What a Fire Marshal Can Actually Ask of a Bed: One Page, Three Tests

A fire marshal asking for “documented low-position evacuation for every occupied bed” is not asking for an evacuation plan. In a hospital built to defend in place, the ask is narrower and far more checkable: proof, bed by bed, that the platform reaches its lowest travel with a patient on it, that it still descends with the mains isolated, and that the occupied bed can be moved through the doors on the route the ward actually uses. Those are three separate functions with three separate tests. One page per serial number, signed by the person who watched the test being run, answers all three. A model certificate signed at the factory answers none of them, and that gap is where most of these requests stall.

Defend in Place Is a Building Strategy, and It Narrows the Ask

The Joint Commission’s environment-of-care standard LS.02.01.10 covers the building and fire protection features that minimise the effects of fire, smoke and heat, and its guidance states the principle without hedging: healthcare occupancies are designed to protect occupants by defending in place, rather than evacuation as in a business occupancy. The same page ties that to NFPA 101-2012, the Life Safety Code, and to keeping construction, suppression and alarm features fully operable. See the LS.02.01.10 general requirements page.

Hold that next to a purchase order line reading low-position evacuation per bed, and the mismatch shows. The building’s strategy is to keep the patient inside a protected compartment. The bed’s job is to stop being the thing that injures the patient or blocks the compartment: get the platform low, get it low with no mains supply, and move the occupied bed out through the door if the compartment has to be left.

What that produces is a per-bed statement of three capabilities, each with a method, a result and a signature. A fire authority can hold a hospital to that file, to the date on it and to the interval before the next one. It cannot certify a bed as evacuation rated; that phrase describes buildings.

Three Functions That Get Confused, and What Each One Proves

A bed can pass one of the three tests below and fail the next, invisibly. Name them separately in the order, on the test sheet and in the conversation with the marshal.

Low-position parking: the deck goes to its floor, loaded

The platform travels to its lowest position with the patient’s weight on it and stays there. On the low-height beds this site builds, the sleeping surface reaches roughly 25 to 30 cm above the floor at the low end, with the full adjustment range retained above it (low-height beds for elderly and rehabilitation wards). The figure that belongs on the page is the measured one, with its conditions: mattress fitted or stripped, patient or test load, and the point on the deck the tape was taken from. Buyer-facing selection guides already treat the lowest deck height as a fall-injury control rather than a comfort setting, and check the central brake and CPR release on the same page (one example).

This proves staff can put a patient close to the floor when staying low is safer than being lifted. It does not prove anyone can get the bed there in an outage, which is a different mechanism. The failure that hides is a bed that reaches low position empty and refuses it loaded, because the actuator end stop is set short or the mattress and rails foul the frame on the way down.

Emergency lowering without mains: the drive fails, the bed still goes down

Mains isolated at the isolator, control system out of service, platform still descends. The beds here carry two layers: a backup battery that takes over the height, backrest and leg raise when the supply drops, and behind it a mechanical or battery-assisted override that works even when the control system has failed (backup battery and emergency lowering). The test proves that a power failure does not leave a patient stranded in a high position.

The failure that hides is an override demonstrated with the bed still plugged in, so the battery does the work and nobody finds that the manual path is seized. The second version of the same mistake is quoting a battery runtime as though it were a lowering test. Runtime is capacity. Lowering is a mechanism. They belong on two different lines.

Occupied lateral transfer: the bed moves with the patient on it

The bed travels with a patient on it, through the doors and the lift on the route that ward actually uses. This is the test the file omits, because it is a geometry and staffing problem rather than a bed problem. It needs published numbers: overall width at low position including rails, castor diameter and tread, whether the brake must be released to push, how the steering castor behaves when locked, the threshold height the bed crosses without a ramp, the ramp slope staff can hold, and the clear width of the narrowest door and lift car on the route.

It also needs a staffing line in the factory’s own words: how many people an occupied move takes on that route. Two, three, or not intended for that route are all usable answers. Silence is not, because the ward will assume two. The failure that hides is a move tested empty, or tested loaded down a flat corridor, or tested with the rails dropped because they fouled the door frame.

Caster seating, brake engagement and rail geometry belong to assembly and commissioning, and the gaps that matter are on the rail entrapment page. The per-bed page needs the recorded result from those procedures, not a copy of the procedure.

Why the Per-Bed File Comes Back: Symptom, Cause, Correction

The same five objections return from fire officers and accreditation surveyors, and none is really about the bed. Each is a missing noun on the page: a serial number, a method, a date, a name.

In every row the equipment was probably compliant and the record was not, so the correction is clerical work for whoever holds the equipment register, not for the service engineer. One fleet-wide template keeps the retest line to something a technician can complete without the manual.

The Page Itself, Field by Field

A page that survives review has seven blocks and fits on one side of A4.

1. Unit block. Model, serial number, hospital asset number, ward and bay, mattress in use, accessories fitted. Accessories change the height reading and the width.
2. Low-position line. Target height, measured height, load used, result, tester.
3. No-mains line. How mains was isolated, which override was used, seconds to reach the low position, result, tester.
4. Transfer line. Route tested, narrowest clear width on that route, staff used, whether the move completed, result.
5. Defect line. What failed, when the bed was withdrawn from use, who repaired it, retest result and date.
6. Signature block. Tester name and role, date, and a separate line for the fire marshal or authority representative who accepted the file.
7. Retest block. Interval, next due date, drill reference.

Set the retest interval from the failure mode, not the calendar. Rails, actuators, battery and brakes are the parts that age, and the battery fails without announcing itself, so it belongs on the maintenance round too.

Then join the two records. Put the drill reference on the bed page and the bed reference in the annual drill record, so a surveyor asking whether the function was exercised this year gets one number in two places rather than two answers. This page is one line in the documentation handover, and it closes out with the rest of the unit record before the acceptance record is signed and payment released.

Six Numbers That Have to Be on the Datasheet First

Every field on that page is downstream of a number settled at order time. Ask for these six per model and the per-bed page becomes clerical work, not a negotiation.

None of these should surface for the first time during a fire inspection. A supplier who cannot put them on one page at quotation stage is telling you something about how the per-bed record will go.

What to Send the Factory This Week

Three short asks, in one email, before delivery rather than after the compartment is signed off. First, the blank per-bed page as an editable file, in the language your ward works in, so your template and the factory’s test record describe the same three functions. Second, the six numbers above per model, on one page. Third, the factory’s own test record for one unit of the model you are buying, because a template the factory does not recognise gets filled in twice.

Send the template to the fire marshal before the equipment arrives. If he accepts the page while the beds are still in transit, the site test becomes data entry. If he sees it after the ward opens, you rebuild the file with a surveyor waiting and forty serial numbers to re-test.

Where the beds are already installed the same page still works: backfill the three lines unit by unit, start the retest clock at the backfill date, and treat a bed that cannot reach its low position under load as a finding, not a paperwork gap.