Risk Communication Planning

Publication Date: 2022

Executive Summary

Overview

This study aims to enhance understanding of how individuals experiencing homelessness or who are precariously housed are considered in emergency and risk communication planning, including how extreme weather risk communication is both delivered to and received by those experiencing homelessness. We identify strategies used to disseminate risk information, as well as challenges and barriers to the implementation of these strategies. This report highlights recommendations put forward by those with lived experience with homelessness to improve emergency and risk communication planning as well as directions for future research.

Research Questions

Four overarching research questions guide this research:

1. How do/have emergency management plans for extreme weather account for the unique risk communication needs of precariously housed individuals?
2. How do emergency management agencies engage and work with organizations serving people experiencing homelessness and people experiencing homelessness toward inclusive, equitable, and culturally appropriate emergency and risk communication planning?
3. What strategies do emergency managers, planners, and organizations serving people experiencing homelessness use to disseminate extreme weather risk information to homeless communities?
4. What are the barriers and facilitators to the equitable implementation of these plans and strategies?

Research Design

To answer our research questions, we conducted in-depth semi-structured interviews with 29 individuals representing organizations serving people experiencing homelessness (n=10), emergency managers and city/county level employees involved in emergency planning for unhoused individuals (n=6), and those with lived experience with homelessness (n=13).

Findings

Preliminary findings identify: (1) communication strategies used by organizations serving people experiencing homelessness and local/county-level agencies involved in emergency planning for unhoused communities to disseminate extreme weather risk, including word-of-mouth and “phone-tree” style communication to relevant individuals and entities; (2) challenges associated with disseminating and receiving extreme weather risk information, such as difficulties communicating with individuals who have limited or no access to a cell phone and providing actionable information and viable options for protective actions to people experiencing homelessness; and (3) recommendations for improving risk communication and emergency planning for individuals experiencing homelessness, including increased inter-organizational coordination, additional sheltering options, and varied modes of risk communication.

Practical Implications

Access to housing or stable shelter supports the adoption of several protective actions from extreme weather. While temporary strategies (e.g., opening emergency shelters, distributing tents and supplies) for reducing risk among those experiencing homelessness are necessary given the current circumstances of housing precarity in the United States, any “solution” short of access to safe and affordable housing will not suffice in reducing disaster vulnerability for these individuals and communities.

In the meantime, there is a dire need for plans to consider ways that individuals experiencing homelessness can reasonably act upon risk information in terms of both the modes of communication and content of communication. Emergency plans should be enhanced and made more equitable by inviting and engaging people with lived experience with homelessness to the planning “table.” Their insights, and those of other historically marginalized communities, should inform emergency plans and strategies so that planning can be made both inclusive and equitable. At the same time, participation in planning must also be appropriate in “meeting people where they are” in order to understand their needs, capacities, and ideas for improving risk and emergency planning. This means identifying multiple modes of participation, considerations for language and access issues, as well as who approaches and how they approach these individuals for participation. Social service organizations, including organizations serving people experiencing homelessness, must also be central collaborators and contributors to emergency plans given their roles in service provision and sheltering (among others) and their potential to mediate or facilitate engagement with people with lived experience with homelessness in emergency planning.

Introduction

People experiencing homelessness and housing precarity are made uniquely vulnerable to the threat of disaster as a result of structural processes, such as systemic marginalization and sociopolitical exclusion, a lack of affordable or available housing, criminalization, and unequal to no access to healthcare (Gaillard et al., 2019¹; Settembrino, 2017²; Vickery, 2019³; Wexler & Smith, 2015⁴). While prior research has demonstrated the need to tailor risk communication strategies and community engagement approaches to reach at-risk populations, critical work remains to determine how (if at all) at-risk communities, particularly those who are unhoused or precariously housed, are engaged and incorporated into emergency and risk communication planning and how they receive risk information.

The current study aims to address this gap in knowledge by assessing current extreme weather risk communication strategies from official sources to people who are unhoused, understanding challenges communicating and receiving risk information, and gathering recommendations from key stakeholders for improving both the modes and content of extreme weather risk communication to individuals experiencing homelessness. Using Boulder and Denver County, Colorado as a context in which to explore these topics, we present preliminary findings from in-depth interviews with people experiencing homelessness, representatives from organizations serving people experiencing homelessness, emergency managers, and local/county-level government employees involved in emergency planning efforts for unhoused communities. Importantly, and as elaborated on later, we engaged (and continue to engage) study advisors with lived experience with homelessness throughout the research process, including interpretation of results and dissemination strategies.

