Working at Height: Hazards, Controls and Requirements

Ask a group of people to picture work at height and almost all of them picture scaffolding. That mental image is one reason so much height work goes unmanaged, because the reality in a facilities or maintenance operation looks nothing like it. It looks like a technician on a stepladder in a plant room reaching for an isolator, or someone on a chair changing a corridor lamp at seven in the evening, or a short walk across a warehouse roof to clear a gutter. Those are all work at height, they carry real risk of serious injury, and they vastly outnumber the scaffolded jobs.

The message up front: work at height is any work in a place where a person could fall a distance liable to cause injury, including falls below ground level and falls through a surface that will not hold them. The dominant pattern in serious falls is not the spectacular job. It is the short, low, routine one. If your height discipline only switches on when scaffolding arrives, it is switched off for most of your actual exposure.

1. What counts as work at height

Work at height is work in any place where, if precautions were not taken, a person could fall a distance liable to cause personal injury. Three consequences follow, and each catches people out. There is no heroic threshold in the concept: the test is whether a fall could injure, not whether the height is impressive. It includes falling below ground level, so working beside an open pit, tank, lift shaft, service trench or excavation counts even though nobody is elevated. And it includes places where the surface underfoot may not support a person, so walking on a roof containing rooflights, cement sheet or corroded decking is height work whether or not an edge is nearby, because the fall route is through the surface rather than off it.

In Great Britain the specific duties for planning, organising and protecting height work sit in the Work at Height Regulations 2005 (SI 2005/735, amended by SI 2007/114), under the framework of the Health and Safety at Work etc. Act 1974 (c. 37). In the United States, federal fall protection duties split by sector: construction in 29 CFR 1926 Subpart M, sections 1926.500 to 1926.503, and general industry in 29 CFR 1910 Subpart D, Walking-Working Surfaces, sections 1910.21 to 1910.30, as rewritten by the rule effective 17 January 2017. In the UAE the binding duties come from Federal Decree-Law No. 33 of 2021, administered by MOHRE with occupational safety and health duties in Article 13, and in Abu Dhabi the applicable framework is ADOSH-SF, the Abu Dhabi Occupational Safety and Health System Framework, Version 4.0, administered by the Abu Dhabi Public Health Centre.

Those regimes are not equivalent and none binds a reader elsewhere. Where a jurisdiction sets a numeric trigger height, a required type of protection or a mandatory inspection regime, that figure comes from its own legislation and differs from its neighbours. This article publishes none of them deliberately. Look yours up in the current instrument, not in a summary.

2. Why low heights and short jobs dominate the harm

This is the corrective that makes the rest of the article useful. The consistent message from safety regulators is that a large proportion of serious and fatal falls occur from low heights: from ladders, stepladders, small platforms, vehicle beds, racking and positions only just above floor level. Not from the scaffolded facade. From the low reach in the plant room.

The reason is not mysterious. Height work that looks dangerous gets planned. Height work that looks trivial gets improvised, because nobody believes a four-minute task warrants the overhead of proper access. So the ladder is footed on a slope, or leaned against a gutter, or stood on at the top step, or the person reaches sideways rather than climbing down to move it. The fall is short, the landing is concrete or plant, and the head or spine takes it. There is a duration bias behind it too: people rate risk by exposure time, so a short job feels safe because it is over quickly. But the probability of a fall is driven mainly by the quality of the access, the stability of the base, whether the person is reaching beyond their balance and whether the surface holds. A short job from bad access is more dangerous than a long job from good access.

3. The hazards, explained rather than listed

Hazard registers for height work tend to collapse into one line: "fall from height". That is a category label, not an analysis. The hazards are distinct, they call for different controls, and treating them as one thing is how arrangements cover the obvious fall route and miss the real one.

Falls of persons happen off an unprotected edge, through an unguarded opening such as a riser, hatch, void or missing floor panel, through a surface that will not bear the person, from a ladder or access platform, and into an opening or excavation at ground level. Those are four or five different routes, and edge protection that handles the first does nothing about the second or third.

