“Especially being homeless, they just think you’re infected with COVID or something”: A qualitative exploration of the COVID-19 pandemic’s impact on people experiencing homelessness with a history of injection drug use in Baltimore, Maryland

Abstract

People experiencing homelessness with a substance use disorder are a highly structurally vulnerable population, facing a unique burden of compounding stigma, discrimination, and adverse health outcomes. Evidence remains mixed on best practices for housing interventions designed to best meet the needs of this population. Ten people with a history of injection drug use experiencing homelessness during the COVID-19 pandemic were interviewed between July, 2021 and February, 2022 in Baltimore, Maryland. Thematic analysis was used to characterize the pandemic’s impact on resource access, housing, and substance use among these participants. Participants highlighted substantial challenges resulting from structural changes during the pandemic, including limits on income-generating opportunities and resource access, and increases in experiences of stigma and discrimination. However, several individuals reported decreased substance use in response to changes to their housing status, with those placed in secure housing attributing their decreased use to this change. These accounts can help guide housing and social support interventions best suited to meet the unique needs of people experiencing homelessness with substance use disorders.

Introduction

The COVID-19 pandemic has disproportionately burdened individuals experiencing homelessness. Persons experiencing homelessness in the United States (US) have faced greater rates of SARS-CoV-2 (COVID-19) infection compared to the general population (Hsu et al., 2020; Mosites et al., 2020). In addition, many COVID-19 prevention measures (e.g., physical distancing) were challenging to maintain among persons experiencing homelessness, particularly within congregate settings such as shelters (Rodriguez et al., 2021). Accordingly, many shelters struggled to contain COVID-19 outbreaks, taking measures to limit transmission by reducing their capacity, moving persons to more private spaces such as single occupancy hotel rooms, or in some cases, closing altogether (Mosites et al., 2020; Recede, 2020; Tobolowsky, 2020). The impact of the COVID-19 pandemic also exacerbated the country’s housing crisis, with job losses and pandemic-related economic strain associated with increased housing instability (Pixley et al., 2022). Increasing rates of homelessness during the pandemic, along with both temporary and permanent closures of many housing support services (particularly due to fears of the spread of COVID-19 within congregate living spaces), led to a heightened need for shelter beds, leaving housing programs throughout the US unable to meet the demand (Rodriguez et al., 2021).

It is well documented that homelessness is a risk factor for substance use (Kaplan et al., 2019; Magwood et al., 2020; Tyler & Ray, 2019). Estimates suggest that over one third of people experiencing homelessness in the US have a substance use disorder (National Coalition for the Homeless, 2017), compared to around 7% of the general adult population (Mental Health America, 2022). People experiencing homelessness with a substance use disorder are a particularly structurally vulnerable group, facing an increased burden of adverse substance use-related outcomes compared to the general population, including significant barriers to treatment access/retention and limited physical and mental healthcare service availability (Swartz et al., 2022; Upshur et al., 2018; Yamamoto et al., 2019). Further, people who use drugs and are experiencing homelessness are known to experience heightened stigmatization and discrimination in response to both their housing and substance use status (Arum et al., 2021; Couto E Cruz et al., 2018), which perpetuates social exclusion, criminalization and other negative treatment (De Las Nueces, D., 2016; Pahwa et al., 2019) and leads to an increased burden of adverse physical and mental health outcomes (Omerov et al., 2020;, Weisz & Quinn, 2018;, Randhawa et al., 2018).

Among individuals with a substance use disorder, people who inject drugs (PWID) are further disadvantaged, facing increased levels of negative treatment compared with people who use drugs by other routes of administration (Ahern et al., 2007). In particular, the perception that PWID are “dangerous,” “immoral,” or responsible for their substance use disorder perpetuates negative attitudes and treatment of PWID (Sabri et al., 2021; Thomas & Menih, 2022). PWID face unique structural barriers to shelter and social service access given these negative societal attitudes which often drive abstinence-only policies at shelters (Bardwell et al., 2018). This also contributes to an increased risk of overdose and infectious disease transmission, as PWID may attempt to rush drug use or hide injection behaviors (Wallace et al., 2018). Accordingly, PWID who are experiencing homelessness are some of society’s most marginalized and merit particular attention to help guide programs and services best equipped to meet their unique mental and physical health needs.