Literature Review

Social vulnerability scholarship has identified a number of factors and processes that create and proliferate homelessness, which must be considered in any work concerning homeless communities. People experiencing homelessness lack access to material or monetary resources, are socially ostracized and stigmatized in society, and experience co-occurring challenges in terms of physical and mental health and wellbeing (Every et al., 2019⁵; Kidd et al., 2021⁶). Based on findings within risk communication and public health literatures, we know that certain individuals and communities, such as people who are unhoused, often face unique challenges when accessing, digesting, and acting upon risk information (Teo et al., 2018⁷). These challenges are the result of structural inequality and intersecting systems of racial and economic oppression that are reflected in programs, policies, and practices (e.g., risk communication strategies) that systematically exclude or limit the equitable provision of services and housing and that stymie effective risk communication.

Inaccessible risk information, or information that is not appropriate for individuals without fixed, regular, or adequate shelter, can have serious negative implications. Exposure to the elements—such as high winds, extreme heat or cold, precipitation, or poor air quality (to name a few examples)—compounds existing health conditions and can result in negative morbidity and mortality outcomes (Every et al., 2019; Kidd et al., 2020; Zhang et al., 2019⁸). Further, individuals experiencing homelessness are diverse in their experiences, capacities, and identities. Failure to account for these differences further marginalizes these individuals through crude homogenization of their existence (Goodling, 2020⁹; Lazrus et al., 2012¹⁰; Vickery, 2018¹¹). This is why it is essential to move beyond language that merely identifies homeless populations as those individuals who need specialized assistance, or “access and functional needs” (Centers for Disease Control, 2021¹²) and instead foster a meaningful, two-way dialogue with people experiencing homelessness in order to create more just and effective emergency operations and risk communication plans.

Past work on risk communication has increasingly identified the need to target and package warning and risk communication in ways that are cognizant of the unique barriers and considerations that socially marginalized communities face (e.g., Campbell, 2018¹³; Trainor & Subbio, 2014¹⁴; Donovan et al., 2019¹⁵). Prior research has found that targeted warning information (Burger & Gochfeld, 2020¹⁶), acknowledgement of barriers to accessing information and the ability to act on such information (Eisenman et al., 2007¹⁷), meaningful engagement with and participation of socially marginalized communities in the risk communication planning process (Donovan et al. 2019; Lee & Chen, 2019¹⁸; MacIntyre et al., 2019¹⁹; Meredith et al., 2009²⁰; Teo et al., 2019²¹), and the need for risk information to be communicated from a trusted, reliable source (Galarce & Viswanath, 2012²²) have all been identified as critical strategies and precursors to effective risk communication. Before an individual can take protective action, however, they must understand the following: “what is the risk?” “how likely is it to affect me?” “can I trust the source?” and “am I able to take the recommended actions?” From past research we know that effective risk communication in the form of warning messages include five key components that must be considered critically: (1) the source of the warning; (2) identification of the hazard; (3) specific areas at risk; (4) protective actions people can take; and (5) how to access additional information (Bostrom et al., 2018²³). Yet, limited literature exists that highlights or evaluates how these risk communication principles are integrated into risk communication practice with marginalized communities such as those experiencing homelessness, including their involvement in planning, how they receive information, and if they are able to act upon such information.

Despite the progress that has been made in highlighting opportunities for engagement and inclusion of socially marginalized and access and functional needs populations into emergency and risk communication planning (e.g., Campbell, 2018; Donovan et al., 2019; Trainor & Subbio, 2014), less is known about the extent to which emergency managers incorporate the perspectives of marginalized communities into planning and how individuals made vulnerable to the threat of disaster want to be engaged in the planning process. The current study works to fill this gap in understanding. Importantly, we draw upon and extend existing work that highlights models for engagement (e.g., Gin et al., 2022²⁴; Klaiman et al., 2010²⁵; and U.S. Department of Housing and Urban Development, 2014²⁶) by incorporating a culturally responsive, participatory research approach (Garcia et al., 2014²⁷; Janes, 2016²⁸; Hood, et al., 2015²⁹).

