Review of evidence - Demographic Factors
Demographic factors
The demographic factors identified in the literature included: age, gender, ethnicity, and physical/mental ill-health (including cognitive impairment, frailty and disability). Age was the most researched factor relating to all types of risk. Age and gender were both among the top three factors that were identified within the evidence.
Age
ADF-related fatalities
The evidence identified that both old age, and young age were associated with greater likelihood of fire fatality due to an ADF. The majority of the evidence was based on assessments of coronial reports. For older age, there was no consensus as to which age an individual becomes more at risk of fire fatality, with some evidence suggesting the majority of fire fatalities were among over 85-years-old (Runefors, Johansson, and van Hees, 2017), and others suggesting that this is as low as 60-years-old and over (Heimdall Consulting Ltd, 2005; Mulvaney, Kendrick, Towner, Brussoni Hayes, et al., 2008). It is also important to acknowledge that the evidence comes from different geographical regions which may impact the social and economic support, health, and housing of the elderly, which in turn may affect the specific age at which an individual becomes more at risk. Despite this, there is consistency in that from age 60 years, the chances of an ADF resulting in fatality increases. In particular, older age tends to be overrepresented in fire fatality statistics relative to population statistics (Holborn, Nolan, and Golt, 2002). For example, Home office (2020) statistics show that over 80 years of age, ADF-related fatalities represent 16.9 per million population, and 65-70 years 8.4 per million population compared to fewer than 5 per million population for those aged under 54 years. This suggests that targeting needs to focus on older age to prevent risk to life in ADF’s.
However, age also interacts with a number of other characteristics that impacts the association with ADF-related fatality in ADF’s. For example, analysis by London Fire Brigade (2013) using Experian Mosaic (MOSAIC is a segmentation tool developed by Experian. For more information see Mosaic type look-up: understanding UK society - Experian Consumer Information (experianmarketingservices.digital) categories identified pensioners in low income, or social housing/care homes with physical and mental impairments as the group to be overrepresented within ADF-related fatality statistics. Other research has linked older age ADF-related fatalities to disability, frailty or inability to escape (Flynn, 2010; Gilbert and Butry, 2018; Heimdall Consulting Ltd, 2005), multiple impairments (Xiong, Bruck, and Ball, 2017), alcohol consumption (Levine and Radford, 1977; Sully, Walker, and Langlois, 2018), smoking (Runefors, Johasson, and van Hees, 2017), gender, in particular older men, (DCLG, 2006; Marshall, Runyan, Bangdiwala, Linzer, Sacks, and Butts, 1998) and SES or deprivation index (Holborn, Nolan, and Golt, 2002; Local Government Association, 2012).
Four sources of evidence cited that being asleep at the time of an ADF was significantly associated with ADF-related fatalities (Flynn, 2010; Greenstreet Berman Ltd, 2014; Xiong, Bruck, and Ball, 2015 and 2017). Flynn (2010) noted that this was as many as 30% of ADF-related fatalities, and that those aged over 75 years were more likely to sleep through high pitched alarms.
Similarly, very young age is also found to be associated with ADF-related fatality. In particular, the majority of evidence suggested that children under the age of 5 years to be most at risk of ADF-related fatality in ADF’s (Flynn, 2010; Hall, 2005; Marshall et al., 1998; O’Shea, 1991; Turner, Johnson, Weightman, Rodgers, Arthurs, Bailey, and Lyons, 2017) largely owing to the inability to self-rescue. However, other evidence also suggested that children under 15 years (Barillo and Goode, 1996; Heimdall Consulting Ltd, 2005) and 10/11 years (Barillo and Goode, 1996; Greenstreet Berman Ltd; Levine and Radford, 1977) were also at risk of ADF-related fatality in ADF’s. A UK report also highlighted the 16-24 age group as high risk (Department for Communities and Local Government, 2006), although this was the only evidence to identify this age group.
Fire risk, and ADF-related injury risk
As identified by Gilbert and Butry (2018), people who are likely to experience an ADF or become a casualty in an ADF are often distinct from those at risk of ADF-related fatality in an ADF. It is therefore to be expected that the age groups identified in this category do not overlap with the ADF fatalities age groups. For example, one study found that although older people were more likely to experience an ADF-related injury in an ADF, they were not the most likely age to experience a fire in the first place (Turner et al., 2017). However, the challenge with these risks is that there is much greater variance in who the evidence suggests is at risk. This is despite there being much more data available for these groups.
