Water mist systems in IRS data are not like-for-like
A report by Optimal Economics has concluded that Ministry of Housing, Communities and Local Government (MHCLG) incident data cannot be used to determine the comparative effectiveness of sprinklers and water mist systems in England.
Commissioned by the British Automatic Fire Sprinkler Association (BAFSA), the analysis drew on the MHCLG Incident Recording System (IRS) dataset of primary fires where a sprinkler or water-based fire suppression system was present between 2018/19 and 2023/24.
It reviewed 2,924 incidents in total.
Sprinkler systems accounted for 2,438 incidents, or 83% of the dataset.
Water mist systems accounted for 486 incidents.
Sprinkler systems are described as fixed, building-wide installations, whereas water mist covers a range of technologies including fixed automatic systems and portable or manually deployed hose-reel and lance equipment.
The dataset does not distinguish between those different water mist system types.
Where fires occurred and how systems were recorded
The analysis identified marked differences in the characteristics of incidents recorded under each system category.
Most sprinkler and water mist incidents occurred in non-residential buildings.
Industrial premises accounted for almost half of non-residential sprinkler incidents, including factories, recycling centres and warehouses.
There were 69 water mist incidents in industrial premises, mainly in food and drinks processing, recycling and factories.
In non-residential settings, 74% of water mist incidents were recorded in custodial environments, including prisons and young offenders’ institutions.
In dwellings, sprinkler incidents were concentrated in purpose-built flats or maisonettes, particularly those with more than ten storeys.
Water mist incidents in dwellings were primarily recorded in low-rise settings, with a concentration in single-occupancy bungalows or self-contained sheltered housing.
The report links these cases to the use of personal protection systems for vulnerable residents under local policies.
Differences were also recorded in the location of systems relative to the fire.
In most sprinkler incidents, the system was located in the room of origin.
In water mist incidents, the system was more often recorded as being on the same floor as the fire.
In 71% of water mist incidents where the system operated, the fire was on the same floor as the system rather than in the room of origin.
Conclusions and recommendations for future data
The report concludes that the IRS dataset does not contain two comparable populations of building-wide automatic suppression systems.
As a result, a robust comparative analysis of operational reliability and performance reliability between sprinklers and water mist is not possible.
It states: “Any claims that MHCLG incident data demonstrate superior effectiveness of water mist systems compared to sprinklers are not supported by the evidence and arise from invalid comparisons between fundamentally different types of fire suppression systems.”
The analysis calls for system-level classification, consistent standards and sufficient sample sizes to enable valid comparisons.
It recommends that future IRS data collection distinguishes fixed building-wide suppression, personal protection systems, local application protection and manually deployed custodial systems.
Without that separation, it warns that “suppression policy risks being shaped by invalid evidence”.