Falling objects endanger everyone below, and tools are only part of it: fixings, offcuts, debris, staged materials and drilling spoil all fall. This is why exclusion zones, toe boards, debris netting and tool restraint exist. It is the hazard most often forgotten when height work happens above an occupied building, a public footway or live plant.

The access equipment itself is a hazard as well as a control. Instability, an inadequate or sloping base, a buried void under an outrigger, incorrect erection, missing components, overreaching from the platform and movement while occupied all produce falls and collapses. Ground condition in particular is routinely assessed by eye on surfaces never designed to carry a point load.

The structure raises two questions frequently unanswered on an existing building: will this support a person and their equipment, and will it support the forces an anchor would put into it. On a new build there is design information. On a thirty-year-old plant room, a roof re-covered twice, or a facade with unknown fixings, the load-bearing condition is unknown until somebody competent establishes it. Fixing an anchor to convenient steelwork because it looks strong is a structural assumption dressed up as a decision.

Weather and environment matter more than people expect. Wind loads both the person and the equipment, and rises with height and around building corners. Rain, ice, dust, algae and oil turn a walkable surface into a slide. Heat degrades judgement, balance and endurance, which in a Gulf summer is not a footnote: a task that is routine in February is a different task in July on a reflective roof.

Overhead services and electrical hazards arrive because going up brings a person and their equipment near things that were harmless at floor level: overhead lines, busbars, exposed terminations, lighting circuits. A conductive ladder or an elevating platform boom extends reach in every direction, not only the intended one.

Suspension after a fall means an arrest system that works has not ended the emergency, it has changed it. The person is suspended, possibly injured by the arrest forces, possibly unconscious, and their condition deteriorates with time in suspension. That is why arrest is one component of an arrangement rather than the arrangement itself.

Combination hazards are the norm in maintenance. Height work coincides with hot work, confined or restricted spaces, live plant and lifting operations, and each pairing creates interactions neither hazard has alone: a fire at height cuts off the escape route, and a casualty at height inside a confined space needs a rescue capability very few sites have. Confined space work carries its own regime, in Great Britain the Confined Spaces Regulations 1997 (SI 1997/1713) with Approved Code of Practice L101, and in the US 29 CFR 1910.146 for general industry and 29 CFR 1926 Subpart AA, sections 1926.1201 to 1926.1213, for construction. Hot work at height brings in the NFPA 51B, 2024 edition framework where a contract or an authority having jurisdiction has adopted that edition.

4. Which hazard calls for which class of control

The value of separating the hazards is that each points at a different class of control. This is a mapping of hazard to control class, not a selection table.

For how these hazards get found in the first place, see the hazard identification methods guide. For the assessment applied to a specific height task, which is a different job from this pillar, see working at height risk assessment.

5. The hierarchy of controls applied to falls

The hierarchy is the structuring idea here, and it is worth being precise about what it is. It is a principle, not a standard in its own right. It is required by ISO 45001:2018, as amended by Amd 1:2024, at clause 8.1.2, which sets out eliminate, substitute, engineering controls and reorganisation of work, administrative controls including training, and adequate personal protective equipment. It is also required by ANSI/ASSP Z10.0-2019 at section 8.4 in the United States, and described by NIOSH on a free public page, which is the origin of the inverted triangle everybody draws. US OSHA does not codify the term in any regulation. The general treatment has its own pillar, the hierarchy of controls complete guide. What follows is the hierarchy applied to falls.

Avoid working at height at all, so far as is reasonably practicable. This is where the real gains sit, yet it is the level most often treated as a box to tick before moving on to harness selection. Taken seriously it is an engineering agenda, and in maintenance it is unusually productive because the same access is needed again and again.