Research Design

Research Questions

Four overarching research questions guide this research:

1. How do/have emergency management plans for extreme weather account for the unique risk communication needs of precariously housed individuals?
2. How do emergency management agencies engage and work with organizations serving people experiencing homelessness and people experiencing homelessness toward inclusive, equitable, and culturally appropriate emergency and risk communication planning?
3. What strategies do emergency managers, planners, and organizations serving people experiencing homelessness use to disseminate extreme weather risk information to homeless communities?
4. What are the barriers and facilitators to the equitable implementation of these plans and strategies?

Study Site Description

This research is geographically focused within Denver and Boulder Counties, Colorado given the prevalence of homelessness within this region, the hazard profile of the area, the contentious history of the area’s approaches to addressing homelessness, the research team’s pre-existing connections and partnerships with key stakeholders within the homeless-serving sector and emergency management, as well as extreme weather events that occurred just prior to and during the study period from 2020-2021 (e.g., poor air quality as a result of wildland fires, extreme temperatures, and winter weather).

Data, Methods, and Procedures

To answer our research questions, we reviewed state and county-level emergency and risk communication plans (Boulder and Denver Counties, Colorado) and conducted in-depth semi-structured interviews with 29 individuals across three groups. Inclusion criteria for interview participation included people with lived experience with homelessness, representatives from organizations serving people experiencing homelessness, and emergency management professionals or officials at city and county levels of governance. Exclusion criteria included individuals who did not speak English, due to our research team’s language capabilities, and those who were under 18 years of age. In terms of participants with lived experience with homelessness, we kept this bound to individuals who typically reside in Boulder and Denver counties. Interview data collection occurred from May 25th, 2021 through December 13th, 2021.

All interviews with organization representatives and emergency management professionals or public officials occurred via Zoom, while all interviews with individuals experiencing homelessness at the time of the interview occurred in-person at food and clothes distribution locations in Denver and Boulder as well as a sheltering location in Boulder. Importantly we followed all local and University of Washington COVID-19 protocols for conducting fieldwork. Interviews with people experiencing homelessness lasted between 6 minutes and 35 minutes (18 minutes on average), whereas for the other two groups (organization and emergency management professionals/public officials) interviews lasted between 29 and 59 minutes (44 minutes on average). Before beginning interview data collection, the study protocol received exempt status from the University of Washington’s Human Subjects Division on February 16th, 2021 (Institutional Review Board (IRB) ID: STUDY00012577). Prior to the start of each interview, we obtained verbal informed consent from all participants as well as permission to record interviews.

Participatory Design and Study Advisors

As part of our study approach, we drew upon existing models for participatory research design (Garcia et al., 2014; Vaughn & Jacquez, 2020³⁰). Specifically, we referenced previous studies and frameworks that highlight strategies for meaningfully engaging relevant stakeholders throughout the research process (e.g., research question formulation, co-creation of study instruments, co-development of sampling frames and locations, and interpretation of findings). In fact, this study was motivated by conversations with staff at organizations serving people experiencing homelessness and emergency management practitioners, two of whom are co-authors in this report (Sweeney and Wendlandt). Before engaging in data collection, we worked with local partners (including Sweeney and Wendlandt), to begin communication with people with lived experience with homelessness who would be willing to serve in a paid advisory capacity for the construction of interview questions, interpretation of findings, dissemination strategies, and the development of a framework for appropriately engaging precariously housed individuals into emergency management and risk communication planning. Outreach to potential advisors was conducted via email and by phone. The advisory group began with four individuals who had all experienced homelessness in the past and resided or worked in Boulder County. One individual withdrew their participation as an advisor mid-summer 2021. During this time, we integrated another advisor from the Denver area. At all points during the 2021 study period, we had four advisors affiliated with the study. See Appendix 1 for an overview of the recruitment flier and payment structure shared with advisors. For those advisors who gave permission to be acknowledged, we include them in our acknowledgements section following the names of research team members.

In addition to our advisory team, we also met with multiple people informally throughout the study period, who represent each of these groups, regarding questions and topics at the focus of this study. These conversations helped us to focus the types of questions asked in interviews and provided additional context to the study.

