Resetting the relationship between local and national government. Read our Local Government White Paper

Applying Behavioural Insights to fire safety - A trial to reduce incidents of accidental house fires

View allFire and rescue articles

East Sussex Fire and Rescue Service wanted to develop a new approach to tackling fires in the home (Accidental Dwelling Fires, or ADFs) using behavioural insights. A randomised controlled trial tested an intervention based on novel messaging to 18-34 year olds living in private rented flats and shared houses in Brighton and Hove. Though the trial sample size was not sufficiently large to reliably determine the effectiveness of the intervention, the study provided considerable learning about the potential use of behavioural insights and using targeted messaging to engage specific ‘at risk’ audiences. The Service views this as encouraging.

The challenge:

Accidental dwelling fires (known as ADFs or fires in the home) are a key area of focus for East Sussex Fire and Rescue Service.

The primary objective of the Service is to reduce injuries, save lives and property. Fires can be very distressing for those involved, with serious fires destroying homes and memories. Fires also mean that firefighters have to put themselves at risk, including travelling on emergency response on public roads and entering buildings which are alight. While these risks are managed, it is much more beneficial to reduce the need for an emergency response in the first place. The human impact of successfully reducing fires is therefore clear.

There is also naturally a financial implication for the Service and the wider community. This is not easy to quantify as each incident is different – some will do little damage, others can make homes uninhabitable.

It costs approximately £250 for a fire engine to attend an incident for an hour. A confirmed fire will always have two fire engines sent out, with more fire engines sent depending on the seriousness of the incident. 

Emergency rehousing costs are most frequently borne by local authorities, which may also be responsible for repairing the property. 39 injuries were reported in 2017/18 primary fire incidents attended by the Service – a fact which will have had an impact on the NHS and ambulances services. Additionally the Association of British Insurers reports that fire is one of the most expensive property insurance claims, with £1.3 billion being paid out to customers during 2018. 

Service targets have been set as follows.

  • 2019/20 - 475 ADFs
  • 2020/21 - 450 ADFs
  • 2021/22 - 425 ADFs

Assuming that at a minimum an ADF costs the Service £500, the achievement of this reduction would be at least £57,500 between 2016/17 and 2021/22.

However progress in reducing the number of these fires in the home has plateaued over recent years and the Service recognised that a new approach was needed.

The Service’s current Home Safety Visit programme has traditionally targeted vulnerable people (e.g. people with reduced mobility, vision or hearing) to provide face to face advice in their own home. While invaluable, home visits are resource-heavy and the Service was keen to reach a larger audience that goes beyond the focus for its traditional interventions.

East Sussex Fire and Rescue Service secured match-funding from the LGA’s behavioural insights programme to test a new approach to reducing incidents of accidental house fires and ‘near misses’ within the city of Brighton and Hove. The Service commissioned Social Engine to support the design, delivery and testing of an intervention intended to reduce the instances of accidental kitchen fires by using behavioural insights.

The solution:

Working together, we designed and delivered a randomised controlled trial to test the effectiveness of interventions on the behaviour of those most ‘at risk’ in order to reduce instances of accidental house fires. We adopted a co-design process to the project, seeking to combine the Service’s own insight and experience with evidence from behavioural science to collaboratively design our interventions and plans to test them. Through the co-design process enthusiasm emerged for the project to focus on a less traditional vulnerable group that had been identified as being ‘at risk’.

In order to have the greatest impact possible in reducing incidents, we were keen to identify those groups in the population which:

  1. Were responsible for a disproportionately high number of incidents relative to their population size
  2. Accounted for the greatest proportion of ADFs
  3. Were larger groups within the overall population.

Our initial analysis identified a population segment – defined as ‘Rental Hubs’ in the Mosaic public sector consumer classification system – that were one of the largest groupings in Brighton and Hove population (14%); accounted for a significant proportion of ADFs (23%) and were responsible for a disproportionately high number of ADFs (1.7 times higher than their population size would suggest). This group therefore became the focus for our intervention.

