What the Last Section Taught Me: Lessons from a Hanging Scaffold Fall into Water

Why I’m writing this

The most dangerous part of a scaffold job is the last section. I learnt that again this month, supporting an investigation where two scaffolders fell into the water while taking down a hanging scaffold under a jetty piperack.

I won’t name the company, the site or the people. They deserve that, and the investigation is still going through its approvals. What I want to share are the lessons, because I’m sure the same gaps exist on other sites right now.

I also want to say this upfront. If an investigation report is full of findings but can’t tell management what in the system needs to change, it’s an audit, not an investigation. So everything below points back to the system, not to the two men who didn’t go home.

Lesson 1: A hanging scaffold comes down upside down

When you strip a normal scaffold, you work from the top and stand on the lift below. A hanging scaffold is the opposite. The platform is at the bottom, and everything above you is what holds you up.

So anything you remove above your head may be what you’re standing on. When the crew reached the last section, they clamped a temporary tube to the steel beam and stood on it to reach the components overhead. That tube came off the beam and took them down with it.

Three things I want every supervisor to check before the last section comes down:

  1. What will they stand on? The method statement must answer this for the last section, not just the first. An improvised tube is not a working platform. Use the deck above, a work boat, a MEWP, or a properly designed temporary platform.
  2. Are the beam clamps fitted correctly? Girder couplers go in pairs, one on each side of the flange. A single clamp per point can slip or rotate. Read the rating too: a scaffold girder coupler is typically rated around 600 kg, while “3 ton” is a figure you see on lifting beam clamps, which are a different product.
  3. Where are they hooked? The harness must be anchored to the permanent structure, not to the scaffold being dismantled. In this case both men were hooked to the tube they were standing on. When it fell, the harness tied them to it.

And if they’re working over water, a harness and a life jacket are not either/or. The harness stops the fall. The life jacket is what keeps you alive if the harness doesn’t.

Lesson 2: An L1 certificate does not cover hung scaffolds

Both men held L1 scaffolding certificates. Most of the crew did. Many people I talk to still assume a certified scaffolder can do any scaffold. That’s not what the syllabus says.

DOSH’s guideline for scaffolder training centres sets the minimum syllabus for each level in its Lampiran C:

LevelScaffold types in the syllabus
L1Free-standing prefabricated and tubular scaffolds
L2Spurred, cantilevered, barrow ramps, loading platforms
L3Hung scaffold, suspended scaffold, and erection and dismantling of advanced scaffolds

Source: JKKP, Garis Panduan Pengiktirafan Sebagai Pusat Pengajar Pengendali Perancah, Lampiran C, p. 18.

Why does this matter legally? BOWEC 1986 was repealed on 1 June 2024. Under the new OSH (Construction Work) (Design and Management) Regulations 2024, regulation 17(1)(e), a contractor must not use anyone without the skills, training and experience for the task. DOSH’s letter of 4 July 2024 also stopped scaffolder registration in MyKKP. Contractors now appoint scaffolders based on their training certificate.

That puts the check squarely on the contractor. If you don’t read the level on the certificate and match it to the scaffold type, nobody else will. In this case the worker list submitted at tender was not the crew who turned up, and nobody compared them.

Lesson 3: The danger you get used to

The hardware failed in seconds. The organisation took months to set it up. Three things stood out for me.

Complacency near completion. The project was almost finished. The last section felt routine. That is exactly when people take shortcuts, and exactly when the scaffold stops supporting itself. The end of a job needs its own hazard review, not the generic JHA written on day one.

Normalisation of deviance. Hooking onto the scaffold itself, when there was no proper anchor point, had become the normal way of working. It had been done before and nothing happened. When the method lives in people’s experience instead of on paper, nobody can see a deviation, because there’s nothing written to compare it against.

“Everyone does it.” Most of the crew wore a harness only, no life jacket. Some leaders believed either one was enough. When a practice is accepted from the client down to the subcontractor, the worker isn’t breaking a rule. He’s following the culture.

Lesson 4: When the alarm goes, who’s in charge?

After the fall, two co-workers jumped into the water to save their friends. Both nearly drowned. I understand why they did it. But no one had told them not to, and no one had shown them what to do instead: raise the alarm, throw the buoy, stay on the jetty.

Then the notification chain broke down. Nobody took command, and the emergency response plan was never activated. When we checked, the plan didn’t name who provides boat rescue for work over water anyway.

The boat did come. Someone on site made the right call and asked the port operator for help. Credit to that person. But a rescue that depends on one person making the right call in a panic is luck, not a plan. Your ERP for work over water should name the rescue boat, the incident commander, and the first phone call, and you should drill it before the work starts.

For my fellow Tripodians: what this case reminded me

A few Tripod Beta points came up again and again while we built the tree.

  • Test every precondition. Ask: if this PC were absent, would the immediate cause have been less likely? We had permit issues that looked important. They failed this test, so they went into Additional Findings, not onto the tree. Honest, not padded.
  • Monitoring is not a barrier. An inspection or a safety walk only helps if someone acts on it. Name the thing that actually stops the hazard.
  • Don’t invent barriers that can’t work. We considered access control at the jetty. But it fails the moment everyone registers, so it came off.
  • An effective barrier carries no causation path. The rescue boat worked, so it stays effective, even though the ERP was never activated. That gap still matters, and it goes in Additional Findings.
  • Facts first, then the tree. One witness statement is a lead, not a finding. We marked every claim that rested on a single statement until it was backed up.
  • One underlying cause, one BRF. In this case most of the underlying causes landed on Organisation. That tells management where to look.

Before your next scaffold comes down

If you manage scaffolding work, especially over water, check these this week:

  • Does the method statement say how the last section comes down and what the crew stands on?
  • Does every scaffolder’s certificate level match the scaffold type? Hung and suspended scaffolds are L3 content.
  • Is the crew on site the same crew you approved?
  • Are beam clamps fitted in pairs, on opposite sides of the flange, with the right rating?
  • Is there an independent anchor point on the permanent structure? If not, does the crew know to stop?
  • Harness and life jacket over water?
  • Does your ERP name the rescue boat, the incident commander and the first call, and has it been drilled?
  • Have the crew been told: raise the alarm and throw the buoy, don’t jump in?

Two young men went to work and didn’t come home. None of the gaps above was new. They were all sitting in the system, waiting. Go and find yours before they find your people.

Asyraf Khalil, Cikgu Barrier

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top