Risk of ADF-related injury was associated with young families (Taylor, Higgins, Lisboa, Jarman, and Hussain, 2015), and generally younger adult age groups, including a range between 20 and 59 years (Flynn, 2010; Gilbert and Butry, 2018; Hall, 2005; Mulvaney et al., 2008; Ministry of Housing and Local Government, 2017). A study that specifically investigated ADF-related burn injuries in winter (in Canada) identified under 25 and over 65-year-olds to be most at risk (Ayoub, Kosatsky, Smarigassi, Bilodeau-Bertrand, and Auger, 2017), although this represents a very specific type of injury. One single study also identified over 75 years to be associated with ADF-related injury, although it wasn’t possible to examine whether this also related to fatality at a later point (Hall, 2005).
There were fewer studies identified that considered risk of fire alone (N= 5), and so some evidence that combined fire risk, with injury and/or fatality risk were used where one type of risk could be reasonably separated from another. Risk of experiencing an ADF was associated much more with young families and younger adults (in combination with other factors such as living alone or SES) (DCLG, 2014; Local Government Association, 2012; London Fire Brigade, 2013; Ministry of Housing, Communities and Local Government; Nilson, Bonander, and Jonsson, 2015; Turner et al., 2015).
Gender
ADF-related fatalities
Men were overrepresented in all three risk areas, but particularly in relation to firefatality in an ADF, which much of the research looking at gender agreed upon (Flynn, 2010; Gilbert and Burty, 2017; Hall, 2005; Heimdall Consulting Ltd, 2005; Holborn, Nolan, and Golt, 2002; Home Office, 2019; Jonsson et al., 2017; Levine and Radford, 1977; Marshall et al., 1998; Sully, Walker, and Langlois, 2018; Turner et al., 2017; Xiong Burck and Ball, 2015; 2017).
Furthermore, Home Office (2019) data showed that men were 1.8 times more likely than women to die in an ADF, but this increased to 2.5 times more likely for the 40–64-year-old age group (also found in Sully, Walker, and Langlois, 2018). In Flynn’s (2010) research 20% of male fatal fire victims are aged 35-49, and 23% 50-64 years, showing that gender alone may not be the risk factor. Males under 5 and over 75 years of age had the highest risk of fatal injury in home fires, but for over 75 years of age, men had a lower risk of death than women over 75 years (Flynn, 2010). In fact, other research identified that male gender in combination with drug/alcohol use (Holborn, Nolan, and Golt, 2002) was a particularly high risk combination.
Fire risk, and ADF-related injury risk
Men were also identified as being more at risk of having an ADF and an ADF-related injury as a result of an ADF (Flynn, 2010; Gilbert and Burty, 2017; Hall, 2005; Levine and Radford, 1977; Turner et al., 2017; Warda, Tenenbein, and Moffatt, 1999; Xiong Burck and Ball, 2015; 2017). As with ADF-related fatality, this was often in combination with other factors. One study that specifically investigated burn injuries in winter found that women were more likely than men to experience burn injuries in an ADF (Ayoub et al., 2017), although as previously stated, the specificity of this
research may mean the findings are an outlier.
Ethnicity
There was very little evidence relating to ethnicity, and that which was found was primarily in relation to risk of fatality in an ADF. This may be due to the information that gets recorded in the IRS data at ADF incidents, and the lack of reporting on certain demographics, which may be assumed by fire-fighters rather than known for certain. Therefore, considering fatalities using coronial reports was likely more accurate.
ADF-related fatalities
Ethnicity was found to be associated with increased risk of ADF-related fatalities among non-white individuals (O’Shea, 1991; Patetta and Cole, 1990; Warda, Tenenbein, and Moffatt, 1999). One study specified that Black individuals were at higher risk of death than Whites and Hispanics, who are greater risk than Asian-Americans (Flynn, 2010). In particular, Black people aged over 65 years were found to be three times more likely to die in an ADF than other age groups, but the same pattern was only seen for those aged over 75 years in White people (Flynn, 2010). This again suggests that a combination of factors need to be considered in risk-based targeting for fire safety. Furthermore, some Black, Asian, and Ethnic Minority (BAME) groups were more likely than other ethnicities to have no working smoke alarms in the property (Greenstreet Berman Ltd, 2014). This may point to a particular demographic segment of the population that may require additional targeting to ensure working smoke alarms are fitted, and awareness of the importance of having working alarms is shared with particular communities.
