What Happens to a Fall Arrest System After It Arrests a Fall

Why this matters

The instant somebody is hanging in a harness, the emergency is the suspension, not the fall. Get eyes on them, get a voice on them, and start the retrieval and the call for medical response in the same breath rather than one after the other. Do not send a second person up an unprotected route to reach them, because a rescue that produces a second casualty has failed twice.

An arrest starts two clocks that run at wildly different speeds. One is measured in minutes and decides whether the worker survives a system that just did its job correctly. The other runs for as long as the equipment exists, and it governs what goes back in the van. Shops write down the second one and almost never write down the first, because the first is difficult and the second fits on a form.

The call

Two techs were walking a warehouse roof to scope a rooftop unit replacement. One stepped on a skylight whose frame had corroded, went through it, and was arrested by a self-retracting lifeline clipped to an overhead beam anchor. He ended up hanging inside the building, roughly 12 feet above the warehouse floor. Nothing failed. The lifeline locked, the anchorage held, and the free fall was short.

Here is what the phone log and the alarm panel reconstructed afterwards.

09:14  fall and arrest
09:15  partner reaches the skylight opening, sees him
09:16  partner calls emergency services
09:17  partner gets no coherent answer from below
09:31  first engine on site
09:38  crew inside the building; the aerial is no use
       indoors, so they go looking for a lift
09:44  first scissor lift found; battery dead
09:54  weight off the harness

Forty minutes. He spent two days in hospital and came back to work. The shop's written fall protection plan did have a rescue section, and the rescue section said, in full, to call emergency services. Everything between 09:16 and 09:54 was improvisation by people who had never walked that building.

That is the finding, and it is not really about the fire department, who did nothing wrong. It is that a rescue plan with no equipment named, no location named, and no elapsed time attached to it is a phone number.

The rescue clock, and what it is measured between

29 CFR 1926.502(d) requires the employer to provide for prompt rescue of employees in the event of a fall, or to assure that employees are able to rescue themselves. 29 CFR 1910.140 carries the equivalent duty for general industry. Neither states a number of minutes, and that is not an oversight: the standard puts the duty on you and leaves the target to your plan.

So set one and measure against it. A workable default is 15 minutes from the moment of arrest to the moment the worker's weight is off the harness, timed by drill, and tuned to your own buildings and crew sizes. Two honest qualifications belong in the same breath. That is an operational target you can manage a crew against, not a physiological guarantee: harness suspension restricts venous return and symptoms can begin sooner, and the onset is not predictable from a worker's fitness. The mechanism is in What a Body Harness Does That a Body Belt Cannot. And the clock starts at the arrest, not at the phone call, which in the case above is a difference of two minutes that nobody in the shop had ever counted.

Never suspend a person to time a drill. Use a weighted dummy of comparable mass on the same rig. A rescue drill that creates a suspension is the exact hazard the drill exists to shorten.

What the shop changed, and what the drills said

They kept the same harnesses, the same lifelines and the same anchorages. Everything they changed was about the forty minutes.

• Suspension relief straps fitted to every harness in the fleet, so a conscious worker can take weight on their feet and buy time. Those buy minutes, they are not a rescue, and a worker who has deployed them is still a live emergency.
• A named retrieval method per building, written on the job file: which lift, where it lives, whose it is, and who has the key. Buildings the crew has not surveyed get a survey before the first roof visit.
• A pre-use battery check on the identified lift, added to the tailboard, because the dead battery cost minutes nobody had budgeted and no one had ever been responsible for that lift.
• Two people minimum for any roof work over an interior space, which is what turns a rescue into something that can start before help arrives.

Then they timed it, twice, with a weighted dummy.

Drill 1, three weeks after the incident   11 minutes
Drill 2, one month later                  13 minutes

Drill 2 was slower, and the shop wrote it down as slower rather than averaging the two into a friendlier number. The cause was mundane: a different two-man crew, one of whom had not run the first drill, spent two minutes locating the lift key. That is the finding a drill exists to produce. A rescue plan degrades every time a crew changes, and the only thing that detects the degradation is running it again with the crew you actually have.

What comes out of service, and what "suitable for reuse" really allows

29 CFR 1926.502(d) requires that personal fall arrest system components subjected to impact loading be immediately removed from service and not used again for employee protection until inspected by a competent person and determined to be undamaged and suitable for reuse.

