The sprinkler contractor closes the riser valve on the third floor at 9 p.m. on a Thursday. The test was supposed to take ninety minutes. By 11 p.m. the valve is still shut, the fire alarm panel shows a supervisory trouble on the same zone, and forty patients are asleep behind fire-rated doors who cannot walk out on their own.
Yes, most hospitals need a fire watch during a sprinkler or alarm outage. NFPA 99, paired with NFPA 101’s life safety provisions and the local authority having jurisdiction (AHJ), requires a documented fire watch once suppression or detection coverage drops below the facility’s approved fire safety plan for longer than the code-allowed window. The trigger is the outage itself, not the size of the building.
The reason is structural, not bureaucratic. Hospitals are built around defend-in-place: patients shelter behind fire-rated compartments instead of evacuating, because moving a ventilator patient down four flights of stairs is its own emergency. When the systems that make defend-in-place safe go dark, a trained set of eyes has to replace them until the repair is done.
NFPA 99 and the Local AHJ Govern the Trigger, Not a Single National Rule
NFPA 99 sets the baseline for fire watch and interim measures in healthcare occupancies, but the exact clock and paperwork depend on which edition your state has adopted and how your local fire marshal enforces it. Some jurisdictions give a facility a set number of hours before a documented fire watch becomes mandatory; others require notification and a watch the moment suppression or alarm coverage drops, full stop.
Don’t assume your sister facility across the state line runs the same clock you do. The code sets the framework. The AHJ decides how strictly the clock runs in your jurisdiction, and your facility’s own fire safety plan sets the specifics your staff already agreed to follow when it was approved.
NFPA 101, the Life Safety Code, works alongside NFPA 99 in most healthcare occupancies. NFPA 99 addresses the building systems — sprinklers, alarms, medical gas, emergency power. NFPA 101 addresses the life safety features those systems protect: compartmentation, egress, and the defend-in-place strategy itself. An impairment to a system covered by NFPA 99 is almost always also an impairment to the life safety strategy covered by NFPA 101, which is part of why healthcare fire watch requirements read stricter than a typical commercial building’s.
A Sprinkler Shutdown or Alarm Outage Starts the Impairment Clock
Three events reliably start a hospital impairment clock: a sprinkler system taken out of service for repair or testing, a fire alarm panel or detection zone in trouble or bypass, and a standpipe or fire pump impairment tied to construction. Any one of these removes a layer of protection the building’s fire safety plan assumed would be there.
The moment maintenance or a contractor reports the system down, someone on staff has to start the clock: when it went down, why, which zones or floors are affected, and when it’s expected back. That notification is what turns an outage into a managed impairment instead of an undocumented gap.
Partial impairments complicate the picture. A single zone going into alarm bypass on one wing is not the same event as a building-wide fire pump failure, and the fire watch plan should scale to match. Treating every impairment as building-wide either wastes money on unnecessary coverage or, worse, leads a facility to assume a partial fix covers a wider gap than it actually does. Confirm scope in writing before staffing the watch, and confirm it again before releasing it.
CMS and Accreditors Turn the Fire Watch Log Into Survey Evidence
If your facility participates in Medicare or Medicaid, the Centers for Medicare & Medicaid Services adopts fire and life safety codes as conditions of participation. An undocumented impairment is not a minor paperwork gap in that world — it’s a finding that can threaten certification. The Joint Commission and DNV layer their own interim life safety measures (ILSM) policies on top, usually with their own hour clocks for notification and watch staffing that run independently of the fire marshal’s.
That means the fire watch log serves two audiences at once: the local fire marshal, and whichever surveyor walks through six months from now asking for the impairment record from that Thursday night. A clean, timestamped log with start time, scope, notifications, and restoration verification is the difference between a routine note in the file and a citation.
Facilities that run frequent construction or renovation projects tend to build a standing impairment packet: a blank notification form, a pre-approved fire watch staffing plan by occupancy type, and a restoration sign-off sheet. Building that packet before the next impairment, rather than during one, is the single biggest factor in whether the resulting survey documentation looks organized or improvised.
The Guard’s Job Is Defend-in-Place Support, Not Evacuation
A hospital fire watch guard does not run the building’s evacuation plan. The guard’s job is early detection, immediate notification, and protecting the compartments the defend-in-place strategy depends on. Four things happen on every round:
- Patrol the impaired zones on the schedule the facility’s fire safety plan sets, not a generic interval pulled from a different occupancy type.
- Detect fire, smoke, or blocked corridors and doors — anything that would compromise a compartment if left unnoticed.
- Notify the facility’s own alarm and fire department notification chain immediately, following the hospital’s protocol rather than improvising one.
- Log every round: time, area covered, guard name, and anything found, in a format a surveyor can read without a follow-up call.
Some facilities train their own staff to stand a fire watch instead of bringing in outside guards. It’s allowed when the staff member is properly trained and freed of other duties for the shift. In practice, most hospitals use dedicated fire watch personnel, because a nurse or engineer pulled off patient-facing work to walk patrol ends up doing neither job well — and surveyors have seen that pattern before.
A guard also has clear limits. The role is not medical, not security enforcement beyond fire-related hazards, and not a substitute for the facility’s own emergency response team. A fire watch guard finds and reports; the hospital’s own procedures decide what happens next. Blurring that line during vendor selection is a common mistake — a proposal that promises the guard will “handle the emergency” is overselling a role that, correctly scoped, is about early warning and documentation.
