Fire Fatalities beyond the dwelling of fire origin in Scotland

2 Phase 1- Review of IRS data

2.1 Phase 1 (IRS)- Methodology

2.1.1 Selected categories for review

For the IRS review of data, the property types (level 1) under Q3.2 were limited to “building” and this was split over nine level 2 categories, namely:

1. Dwelling

2. Other Residential (institutional)

3. Non Residential Buildings

4. Road Vehicle

5. Rail Vehicle

6. Aircraft

7. Boat

8. Outdoor

9. Outdoor Structures

Of these nine level 2 categories, only “Dwellings” and “Other Residential (institutional)” were selected for scrutiny. There are a number of options for the associated level 3 sub-categories which are shown in Table 1 (for dwellings) and Table 2 (for Other Residential (institutional).

The sub-categories, identified in Tables 1 and 2 (below) were discussed amongst members of the TSG, the SFRS and the BSD on 1 February 2023. During this meeting it was agreed that BRE Global would be tasked with identifying (from the IRS) as many sub-categories of accommodation in which fatalities occurred, and then, to consider whether those fatalities occurred beyond the location of fire origin.

Whilst the intention was to identify, from the IRS dataset, those incidents where one or more fatalities had occurred in a block of flats, BRE Global was asked by the BSD to broaden the scope of the initial sift to include additional building types – see the categories and sub-categories in Tables 1 and 2 (below).

The TSG were aware of at least two fatal fires in sheltered housing in England, where fatalities occurred beyond the flat of fire origin. Accordingly, the TSG wanted the SFRS’ IRS data scrutinised to determine if similar fire deaths had occurred in sheltered housing in Scotland, over the 14-year study period.

Tenement buildings were included because the TSG were aware these were often observed to have poor compartmentation, so any constructional failures that may have been recorded at fatal fires (and particularly issues relating to cladding) would be of interest to the group. It should be remembered however, that BRE Global was reliant on the SFRS incident reports being correctly sub-classified, at the time of the records being entered onto the IRS. BRE Global was not in a position to check or challenge these entries.

Indeed, one of the reasons cited for broadening the scope of the initial sift, to include the categories and sub-categories in the tables below, was because the TSG expressed concern that sometimes fire officers entering the records on the IRS inadvertently misreport the dwelling type. In order to ensure that any information relevant to this research was not inadvertently excluded, the approach was taken to include as many dwelling types as possible that could potentially be of interest in this study. Hence, some property types, such as care homes, were also requested to be included by the TSG.

2.1.2 Data supplied by SFRS

The Home Office and SFRS were both contacted with a request to provide at least 20 years’ worth of data relating for the categories listed in Tables 1 and 2 for analysis. Both reported back that there were no electronic data available for fire statistics prior to 01/04/2009, since before this date, the data were collated on paper using the Fire Data Report system (FDR1). Since that date, the IRS was adopted by the SFRS and hence this data is available electronically for review. The Home Office reported in a private communication that, since fire is a devolved issue, they no longer report on Scottish incidents and requested that this information be obtained directly from SFRS.

After a few iterations of the required responses to specific IRS questions, SFRS provided data on 26^th April 2023, covering the 14-year period from 01/04/2009 to 24/04/2023, in two electronic worksheets. The two worksheets titled “Incidents” and “Victims” contained responses to common headings, namely:

• Incident Id
• Time Of Call
• Time Stopped Mobilising
• Property Type Full
• Ignition To Discovery
• Main Source Of Ignition
• Special Building Construction Involved (Level 2 - Buildings Only)
• Estimated Fire Size On Arrival of FRS
• Extent Of Damage
• Floors Or Decks Above Ground
• Fire Floor Or Deck Of Origin

The “Incidents” worksheet contained basic details of 57,325 fatal and non-fatal incidents. The data in the “Incidents” worksheet was used to gather information about all the incidents involving cladding – see Section 2.3 (below).

The “Victims” worksheet contained 15,814 incidents, focussing on details of injuries and fatalities. The data in the “Victims” worksheet was used to gather information about all the fire fatality/injury incidents – see Section 2.2 (below), property types see Section 2.4 (below) and the extent of damage see Section 2.5 (below). The “Victims” worksheet included the following:

• Victim Id
• IRSQ9_6 Victim Type
• IRSQ9_24 Severity Of Injury
• IRSQ9_10 Location At Fire Start
• IRSQ9_11 Where Victim Found

No detailed analysis of the data associated with property types has been performed, other than an overview of the types of dwellings in which fatal fires occurred – see Section 2.4 (below).

2.2 Phase 1 (IRS)- Analysis of total fatality/injury data

There were 57,325 fire incidents for the 14-year period from 01/04/2009 to 24/04/2023 meeting the selected sub-categories shown in Tables 1 and 2. Of this number 83.0% were recorded as “no injury” and the remaining 17.0% (9739 incidents) are shown below in Figure 1. These were the incidents resulting from the combined analysis of responses of the IRS data to Q9.6 Victim Type and Q9.24 Severity Of Injury.