Data Analysis

All but one interview was audio recorded and transcribed. All transcriptions were checked for accuracy and consistency with the recordings. The one interview that was not audio recorded was conducted outside during a rainfall event, prohibiting the use of audio recorders. Instead, detailed notes were taken and used in the analysis. As part of the formal coding process, we determined preliminary findings using a structured approach in which we summarized interviews and synthesized key themes across interviews and by interview questions. Questions centered around how risk communication is both disseminated to and received by those experiencing homelessness, risk communication challenges across interviewee groups, and recommendations for improving emergency and risk communication planning for unhoused communities. Three publicly available state and county-level (Denver and Boulder) emergency operations plans were reviewed in their entirety; however, we are in the process of gathering additional plans and annexes not readily available online for review and formal content analysis.

Dissemination of Findings

Findings from this analysis will be shared in an accessible way to all individuals at the focus of this study, including those with lived experience of homelessness, organizations serving people experiencing homelessness, and local/county officials responsible for emergency planning for precariously housed communities. We will ask these individuals and groups the most appropriate ways for communicating findings and ensure that our dissemination strategies are designed accordingly (e.g., presenting through in-person or virtual meetings, shareable Google docs). Importantly, findings will continue to be shared and discussed with local partners and advisors to integrate their interpretation of key themes into broader dissemination materials and to identify future research directions. Informal dissemination and discussion of high-level preliminary findings among research and advisory team members are currently underway.

Ethical Considerations, Researcher Positionality, Reciprocity, and Other Considerations

Throughout the development and implementation of this study, members of the research team made the ethical conduct of research, including considerations of positionality, cultural responsiveness, and reciprocity, central in the study design. Drawing upon key ethical standards put forward by the 2020 CONVERGE Working Group on Homelessness, Housing Precarity, and Disasters (CONVERGE, 2020³¹), we began the development of this study by first engaging with key stakeholders who work closely with precariously housed individuals and communities (e.g., co-authors Sweeney and Wendlandt) about the appropriateness and necessity of the current study. From these discussions and our intention to develop a participatory research design, we invited individuals with lived experience with homelessness to participate in a paid advisory group to support the research design (e.g., instrument development, sampling, data collection) and interpretation of findings. We will continue to work with the advisory group for interpretation of additional analyses of interview and content analysis data and to ensure that dissemination strategies are appropriate and accessible.

Preliminary Findings

Interviewees represented (1) people with lived experience with homelessness (n=13); (2) representatives from organizations serving people experiencing homelessness (n=10); and (3) emergency managers and city/county level employees involved in emergency planning for unhoused individuals (n=6). Some interviewees representing organizations had previously experienced homelessness or housing precarity. In order to maintain anonymity, and in line with our informed consent agreement, we do not disclose the exact titles or organizations/agencies of individuals interviewed. However, the types of agencies represented outside of emergency management include city and county entities directly engaged in work pertaining to homelessness, human services, and emergency planning. Organizations serving people experiencing homelessness provide varying services (e.g., day or nighttime shelter, food provision, housing assistance, and advocacy) to particular groups of individuals living unhoused (e.g., single adults, families and youth). Eighteen interviewees represented Boulder County, nine represented Denver County, and two were county-level employees operating within the broader Denver metro area but outside of Denver County.

High-level findings from the interview data are summarized in three sections: (1) strategies used by local and county agencies and organizations to share risk information with unhoused or precariously housed members of the community; (2) challenges and barriers to providing and receiving risk communication for extreme weather; and (3) recommendations given by people with lived experience with homelessness for how to improve risk communication and emergency planning for unhoused communities. In the first two sections, responses are broken out by interviewee group.

Communication Strategies

Of the 10 representatives we interviewed from organizations serving people experiencing homelessness, all reported sharing risk information for extreme weather to their clients and/or people living unhoused in Boulder and Denver. The methods they reported for communicating risk primarily include in-person notifications at the physical sites of the organization. However, depending on the organizational structure, staff capacity, and organizational mission, some organizations conduct street outreach to those camping outside to communicate extreme weather risk and provide supplies and resources for taking protective action (e.g., hotel vouchers, tents, sleeping bags). Some organizations in Boulder County, specifically, provide resource sheets with information for where to access resources and services (e.g., day centers, overnight shelters, free meals, mental health services, dental services), including contact information for the severe weather shelter.