One promising strategy for promoting housing access to individuals experiencing homelessness is the Housing First (HF) model. Originating in the 1990s, this approach was founded on the premise that housing is a basic human right (National Alliance to End Homelessness, 2022). Program participants are provided with immediate low-barrier housing support not contingent upon sobriety, employment, or other restrictions or parameters (Tsemberis & Eisenberg, 2000). While HF has shown great promise as a means to increase housing stability among people experiencing homelessness, evidence on the effectiveness of this model among individuals with a substance use disorder remains mixed (Baxter et al., 2019; Woodhall-Melnik & Dunn, 2016). Evaluations of HF programs have called for a better understanding of the impact of the program on sub-populations within the broader community of people experiencing homelessness, particularly among those experiencing a severe or active addiction (Kertesz et al., 2009). However, conducting such research is challenging, as individuals with severe or active addiction who are experiencing homelessness are historically difficult to reach with research and programmatic efforts.

Given this context, this study aimed to elicit first-hand accounts from individuals with a history of recent injection drug use who were experiencing homelessness at the time of the pandemic to 1) explore how societal and community-level changes due to the COVID-19 pandemic impacted their daily lives, housing status, and drug use, and 2) assess how the impacts of these changes can be used to inform housing policies and interventions designed to best support this uniquely vulnerable sub-population of people experiencing homelessness.

Literature Review: Mixed Findings on the Impact of Housing Interventions Among People with Substance Use Disorders

Two distinct approaches to providing housing interventions to people experiencing homelessness have been systematically studied. The Housing First (HF) model, originally intended to support people with mental illness who were experiencing homelessness, is designed to enable people experiencing homelessness to access permanent housing that is not reliant upon compliance with specific requirements such as treatment engagement or abstinence from substance use (Tsemberis et al., 2004). The program offers a contrast to traditional Treatment First (TF) housing programs, which provide tiered housing support (e.g., a transition from temporary accommodations to transitional housing to permanent housing), the progression of which is conditioned upon successful mental health and/or substance use treatment adherence (Tsemberis, 2011). TF programs have often been implemented among individuals with substance use disorders, requiring participants to remain abstinent from drugs or alcohol and successfully engage in treatment before entering transitional and eventually permanent housing (Tsemberis et al., 2004). This model has been criticized for its exclusion of more structurally vulnerable individuals, particularly those who are unwilling or unable to engage in treatment programs or maintain abstinence. HF has been viewed as a less punitive alternative, though there remains a lack of consensus on one explicit definition or operationalization of the HF model (Woodhall-Melnik & Dunn, 2016). However, there is general agreement that important components include 1) no requirement for participants to demonstrate housing readiness, 2) the availability of individualized support for those who desire it, and 3) a grounding in the principle of self-determination (Atherton & Nicholls, 2008; Goering et al., 2011).

Several attempts have been made to systematically evaluate HF interventions in effort to generate an evidence base for future program implementation. Among those evaluating housing stability as an outcome (often operationalized by a measure of the length of time participants remain placed in housing), there is clear evidence that HF participants maintain increased levels of housing retention and stability compared to individuals in TF models (Aubry et al., 2020; Baxter et al., 2019; Desilva et al., 2011; Montgomery et al., 2013; Palepu et al., 2013). HF participants also consistently demonstrate decreased participation in the criminal justice system, fewer hospitalizations, and less visits to emergency departments or time spent in shelters (Baxter et al., 2019; Hanratty, 2011; Padgett et al., 2006; Srebnik et al., 2013).

However, evidence for the impact of HF programming on substance use outcomes remains mixed. Some observational studies have found that participants engaging in a HF intervention demonstrate decreased rates of substance use compared to those in TF programs (Collins et al., 2012; Padgett et al., 2011; Tsemberis et al., 2012). One study, for example, found that HF participants reported significantly lower rates of substance use (assessed as any illicit drug use and/or frequent and heavy alcohol use) within one year of program engagement compared to individuals in TF programs (Padgett et al., 2011). Others, however, have found that TF participants in fact reduce their substance use more significantly than those engaged in HF (e.g., one study which found that veterans placed in a contingent-free housing program were more likely to escalate drinking and/or drug use within one year post placement than those in a contingency-based approach [Westermeyer & Lee, 2013]). A 2019 systematic review of randomized controlled HF trials found no clear difference in substance use outcomes among intervention and control group participants, though in this case only two studies had included substance use measures as outcomes (Baxter et al., 2019). Another published in 2020 found that among nine studies evaluating substance use outcomes after implementation of permanent supportive housing interventions, there was no measurable differences in substance use among treated and control participants (Aubry et al., 2020).