The data shows that most frequently (3,690 persons) were given first aid at the scene. The next most frequent was 3,007 persons taken to hospital for a slight injury. There were 445 persons recorded as having a serious injury, requiring hospitalisation. There were 291 fatalities averaging at 20.7 per year over the last 14 years for the selected building types.

During 2009 - 2010 there were 32 fatalities and during 2022 - 2023 there were 17 fatalities.

For the 291 fire fatalities, further interrogation of the IRS data (using responses to Q9.11 Where was the Victim Found) was performed to identify where the fatalities were found. There are 11 possible responses to this question which, together with the associated numbers for each, are shown in Figure 2.

The bar chart in figure 2 shows the number of fatalities for each of 11 locations in either red or blue.

As the intention of this work is to investigate fatalities beyond the dwelling of fire origin*, those blocks highlighted in blue (166 in total) are not considered further as they are in the same dwelling or outside the building or otherwise not applicable. However, those shown in red (125) are considered further in Section 2.5 (below).

* Note that in the case of maisonettes, where fatalities could be on a different floor, but still within the dwelling of fire origin, the final three options and “Floor above origin” on the x-axis of the chart would capture these.

2.3 Phase 1 (IRS)- Analysis of cladding data

From the data supplied by SFRS, no instances were recorded where cladding was involved in a fatal fire event. It was not possible to identify specifically how many injuries resulted from fires in which the cladding played a part so these could not be reviewed further.

Cladding is not listed as a named option, in the IRS list of questions. However, BRE Global notes Question 8.8. which appears to be the logical place where any reference to cladding should have been found. Question 8.8 asks of the person recording the information on the IRS:

• If there was any rapid fire growth, how was this caused?

It is notable that, since the Grenfell Tower fire on 14 June 2017, data, in response to question 8.8, have been recorded on 112 occasions. In the six years prior to the Grenfell Tower fire, data in response to question 8.8 were recorded on 53 occasions. It is unknown whether this apparent ‘change’ is due to greater vigilance on the part of firefighters completing IRS forms, or a greater incidence of fires involving external cladding.

There is only one use of the term “cladding” within the IRS which is as one of the nine options to Q8.18 – Was there any special method of building construction involved? The responses to this question were analysed, since, if it was reported that cladding may have played a part in the fatality, the responses to this question were reviewed further during Phase 2.

Of the 57,325 fire incidents in this dataset, 191 (0.33%) were listed as having “Cladding” under “Special Building construction involved”. Of these 191 "Cladding" incidents - there were 30 victims (injuries) and 0 fatalities associated (see Table 3). Note that these injuries were not necessarily the result of the cladding, but just that cladding was present in the building construction in which those injuries occurred. From the IRS data provided it could not be determined whether cladding contributed towards any of these 30 injuries.

For the cladding related data, the extent of damage for all 191 incidents was analysed using the responses to Q8.20 – What was the extent of flame and heat damage on arrival? shown in Table 3.

The responses highlighted in bold italics represent those that are of interest as they are injuries that occurred beyond the room of origin. For this reason, these were analysed further and were compared with the data for the extent of the injury (Q9.24– What is your understanding of the cause of death / nature of injury?) which may indicate if cladding was in any way involved.

Note that the ‘Null’ entries that appear in Table 3 (and in Table 5) may be different from a ‘not known’ value. Certain questions are noted to have been responded to on the basis of earlier entries on the IRS and, in BRE Global’s opinion, Null is unhelpful in eliciting meaningful information. By way of example, ‘Null’ may mean the question was not specifically asked. ‘Null’ should not be taken to mean that there was no damage, but that there is no associated information and so is equally important. The ‘Null’ entries were investigated further.

The data for injuries to persons show that there were 18 injured persons beyond the floor of fire origin. Of these, one was classified as a serious injury, seven were classified as slight injuries (requiring a hospital check-up or treatment), six were treated at the scene and four were given a precautionary check at the scene. There is no meaningful correlation between the extent of fire and smoke spread, as was evidenced by arriving firefighters, and the reported levels of injury. Given the limited amount of data this is not examined further.

The data provided by SFRS for each Incident ID was further analysed for additional information. BRE Global concluded that it would not be normal practice for the SFRS to determine or provide an opinion on the contribution of any part of the building structure or its contents on reported injuries.

From the data provided thus far by SFRS, it can be concluded that there were no instances (during the preceding 14 years) where cladding was recorded on the IRS, by the SFRS, as being involved during a fatal fire event.

In BRE Global’s opinion it would never be reasonable or appropriate for a SFRS incident commander, entering data on the IRS, to ‘opine’ on likely cause of death at a fatal fire event. The Procurator Fiscal is responsible for considering such matters.

2.4 Phase 1 (IRS)- Analysis of property type data

Over the 14-year period, the numbers of fatalities associated with each property type was collated and is shown in Figure 3.

Over two-thirds of all fatalities occur in multiple occupancy Purpose Built Flat/Maisonettes.