The content of the messaging and information they shared is primarily relayed from official sources about the characteristics of the threat and how to take protective action. All interviewees from organizations mentioned the importance of word-of-mouth or informal communication channels about extreme weather across and within unhoused populations in Boulder and Denver. However, according to many interviewees from organizations, there was often not necessarily a formal organizational plan in place for extreme weather risk communication. Instead, their communication and information dissemination strategies included “phone tree” style notifications to their leadership teams, case managers, and partner organizations that work closely with unhoused members of the community. All noted the significance of word-of-mouth communication within unhoused communities. As a representative from an organization based out of Denver County explained:

We find that we can write all these plans, which are helpful, and do all this communication thing, but it's always the word-of-mouth. So we always have some folks that come to our shelter who have kind of made themselves ambassadors of the greater community of homeless individuals. And man, they're quick on the bicycle, and they can take a bike apart in 20 minutes and put it back together. And so they like that their role is spreading the word of where the resources are.

One representative from an organizaiton in Boulder County reported having thresholds in place for severe weather sheltering so that communication is consistent about when these shelter options and other resources are available to people (e.g., when temperature drops below 32 degrees Fahrenheit or when there are lower temps above freezing combined with precipitation):

So severe weather situations, our policies have morphed and they'll probably change again, but we have policies in place regarding primarily cold weather extremes. We don't really have policies in place around heat or other weather events. We do have policies around cold and moisture during the winter months. So those reflect just what resources we have available and how we try to allocate those resources in the most fair way we think we can. So those policies are stated … They're well known procedures around when certain weather triggers hit, what happens and those kinds of things.

Emergency managers and city/county agency employees represented in this study shared similar strategies for communicating risks to unhoused and other difficult to reach communities by drawing upon networks and preexisting relationships with homeless and social-serving organizations to disseminate messaging. All referenced either their relationship with law enforcement or partnering with law enforcement to communicate extreme weather risks to individuals camping or residing outside, as well as utilizing cell phone alert systems. However, they noted challenges with these approaches (described in the next section). A key finding across the six individuals representing this subgroup was the significance of partnerships and interorganizational coordination with agencies, organizations, and groups that work with unhoused or precariously housed members of the community to spread extreme weather and other risk messaging out to unhoused individuals. Notably, all emergency manager and city/county agency interviewees reported either the desire to improve, or being in the process of improving, their plans by incorporating perspectives of those with lived experience with homelessness. As one emergency manager explained,

I don't know that the city [anonymized] has actually sat down and had an advisory committee of people who have experienced being homeless to be able to help kind of inform their best practices about how to deliver that communication or create that kind of communication, in general. And I think that that would be extremely useful for us to have.

People experiencing homelessness represented in this study shared ways that they often receive risk communication messaging, which primarily included word-of-mouth, environmental cues (e.g., cloud cover indicating incoming weather fronts, smoke), automated or app-based alerts via cell phones (for those who have access to cell phones or power sources), and direct communication with staff at organizations (for those who use organizational services), clinical or mental health services, and at “feeds” or areas where food distribution occurs. As one interviewee based out of Boulder County explained, they receive emergency alerts, “Mainly through my phone, checking on the weather. If I hear somebody talking, maybe one of the other people in the shelter. We talk about weather all the time. It's a daily thing for us because we're out in it.”. Another interviewee shared that they primarily receive their information through the shelter they frequent:

Online or through the shelter. They have a bulletin board. They post every day what the weather's going to be like. As far as weather and emergencies like that, they let us know. I think they do a great job letting people know that something's coming. You know?

However, the following quote from a participant who was experiencing homelessness in Denver at the time of their interview both highlighted the challenges for many in the community for receiving risk communication and the way they rely on environmental cues to assess extreme weather risks.

When they happen… Honestly, it's hard to communicate that to us out here because we don't carry phones, and we don't watch the news, and we don't go in for things. Short of having a town crier, I don't know—I can't imagine how to get that information … I watch the mountain range and watch the rains move in. That's what I do … wake up in the morning, you smell it. There’s a scent in the air. I don’t know how to explain that. It’s clean.

Communication Barriers and Challenges

In order to gauge perceptions about the extent to which risk communication or emergency planning incorporates unhoused communities, we asked representatives at organizations serving people experiencing homelessness, “[t]o your knowledge, are homeless communities meaningfully considered in local and county emergency management and response plans in the event of extreme weather?” A majority of interviewees in this subgroup responded that they feel these communities are not thought of at all or that they are considered an afterthought when an extreme weather event is impending or unfolding (n=6). As one representative from an organization in Boulder explained, “I think the people who work with people experiencing homelessness always think of them. Then for people who don't, it's an afterthought.”