In a facilities or plant estate it looks like this: relocate the things needing routine attention so they can be reached from the floor, including filters, gauges, sensors, dampers, lubrication points, valves and isolators; specify equipment that can be serviced from ground level rather than equipment somebody must climb to monthly; use extending tools; use remote inspection, fixed cameras and instrumentation so the reason to climb disappears, since a great deal of routine height access is somebody going up to look; bring high-level lighting down on a lowering mechanism; and get access into the specification at design stage, because permanent engineered access designed in at build costs a fraction of retrofitting it and removes the exposure for the life of the asset.

That is where a maintenance organisation has leverage a safety function does not. Every recurring height task in the PM programme is a candidate for elimination by design. A job appearing twelve times a year with a ladder each time is twelve annual exposures an engineering change could remove permanently, and that access is visible in the work order history, which makes it one of the few safety improvements you can target with data you already hold. See work order types in a CMMS and facilities maintenance management.

Where height work is unavoidable, prevent the fall. The first preference is an existing safe place of work: a permanent platform, walkway, plant room floor or a roof with permanent edge protection, where the person is simply not exposed. Failing that, collective physical protection comes next, meaning permanent or temporary edge protection, guarded platforms, and covers over openings. Collective means it protects everybody present without each person having to do anything correctly, which is why it outranks personal equipment. Then collective fall mitigation, which limits the distance and consequence of a fall area-wide rather than person by person. Then personal fall protection, where one distinction matters more than any other here.

Then administrative measures: planning, permits, authorisation, exclusion zones, supervision, competence assurance and briefing. These matter enormously but sit low because they control behaviour rather than the hazard, and behaviour degrades under time pressure. Height work is frequently brought into a permit system: see the permit to work complete guide and types of permit to work. The widely used reference is UK HSE guidance HSG250, published 2005, which is guidance and creates no duties itself; no international standard specifies a permit to work system.

And last, PPE: head, foot, hand and eye protection and high-visibility clothing. Last does not mean unimportant, it means the final layer and never the primary one. See the PPE complete guide. In the US, general industry PPE duties sit in 29 CFR 1910 Subpart I, sections 1910.132 to 1910.140, with personal fall protection systems at 1910.140 added by the 2016/2017 rule; construction PPE is 29 CFR 1926 Subpart E, and the 2024 construction PPE fit requirement is 1926.95, effective 13 January 2025, which has drawn legal challenge.

6. Fragile surfaces, which deserve their own section

Fragile surfaces earn separate treatment for two reasons: they are a leading cause of fatal falls in building work and maintenance, and the hazard is essentially invisible. A surface that will not carry a person can look identical to one that will, because age, weathering, corrosion, previous repairs, coatings and dirt all obscure the distinction. A rooflight under years of dust looks like the sheet beside it.

The behavioural trap compounds. Somebody crosses a roof, nothing happens, and that becomes evidence the roof is safe. It is evidence of nothing except that this route carried this person on this day. Degradation continues invisibly, and a surface that held last year may not hold now. Falls through fragile surfaces characteristically happen to experienced people on routes they have walked before.

The control principles: fragile surfaces must be identified before anyone goes near them rather than judged from the roof; they must be clearly marked so the next person is not repeating the same judgement; the preferred control is not going on them at all; and where access is genuinely unavoidable the arrangement must be designed by someone competent rather than improvised on site. Every specific of that arrangement belongs to competent design under the applicable standards, not to this article.

7. The equipment families, described functionally

A large share of bad height decisions come from using whatever was available rather than what suited the work, so it is worth knowing what the broad families are for. What follows is functional description only. It names nothing as suitable for any particular situation and publishes no rating, capacity, clearance, distance or standard designation, because selection is a competence matter and those parameters come from the equipment's own standards and the manufacturer.