Based on our review of service data and research evidence, we therefore focused on this group of younger people (aged 18-35) renting flats in urban neighbourhoods, who either live on their own or share with friends. This group, whilst having relatively modest incomes, were felt likely to prioritise socialising and going out.

A common scenario (behaviours) that was identified through the co-design process and the evidence review was of our target audience going out for the evening and returning home after having been drinking alcohol thereby being at risk of causing a fire when trying to cook whilst intoxicated.

Specific behaviours we sought to encourage through our intervention:

  1. Buy / eat some food on the way home
  2. Not cooking having drunk alcohol
  3. Preparing food in advance, ready to eat without further cooking when returning home.

Our behavioural goal was therefore to stop people cooking food at home after having drunk alcohol.

Our intervention involved sending novel messages to engage our target audience which were noticeably different from traditional fire safety messages (whilst retaining the credible, authoritative identity that fire services enjoy). They included light-hearted photos, animated gifs and video clips which intend to gently emphasise our underlying intervention message – ‘don’t cook yourself, enjoy a takeaway instead’.

Fireman smiling

Messages were based on the Transtheoretical approach to behaviour change - supporting people on a journey from pre-contemplation to contemplation, preparation and action. They were intended to encourage people to consider and adopt alternatives to cooking after a night out.

The relatively small number of ADFs within a fairly large population meant that in order to reliably evaluate the impact of our intervention a large sample size for the trial was required. Our initial calculations suggested a sample of approximately 6,000 participants was required – though lower than anticipated levels of attrition allowed us to subsequently revise this downwards to 774 participants.

In order to secure the large number of people required to deliver the trial, the first phase of the project involved a public facing campaign, inviting people to join ‘the Brighton Crew’ from which eligible participants could then be randomly assigned to control and intervention groups in order to conduct our trial.

Copy of #Welcometofiretribe


To engage our target audience, our messaging – in keeping with the tone of our campaign overall – aimed to be fun, appealing, novel and useful. It was important to avoid being seen as preaching or slipping into expected behaviour or messages that people might associate with the Fire Service.

There were two main strands to our messaging; the first sought to incentivise our desired behaviour – making it tempting to eat pre-prepared or take away food after a night out. The second strand sought to subtly make people aware of the risks of cooking after drinking, whilst doing so in a manner consistent with our desired tone/approach. At all points we were mindful of the need to avoid the perception that we were recommending unhealthy foods. The Service approached a local public health expert who reflected that positioning was key i.e. that anything with excessive fat or salt should be regarded as a treat not a regular part of diets. 

The impact (including cost savings/income generated if applicable):         

At the conclusion of the trial we compared instances of ADFs within our treatment group with the instances within the control group to determine whether our intervention had been effective. Qualitative research – telephone interviews and an online survey – was also conducted to capture behavioural and attitudinal information on participants’ experience of the intervention and views of fire safety more generally.

Due to the challenges of securing trial participants in sufficient numbers (the final sample size was 212 people) to meet our sample size calculations, we found no incidents of ADFs within either our intervention or control groups. As such it has not been possible to accurately determine the effectiveness of our intervention.

Having said this, in the qualitative research, participants in our intervention group within the 18-34 target age range did report higher awareness of fire safety than was reported in the control group. While 38% of respondents aged 18-34 in our intervention group said they felt their awareness of fire safety had improved, only 22% of those in our control groups felt their awareness had increased.

We also found that almost three quarters of the treatment cohort reported thinking about fire safety ‘regularly’ (72%), whilst only a quarter of the control cohort did. However participants did not report any conscious modification of their behaviour as a result of the messaging – though some acknowledged (unprompted) that it may have had an unconscious effect. In general, the intervention group interviewees did recall receiving messages about avoiding cooking after a night out and around half praised the content and style of the reminders.

This finding suggests that whilst the treatment group may not have consciously taken steps to alter their behaviour in response to the fire safety messaging, they were significantly more likely to have thought about fire safety more regularly. However, without a baseline measure of how concerned participants were about fire safety before the trial, it is not possible to confidently determine the extent to which this can be attributed to our intervention. Nonetheless this finding suggests that the intervention may have moved our intervention group from ‘pre-contemplation’ to ‘contemplation’ within the transtheoretical behaviour change process. 

