Work at height is any task where a person can fall to a lower level and be hurt by the fall. Where the legal line sits depends on the country: in Poland, § 105 of the general OHS regulation (Rozporządzenie MPiPS of 26 September 1997) defines it as work performed on a surface at least 1.0 m above the floor or ground level (lexlege.pl, § 105). Working at height training has to start with that definition, then get quickly to the part that keeps people alive.
That part is judgement, not equipment trivia. A worker who can name every component of a harness and still steps onto a fragile roof to save ten minutes has not been trained. The outcome you want from a course is a person who decides whether the job has to happen up there at all, who picks protection in the right order, who inspects what they clip into, and who knows what happens in the minutes after a colleague ends up hanging in mid-air.
Why falls stay near the top of the injury statistics
Falls are not a niche hazard. Eurostat's 2023 EU data on accidents at work shows that "slipping, stumbling and falling" caused 20.0% of non-fatal accidents and 15.4% of fatal accidents at work — a fifth of everything reported. The category is close behind the largest one, and it is spread across almost every industry rather than concentrated in construction.
For comparison, the same Eurostat dataset puts "loss of control of machine, means of transport" at 21.7% of non-fatal and 27.1% of fatal accidents at work, making it the single largest cause, with falls following (Eurostat, 2023 data). Two categories therefore account for a very large share of the harm. One of them you address with traffic management and machine safety. The other you address with how people work above ground level.
What counts as work at height
There is no single European number. Countries define the trigger differently, and some define it by risk rather than by metres. In Poland the threshold is explicit: § 105 sets it at a working surface at least 1.0 m above floor or ground level, which pulls in far more everyday work than most people assume.
The same paragraph carves out an exception. A surface is not treated as work at height if it is enclosed on all sides to a height of at least 1.5 m by solid walls or by walls with glazed windows, or fitted with other permanent structures or devices protecting the worker against a fall. In practice that is the difference between a properly guarded platform and a bare edge.
In Germany the duty is framed differently. The Arbeitsschutzgesetz and the DGUV rules put the obligation on the employer to assess the risk and instruct workers accordingly, without a single universal height figure doing the work. If you operate in both countries, do not export the Polish metre as if it were an EU rule; assess the exposure and protect against it.
What the training has to achieve
Good training changes four behaviours: the decision to go up, the choice of protection, the pre-use check, and the response when something fails. Everything else, from knot names to the anatomy of a lanyard, supports those four. If a course does not visibly move them, it is a lecture with a certificate attached.
Written as outcomes, a competent worker leaving your course should be able to:
| Outcome | What it looks like on site |
|---|---|
| Question the task | Asks whether the work can be done from the ground before reaching for a ladder |
| Apply controls in order | Reaches for collective protection first, personal fall arrest last |
| Inspect before trusting | Checks harness, lanyard and anchor every single time, not on the first day of the job |
| Know the rescue route | Can say who gets the worker down, with what, and how quickly |
The hierarchy of control, in the order it matters
The order is not decoration. Eliminate the need to go up where you can, by designing the task so it happens at ground level. Where you cannot, use collective protection that works without anyone remembering anything: guardrails, edge protection, covered openings. Only then reach for personal fall arrest, which depends on a person clipping on correctly, every time.
That last dependency is why fall arrest sits at the bottom. Guardrails protect the worker who is distracted, in a hurry, or new. A harness protects the worker who remembered. Sites that treat harnesses as their primary control have effectively made human memory their safety barrier, which is a hard thing to defend after an incident.
There is also a cost argument worth making to management. Fitting an anchor line or a permanent walkway once removes a recurring risk from every future job on that roof. Personal protection has to be bought, inspected, replaced and re-trained forever.
Inspecting what you clip into
A pre-use check takes under two minutes and is the last opportunity to catch a harness that will not hold. Run the webbing through your hands looking for cuts, fraying, chemical staining, heat glazing or UV chalkiness. Check stitching, buckles and D-rings for distortion. Then check the thing you attach to, which is where most attention runs out.
Anchor points are the quiet failure mode. Workers often clip into whatever is nearby and looks solid: a pipe, a handrail, a ventilation duct. Solid-looking is not the same as rated, and an anchor that cannot take a fall load will simply come away with the person attached to it. Training should force the question out loud — is this anchor specified for fall arrest, and do I know it, or am I guessing?
Anything that has arrested a fall goes out of service. That instinct is easy to state in a classroom and easy to ignore on a site where the harness looks fine and the next job starts in an hour.
The plan almost nobody writes: rescue
A worker hanging in a harness after a successful fall arrest is not safe. They are suspended, possibly unconscious, possibly injured by the arrest itself, and the clock is running. Suspension trauma is the reason a rescue plan is part of the work plan, not something to improvise while someone dangles.
Write it before the job starts, not after the alarm. The plan should name who performs the rescue, what equipment is on site and reachable, how the person will be brought to a stable surface, and what happens while the emergency services are still on their way. "We will call 112" is not a rescue plan; it is a phone call.
Rehearse it. A rescue that exists only on paper has never been timed, and time is the whole point.
Where people actually get complacent
The falls that hurt experienced people are rarely the dramatic ones. They happen in the first few metres of a climb, on the short job that "will only take a minute", and on the fourth visit to a roof someone already knows. Familiarity lowers the felt exposure long before it lowers the real one.
Two patterns are worth naming in every course. First, the effort of setting up protection feels disproportionate for a two-minute task, so people skip it in exactly the situation where they have the least margin. Second, low heights read as safe. A fall from three metres onto concrete does not care that the number sounded modest.
How VR fits into height training
VR puts the trainee at the edge, in a body that reacts to it, without any possibility of a real fall. That combination is unusual. They can survey a roof, choose an anchor, discover that the convenient pipe is not rated, clip on badly and see the consequence — repeatedly, in a morning, across scenarios you could never stage safely on a live site.
Our Working at Height scenario runs that loop directly, and it pairs naturally with Hazard Identification for crews who need to get better at spotting the problem before choosing the control. The wider evidence on the method is covered in why VR training is more effective than traditional training, and the same procedural logic drives lockout/tagout training: what it must cover.
Be straight about the limits. VR does not fit a harness to a real body, does not teach someone how tight a leg strap should be, and does not substitute for a physical rescue drill with real weight on a real rope. Those stay hands-on. What VR removes is the need to choose between realistic exposure and safe practice, which is the trade-off that has always constrained height training.
A blended programme that holds up
The strongest setup layers three things. Classroom or e-learning for the legal definitions and your site rules. VR for judgement, hazard spotting, inspection habits and the consequence of a wrong decision, repeated until it is automatic. Hands-on sessions for harness fitting, equipment handling and a timed rescue drill with your own kit.
Then verify by watching people work, not by scoring a quiz. Browse the VR course catalog to match scenarios to the work your crews actually do, and keep the records regulators ask for: who was trained, on what, and when.