In Great Britain the general work equipment duties underneath all of this come from the Provision and Use of Work Equipment Regulations 1998 (SI 1998/2306), with lifting equipment under the Lifting Operations and Lifting Equipment Regulations 1998 (SI 1998/2307). Construction projects add the Construction (Design and Management) Regulations 2015 (SI 2015/51), and it is worth noting that CDM is where design-stage access decisions acquire legal weight, which is the avoidance level of the hierarchy expressed as a duty. None of these apply outside Great Britain, and Northern Ireland has separate instruments with different years.

8. Rescue and recovery, in principle

A fall-arrest arrangement without a plan to recover the person is incomplete. This is the most frequently missing piece in otherwise credible arrangements, and it goes missing because the arrest system feels like the answer. It stopped the fall. But a person in suspension after an arrest is in a deteriorating situation: possibly injured by the arrest itself, possibly unconscious, and often hard to reach from wherever anyone else is standing.

Two principles follow. The recovery plan must exist before the work starts and must match the actual location, including how a casualty is reached, how they are brought to safety and by whom; a sentence in a method statement saying the team will effect a rescue is not a plan. And it must not depend on the emergency services arriving in time, because response times vary, access to a roof, shaft or plant room is often difficult, and the clock on a suspended casualty does not wait. The logic applies with more force where height work combines with confined space or restricted access, since a casualty is harder to reach in both. See the confined space hazards and requirements guide for how that regime treats rescue.

9. Inspection, examination and the asset-management angle

Access and fall-protection equipment is safety-critical equipment whose condition cannot be assumed, and in most jurisdictions it is subject to formal inspection and, for some categories, thorough examination by a competent person. What the regime is, who may carry it out, what must be recorded and how often it happens all come from the applicable legislation and the manufacturer's instructions. This article publishes no interval, because intervals are jurisdiction and equipment specific and a generic figure would mislead.

The principles are stable even where the specifics are not. Equipment is checked before use by the person using it, inspected at defined points by someone competent, and taken out of service and physically prevented from returning if defective. Records are traceable to the specific item, which means items must be individually identifiable, and the regime is not negotiable against a schedule.

This is where height work meets maintenance management. Access and fall-protection equipment is itself a population of assets with a compliance history: individually identified, with an owner, a location, an inspection due date and a defect history. Managed as a register it is auditable. Managed as a pile of kit in a store with a folder of certificates somewhere, it drifts, and the first sign is usually an item in use with an expired record.

The version buildings lose track of most reliably is permanently installed fall protection. Anchors, eyebolts, mansafe systems, davits, restraint lines and their fixings, installed on a roof during a project, handed over in a document pack and then forgotten. Five years later nobody in the facilities team can say who installed them, to what design, on what structural basis, or when they were last examined, and somebody is about to clip onto one. Those installations are assets. They belong on a register, with the design and structural basis attached, a periodic examination regime and a named owner. A building that cannot answer basic questions about its own anchors has a live exposure sitting on its roof, and the people most likely to discover it are its own technicians. The mechanics of keeping regimes like this visible are covered in permit to work integration with a CMMS.

10. How height work actually goes wrong

The failure patterns are consistent enough to state plainly, and almost none are exotic.

• The short job treated as not needing planning. The dominant pattern, and most of the rest of this list sits downstream of it. Brief task, modest height, equipment not to hand, somebody improvises.
• The wrong equipment used because it was available. Selection by proximity rather than suitability. A ladder used as a platform for an hour and a large machine hired for two minutes are the same error in opposite directions.
• Fragile surfaces walked on. Usually by experienced people, on a route walked before, because nothing happened last time.
• Anchors used whose strength nobody established. Clipping onto convenient steelwork, pipework, handrail or an unlabelled eyebolt. The structural question was assumed rather than asked.
• Equipment moved while occupied. Nobody wants to climb down and back up, so the tower gets nudged along with a person on it.
• Harnesses worn but not connected, or connected to something unsuitable. Treated as compliance clothing rather than one component of a system that only works correctly connected to a verified anchor.
• Overreaching rather than repositioning. The fixing is just out of reach, moving the access takes two minutes, and the person leans. This is the mechanism behind a large share of low-height falls.
• Edge protection removed for access and not replaced. Taken down to get materials through and left down, because the person who removed it was not the person who needed it.
• Work above openings created by other trades. A hole covered yesterday is open today because someone else needed it open, and nobody told the person working above. A coordination failure, not an individual one.
• The cultural tell. The reliable indicator is what people do when the right equipment is not to hand and the job is one bolt. Where the discipline is genuine they stop and get the right access, and nobody treats that as obstruction. Where it is decorative they improvise, and everybody present understands that this is how it is actually done. Briefings that make this specific and local rather than generic are where toolbox talks earn their place.