However, one interviewee from an organizaiton in Denver shared that based on their experience working with local and county emergency management over time that emergency management planning has increasingly considered the needs of homeless communities in disaster:

[I]t's been raised to the level that now people automatically think, "Okay, what's going to happen with the people in the shelters but also the people who are outside camping on the streets?" so I would say it's very much acknowledged. People know it. I'm not sure the response is always adequate, but it's considered for sure.

Interviewees from organizations reported serveral challenges for communicating extreme weather risk, including how some unhoused individuals do not feel comfortable sharing personal information to receive cell phone alerts, the difficulty navigating communication and interactions with those experiencing mental health issues, and how many individuals have limited to no access to a cell phone. One interviewee discussed how information dissemination is stymied by last-minute severe weather sheltering options, including limited to no transportation options, and difficulty communicating details when sheltering situations quickly evolve:

[H]ow information is spread or people outside are informed of things, this is a real issue. Because there were a few days where the city decided to, essentially at the last minute, create a severe weather shelter during the day … And the way people were finding out about it was through two officers who sit on… the Homeless Outreach Team. And then, they also informed the shelter and service providers, but none of them actually do any outreach. So it was like, "How are people outside even knowing about this?" And then here comes the other large barrier, besides not even knowing about it, is how do people get there?... So I think they operated that day shelter, kind of slapped it together. It wasn't central to downtown, it was on the outskirts. Not that far, but far when you're homeless, and you have a backpack that weighs 80 pounds and dragging a suitcase, and it's freezing out … I would say the way that they communicate with these last-minute details – because that's basically how they operate when it comes to considering anything for the unhoused community, last minute or they don't think about it at all – It's completely ineffective.

Additionally, as touched upon in the previous quote, participants noted that who communicates these messages plays an important role in how people receive and act upon risk information. For example, multiple interviewees across groups explained a sense of distrust among unhoused individuals toward law enforcement. Two interviewees living unhoused in Denver explained, “most people, when you see cops… you immediately freeze up and think, "Oh, they're going to check IDs." Or they're going to be doing arrests." So anytime you see a cop, it stresses you the f—k out.”

In an interview with two Boulder County-level agency employees also noted staff turnover as being a constraint for effectively communicating risk to partner organizations, and that this has worsened over the course of the pandemic: “I think the other piece that's really a challenge in this population and these service providers… is staff turnover. Always trying to stay on top of who is the right person at each organization is a full-time job within itself.”

In addition to logistical barriers for ensuring that a majority of unhoused individuals and communities are able to receive extreme weather risk information, a notable challenge shared by both city/county agency and interviewees from organizations included the fact that many organizations reported not being able to provide viable alternatives or opportunities for unhoused individuals to actually act upon risk. For example, some interviewees described how people with pets may be less willing (or unlikely) to go to a shelter that does not allow them to bring their pets. Multiple interviewees shared the dilemma of communicating risk without necessarily having resources or alternatives to provide to people camping outside (e.g., hotel rooms during inclement weather, tents, necessary gear). An executive director at an organization in Boulder County highlighted this issue, stating that:

[T]here has [sic] to be alternatives … “You should go to the severe weather shelter. Make sure that you stay warm," and they're like, "I got kicked out of the severe weather shelter, so what else should I do?" And as someone with housing, it makes me uncomfortable to get that question because the reality is – and not that I'm concerned about my feelings in this situation – more so, the reality is I don't have anything to tell them. "Do you have a friend you can stay with? Do you have access to a car? Can you afford a hotel room tonight?" It's just like, if there aren't other options for people, then what you're saying to people is, "Hey, there's going to be this thing that's going to impact you greatly. And there aren't a ton of options for you to mitigate it. But just want you to know, you're going to have a hard time breathing. You might freeze to death." That's the piece that's really, I think, hard and frustrating for everyone.

Some interviewees living unhoused explained challenges they had experienced in receiving and acting upon risk information in previous extreme weather events (e.g., extreme cold and blizzards), such as inconsistent or out-of-date information as to where to seek shelter or receiving information that they are unable to act upon to protect themselves. One woman who recently became homeless along with her teenage daughter described barriers she experienced while navigating the shelter system in the Denver metro area during a winter weather event in February 2021, explaining:

We were in Aurora when that was going on … There wasn't anywhere to go. There was just a lot of worry. A lot of people talking about where they were going to be able find shelter, what was going on, what was available. Things that were supposed to be, weren't. We were sent to a couple of places that, once we end up getting there, there was nothing there. There was nothing available. So when you're out in the cold and you're trying to take a bus, or you're trying to walk, that information, for it to be accurate, is really kind of important. Because some of us, that's the last bus fare. Or that's all the energy I've got the rest of the day can't go any further than this now.

