Review of evidence - Socio-economic factors
Socio-Economic Status (SES) was the most researched socio-economic factor within the evidence found (N=12), followed by type of property and household composition (N=10), and finally some research indicated that employment status posed an additional risk factor (N=2). However, it is important to note that these factors are likely all influenced by SES, and strongly interrelated. Therefore a summary across these categories may be a better assessment of impact than taking the research in these areas separately in stratifying the population.
Socio-Economic Status (SES)
As highlighted above, SES encompasses a number of factors relating to social standing, and in some research only a measure of SES was used, whereas others identified factors relating specifically to educational attainment or employment status and income levels, and so these were included here.
ADF-related fatalities
Data from the U.K. showed that there was a very strong correlation between deprivation and fire fatality rate at ward level (Holborn, Nolan, and Golt, 2002), as well as injury rates (Mulvaney et al., 2008). In both cases, as deprivation score increased, so did fatality and injury rates. The relationship between lower SES and greater risk of fatality has also been found in international research (Heimdall Consulting Ltd, 2005; Jordan, Squires, and Busittul, 1999; O’Shea, 1991) and appears to have remained consistent over time. A Swedish study using national security numbers linked to multiple other datasets containing detail of disposable income, educational attainment, employment, social allowance, unemployment and disability benefits (among others) found that increased risk was associated with having only primary-level education, social allowance benefits, and living in rented housing or a low-income household. In fact, the strongest predictor was found to be social allowance (Jonsson and Jaldell, 2020).
Fire risk, and ADF-related injury risk
Although some research did not specifically differentiate between ADF-related fatality or injury and risk of fire, similar patterns were observed. For example, suggesting that the combination of number of residents in a property, being unemployed, low-income, non-privately owned property, properties in poor condition, and mobile homes, were more at risk which suggests a link to SES (Turner et al., 2017; Warda, Tenenbein, and Moffatt, 1999). Additionally, burn injuries were associated with below high school education, low SES, rented housing, age of property, and property value (Lehna, Speller, Hanchette, Fahey, and Coty, 2015). In relation to risk of experiencing a fire, the Local Government Association (2012) combined health, deprivation, and disability score, which showed that pensioners aged 80 years or older, adult social care users, and tax/benefit claimants were more likely to experience a fire. Furthermore, the proportion of social renters to have had a fire within the last two years was greater than owner occupiers, and parents or couples with children and low-income households were more likely to live in a house with fire hazards (Ministry of Housing, Communities and Local Government, 2017). Lower SES groups were more likely to experience an ADF than higher SES groups (based on Acorn (Acorn classification is produced by Consolidated Analysis Centres Inc. (CACI). For more information see Acorn consumer classification (CACI) - GOV.UK (www.gov.uk)) groups; Ministry of Housing, Communities and Local Government, 2017), and a similar pattern emerged in the data using Experian Mosaic groups (London Fire Brigade, 2013).
Employment/Income
Low-income households were found to be at greater risk of experiencing a fire, and ADF-related injury or fatality (Turner et al., 2017). A study in Sweden identified socio-economic factors that predicted ADF-related fatality and found that risk was higher among those receiving unemployment benefits and/or with low household income, but risk was significantly lower for those in work and/or with higher household income (Jonsson and Jaldell, 2020).
Education
Lower levels of education were found to be associated with greater risk of ADF-related fatality (Jonsson, Bonander, Nilson, and Huss, 2017; Jonsson and Jaldell, 2020), and ADF-related injury (Lehna, Speller, Hanchette, Fahey, and Coty, 2015). However, Nilson, Bonander, and Jonsson (2015) found that higher educational attainment was associated with greater risk of experiencing an ADF, but that these fires were often less severe and did not result in injury of fatality. One explanation for this difference in risk of fire may be that higher educational attainment was correlated with income, which may suggest that these households owned more equipment (e.g., electricals) that pose a fire risk (Nilson, Bonander, and Jonsson, 2015). This highlights the need to take into consideration behaviours and attitudes towards fire safety and risk, since the people identified in this research are still experiencing fires, albeit less severe.
Household composition
Household composition primarily focused on the number of people living in a property and/or the relationship between residents, and was included in 12 pieces of evidence.
ADF-related fatalities
The majority of evidence suggested that living or being alone in the property at the time of the fire, and being a single parent was associated with ADF-related fatality (Department for Communities and Local Government, 2006; Greenstreet Berman Ltd, 2014; Marshall et al., 1998; Runefors, Johasson, and van Hees, 2017; Xiong, Bruck, and Ball, 2015). Marshall et al., (1998) suggested that this could be in as many as 41% of ADF-related fatalities. Supporting this, Xiong, Bruck, and Ball (2017) found that survival was more associated with one- or two-family dwellings- that is where there were likely to be more than one person in the household.
