Satellite imagery: Esri, Maxar, Earthstar Geographics
Fire safety
Smoke alarms that were fitted, and the fires they never saw
The Home Office's detailed analysis of fires in England for the year to March 2026 found a smoke alarm fitted but failing to operate in 19% of dwelling fires. In two thirds of those failures the alarm worked. The fire was somewhere the detector could not see. In other buildings, 46% of fires had no smoke alarm at all.
A smoke alarm on the ceiling is easy to tick off. It is there, the test button beeps, the box on the checklist gets a mark. The latest national fire figures suggest that is not the question that matters. The question is whether the alarm is in a place where it will notice the fire.
What the figures show
The Home Office published its detailed analysis of fires attended by fire and rescue services in England, year ending March 2026, on 19 August 2026. Fire and rescue services attended 26,180 dwelling fires in that year, up 2.8% on the year before, and 13,679 fires in other buildings, up 3.7%.
In dwelling fires, a smoke alarm was present and raised the alarm in 48% of cases. No alarm was present in 23%. In 19% of dwelling fires an alarm was fitted and did not operate, virtually unchanged on the previous year.
The detail is in why they did not operate. Two reasons, fire products not reaching the detector and the fire being in an area not covered by the system, together accounted for 66% of failures. Missing and defective batteries, the reasons most people would guess, were far smaller: a missing battery in 1.8% of fires and a defective one in 3.1%, though both were more common where someone was hurt.
So in most failures the alarm was not broken. It was in the wrong room, or behind a closed door, or too far from where the fire started.
In other buildings the picture is starker. No smoke alarm was present in 46% of other building fires, up from 44% the year before, and those fires accounted for 13 of the 17 fire-related deaths in other buildings.
Why coverage drifts
Detection is usually designed once, when a building is fitted out or converted, and then left. The building does not stay the same.
A flat gets a kitchen moved into the living room. An office floor is split into meeting rooms with full-height partitions. A storeroom becomes a staff room with a kettle and a microwave. A loft gets converted. Each change can put a room with a real fire risk outside the reach of a detector that was sited correctly for the old layout.
For homes, BS 5839-6 sets out grades and categories of system, from alarms on the escape route only through to detection throughout. The category that was right for a flat as built may not be right for the flat as it is lived in now. For other buildings, the Regulatory Reform (Fire Safety) Order 2005 requires appropriate fire detection and warning, and BS 5839-1 is the usual reference. In both cases the fire risk assessment is where the question should be asked again.
In purpose-built blocks of flats with a stay put strategy, communal areas are often deliberately not fitted with a common alarm, so the detection that matters most is inside each flat. That makes the flat's own coverage the thing to check.
What to check
On the next fire risk assessment review, or the next walk round, look at detection against the current layout rather than the original drawings.
Check every room where a fire could plausibly start, particularly kitchens, rooms with heaters or charging, and any room that has changed use. Ask whether a detector would see a fire there with the door shut.
Check that the type of detector suits the room. A heat detector in a kitchen, not an optical smoke alarm that gets disabled after the third burnt slice of toast.
Check for alarms that have been covered, bagged during decorating and never unbagged, or removed because they kept going off.
In rented homes in England, the Smoke and Carbon Monoxide Alarm (England) Regulations 2015, as amended in 2022, require at least one smoke alarm on every storey with a room used as living accommodation, checked working at the start of each tenancy and repaired or replaced once a fault is reported. That is the legal floor. It does not tell you whether the alarm on the landing will hear a fire in a back bedroom.
Why the record matters
After a fire, the questions are predictable. What detection was fitted, where, to what standard, when it was last tested, and whether anyone had flagged that the layout had changed. If the answers sit in an installer's certificate from years ago and a test log on a clipboard, they will not show that anyone looked at coverage since.
Where SAMRISK puts it
SAMRISK holds the building and the land it sits on across thirty connected systems, twenty for the building and ten for the land.
Detectors, alarm panels and their test records sit in the fire safety register against each building, alongside the fire risk assessment and its actions. Layout changes and detector positions are marked on the building plans, so the drawing matches the building as it is used now. Testing and servicing go through maintenance, and the review dates sit on the compliance calendar with a name against each one.
When the layout changes, the question of whether the detection still covers it has somewhere to be asked, and somewhere for the answer to be kept.
Sources: Home Office, Detailed analysis of fires and response times to fires attended by fire and rescue services, England, year ending March 2026, published 19 August 2026; Regulatory Reform (Fire Safety) Order 2005; Smoke and Carbon Monoxide Alarm (England) Regulations 2015 and (Amendment) Regulations 2022; BS 5839-1 and BS 5839-6.
