Case Study: Incorporating Behavioural Insights into Fire Safety

This case study looks at the incorporation of behavioural insights into fire safety, trialled by the East Sussex Fire and Rescue Service [1].

While the results of the initiative were somewhat inconclusive, there were some important lessons learned regarding the messaging and brand positioning of the fire service.

Challenge

The majority of accidental dwelling fires are caused by people, rather than faulty equipment. Incidents can be caused by acts such as smoking in bed, lighting candles, or cooking under the influence of alcohol and/or drugs.

However, during fire safety visits, efforts tend to be aimed towards looking at faulty equipment and appliances.

Alongside this, most of the time and resources are spent on those who are less likely to escape during a fire. This would include the elderly, physically impaired, or those not able to recognise the immediate danger they are in.

Over the past 20 years, numbers of deaths have halved due to these targeted tactics, as well as better data sharing amongst adult social care, GPs, and other relevant agencies.

Although deaths are down, there remains a high number of fires. East Sussex having had the 3^rd highest nationally per 10,000 population in 2017.

The Scheme

The Local Government Association (LGA) announced a project targeted at reducing accidental dwelling fires by incorporating behavioural insights into fire safety.

East Sussex was selected for the project and granted £25,000 which they matched, giving the project a budget of £50,000.

The focus was:

• Geographically limited
• Demographically limited
• On fires that don’t cause injuries
• Aimed at those not targeted by traditional fire safety messaging

With this as the guide, East Sussex Fire and Rescue Service sought out partners in the form of social enterprises and behavioural insight businesses to scope out the project.

This partnership was tasked with:

• Finding which behaviours needed to change
• What could be put in place to change these behaviours
• Sending monthly reports to the LGA
• Recommending next steps

Operating within those parameters, it was decided Brighton and Hove would be the restricted geographical area and rental group J would be the demographic.

Key features of rental group J:

• 18-35
• Private renting
• Singles and sharers
• Urban locations
• Young neighbourhoods
• High use of smartphones

The Method

Once the location and demographic was decided upon, the methodology and theory that was going to inform the initiative were decided.

The group used the principles of the transtheoretical model of behaviour change. This framework maps out the stages of changed behaviour as:

• Precontemplation
• Contemplation
• Preparation
• Action
• Maintenance

Working on that theory, the recruitment process began with an aim of 6000 volunteers. However, this almost immediately derailed the project, and the final number was closer to 800 volunteers.

Among the key tenets of the project was a desire to delve deeper into the causes of fires amongst this sub-section. There was thought to have been a reluctance of operational crews to record what they thought actually caused the fire, with those filling out the forms not wanting to be confronted with their findings in court.

This stems from the ‘fact’ that only 3% of fires within this group are caused by drug or alcohol-induced reasons according to the post-fire forms. Anecdotal evidence suggests reasons pertaining to the consumption of drugs and alcohol should be higher.

The first suggestion to reduce fires being caused by food leaving on the stove was to deter cooking after a night out, with many of the volunteers going out regularly then cooking once home.

These suggestions were communicated to the group, prompting them to buy a takeaway on the way home, or cook prior to going out and eating the pre-made meal once home.

This was rejected by Public Health England as it promoted an additionally unhealthy lifestyle. As the fire service and cross-agency partnership would conclude, it is hard to convince drunk people to eat a salad as opposed to something fried.

To try and get a broader understanding and better results, the messaging was expanded beyond the core group of graduate-aged adults, including students, who they had been targeting.

This new control group was made up of 35-44-year-olds and was used as a rough benchmark to test other messaging.  

The messaging for the two J groups, known as the intervention groups, needed to be narrative in nature and take the individuals on a journey. This narrative messaging would then hopefully take them from the pre-contemplation phase, through to action.

To do this the messaging need to disrupt normal and damaging behaviours, such as cooking under the influence of alcohol and drugs, and provide tools to stop this behaviour.

The Solution

For the intervention groups, discounts for takeaways were an appealing tactic. Coupled with an informative campaign to highlight the dangers of cooking while intoxicated, this became the approach.

For the slightly older group, more traditional messaging resonated better such as checking your appliances and a more formal tone.

The tone of voice used in the messaging became more important as the project went on. Creeping away from the traditional and formal approach, the messaging was intended to be:

• Fun
• Engaging
• Novel
• Useful
• Healthy

While maintaining the authoritative position the fire service usually is associated with, making the messaging appropriate for the life stage of the target audience was key.

The messaging was sent weekly over a 3-month period, with a randomisation process to keep a scientific method throughout. This process was then repeated and randomised again over the course of 12 months.

The feedback was that the messaging and tactics specifically deals and discounts on takeaways appealed to the younger group.

The messaging was described as friendly, informative, and fun.

Challenges and Learning

Despite the positive reception to the campaign, the intervention groups did not report a change in behaviour but did at least consider fire safety more frequently. This signalled a shift from the pre-contemplation to the contemplation phase of the framework.

The key findings were:

• Engaging different audiences requires different messaging and approaches
• Partnerships are valuable but need time and resources to be effective
• Influences and motivations are vital to consider when attempting to change behaviour
• To truly change behaviour you need to offer real value
• Change requires strong leadership and organisational support

In 2019-20 East Sussex was down to 6^th per 10,000 population on the chart of accidental dwelling fire frequency.

Roughly speaking, each of these non-injury fires costs the fire service a minimum of £500 each time. Since the scheme there have been 74 fewer incidences, meaning a saving of over £30,000.

It is hard to quantify, and difficult to tell for certain if it was the project that achieved this reduction, but there is every likelihood it helped.

In hindsight, it was more of a generalisation that young people love to drink but there is a lack of data to support this, which may have left the project working on assumptions rather than a data-driven approach.

This is partly because no data is gathered nationally on non-injury fires.

The majority of incidents are recorded as a person or persons being distracted while cooking, but no information is given on what the distraction was.

With more time and planning, there could be a better understanding of why fires occur within this group. Whilst there was the generalised assumption that drink was the cause, there was no time to test another messaging targeting other potential causes.

Some key suggestions after evaluating the project are:

• More data and information needs to be captured on a granular level on why fires happen
• The form used needs to be more detailed
• Staff need to be more inclined and reassured they will be supported when asking potentially invasive questions
• On a national level, better data sharing and consolidation could help targeted messaging

While the results were ultimately inconclusive, the project has introduced new tools and theories to the service.

East Sussex Fire and Rescue are looking to build on these improvements, with greater investment in targeted, insight-led messaging.

The service is also conducting more, in-depth surveys to get a greater amount of better quality data.

There is also greater consideration of behaviours and behavioural metrics being incorporated into methods to reduce risks, fuelled by a better understanding of how people change (or can change) when confronted with fire safety messages.

Shifting from the traditional outlets, fire services are now working with local partners, including local digital influencers to spread their more informal yet informative message.

Partnerships should be seen as conduits to these groups of people, rather than individual agencies acting in silos.

The greatest impact of the project has been a modernisation of the fire service’s approach to messaging and brand positioning. Having a brand as credible as the fire service has meant the relaxation and more engaging tone of their social media presence are resonating well with their target audience.

---

[1] Kemp, David. 2021. Incorporating Behavioural Insights into Fire Safety