Physical/Mental health and Disability
ADF-related fatalities
Impairments that mean a person is less likely to be able to escape a house fire were cited as risk factors of ADF-related fatalities (Hall, 2005; Heimdall Consulting Ltd, 2005). Some research suggested that this was primarily associated with physical disabilities, but that mental ill-health was also a high-risk factor associated with ADF-related fatality (Holborn, Nolan, and Golt 2002; Marshall, Runyan, Banddiwala, et al., 1998; Xiong, Bruck, and Ball, 2017). In a UK report, mental ill-health accounted for 15% of fatal fires, whereas physical ill-health accounted for 30%, of which most (19%) were due to limited mobility and 2% were age related (Department for Communities and Local Government 2006). Similarly, in Japan, 69% of ADF-related fatalities were associated with a health or age-related condition, particularly, being disabled, and over 65 years of age (Sekizawa, 2005). Multiple impairments may also
pose a greater risk for ADF-related fatality as an Australian study showed- in 47% of single-fatality fires the victim presented with three or four impairments compared to 28% in non-single fatality fires (Xiong, Bruck, and Ball, 2017).
Age was cited as a particular intersection with physical impairment and mobility as a risk factor. For example, Flynn (2010) noted that as age increased, disability and impairment were much more likely to feature as risks associated with ADF-related fatality. Additionally, the analysis using Experian Mosaic groups in London highlighted group M as greater risk, which included people with lower physical or mental ability, and pensioners living in social housing or care homes (London Fire Brigade, 2013). This also highlighted an intersection with socio-economic status that was also found in a Swedish study using socio-demographic data which identified that those receiving disability allowance, and/or having health-related early retirement pension were more at risk of ADF-related fatality (although the dataset included deliberate fires and not only ADF’s; Jonsson and Jaldell, 2020).
Fire risk, and ADF-related injury risk
There was much less evidence linking physical/mental health and disability to risk of fire, and none relating to ADF-related injury. Three sources of evidence from the UK suggested that people with a long-term illness, registered disability, and not working were more likely to experience an ADF (Department for Communities and Local Government, 2014; Ministry of Housing, Communities and Local Government, 2017), but that this may be a combination of factors, such as combined health, deprivation and disability score (Local Government Association, 2012).
A single piece of research was identified that related specifically to people living with dementia (Heward and Kelly, 2015), and although there is no comparison group, it identified that FRS and dementia care professionals felt that a person with dementia was at greater risk of fire due to mis-use of appliances, memory impairment, and home environment. People with dementia were much less likely to identify specific ADF-related risks in their homes, although with probing some risks were identified which did not overlap with the FRS and care professionals assessment (such as sitting too close to or falling into open and electric fires, using metal containers in microwaves, forgetting to put cigarettes out, leaving cooking unattended or forgetting to turn the gas/cooker off, and putting an electric kettle on a gas hob). This identified some potential risks, but also highlighted a clear need for more awareness and understanding of the behaviour of people with dementia in relation to fire-risk.
Summary
In summary, demographic factors interact with each other, and may overlap with other behavioural and socio-economic factors that predict the likelihood of a person experiencing a fire or ADF-related injury/fatality. The clearest picture is formed for age and ADF-related fatalities, where it may be effective to use age to target people who require home fire safety support. However, the picture for ADF-related injury and risk of experiencing an ADF is much more complex, and targeting would need to use age in combination with other factors. Furthermore, more research and/or data is needed to gain a more nuanced understanding of people at risk of having an ADF. One such example is to collect self-report data of smaller fire incidents where people have dealt with the fire themselves and not called the FRS. Additionally, new technology in smoke alarms may allow for tracking of activations which indicate near-misses which could also predict actual incidents, but it would also be important to consider how to capture information about individuals, such as demographics to support targeting.
Data availability and the type of information that can be drawn from public data is not always nuanced enough to support fire risk targeting. For example, within physical and mental health data it may not be possible to draw out specific conditions that put someone at more risk than others. As was seen in the research evidence, most mental health factors were grouped very broadly. Although it is thought that dementia probably poses a greater risk than some other cognitive or mental health impairments, there is as yet little-no evidence to specifically support this. To our knowledge, the only research on this was conducted in 2015 and relied on focus group feedback from FRS and dementia care professionals (N=16), and people living with dementia (N=8) and their family carers (N=8), and so more work is needed in this area (Heward and Kelly, 2015).