Improved public messaging for evacuation and shelter in place.

FEMA have just published on the internet a very interesting paper entitled “Improving Public Messaging for Evacuation and Shelter‐in‐Place Findings and Recommendations for Emergency Managers from Peer-Reviewed Research” (April 2021) (link here). It reports the findings of a comprehensive literature review on the factors that affect the level of compliance with advice on personal protective actions (Shelter in place and Evacuation) in the event of a storm, flood or wild-fire. While these situations are not entirely analogous to radiation emergencies and there may be differences in the behaviour of UK and US populations when faced with an external event, there may be some important messages for UK planners to be gleaned from this work.

The work is related to the Protective Action Decision Model proposed in the literature (see figure) which attempts to identify the cues people may be sensitive to, the level and nature of prior consideration, and the on-the-day perception of the threat, possible responses and how others are responding which may influence decision making.

The key advice to those planning systems to warn and inform the public is to understand the potential impediments to action and take steps to address these barriers in advance, provide consistent advice through multiple trusted channels and to provide frequent updates.

Among many observations, those that seemed most relevant to the UK nuclear industry include:

• Individuals find environmental cues such as sights, sounds or smells that indicate an impending threat an aid to decision making. This puts the nuclear industry at a disadvantage because we cannot show pictures of storm clouds, fast flowing rivers about to burst their banks or raging forest fires. On the other hand, we have the dread many people feel about radiation helping to focus minds.
• It was noted that, in response to a wildfire, individuals could be categorised into three broad groups:
  • • Wait and see (the largest group);
    • Stay and defend;
    • Likely to evacuate.

• Seeing neighbours evacuate or other leave was a predictor of increased evacuation across a number of hazard types. This is a well-documented response to an alarm and most people will have observed it where they have been in a building when the fire alarm is tested; many people look to others and copy their behaviour.

• Receiving messages from family and friends in addition to the local authority influenced decision making. Many people will seek confirmation of the preferred path of action from their social circles before acting.
• Local governments and businesses provide important social cues that can impact on risk perception. Advice to evacuate an area, or even to shelter in place, could be undermined if council employees continued activities such as cutting grass and collecting waste in those areas.
• Receiving several consistent warning messages from multiple, credible and trusted sources increases the rates of compliance.
• People tend to use social media as a complementary rather than their primary source of information. Social media was also often used to amplify or share information with others.
• As mobile phone ownership is now more prevalent than home landlines, public alert and warning calls to landline phone numbers are becoming less effective. The increased reliance on mobile phones may also result in bandwidth congestion during an incident. In the UK where fraction of homes and offices with land lines is falling while the procession of mobile phones is increasing. (See Table A45 in ONS report here which implies that 82% of households have land lines and 90% have mobile phones).
• Households with multiple vehicles evacuated in multiple vehicles often with staggered leaving times. This is in the context of an impending storm but is plausible in a radiation accident if only as a mechanism for protecting their vehicles. This would add to traffic congestion.
• There is strong agreement across studies and hazards that women are more likely to take appropriate protective action (SIP or evacuate) than men.
• Parents with children in the household tended to have more difficulties with making the decision to stay or to leave for hurricanes and flooding. While some of their concerns may be similar to those of other households (e.g., traffic congestion, fuel availability, uncertainty regarding destination, cost), children in the household, especially younger children and larger numbers of children, raised the anxiety level and increased logistical challenges, which caused delays in decision making. Again the dread of nuclear may balance these concerns.
• Having a pet, especially where there is a strong attachment to the pet, decreased the likelihood of evacuation. Many studies highlighted concerns about shelters accepting pets, the added cost of evacuating with pets and the logistics of having a pet at a shelter as impediments to evacuation.
• Adults who have dementia or other cognitive disabilities and a caregiver(s) who would evacuate with them have evacuation rates that are the same as, or lower than, others. Caregivers were concerned with the potential for those in their care to be exposed to stigma and lack of privacy in a shelter. They were also concerned that unfamiliar settings would exacerbate their symptoms. Family and friends (the social network) tended to play an important role in determining whether to evacuate or not.
• Adults with dementia and their caregivers who did go to shelters experienced a range of difficulties, including increased agitation, emotional distress and disorientation. It was challenging for caregivers to provide normal levels of care and comfort in this environment.
• Care facilities and their caregivers were challenged in making the decision whether to evacuate or not, given their sense of responsibility to their residents. This research also indicated the importance of care facility residents and their families deciding (and documenting) who would care for them in a disaster (e.g., whether or not they would evacuate to a family’s residence) and then not changing that decision as the threat neared.
• Having a household plan increased the likelihood of taking the appropriate Shelter in Place protective action for a tornado. This may be presumed to apply for any threat, underlining the importance of prior information that encourages preparation.
• Studies found that individuals grapple with many concerns when deciding to evacuate. According to these studies, the following concerns delayed or negatively influenced the decision:
  • • Traffic congestion and the availability of fuel;
    • The ability and cost of evacuating with pets;
    • Costs of evacuation, including travel costs;
    • Potential issues around the legal status of undocumented immigrants;
    • Individuals faced with a public shelter as their primary destination had more reluctance to evacuate. Their concerns include crowding with strangers and being located farther away from social networks.

On the basis of the observations a number of recommendations were made:

• Use websites and social media platforms and work with local media to provide authoritative, time-stamped, geo-tagged photos and videos of hazards such as rising waters and wildfires. Encourage individuals to share those visuals with friends and family, including via social media. Again, there are differences between these events and radiation emergencies to take into account but there is something to take away from this recommendation.
• Warning messages should be clear, consistent and strong but not overly dramatic. Mandatory evacuation orders had more weight than voluntary ones and also carried increased media coverage.
• Changing the geographic areas subject to advice can cause confusion and a resulting drop in compliance. This should be minimised where practical.
• Messages that clearly described the probable personal impact of the hazard helped individuals realise that they would be personally impacted which motivated protection action.
• Visuals such as maps and photos improve message comprehension and support decision making.
• Authority figures acting as role models and being seen to comply is helpful.
• Tourists who sought information from tourist offices rather than hotel staff were more likely to evacuate (this was in the context of a major storm brewing).
• The current event should be compared to those that have posed similar threats. Hopefully nuclear industry will ever have a good back catalogue.
• In the preparation stage relationships should be built with television and radio forecasters and other journalists likely to cover the story should it arise.
• People should be encouraged to sign up to relevant alert and news feeds, including during the event.
• There should be a mechanism in place to follow and monitor the social media of authoritative sources to keep information consistent and address inconsistencies and inaccuracies if they occur. In the UK we also try to coordinate the media lines taken before media releases are issued.
• Communications strategies should be tailored to gender differences. For example, given that women are more likely than men to take protective actions, messaging on preparedness should consider the use of outreach channels geared toward women.
• Include individuals with disabilities, access and functional needs, and associated advocacy organizations in developing and reviewing community plans for evacuation.
• If an evacuation may be called for then consider breaking the news at a time that allows the travel to be completed during daylight hours.
• When issuing evacuation orders, explain the risks that led to the decision to evacuate some zones and why other zones are not evacuating.
• Provide information about public shelters, including items associated with comfort (e.g., availability of power, air conditioning, rest rooms, and space for families and pets) as well as services for individuals with disabilities and access and functional needs.

The slide library available here is a very good way to assimilate the information given in this report.