Guard Count and Patrol Frequency Scale With the Size of the Impairment
A single impaired zone on one wing rarely needs the same coverage as a building-wide alarm outage. Three variables set the guard count and patrol interval:
Zones affected — one guard can usually patrol a single impaired zone on a reasonable rotation; multiple zones or floors typically need either more guards or a tighter patrol interval to keep rounds current everywhere.
Occupancy risk — an ICU or a locked behavioral health unit with immobile or non-ambulatory patients gets tighter patrol intervals than an administrative wing, because the consequences of a missed detection are higher.
Duration — a two-hour test window rarely needs the same staffing plan as a multi-day parts-on-order repair. Longer impairments usually mean shift relief planning, not just an initial assignment.
A basic staffing formula a facilities director can sanity-check against a proposal: number of impaired zones × patrols per hour required by the fire safety plan × hours of expected impairment = total patrol-hours needed. If a vendor’s quote comes in well under that number, ask what patrol interval they’re actually running, and ask them to show it in writing against your facility’s specific fire safety plan rather than a generic industry number.
Overnight and weekend impairments deserve extra scrutiny. A repair that starts on a Friday afternoon and can’t be finished until Monday morning changes a two-shift assignment into a multi-day one, with relief guards, meal breaks, and shift-change log continuity all needing a plan before the first patrol starts, not after.
What a Hospital Facilities Team Should Do in the First Hour
The decision that matters most happens before the first patrol ever walks a hallway. In the first hour after a sprinkler or alarm outage is confirmed:
- Notify the fire marshal or AHJ per your facility’s fire safety plan, and log the exact time.
- Notify your property insurer if the policy requires it — some carriers issue their own impairment tags separate from the fire marshal’s.
- Confirm which zones and floors are actually affected; don’t assume the outage is building-wide until verified by the contractor or your own facilities engineer.
- Start the fire watch log the moment coverage drops, even before guards arrive on-site, so the record shows continuous management of the gap from minute one.
- Set the restoration verification step now — who signs off that the system is back in service, and how that gets documented — so the watch has a defined end point instead of running indefinitely.
A hospital that treats the first hour as a checklist instead of a scramble ends up with a survey-ready file instead of a gap in the record. The facilities that struggle most with this are usually the ones that only think through the process the first time they’re already living it — by the second or third impairment, a well-run facility has a written procedure that turns a 2 a.m. call into a five-step routine instead of an improvised judgment call.
Common Mistakes Turn a Routine Impairment Into a Survey Finding
The most common failure isn’t a missed patrol. It’s a missing record. A facility calls a fire watch verbally, guards walk the halls correctly for three days, the system comes back online, and nobody wrote down when the watch started, who staffed it, or when it ended. To a surveyor, that impairment did not happen in a documented way — which, for compliance purposes, is close to the same thing as not happening at all.
A second common mistake is scope creep in the wrong direction: staffing a watch for the specific impaired zone but skipping an adjacent area that shares the same fire alarm circuit or sprinkler zone valve. Confirm the actual protected boundary with your fire protection contractor before finalizing the patrol map, not after a surveyor asks why an adjoining corridor wasn’t covered.
A third mistake is releasing the watch the moment a contractor says the system is “back on.” Restoration verification is its own step — someone with authority needs to confirm the system tested functional, not just reconnected, before the watch stands down. Ending coverage on a verbal assurance instead of a signed restoration record is the gap that shows up during the next life safety survey.
ICU and Behavioral Health Units Need Tighter Patrol Coordination
Not every unit in a hospital carries the same risk during an impairment. An intensive care unit houses patients who cannot self-evacuate under any circumstance and often depend on equipment that cannot be quickly disconnected. A locked behavioral health unit adds a different complication: doors that are locked for patient safety reasons can slow a guard’s own patrol route, and coordination with unit staff on access is not optional.
For both unit types, the patrol interval set in the facility’s fire safety plan should reflect that higher risk, and the guard assigned should be briefed specifically on the unit’s access procedures before the first round, not figuring it out during the shift. A general facility-wide patrol plan that doesn’t call out these units by name is missing the detail a surveyor is most likely to ask about.
How the Fire Watch Ends
A fire watch should have a defined end point built in from the start, not an open-ended assignment that continues until someone remembers to cancel it. Ending the watch correctly involves the same rigor as starting it:
- The system that was impaired gets tested and verified functional by a qualified technician — not just visually reconnected.
- Someone with the authority to release the watch signs off, with a timestamp, confirming the test passed.
- The fire marshal or AHJ is notified of restoration if your jurisdiction’s notification requirement includes a close-out step.
- The complete log — start time, scope, staffing, patrol records, and restoration sign-off — gets filed as one packet, not scattered across separate notes.
Facilities that build this close-out step into their standard procedure rarely have trouble producing a clean record when a surveyor asks for it months later. Facilities that treat the end of the watch as an afterthought usually can’t reconstruct exactly when coverage stopped, which reopens questions about the entire impairment.
The Fast Fire Watch Company provides firefighter-run fire watch services nationwide, with certified guards on site in under 3 hours, 24/7. If your facility is in Iowa, review the Iowa fire watch requirements page for the state-specific adoption details your AHJ enforces, or call 1-800-899-7524 to get a hospital impairment watch staffed before the next shift change.
