Fire and Rescue Statistics, Scotland, 2013-14

6. Notes on statistics used in this publication

6.1 Background

6.1.1 The statistics in this publication are compiled from reports submitted to the Department of Communities and Local Government (DCLG) on fires attended by the Scottish Fire and Rescue Service. Although a variety of Scottish fire statistics are published annually by DCLG within Fire Statistics Great Britain (FSGB), this publication is intended to provide a detailed overview of Scottish fire and rescue statistics from the past decade at Scotland level and for the most recent year at Local Authority level. Prior to 1^st April 2009, detailed information was provided on all fires in buildings, vehicles and outdoor structures and any fires involving casualties on the 'fire damage report' form FDR1 (94). Since 1^st April 2009, this information was provided via the Incident Recording System (IRS). Before IRS, only aggregate data on 'secondary' and 'chimney' fires was collected on the FDR3 form. As a result, analysis of these types of fire prior to IRS is limited.

6.1.2. Analysis in this publication is based on provisional data for 2013-14 which was extracted from IRS by DCLG in May 2014.

6.1.3 On 1^st April 2013, the Scottish Fire and Rescue Service was established as the national fire and rescue service for Scotland, combining the eight predecessor fire and rescue services and the Scottish Fire Services College. For service delivery purposes, the SFRS is organised into three Service Delivery Areas (SDAs); East, North and West. Each SDA has a number of Local Senior Officers (LSOs) responsible for managing the resources within their area to provide engagement with the local authority, and to deliver response and community safety strategies. There are 17 LSO areas across the three SDAs, each containing one or more of the 32 local authority areas. Table 29 in the workbook published alongside this bulletin provides details of the Local Authorities in each LSO and SDA area, as well as former FRS areas for reference.

6.2 Changes to this publication

6.2.1 Following the creation of the Scottish Fire and Rescue Service on 1^st April 2013, a public user consultation was undertaken by the Scottish Government. The purpose of the consultation was to collect views from users of Scottish fire and rescue statistics on the redesign of Scottish Government fire and rescue statistics publications to account for the organisational change brought about by fire service reform. Based on the results of this consultation and engagement between SFRS and Local Authorities, SG agreed to produce statistics at Local Authority level. In this way, the statistics can be aggregated to LSO or SDA level or, for comparison purposes, to former Fire and Rescue Service (FRS) level. Table 29 in the accompanying dataset to this publication shows how the 32 Local Authority areas in Scotland map to the Local Senior Officer and Service Delivery areas of the SFRS, as well as former FRS areas.

The statistics in the 2013-14 publication cover the first year following the establishment of the SFRS. As such, this is the first year that this publication includes statistics at Local Authority level, in place of former FRS level breakdowns.

Responses to the consultation also indicated that users were not clear of the distinction between this publication, formerly Fire Statistics Scotland, and the other annual statistical publication on fire and rescue in Scotland, the former Fire and Rescue Service Statistics Scotland. In order to clarify, the publications were renamed in consultation with users. This publication is now named Fire and Rescue Statistics Scotland, covering statistics on all incidents attended

by the Scottish Fire and Rescue Service (fires, special service incidents, false alarms and casualties). The second annual statistical publication has been renamed Scottish Fire and Rescue Service Fire Safety and Organisational Statistics, covering statistics on Community Fire Safety, SFRS Workforce and Stations and Appliances.

6.2.2 Prior to 2009-10, DCLG (and predecessor government departments with responsibility for fire statistics) extracted data on all fires involving but only a sample of data from other fires. Other than for fires involving casualties, the detailed analysis of data for this period is based on the sampled data weighted to agreed FRS totals.

In April 2009, Scotland's former FRSs switched from the paper-based forms FDR1 and FDR3 to a new electronic recording system - the Incident Recording System (IRS). The benefit of this new system is that the data collection process is more streamlined and allows the statistics to be calculated on census data rather than using sampled data. It broadens the data collected to include all incidents attended by the SFRS, rather than just fire incidents. The change in collection caused discontinuities to occur in some of the longer-term time series. These include: non-fatal casualties, subcategories of other buildings, outdoor primary fires, secondary fires, smoke alarms and spread of fire, please see section 6.5.2 and 6.5.3 below.

6.2.3 Since 2008-09, data in the publication has been produced in financial years rather than calendar years. The 10 year trend data has been recalculated to this new time span and comparisons to the previous Fire Statistics Scotland series should not be made as the time periods are not aligned. This important change will allow us to compare the fire statistics at GB level and aligns this publication with the rest of Justice Statistical Bulletins.