11. The maintenance-specific reality

A facilities team's height exposure has a particular shape, and recognising it is the most useful thing a maintenance leader can do with this article. The locations are familiar: roof-mounted plant such as air handling units, condensers, chillers and fans; gutters and rainwater goods; rooflights and roof access routes; high-level lighting in warehouses, atria and workshops; high-level ductwork, pipework, valves and dampers; facades, cladding, signage and window systems; lift shafts and plant rooms; and the tops of tanks and vessels.

What characterises almost all of it is short-duration recurring access. Not a project. A twenty-minute job that happens monthly, quarterly or on demand, dozens or hundreds of times a year across an estate, by small teams working alone or in pairs, often out of hours and under pressure to restore a service. That is exactly the category this article opened with, and it is the category least likely to attract a plan.

It is also the category that rewards permanent engineered access more than any other. A location visited once in a building's life does not justify installing a platform. A location visited forty times a year for twenty years does, comfortably, on both safety and cost grounds, and yet it is routinely left to a ladder because each individual visit is too small to trigger a capital conversation. Anyone who can query which PM tasks require height access, and how often, can build a genuine engineering case for removing the exposure. That is the avoidance level of the hierarchy applied with data a maintenance function already owns, and it is the highest-value thing in this article.

The idea to walk away with

Work at height is not a category defined by drama. It is any work where a fall could injure, including falls through surfaces and falls below ground level, and the harm concentrates in exactly the jobs that do not look like height work: low, short, routine, unplanned. The structuring idea is the hierarchy, and its top level is not a formality. Avoiding the work altogether, by designing plant and buildings so people do not have to go up, is where the permanent gains sit, and it is the level a maintenance organisation is uniquely placed to push because it can see its own recurring height access in its own work history.

Below that, collective protection beats personal equipment because it does not depend on anybody doing anything correctly. Within personal equipment, restraint beats arrest because it prevents the fall instead of surviving it. And an arrest arrangement without a recovery plan is not finished. Everything numeric belongs to the applicable law, the equipment standards and competent persons, which is why none of it appears here.

Final thoughts

If you take one thing from this into Monday, make it the short job. The scaffolded facade will get a plan whether or not anybody reads an article about it. The two-minute reach for a damper in a plant room will not, and that is where the harm lives. A height regime that only activates above a threshold of apparent seriousness is inactive for most of its own exposure.

The wider framing sits across the related pillars: what HSE means in practice, the risk assessment complete guide, and the assessment applied specifically to a height task in working at height risk assessment. General risk-management guidance is ISO 31000:2018, which is guidance and not certifiable, with techniques catalogued in IEC 31010:2019. Primary regulator material comes from UK HSE for Great Britain and US OSHA for the United States, with research from NIOSH , which describes the hierarchy of controls but has no regulatory power, and standards catalogues at ISO . Neither US nor UK material has legal force in the UAE, where the binding law is Federal Decree-Law No. 33 of 2021 plus the emirate frameworks. The current published instrument for your own jurisdiction is the only version that counts.

Related reading: Working at height risk assessment, Hierarchy of controls, Risk assessment complete guide, Permit to work complete guide, PPE complete guide, Confined space hazards and requirements, Facilities maintenance management.