Recommendations for Improving Risk and Emergency Planning for Unhoused Communities

Nearly all individuals who were experiencing homelessness at the time of the interview shared recommendations for how organizations serving people experiencing homelessness, county and/or local government agencies could communicate and plan for the needs of homeless and precariously housed communities leading up to and during extreme weather. They offered a range of suggestions, although many of them centered around needing additional and multiple modes of communication for effectively receiving information about weather risks, additional spaces and options for sheltering (not necessarily “traditional” emergency shelters, but provision of tents and supplies), providing people with access to housing or resources they could use to act upon risks, and increased coordination across service providers. To center the voices of those at the heart of this study, we present recommendations in their own words for improving risk communication (see Table 1).

Table 1. Recommendations Shared by Interviewees Experiencing Homelessness

A couple of interviewees were not sure what else could be done to improve risk communication, but noted the need for those involved in emergency planning to directly engage people with lived experience with homelessness, as one interviewee shared:

I have no idea [how to improve risk communication]. That's a tough one other than just frequenting the places we come to eat and things like this. It's really the only time you get a hold of us enough to get our attention. We just don't pay attention to anything else. We just don't … word-of-mouth will spread quicker than anything else around us.

Others also echoed this sentiment regarding the need for those with lived experience to be engaged intentionally in emergency planning by providing space for them to participate and hear their ideas and concerns. An interviewee based out of Denver asserted, “listen to us and not do what they [decision makers and planners] want and what they think we need.” Another echoed this request, stating “[d]efinitely get us involved. I mean, that might help. Ask us what we can do to help. There's some of us out here that will volunteer. We just don't know where to start.” Highlighting some examples of issues associated with not asking people what they need, what they would like to see, and how they would like to be involved in emergency planning, an interviewee based out of Denver explained in frustration that:

There's times where it's—like random people, just people not with any organization or with the government or anything like that, they're just random people. They'll come out and they'll give you something … And so the thought is nice but they'll literally bring—you can tell like, okay, this is just crap that they don't want to throw away. So instead of throwing it away, they're going to, "Let me go give it to a homeless person." And they'll bring you something that you have—it requires a stove and water and electricity or gas. It's like, "Are you f–king kidding me? Where in the hell am I going to cook this at? This is in no way helpful to me… So it's kind of baffling so it's like sometimes you look at government officials or government organizations and it's like you have that same thought like, "Are you kidding me?"

Other specific ideas included the need for secure storage space for belongings as opposed to people having to leave their tents and supplies outside in order to use emergency shelter services, as well as a suggestion that information should be communicated about where people can go to safely set up tents and get supplies for camping outside during extreme weather.

Conclusions

This high-level analysis highlights a number of challenges related to disseminating and receiving extreme weather risk communication for individuals experiencing homelessness. Representatives from organizations described challenges resulting from unhoused communities being an afterthought or not considered in emergency planning. They also noted difficulties accessing and communicating with people experiencing homelessness since many do not have phones or access to power sources, or are in hard-to-reach locations. Multiple interviewees from local/county level agencies and organizations described an inability to provide alternative recommendations for protective actions so that people experiencing homelessness can actually act upon risk.

Those with lived experience with homelessness, who often rely on word-of-mouth communication and/or information from shelter and service providers, described challenges associated with the rapidly changing nature of information during an extreme weather event. Thus, it is difficult to ensure they are receiving timely and up-to-date information in order to take appropriate protective action. Interviewees with lived experience with homelessness also shared ways they believe risk communication and emergency planning can be improved so that the unique needs and capacities of precariously housed individuals are intentionally and meaningfully considered in emergency planning. Their suggestions include improving inter-organizational coordination across agencies and organizations serving people experiencing homelessness, providing additional sheltering options and space, and increasing the modes of risk communication.