6.2.4 Additional datasets are provided online and will be updated after the release of this publication. These include long-term and monthly trend data at Scotland level. The long-term data covers at least the last ten years and the monthly trend data is for 2009-10 onwards.

Local Authority level data for each year since 2009-10 will be produced in 2015, on the website address above. An email notification will be sent to all Scotstat users when this becomes available.

Datazone level fire data is provided on the Scottish Neighbourhood Statistics (SNS) website for three key indicators:

Number of all fires;

Accidental dwelling fires per 100,000 population; and

Deliberate fires (excluding chimney fires) per 100,000 population.

The SNS website will be updated to include provisional figures for 2013-14 and revised figures for 2011-12 and 2012-13 following the publication of this main report.

Trend data at former Scottish FRS level up to and including 2012-13 will remain available but will not be updated following the creation of the single Scottish Fire and Rescue Service in place of the 8 former FRSs on 1^st April 2013.

Additional datasets include revisions to key statistics as set out in our revision policy in section 6.3.

6.3 Revision Policy

6.3.1 Since the introduction of IRS, this report has been published using the most current year of data in provisional form. This allows timely data to be provided, which is then finalised in future publications. Prior to IRS data, finalising the dataset took up to two years, mainly due to investigations into fire casualties. A consultation of our data users in 2011-12 established that they felt that the benefits of a more timely publication on provisional data outweighed the provisional status of the data^[12]. For this reason, provisional data is published here and will be revised in line with our revision policy (this is explained further in section 6.3.3). There can be notable revisions in casualty figures but revisions to incident numbers are typically negligible at Scotland level (see Table B).

6.3.2 The data for 2011-12 and 2012-13 in this publication is revised as of May 2014. IRS is a live system where fire and rescue services can amend previously recorded information e.g. fire casualties. The data for 2013-14 is provisional and will be subject to future revisions. Data for 2012-13 will be revised for the last time and finalised in the 2014-15 version of this publication. The dataset for 2011-12 is now finalised and will not be subject to any further revisions other than in exceptional circumstances.

In this publication, provisional data in tables and charts are annotated with a ^p. Where the data has been revised since last published, an ^r is used to annotate tables and charts. All 2013-14 figures referred to in the body of the text are provisional, though for ease of reading this will not be stated at each mention.

6.3.3. Revisions to 2012-13 data

The change between provisional figures published last year for incidents and casualties at Scotland level and revised figures published this year was small. Table B provides further detail of revisions to these figures.

Table B: Changes due to revision of 2012-13 data

	Difference between first published and revised1
	Number	Percentage
Primary fire	23	0.2%
Secondary fires	80	0.6%
False alarms	246	0.5%
Chimney fires	3	0.2%
Special service incidents	28	0.3%
Fatal casualties (in fires)	0	0.0%
Non-fatal casualties (in fires)	8	0.6%

1 - In each case above, the revised figure is larger than figure first published, excluding fatal casualties where no revision was necessary

6.3.4 Revisions to 2011-12 data

Last year's edition of this publication reported a provisional figure of 60 fire fatalities in 2011-12. This figure has been revised to 59 in this publication and is now finalised. Typically there are a number of incidents where fire investigations, including toxicology reports, are on-going at the time of publication. The results of these explain why there are sometimes changes to the provisional number of fire fatalities. There has been no revision to the number of fatal casualties in 2012-13 as reported in last year's publication (46).

6.3.4 Other revisions may still be made and will be handled according to the Scottish Government's 'Corporate Policy Statement on Revisions and Corrections' with the following clarifications:

Data to be revised	When revisions are to be made and where
Scotland figures	Revisions to 2011-12 and 2012-13 figures appear in this publication (2013-14)
	Final revised figures for 2011-12 (barring exceptional circumstances) appear in this publication (2013-14)
	Final revised figures for 2012-13 (barring exceptional circumstances) to appear in 2014-15 publication (next year's publication)
	Revisions to 2013-14 figures appear in next year's publication (2014-15), to be finalised in 2015-16 publication
Local Authority figures	Revisions to Local Authority level statistics will be published in future publications and in the additional datasets on the same basis as above.
Former FRS Level	Last revisions made at time of 2012-13 publication

The revision policy for Scotland allows for data to be provisional for longer than DCLG and Welsh Assembly Government (WAG). This is because Fire and Rescue Statistics Scotland is published on an annual basis, unlike the other nations' publications.