The data were further interrogated to identify the likelihood of fatalities for each property type, over the 14‑year period (see Appendix A). This was done by generating the ratio of fatalities : total incidents for each property type over that same time period. These all generally sat between 2 and 4.5% with the outliers being Residential Home, Retirement/Elderly which had a ratio of 0.033% and Licensed HMO, Up to 2 storeys having a ratio of 6.7%.

BRE Global has not been able to determine any meaningful interpretation of these figures. To do so would be to speculate too widely and too much.

2.5 Phase 1 (IRS) - Analysis of extent of damage data

The Victims spreadsheet provided by SFRS (containing 15814 records over the last 14 years) was analysed to determine the extent of the fire damage and the location of where the fatalities were found. A review of these data was expected to reveal those incidents in which fires spread beyond the floor of origin and fatalities that were discovered on the floor of fire origin, but not in the room of fire origin. This would support the identification of incidents in which fatalities were found beyond the room or cabin or compartment of origin, for further analysis.

The data for responses to the IRS Q8.22 – What was the extent of flame and heat damage (at stop)? were noted and compared with the associated data of Q9.11 Where was the Victim Found. Note there were no incidents recorded for the three options to Q8.22 that involve roof spaces (External roof only, Roof space and other floors(s) and Roof space only). The remaining options from Q8.22 are shown in Table 5 together with the injury and fatality data.

As with Table 3, only the responses highlighted in bold italics are considered because they are fatalities that occurred beyond the room of fire origin and, in the case of the Null responses, these are ruled out. The injury data is provided for information but is not examined further. For fatalities, there is no correlation between the spread of the fire with the number of fatalities.

The options for responses to Q9.11 Where was the Victim Found are:

• Room, cabin or compartment of origin
• Different room, cabin or compartment on floor of origin
• Floor above origin
• Two or more floors above origin
• One floor below origin
• Two or more floors below origin
• Outside building, vehicle etc. of origin
• Seat of fire unknown or multi-seated (and above not applicable)
• Location of person unknown
• Not applicable
• Other location

As with responses to previous questions, some of the options are not of interest and entries for Q9.11: Room, cabin or compartment of origin, incidents outside the building and some that are not applicable, were excluded from further analysis.

Only those options highlighted in bold italics (in bullet list above) are of interest and fatality data were extracted for each of the Q8.22 options highlighted in bold italics (in Table 5). By comparing these two it enabled individual FI reports to be identified (column 1 in Tables 6-10) in which the victim was found beyond the room/cabin/compartment of origin and where the damage extended beyond the room of origin. This is why the number of cases shortlisted for further review in 2.5.1 to 2.5.5 are less than those shown in Table 5 (as we are only interested in a subset of the data). In total, 105 fatalities (from 99 FI reports) were identified for further review in the next stage of the project.

2.5.1 Affecting more than 2 floors (not whole building)

The fatality data associated with the option for Q8.22, Affecting more than 2 floors (not whole building) are shown in Table 6 below with all options of interest for Q9.11 Where was the Victim Found.

There were no cases of the victim being found beyond the floor of fire origin. For the above 23 incidents, it could not be confirmed whether the fatality was in a different flat on the same floor of fire origin and therefore a review of Fire Investigation reports for these fires has been performed, during the Phase 2 work.

2.5.2 Limited to 2 floors (not whole building)

The fatality data associated with the option for Q8.22, Limited to 2 floors (not whole building) are shown in Table 7 below with all options of interest for Q9.11 Where was the Victim Found.

There were no cases of the victim being found beyond the floor of fire origin. For the above 34 incidents it could not be confirmed whether the fatality was in a different flat on the same floor of origin and therefore a review of Fire Investigation reports for these fires has been performed, during the Phase 2 work.

2.5.3 Limited to floor of origin (not whole building)

The fatality data associated with the option for Q8.22, Limited to floor of origin (not whole building) are shown in Table 8 below with all options of interest for Q9.11 Where was the Victim Found.

There were two cases (in bold) of the victim being found above the floor of fire origin. Additionally for 14 incidents it could not be confirmed whether the fatality was in a different flat on the same floor of origin and therefore a review of Fire Investigation reports for these fires has been performed, during the Phase 2 work.

2.5.4 Whole building

The fatality data associated with the option for Q8.22, Whole building are shown in Table 9 below with all options of interest for Q9.11 Where was the Victim Found.

There were four cases (in bold) of the victim being found beyond the floor of fire origin. Additionally, for 9 incidents, it could not be confirmed whether the fatality was in a different flat on the same floor of origin and therefore a review of Fire Investigation reports for these fires has been performed, during the Phase 2 work.

2.5.5 Null responses

The fatality data associated with the option for Q8.22, Null are shown in Table 10 below with all options of interest for Q9.11 Where was the Victim Found. There were 19 incidents that require further review.

There was one case (in bold) of the victim being found beyond the floor of fire origin. Additionally, for 18 incidents, it could not be confirmed whether the fatality was in a different flat on the same floor of origin and therefore a review of Fire Investigation reports for these fires has been performed, during the Phase 2 work.