These findings suggest that people experiencing homelessness are not equitably considered in emergency and risk communication planning, and that this is an extension of the everyday precarity and systematic marginalization experienced by these communities (Gaillard et al., 2019; Vickery, 2019). For example, while key informants noted that pets were often prohibited in emergency shelters, this is in fact, a facial violation of the Fair Housing Act, demonstrating need for education and enforcement (Office of Fair Housing & Equal Opportunity, 2020³²). Additionally, some organization and local/county agency interviewees shared that they may not be able offer viable alternatives to individuals experiencing homelessness when faced with extreme weather. This is a structural problem, which we suggest as a likely result of a lack of political will, compounded by measures put in place within both study areas that effectively attempt to criminalize or “push out” people experiencing homelessness. Further, the limited amount of research focused on understanding how risk is communicated to marginalized populations is an equity issue in and of itself. Individuals and communities experiencing homeless and housing precarity must be meaningfully represented in research agendas to advance knowledge and research-guided recommendations for improving disaster outcomes.

Limitations

Limitations of the study include the limited sample size, both in terms of the full study sample and subsamples specific to interviewee group, the geographic scope of the study, as well as the lack of ethnic and racial diversity across the study sample, which is predominantly white. We thus are limited in understanding the full extent of challenges to and recommendations for extreme weather risk communication without critical insights from individuals across gender, race, ethnicity, and other sociodemographic considerations. The relatively small and racially homogenous (predominantly white) sample leads us to findings that are likely not generalizable nor representative of the interviewee groups at the heart of this study, warranting additional research in other contexts and with larger, more diverse study samples.

Other limitations and challenges throughout the award period emerged, including how and when to safely engage and conduct interviews with individuals experiencing homelessness given that remote recruitment and data collection (e.g., telephone or video call) is a notable challenge in this population. While our study design took into consideration both current (at the time) and potential barriers to data collection due to COVID-19, we experienced a delay in in-person data collection activities until (1) the request for fieldwork was approved through the University of Washington Fieldwork Health and Safety Plan for COVID Prevention (a step following IRB exempt status) and (2) a member of the research team could become vaccinated in order to conduct interviews safely. This delayed and truncated the data collection period for the award timeline and may have resulted in recall bias among interviewees (across groups), especially when interviewees were asked to reflect upon extreme weather events from winter 2020-2021 and wildfire smoke events from the previous summer and fall 2020. We adapted the interview guide accordingly from anchoring to specific weather events to more general questions concerning extreme weather risk communication and planning as the data collection period progressed.

Next Steps and Directions for Future Research

Given that the findings presented in this report are preliminary, a next step in the data analysis process will involve the creation of a formalized codebook that will be integrated into NVivo qualitative analysis software. Using NVivo, we will systematically code transcripts, notes, and emergency and risk communication plans. We continue to work with an Master’s of Public Health student at the University of Washington as part of their thesis development to analyze and evaluate emergency and risk communication plans via content analysis. We will upload interview instruments and our IRB protocol into DesignSafe for other researchers to adapt and use. Finally, we intend to conduct additional interviews and focused discussions with individuals across interviewee groups, specifically those with lived experience with homelessness and those involved in emergency planning. In addition to co-determined dissemination strategies and outlets mentioned earlier in the report, we will present preliminary findings at the American Meteorological Society’s Annual Meeting in January 2022 and will submit a research highlight and abstract to the 2022 Natural Hazards Workshop and Researcher’s Meeting, respectively. We will build on the current study to inform the development of future grant proposals and to develop a policy brief to be shared with community partners, emergency management associations, and homeless advocacy organizations.

Future research could replicate the current study in other contexts to understand the needs, challenges, barriers, and facilitators for people experiencing homelessness in receiving and acting upon extreme weather risk communication. However, especially in working with marginalized communities in which information has historically been extracted, careful attention to study design, purpose, participant compensation, and justification must be considered. More broadly, best practices for equitable emergency management practice need to be investigated and co-determined by the communities central to emergency management planning, emergency managers, and other key stakeholders involved in service provision to or care for marginalized communities.

Acknowledgements. We would like to offer our sincere gratitude to the following individuals for their assistance and support throughout this project: Kattie Forbus, Study Advisor; Caitlin Hall, PhD Student, University of Durham; and Joanne Medina, MPH Student, University of Washington.

We would also like to acknowledge the other members of the advisory team, whose knowledge, experience and guidance have been invaluable throughout the study. We are grateful for their time and for the time of all interview participants who shared their valuable insights with us.

References