Living alone and being over 50 years of age was highlighted as a particular risk (Greenstreet Berman Ltd, 2014; Runefors, Johasson, and van Hees, 2017). Although the Department for Communities and Local Government (2006) further specified that ADF-related fatalities were especially associated with single parent households, males aged 46-60 who lived alone (and smoked and/or drank alcohol). This was based on U.K. data from fatal fire investigations and the British Crime Survey 2004, and so the data may be somewhat out of date, but is based on the U.K. population, making it highly relevant for the purpose of stratifying risk in the context of this report. There was also some evidence that living alone was linked to lower prevalence of having a smoke alarm (Zhang, Lee, Lee, and Clinton, 2006), which may put single occupants at greater risk since there are no other occupants to alert the presence of fire. ADF-related fatalities found to be asleep at the time of fire was also associated with lone adults or couples without children (Greenstreet Berman Ltd, 2014). It is also possible that being asleep may be impacted by other factors such as alcohol and drug (especially medication) use although this was not investigated in the current
evidence base, analysis of causes and ignition sources could be used in combination with human factors to make tentative assumptions, although that is beyond the scope of the current report.
Fire risk, and ADF-related injury risk
The evidence was less clear in relation to fire and ADF-related injury risk. Some research suggested that adults in multiple person households were more likely to experience a fire and/or ADF-related injury (London Fire Brigade, 2013; Turner et al., 2017; Xiong Bruck, and Ball, 2007). However, there was also evidence suggesting that fire and ADF-related injury risk were associated with households where more than five people resided, and where children (especially aged 6-17 years) were present (Nilson, Bonander, and Jonsson, 2015). Furthermore, couples or lone parents with and without children at home, or young families were also found to me more at risk of experiencing a fire or ADF-related injury (Greenstreet Berman Lts, 2014; Ministry of Housing, Communities, and Local Government, 2017; Taylor et al., 2015). U.K. research also identified that lone parent households were most likely to have removed a smoke alarm battery, whereas lone pensioners and managed accommodation were the least likely to have removed the battery, but smoke alarms were more likely to fail due to being near the fire (Greenstreet Berman Ltd, 2014). This makes targeting more challenging since it suggests that there are a combination of factors that impact fire and ADF-related injury risk, and that household composition is not likely the driving factor.
Type of property
Type of property refers to research (N= 10 papers) that identified rented vs. privately owned properties, or structural differences such as detached, terraced, or mobile homes and associated risks.
ADF-related fatalities
Mobile homes were identified in the evidence as being particularly high-risk properties, linked to ADF-related fatalities (Runyan et al., 1992; Turner et al., 2015; Warda, Tenenbein, and Moffatt, 1999). In particular, Patetta and Cole (1990) found that although 9% of the US population lived in mobile homes, 19% of mobile home fires resulted in fatalities, which was 2.7 times greater than single-family detached homes. Other data from the USA found similar results in that mobile homes were more than twice as likely to be associated with ADF-related fatalities, which increased to four times more likely in mobile homes with no smoke alarms (Barillo and Goode, 1996). It is important to acknowledge that this data is based on homes in the USA and is somewhat outdated, however little evidence from the UK included mobile homes as an independent category.
Data from the U.K. identified terrace houses as being slightly more likely to be associated with ADF-related fatality (Department for Communities and Local Government, 2006). Although since the dataset that this is based on does not separate mobile homes as a single category it is possible that some nuance in the data is lost, and that U.K. figures may align with the US data.
Fire risk, and ADF-related injury risk
There was a clearer pattern in the evidence relating to likelihood of experiencing an ADF. Non-privately owned, or socially rented properties, and flats/apartments were more likely to experience an ADF (Department for Communities and Local Government, 2014; London Fire Brigade, 2013; Ministry of Housing, Communities and Local Government, 2017; Turner et al., 2015). UK data identified flats and social rented properties to be overrepresented in fire statistics, meaning that despite representing a smaller proportion of the houses in the UK, they experienced a high proportion of all ADF’s (Department for Communities and Local Government, 2014). Furthermore, the Experian Mosaic groups identified in London Fire Brigade’s (2013) analysis highlighted low income/social housing or care homes and those on limited income renting small houses or flats from Local Authority or Housing Association. However, it was found that private rentals were less likely than social rentals or owned properties to have smoke alarms fitted, and longer tenancy was associated with lower likelihood of having fire safety devices in the property (Zhang, Lee, Lee, and Clinton, 2006). Properties in poorer condition or fewer security features were also highlighted as higher risk properties for experiencing a fire (Turner et al., 2015; Warda, Tenenbein, and Moffatt 1999).