Despite being unable to determine from the trial data whether the intervention was effective, the project has introduced new theories, tools and techniques that can be applied to the Service’s working practice.

How is the new approach being sustained?:

The Fire Service now plan to build on the experience to apply the learning from the project beyond solely focussing on ADFs and 18-35 year olds, to service improvement:

  • Investment in targeted, insight-led messaging aimed at engaging particular non-traditional audiences. The data are available and insights are being developed further with work such as a post-incident survey.
  • Greater consideration of behaviours – both current and desired – and behavioural metrics to incorporate into plans aimed at reducing risks among a wide range of audiences. This could include considering how people’s attitudes and behaviours change as a result of fire safety activity which the Service undertakes – as well as measuring changes in the frequency of fires and other incidents.
  • Working through local partners – and the untapped opportunity to engage local digital influencers – as an effective way for the Service to engage different audiences.
  • Partnerships should be seen as conduits to wider/key audiences rather than simply as individual organisations themselves. Developing this approach would build on the ‘on your side’ idea and enhance Service’s existing work in the community.
  • A shift in approach to digital communications and opportunities to further develop the Service’s social media presence and use as an engagement tool. This shift in positioning – firmly placing the Service as a friend, not just an emergency response - has the potential to engage wider audiences in a way that resonates with their perceptions of personal risk and in a way that is relevant to their lives.

Our findings suggest that it would be worth seeking to evolve practice so that a more nuanced or segmented approach to fire prevention and safety messaging is embraced by firefighters as well as leaders and support staff. Firefighters hold a position of authority amongst those in target audiences and we suggest that increasing their understanding of the lives and lifestyles of non-traditional audiences – such as by using focus/discussion groups or creating opportunities to listen (not just tell people about safety) will increase awareness and understanding of the different approaches needed.

As such the project’s long term impact will be felt in how the experience is built on to develop policy and practice which extends beyond the initial trial.

After considering a number of options, the Service plans to continue with the project, targeting two or three key partners to get more people signed up in the city of Brighton and Hove. There will be less of an emphasis on deals, and more on the monthly prize draws and local events taking place which people might want to get involved with. This will continue for a period of 6 months when another review will take place. It is hoped that the project will then be rolled out to other areas.

Lessons learned:

Despite the absence of robust quantitative trial results to determine whether or not our intervention was effective, there has been considerable learning from the project both for the Service and other local authorities and fire services.

Focusing on non-traditional ‘at risk’ audiences

The majority of the Service’s work is focused on traditional vulnerable audiences and using well-established fire safety approaches to reducing incidents. However different audiences – such as the younger flat-sharers targeted through the project – are also vulnerable to the risks of fires and are therefore of interest to the Service.

The Service’s decision to undertake a behavioural insights trial looking at this “new” audience demonstrates its commitment to approaching the issue of ADFs in new and untested ways. The Service had also recently set up an ADF Working Group to assess the potential for using data in different ways to inform Service improvement, particularly in identifying specific audiences to target with tailored engagement and communications.

The project’s emphasis on understanding why people behave the way they do is important in order to develop appropriate interventions and approaches that are effective at encouraging behaviour change. The current approach draws heavily on segmentation frameworks such as MOSAIC that are useful in identifying population sub-groups that share common characteristics, but also have significant limitations. The absence of behavioural factors and influences can unwittingly group people with very different motivations into single cohorts based on demographic commonalities.

This project is part of wider efforts to give greater consideration of behaviours – both current and desired – and behavioural metrics to incorporate into plans aimed at reducing risks among a wide range of audiences and how their attitudes and behaviours change as a result of planned fire safety activity.

The value of partnership working

The ambition to recruit 6,000 trial participants was made in good faith on both sides and a belief that this could be achieved through local partnerships with major employers and local influencers. In reality recruitment was far harder to achieve, despite considerable effort. Nonetheless it is worth reflecting that when large not-for-profit employers, in particular Brighton & Hove City Council, did actively engage with and support the project there was a big uplift in participant recruitment.

