
On the facts given, there was no injury, so this is not a reportable accident. It is also not, without more, a reportable dangerous occurrence: a load sliding off the forks is not the same as a failure of a load-bearing part of lifting equipment. If the evidence later showed that the forks, mast or lifting attachment had failed, the analysis would change and the specified dangerous occurrence provisions would need to be checked.
The correct action is still to record it internally as a near miss, because internal recording is a risk-management function, not just a legal reporting function. It should capture the date, location, activity, equipment, driver, load, witnesses, immediate cause and immediate actions. The supervisor's 'no action' entry is inadequate: it suppresses the signal from a high-potential event.
The distinction matters because:
- Legal reporting is a narrow duty to the regulator; internal recording is how the organisation learns.
- Treating a near miss as 'nothing happened' loses the chance to prevent a fatality. A 900 kg load falling near a person is a high-potential event even without injury.
- The two earlier similar near misses on the same corner show a pattern that a competent investigation should have detected.
If a person had been injured, reportability would need to be assessed against the injury and lost-time criteria. On these facts, the priority is investigation, not a RIDDOR submission.
2. Plan the investigation. What level is appropriate, who should be involved, what evidence is needed, and what are the likely root causes?
Read the worked answer
Investigation proportionate to potential, not to harm
Level. A formal investigation is justified because the potential consequence was serious, there is a recurrence pattern, and an agency worker is involved. A quick supervisor enquiry is not enough.
Team. Include a trained investigator or safety adviser, the warehouse supervisor, the forklift driver (with support), an agency representative, and someone with knowledge of traffic management. The safety manager can review remotely but should not be bypassed.
Evidence. Secure CCTV before it is overwritten; check forklift pre-use checks and maintenance records; photograph the corner, floor, mirrors, signage and pallet banding; review the load's stability and banding method; interview the picker and witnesses; review induction records for both the agency and the site; check the near-miss log for the previous two events; examine speed-limit enforcement and telematics if fitted.
Likely root causes. These should be tested, not assumed:
- Blind corner design with no mirrors or convex aids, so the driver could not see the picker.
- Load instability: banding only two sides, and the driver's comment that the load 'felt light' suggests poor load presentation or a pallet problem.
- Agency induction gap: one hour versus three hours for site-specific traffic rules.
- Supervision and culture: near misses logged but not investigated, and a supervisor willing to close an event as 'no action'.
- Speed and segregation: a 5 mph limit is ineffective without enforcement or physical measures.
The investigation should distinguish immediate causes from underlying management-system causes, then recommend actions that address the corner, the load, the induction and the near-miss process.
3. Recommend corrective actions and explain how their effectiveness should be verified.
Read the worked answer
Corrective actions should be ranked and verified
Engineering and design
- Install convex mirrors and warning lights at the blind corner, or redesign the traffic route to remove the blind corner.
- Separate pedestrian and forklift routes with barriers and marked walkways; consider one-way flow.
- Introduce speed-control measures such as bumps or chicanes, and enforce the limit through supervision and telematics.
- Review pallet banding: all sides or a load-stability standard, with a pre-move check by the driver.
Management and training
- Bring the agency induction up to the site standard, with a documented assessment before driving on site.
- Train drivers and pickers on the revised traffic rules and load checks.
- Introduce a near-miss review threshold: any load shift near a person is investigated, not closed at supervisor level.
- Clarify the agency's and host's responsibilities in the contract and in practice.
Verification
Effectiveness is not the same as completion. Verify by:
- observing the corner at shift change and peak times;
- auditing a sample of pre-use checks and banding compliance;
- tracking near-miss reports and checking that they are investigated and closed with evidence;
- reviewing telematics or spot speed checks;
- re-interviewing drivers and pickers after a few weeks to see whether behaviour has changed;
- setting a review date and reporting results to the safety committee.
If near misses on that corner continue, the engineering solution has not worked and should be revisited.