Summary
There is overlap in the factors identified under the socio-economic theme, however since these factors were identified in separate pieces of evidence or investigated as distinct predictors they were discussed as such in this report. The clearest pattern of results across the evidence suggests that those from low SES groups, who live alone and in poor quality accommodation are more at risk of ADF-related fatality, although these factors are also influenced by demographic and behavioural predictors of risk as well. The pattern is similar for risk of fire or ADF-related injury, but the intersecting factors differed, suggesting that whilst socio-economic factors are important, at a more granular level these must be considered in combination with individual human factors to effectively target fire safety.
Additionally, the English Housing Survey ‘variations in housing circumstances, 2016-17’ report showed that certain demographic groups were more likely to live in particular types of accommodation or household compositions. In particular, there may be stronger associations with disability, age, and marital status and the type of housing one occupies (Ministry of Housing, Communities and Local Government, 2018). This further highlights the need to take into account a combination of factors when segmenting the population. In order to effectively link this to fire risk and fire incidents, more information is needed about population characteristics that cannot be achieved through higher level variables such as socio-economic factors relating to housing and employment.
Summary
In summary, the review of available evidence highlighted several demographic, behavioural and socio-economic risk factors that contribute to experiencing an ADF, ADF-related injury, or ADF-related fatality. There are some risk factors that clearly intersect with others, and as such it is impossible to define any singular risk factors that predict risk of fire or ADF-related injury/fatality, and as yet more research is needed to understand more clearly how risk factors interact, and whether particular combinations of factors pose a greater risk than others. What can be determined from the evidence review, is that ADF-related fatalities appear to be associated with particular factors including age, gender, ethnicity, frailty and disability; behaviours including smoking, alcohol or drug use, and maintaining smoke alarms; as well as socio-economic status, which was linked to housing and use of safety devices. These factors interact and there is some nuance in specific factors for some groups that put them at higher risk of ADF-related fatality. In relation to fire risk, and ADF-related injury risk there was some overlap with ADF-related fatality risk, but the evidence suggests that these groups appear to be distinct from ADF-related fatalities, particularly in how the predictors of risk interact. Fire and ADF-related injury risk was found to be associated with younger adults (young families), being male, physical disability, low SES, multi-person households, and non-privately owned or social housing, but these factors may overlap in different ways than for ADF-related fatalities.
Limitations
This report draws together international empirical and review evidence on factors that influence risks associated with experiencing an ADF or an ADF-related injury/fatality. It is not an exhaustive list of factors, but rather sought to examine broad themes that could be used to identify human factors that can be targeted through fire safety intervention by FRS’s in the U.K. in order to reduce ADF incidents and prevent ADF-related fatalities. The report is not, however, without limitations. Firstly, given the dearth of high quality research relating to fire safety and predictors of fire incidents a snowballing search approach was adopted. The search was therefore very broad, but unsystematic which means that other relevant research may not have been captured in this process. However, the report is intended to remain ‘live’ and be updated as new evidence and knowledge becomes available.
A second limitation that this approach gives rise to, is that the evidence included ranges from research published in the 1970’s up to present day, and is geographically dispersed across global locations. This means that some of the
findings may be outdated or less relevant to the current U.K. social, economic, and technological climate. However, where possible the evidence has been discussed in relation to up-to-date Home Office data to highlight consistencies and disparities with the relevant context for U.K. FRS’s.
Finally, the quality of the evidence found varies. In most cases of research investigating ADF-related fatalities, coronial reports and IRS (or equivalent) data are used. Both of these sources are prone to human error and/or rely on the depth and clarity of information provided by the person completing incident related paperwork. This is open to a high degree of variability and may only be as reliable as the information available at the time. However, this may be a more accurate reflection of incident information than for fire risk and ADF-related injuries whereby information
linking human factors to the incident itself vary to an even greater degree. In some cases information is either not gathered at all, or is estimated by responders at the incident, or assumptions are used to link human factors to incident details. This means that there are a number of unknowns or estimations made that may not accurately reflect the features of the incident, thus making future targeting less accurate. What this limitation identifies is a strong need for more research and greater focus on the types of data collected and used by FRS’s at incidents and in order to efficiently target intervention resources for the important need of reducing ADF incidents and protecting life.