6.3.5 During the production of the 2013-14 publication it came to light that figures for England and Wales for the number of non-fatal casualties excluding precautionary checks, published in Table 10b in previous versions of this bulletin, were incorrect. Table 10b provides a comparison of the number of non-fatal casualties excluding precautionary checks and the rate per million population for England, Scotland and Wales. The numbers and rates for England and Wales in previous publications incorrectly excluded casualties who received first aid at the scene as well as those who were recommended a precautionary check. The affected data is for England and Wales only and for the years 2009-10 to 2012-13 inclusive. Corrected figures are shown in Table C.

Table C: Non-fatal casualties excluding precautionary checks per million population, before and after correction, Great Britain, 2009-10 to 2013-14

	2009-10	2010-11r	2011-12r	2012-13p
England (before correction)	79.6	83.1	80.9	71.7
England (after correction)	139.5	142.3	138.9	122.8
Wales (before correction)	82.3	91.1	78.3	72.9
Wales (after correction)	155.0	150.2	134.8	132.7
Scotland (no correction needed)	182.7	208.9	207.4	186.9

The result of the corrections is that the rate of non-fatal casualties excluding precautionary checks for Scotland is still higher than that for England and Wales, but to a lesser extent. Chart 30 shows the changes in more detail.

Erratum will be published to correct older versions of this publication in light of these findings.

Chart 30 - Rate of non-fatal casualties excluding precautionary checks per million population, Great Britain, 2009-10 to 2012-13 (before and after corrections)

Notes

1- 2013-14 data not affected

6.3.6 Mid-year population estimates are used within this publication to calculate population rates for Scotland compared with other countries and for Local Authorities. Population data are derived from relevant mid-year population estimates prepared by the National Records of Scotland (NRS).

Following the 2011 census, the mid-year population estimates for 2004 to 2011 were revised by NRS. This is the first year that it has been possible to use the revised mid-year population estimates to calculate population rates within this bulletin. Population rate figures for 2004-05 to 2011-12 included in this publication have been calculated using the revised figures, meaning there are some differences with those based on unrevised population figures in previous versions of this publication.

6.3.7 There was an exceptional revision of data in 2009-10, where Highlands and Islands FRS was investigating a discrepancy between its local management information system and IRS. The investigation showed that revisions to categories of fires were required. The investigation was completed in time for the 2011-12 publication and the data within "Additional datasets - trends" has been updated to reflect the revisions.

6.4 Disclosure

6.4.1 The disclosure policy for Scottish fire data has been updated and is published alongside the additional datasets. It covers geographies for Scotland, FRS level and Local Authority level data.

6.5 Accuracy of Statistics

6.5.1 Industrial action by FRSs across the UK took place in November 2002 and January and February 2003. No information was recorded for the 15 days when the industrial action took place. Only the long-term trend data in the additional datasets to this publication are affected by this.

6.5.2 Following the introduction of the new IRS in 2009-10, DCLG identified the following two areas of potential discontinuity arising from the switchover from the old largely paper-based FDR system to the new IRS questions (i.e. between 2008-09 and 2009-10).

i. The first area relates to increases in the numbers of certain incident types within the data of a handful of Fire and Rescue Services, notably in numbers of primary outdoor fires (Table 1: Road Vehicles and Other Primary Fires). These are not believed to have been real increases, but rather the result of a small proportion of incidents in the past having been incorrectly reported as being 'secondary fires' rather than 'primary fires'. The following conclusions can be drawn:

• it appears that these differences follow from incorrect reporting under the old FDR system
• the effect on national totals appears to be minimal

ii. The second area was a discontinuity in the number of non-fatal casualties. For Scotland in particular, the changes in recording as a result of the introduction of the IRS have affected the trend data for non-fatal casualties. It is recommended that data prior to 2009-10 should not be compared to that from 2009-10 onwards. The changes follow from two improvements to the way in which non-fatal casualties have been recorded since the introduction of the IRS:

• The first change is that each casualty or fatality can be marked in IRS as 'not fire-related'. Around 10 per cent of non-fatal casualties in the UK were marked as such in 2009-10. In fire incidents however, almost all non-fatal casualties can be expected to be 'fire-related', since very few would have occurred if there had not been a fire. As a result of these concerns, non-fatal casualties marked 'not fire-related' were not excluded from non-fatal fire casualty figures. It is worth noting that excluding them would have introduced an even larger discontinuity when comparing data from before and after the introduction of the new IRS.