The Service has strong relationships with many key local organisations; however, these are often focused on supporting activities aimed at the traditional target audience of older generations. We feel there are opportunities to build strategic partnerships that go beyond this and the current regulatory and enforcement role to extend to broader themes and priorities.

Working with businesses

There is limited evidence that local businesses are willing to engage with the Service if asked. However, as there is potentially mutual benefit to such associations with commercial enterprises (as opposed to the third sector and public services), the Service may wish to invest in developing these partnerships – it could prove to be efficient and cost-effective in the long-term. This is an area which the Service has already expressed an interest in following up.

Making social media more ‘social’

At the start of the trial, the use of social media by ESFRS was somewhat traditional and tended to adopt ‘broadcasting’ to traditional audiences with conventional messages of fire safety. The change in tone and increased emphasis on interaction, engagement and content curation required something of a cultural shift and upskilling of the team. Capacity was also an issue in being able to dedicate time to this engagement. ­  

Encouraged by the project, the Service has been shifting towards a more informal approach to social media, with a noticeable increase in the use of animated gifs and a more relaxed and informal tone on the Service twitter account that extends beyond tweets relating to the project. There remain opportunities to further develop the Service’s social media presence and use, particularly across platforms such as Instagram.

Brand positioning

The Service, as a blue light public service, has a strong and credible brand and reputation that might be expected to be formal and serious – often summarised as “friendly and authoritative”. As such Social Engine’s suggestions to adopt a more light-hearted and fun approach was at first felt to go against that identity, credibility and tone of voice.

That being said, the Service agreed that it should employ such an approach, particularly in seeking to engage less traditional audiences for fire safety messages.

Copy of firefighter promotional prints

The ideas and activity undertaken as part of the project have the potential to form part of a bigger brand positioning for the Service and therefore enhanced by other activities and communications, rather than something that stands alone as a specific initiative. The Service currently undertakes a wide range of preventative work that largely goes unseen – such as fire safety audits on venues and events - and we suggest that this legitimate ‘keeping things safe for you’ builds on the Service’s main stated purpose of Making Communities Safer and is a useful brand position and platform for the future.

There is the opportunity to retain the position of authority whilst building awareness of the fire service as ‘your ally in keeping safe’ – this could in turn promote their often unseen role in inspecting /licensing premises to operate safely, but also in providing more youth-centred advice on behaviours likely to reduce ADFs. The Service already currently actively promotes activities such as Pride and comments on events like the Lewes annual fireworks and this could be developed further in order that an ongoing relationship can be embedded.

Culture change

The project has benefitted from having a number of key advocates within the Service who have shown considerable enthusiasm, commitment and a desire to grasp the opportunity to apply behavioural insights within the Service. There has also been good support for the project from senior leaders which was evident from the outset. It seems likely that the traditional approach adopted by the Service – targeting vulnerable groups and using well established fire safety messages – is deeply engrained in the organisation’s culture and social norms.  This project has allowed the Service to explore a new way of doing things, and was deliberately chosen by the Service to help enable cultural change.

The project team focused on helping frontline and communications team members of staff in particular understand the proposed approach and the inherent value of applying behavioural insights to encourage behaviour change or to focus on non-traditional audiences. On reflection, this project has been a milestone in beginning discussions about the benefits of behavioural insights. Since it began, a number of key members of staff have attended conferences and events to learn more.

We found some reluctance to change how and what the Service communicates, and a tendency to revert to known and default ways of working (even within the context of seeking to engage a different audience).

Ultimately this has been of great benefit to the Service as it has demonstrated the value of innovation and experimenting with new approaches, giving staff more freedom to step out of their comfort zone and the traditional way of doing things.  It has led to discussions at senior levels about community engagement and how to build more effective networks. The Communications team is also seeking to resource new technology include SMS capability to use for marketing purposes.

Contact: [email protected]