That sentence gets read as discretion, and it mostly is not. Everything in the load path comes out, not just the harness: the harness, the lanyard or self-retracting lifeline, every connector, and the anchorage connector or strap.

Then the competent person's determination runs into the boundary that makes this simple in practice. The determination is bounded by the manufacturer's instructions, and for a deployed energy absorber and for a harness that has taken an arrest, essentially every manufacturer's instruction says destroy and replace. A competent person cannot determine a component suitable for reuse against instructions that say it is not. Self-retracting lifelines are the one family where the answer is sometimes different: many require return to the manufacturer or an authorised service centre for inspection and recertification after an arrest, which is a real path and is not the same as a competent person looking at it on a bench.

Destroy the removed components the same way a failed inspection is closed out, by cutting through the webbing in several places with a sharp blade on a bench, cutting away from your body, so nothing can be recovered from a bin. How to Inspect a Harness and What Takes It Out of Service carries the routine criteria; this is the trigger that overrides all of them.

The component nobody quarantines

The anchorage took the load too. It is the one part of the system that stays behind when the crew leaves, because it is part of somebody's building, and it is the part most likely to be used again next week by a crew that knows nothing about the incident.

An anchorage that has arrested a fall needs evaluation before it is used again, by a qualified person, and where the question is whether the structure itself was damaged, by a registered professional engineer. That is the same routing What an Anchorage Has to Be Rated For and Who Decides gives for accepting an anchorage in the first place, and for the same reason: the load path runs through a building, and fall protection expertise is not structural expertise.

Two extensions people miss. If a horizontal lifeline was involved, the whole line and both end anchorages took the event, not just the point nearest the fall. And if a second worker was attached to the same anchorage or the same line at the time, their system was loaded as well, however little they felt.

What the incident owes in writing

Two separate obligations, and they have different deadlines and different audiences.

Reporting. 29 CFR 1904.39 requires an employer to report a work-related in-patient hospitalization, amputation or loss of an eye to OSHA within 24 hours, and a work-related fatality within 8 hours. The case above met the hospitalization trigger. Those clocks run from when you learn of the event, and they are the shortest deadlines in this whole card.

The configuration record, which nobody requires and which is the only thing that improves the next plan. Capture the anchorage location and type, the device model and serial, the harness ID, the free fall and clearance as they had been computed beforehand, the actual offset from the anchorage, and the elapsed times from the arrest to each step of the retrieval. Take photographs of the rig before anything is disturbed, once nobody is in danger.

What that record gets you is an operational fact you can bank: a dated, specific account of what the configuration was, written while people still remembered. What it means legally, and how it would be treated in any dispute or claim later, is a question for your own attorney and your insurer, not something to settle from a knowledge base card.

How to tell whether your rescue plan is real

Three tests, and a plan that fails any of them is an intention.

Name the equipment and its location for this building. Not "a lift." Which lift, where, whose, and who checked it this morning. A plan that works at your shop and not at the customer's building is not a plan for the job you are on.

Attach an elapsed time you have actually measured. Not estimated. Drilled, with a weighted dummy, with the crew size you send. If you have never timed it you do not have a number, and the case above ran 40 minutes on a plan that had never named one.

Run it again when the crew changes. The second drill above was slower than the first for a reason that had nothing to do with equipment. Plans do not decay, but the knowledge of them does, and the only detector is a repeat.

The failure mode is not carelessness and it does not look like carelessness. It is a shop that buys good harnesses, computes clearance properly, inspects on schedule, and treats the rescue section as the paperwork at the end. Every dollar of that effort is spent making the arrest survivable, and then the survival is decided by whether anyone knew where the lift was.

References

• 29 CFR 1926.502(d), personal fall arrest system criteria for construction, including the prompt rescue duty and the immediate removal of components subjected to impact loading.
• 29 CFR 1910.140, personal fall protection system criteria for general industry.
• 29 CFR 1904.39, reporting fatalities, hospitalizations, amputations and losses of an eye to OSHA, with the 8-hour and 24-hour deadlines.
• Manufacturer instructions for the harness, energy absorber and self-retracting lifeline, which govern whether any impact-loaded component can be returned to service and by what route.
• See related: What a Body Harness Does That a Body Belt Cannot, How to Inspect a Harness and What Takes It Out of Service, What an Anchorage Has to Be Rated For and Who Decides.