• The other issue arises since the IRS collects details of the injury of non-fatal casualties in two questions. The first asks for the nature of the injury and the second, follow-up question categorises the severity of the injury (the nature of treatment received) as either 'precautionary check recommended', 'first aid at scene', 'hospital (slight injuries)' or 'hospital (severe injuries)'. Whereas precautionary checks were an answer option for the first question in FDR1, IRS requires the FRS to state the nature of injury before they can enter a precautionary check as an answer for the second question. The effect was a reduction in the number of precautionary checks selected as a response between the two recording systems.

Before 2009-10, precautionary checks^[13] accounted for about 50 per cent of all non-fatal casualties. This reduced following the introduction of IRS.

IRS also introduced a new recording category under 'Severity of injury' which is 'First aid given at scene'. Chart 31 illustrates that when precautionary checks are removed from the ten year data the trend increases slightly in 2009-10. This is due to the inclusion of the new first aid category, following which an increase was expected. The data including precautionary checks shows a distinct drop and should not be compared with data before 2009-10.

As non-fatal casualties are an important statistic for community fire safety policies, it was decided not to exclude the previous seven years' worth of data but to mark the tables with a break in trend. Also, to help with comparability, Table 2 now also includes the subset "Non-fatal casualties in primary fires, excluding precautionary check-ups".

Chart 31- Non-fatal casualties from primary fires including/excluding precautionary checks, Scotland 2003-04 to 2013-14^p

6.5.3 Further quality assurance performed by the Scottish Government has highlighted other discontinuities in trend data following the introduction of IRS (2009-10 onwards):

i. 'Other building' categories prior to 2009-10 were related to industry classifications. When IRS was created more relevant categories were developed, although 10 year trend has been provided where possible(if IRS and FDR1 definitions are the same).

ii. Outdoor primary fires, except road vehicles, had 27 main category types when recorded in FDR1. Within IRS this increased to 42 types. As a result of increased accuracy in the recording of this information and the introduction of the new categories in IRS not all trend data could be aligned. The new categories in Table 9 better support environmental (e.g. woodland and grassland fires) and anti-social behaviour (e.g. deliberate fires) policies.

iii. Secondary fires were previously recorded on an FDR3 form, where there were 6 sub-categories for these types of fires. Information was totalled monthly and recorded on this form. Within IRS there are over 100-sub categories of secondary fires. These have been grouped to align with previous categories where possible but some of the trends have not continued, in particular for refuse fires. Before IRS the total number of refuse fires always exceeded 18,000 fires. Now, with the 4 specific 'Refuse' fire recording categories in IRS, this has reduced to 12,000 fires. It is believed that the IRS is providing wider and more accurate recording of secondary fires and that some fires that would previously have been categorised as 'Refuse' fires are now being more accurately recorded in another category. Overall secondary fires are following the expected trends, but there has been a redistribution of what was previously recorded on FDR3 forms as refuse fires.

iv. On FDR1, the question of whether there was a smoke alarm present offered a 'Yes' or 'No' response, whereas on IRS this option has been expanded to include 'Don't know'. The introduction of this new category has affected the trend data. It was felt important that the previous trend data be displayed, as this information is valuable in relation to community fire safety policies. It is advised that smoke alarm data for 2009-10 is not compared with 2008-09 and earlier.

v. FDR1 allowed only one smoke alarm per incident to be recorded, whereas IRS allows for the recording of multiple smoke alarms at a fire. In order to maintain continuity, if more than 1 smoke alarm has been recorded at a single incident within IRS then the following hierarchy has been applied to the smoke alarm operation:

1. Present, operated and raised the alarm

2. Present, operated but didn't raise alarm

3. Present but didn't operate

So an alarm which operated and raised the alarm 'outranks' one which operated but didn't raise the alarm and so on.

vi. Before the IRS the type of damage caused by a fire was determined using a grid where the percentage of damage caused by each of the 4 causes (fire, heat, smoke and other) was entered. If smoke and/or heat were the only categories marked, the damage from the fire was considered as 'Heat and/or smoke damage only' and the question on spread of fire was not applicable. IRS now asks 'was there heat and/or smoke damage only?' as a specific question. If 'Yes' is answered to this question, the spread of fire is not applicable. Prior to 2009-10 heat and/or smoke damage only fires never accounted for more than 6 per cent of all dwelling fires. With the introduction of IRS, this jumped to 45 per cent in 2009-10. It is felt that further investigation is needed and it may be that another variable in IRS may be used in the future to estimate size of damage, including smoke and/or heat damage.

