Preventing falls while erecting and dismantling

A completed scaffold is a working platform with edge protection. A scaffold halfway up is neither, and the people most exposed to a fall on any site are the ones building the thing everyone else will later work from safely. That asymmetry is the whole subject.

The control has to exist before the lift does

The practical problem in erection is sequencing: protection for the lift being built has to be established from the lift below, before anybody works at the new level unprotected. Where that sequence is understood and followed, the risk is managed. Where it is treated as something to sort out once the lift is up, there is a window — sometimes minutes, sometimes a whole shift — in which somebody is working at height with nothing between them and the ground.

Which means the method is not a document to be produced for a client. It is a description of the order in which things happen, and the test of whether it is real is whether a gang can describe it without reading it. The current NASC guidance on this, and the method your own designer specifies, are the references that matter — they are revised, and they are the authority rather than any summary.

Harnesses are the last line, not the plan

Where personal fall protection is used, two things go wrong regularly. The first is treating it as the primary control when the sequence should be: the harness catches a failure of the method, it is not the method. The second is anchorage — a harness clipped to something that will not hold a fall arrest load is worse than no harness, because it produces confidence without protection.

Worth establishing explicitly for each structure type: what a scaffolder is expected to clip to, at what stage, and who has confirmed that anchorage is suitable. Left to judgement on the day, it becomes whatever is nearest.

In practice: the lift nobody planned for

A gang erecting a straightforward independent hits an obstruction at third lift — a projecting canopy the survey did not show. They work round it, which takes an extra hour and involves two operatives working from a lift with the guardrail temporarily removed to get tube past the obstruction.

Nothing in the method covered that, because the method described a structure without a canopy in it. The gang improvised competently and the protection was absent for an hour. The failure was not the improvisation — it was that no route existed to pause and ask, and that removing edge protection to solve a problem felt like a normal part of the job.

Common mistakes

• –Treating the method statement as a document for the client rather than the order work actually happens in
• –Relying on harnesses as the primary control rather than as the last line
• –No agreed anchorage points, so operatives clip to whatever is nearest
• –Removing edge protection to solve an obstruction, with no pause and no record
• –Gangs unable to describe the sequence without reading it
• –No route to stop and ask when the structure differs from the drawing

What tends to work

• –A sequence the gang can state from memory, because they helped set it
• –Anchorage confirmed per structure type before work starts, not chosen on the day
• –A fast route to a supervisor or designer when the site differs from the drawing
• –Treating any removal of edge protection as an event to be recorded and reinstated deliberately
• –Reviewing near misses on erection specifically, rather than lumping them with general site incidents

Advance guardrail, and why the argument is about sequence

Systems that let edge protection be placed from the lift below, before anybody works at the new level, are the practical expression of the hierarchy: the fall is prevented rather than arrested. Whether a particular system suits a particular structure is a question for the designer and current guidance, not a universal answer.

The principle underneath is universal though. Whatever method is used, the question worth asking of it is whether anybody has to work at an unprotected edge in order to establish the protection. If the answer is yes, that is the part of the method to examine rather than accept.

Getting this right

Near misses during erection are the cheapest information available and the least often captured, because the gang solves the problem and moves on. A route to report that something was tighter than it should have been is what turns it into a planning change, and it needs a supervisor able to act on it.

• –Ask of any method: does establishing the protection require working unprotected
• –Confirm anchorage per structure type before work starts, not on the day
• –Treat removing edge protection as an event to record and reinstate deliberately
• –Give gangs a fast route to pause when the site differs from the drawing
• –Capture erection near misses separately from general site incidents

Where this connects: what a near miss is worth

Erection near misses are the cheapest safety information a scaffolding firm can get and the least often captured, because the gang solves the problem and moves on. Nobody was hurt, the structure went up, and the fact that somebody worked for an hour in a way nobody planned disappears with the shift.

A route to report that something was tighter than it should have been — with no penalty attached — is what converts that into a planning or design change. It requires a supervisor able to act on it and a firm that treats the report as useful rather than as an admission, which is a cultural question more than a procedural one.

• –Capture erection near misses separately from general site incidents
• –Make reporting cost the reporter nothing
• –Feed them back into method and design, not just into a log
• –Ask gangs what they had to work around, rather than waiting to be told
• –Review whether the same obstruction recurs across jobs
• –Treat a site with no reports as unlikely rather than exemplary