vii. In IRS, the motive of fire is reported as 'accidental' or 'deliberate'. The cause of fire also has 'deliberate' as a possible option. In FDR1 the 'deliberate' motive was the same as the 'deliberate' cause. This is not the same in IRS. In the IRS data, there is a small percentage of building fires where the motive and cause are different. The cause has been reported as 'Other' or 'Unspecified' although the motive was 'deliberate'. To maintain trend data, when the motive of fire is 'deliberate', the motive over-rides the cause of fire.

6.5.4 The databases before 2009-10 contain details of all fires with casualties, but only a sample of other fires. The data were all weighted to agreed Fire and Rescue Services' totals. The detailed analysis of data other than for casualties, or fires involving casualties was based on the sampled data weighted to agreed Fire and Rescue Service totals. Data in the tables from before 2009-10 may not summate to the total shown, due to the rounding of the sample data within sub-categories.

6.5.5 In April 2012, there was an upgrade to the IRS. At this time changes were made to some of the lower sub-categories for location of fires and special service incidents.

The details of these changes can be found at the Department of Communities and Local Government website.

In table 6, a number of the new special service categories were combined to align with previous categories and enable comparisons with previous years. These categories were:

i. 'Medical incident co-responder' and 'Medical incident first responder' have been combined to form the original category 'Medical incident co-responder/first responder'

ii. 'Removal of objects from people' and 'Removal of people from objects' (new categories) have been combined to form the original category 'Removal of objects'.

iii. In tables 9, 9a, 19 and 19a the 'Outdoor' location categories have been aligned with previous categories. For example, 'Wheelie bins' is now a separate recording category within IRS, but has been included under 'Refuse - small/rubbish container' to preserve the trend data.

Other changes resulting from the IRS amendments were:

iv. In table 6b and 6c, a number of new sub-categories have been introduced for RTCs and Flooding. Data for 2011-12 has been provided where it could be aligned with these new categories.

6.6 Additional Information

6.6.1 Links to fire statistics for Great Britain, England, Wales and Northern Ireland are provided below.

DCLG - Great Britain

https://www.gov.uk/government/collections/fire-statistics-great-britain

DCLG - England

https://www.gov.uk/government/collections/fire-statistics-monitor

Welsh Assembly Government (WAG) - Wales

http://wales.gov.uk/statistics-and-research/fire-statistics/?lang=en

Northern Ireland Fire and Rescue Services - Northern Ireland

http://www.nifrs.org/statistics/

6.6.2 Summary information on the quality of these statistics and the potential use of the fire statistics can be found at: http://www.scotland.gov.uk/Topics/Statistics/Browse/Crime-Justice/DataSourcesFire

6.6.3 In 2013-14, a public user consultation on fire and rescue statistics in Scotland was undertaken. The report has now been published and the findings used to shape the future of statistical publications on fire and rescue statistics in Scotland.

If you wish to comment on the contents of this publication, please contact us on justicestatistics@scotland.gsi.gov.uk

6.6.4 If you have an interest in fire statistics, please sign up to the Scotstat register. The register allows users and providers of Scottish Statistics to contact each other and consult on specific topics of interest. Registered users can consult on new collections, seek views on changes to existing collections and receive notifications for new or forthcoming publications. To register an interest in fire, please tick this topic under the heading 'Justice'.

6.7 Forthcoming changes

6.7.1 DCLG are in the process of reviewing IRS at the time of this publication. Any forthcoming changes from this review will be passed on to users of this publication through ScotStat and the Scottish Government website

6.8 UK Statistics Authority - Assessment Report

The United Kingdom Statistics Authority has designated these statistics as National Statistics, in accordance with the Statistics and Registration Service Act 2007 and signifying compliance with the Code of Practice for Official Statistics.

Designation can be broadly interpreted to mean that the statistics:

• meet identified user needs
• are well explained and readily accessible;
• are produced according to sound methods; and
• are managed impartially and objectively in the public interest.

Once statistics have been designated as National Statistics it is a statutory requirement that the Code of Practice shall continue to be observed.

The assessment report was published in June 2012.

6.9 Symbols

6.9.1 Symbols used in the tables are:

- Nil or less than half the final digit shown
^pProvisional
^r Revised
* Trend data not comparable to current information
** Data not disclosed