One evaluation of HF programming across Europe similarly found mixed impacts of the program on substance use outcomes. However, the authors identify that participants with substance use disorders who engage in HF programs report increased satisfaction, an increase in time spent in independent housing, fewer psychiatric symptoms, and better community integration, leading the authors to conclude that HF consistently predicts “greater recovery” than TF across a range of other important socioeconomic outcomes (Greenwood et al., 2020). Other improved outcomes among individuals with substance use disorders engaged in HF interventions include decreased emergency healthcare utilization and improvements in methadone treatment retention (Appel et al., 2012). Yet past evaluations have cautioned against over-generalizing about any positive impacts of HF programming on individuals with substance use disorders, noting a lack of evaluations of the model’s impact on individuals with ‘severe addiction.’ It has therefore been suggested that the evidence base for the impact of HF on individuals with active addiction remains incomplete (Kertesz et al., 2009).

Given these inconclusive findings, there have been consistent calls for more research on the role of contingent-free housing among sub-populations of people experiencing homelessness, particularly those with substance use disorders. These varied outcomes highlight the importance of understanding the specific needs of each unique client population prior to any future HF program implementation in order to maximize its effectiveness (Woodhall-Melnik & Dunn, 2016). PWID face unique risks and vulnerabilities, and to our knowledge, no evaluations have specifically explored the experiences of this sub-population within HF interventions. Given their unique risk profiles, it is highly likely that tailored supports are merited within any housing program to best accommodate these needs and maximize positive outcomes (Kirst et al., 2020). We strive to address this gap in the available evidence on housing programs uniquely designed to support this hard-to-reach population through first-person accounts of people experiencing homelessness with a history of recent injection drug use. We believe that characterizing their experiences can help build an evidence-base for effective housing interventions designed to maximize mental and physical health outcomes among this vulnerable sub-population of people experiencing homelessness.

Methods

This paper presents findings from a qualitative study conducted with data from CHANGES, the COVID Health And Network Group Experience Study, which was conducted with 28 individuals between July, 2021 and February, 2022. The CHANGES study aimed to engage the socio-ecological model to examine the ways in which multilevel changes and shifting social and community-level conditions resulting from the COVID-19 pandemic impacted the individual-level experiences of this uniquely structurally vulnerable population. Specifically, we sought to understand the impact of the COVID-19 pandemic on the day-to-day lives, social networks, and drug-related behaviors of people with a history of recent injection drug use in Baltimore, Maryland, a mid-sized urban city on the east coast of the US. CHANGES participants were recruited from the AIDS Linked to IntraVenous Experience (ALIVE) study, an ongoing longitudinal cohort of over 5,500 community-recruited current and former PWID of mixed HIV status in Baltimore (Vlahov et al., 1991). The study initially enrolled 2939 individuals in 1988–1989, with additional enrollment periods occurring in 1994–1995 (n=434), 1998 (n=295), 2005–2008 (n=1,009), and 2015–2018 (n=829). Eligibility for enrollment included being 18 years of age or older and self-reporting a history of injection drug use. Participants were recruited through various community-based recruitment methods, including street outreach in city locations with known drug markets, referrals from locations with services catering to people who inject drugs (e.g., syringe service programs, drug treatment programs, community-based organizations, specialty care clinics), flyer distribution at community-based locations (e.g., libraries, senior centers, housing shelters, and community centers), and word-of-mouth and peer referrals. This strategy facilitated the recruitment of participants with a range of housing statuses and differing levels of connectivity to social services and health care. A report of the ALIVE participants who had reported active injection drug use within twelve months of their most recent ALIVE network survey in 2018–19 was generated (n=237). We randomly sorted the list, and a research assistant (A.K.W. or J.C.) contacted participants over the telephone to ask whether they were interested in completing a one-time qualitative interview for the CHANGES study. Individuals who wished to participate were read an informed consent document outlining the CHANGES study’s goals, provided oral informed consent, and scheduled a time to complete a phone-based interview.

In order to attempt to reach a more vulnerable subset of ALIVE study participants, purposive sampling techniques were employed near the halfway point of the study to oversample individuals with recent experiences of homelessness and/or reports of current injection drug use. This aligned with the resumption of in-person data collection activities following the onset of the COVID-19 pandemic, facilitating the recruitment of individuals who may not have had a cell phone but presented in person for an ALIVE study visit. Participants who responded “yes” to any of the following questions were referred by an ALIVE study team member to the CHANGES study: “In the past six months, have you been homeless?”; “In the last six months, have you slept in a shelter, park, abandoned building, in the street, or in a bus or train station for at least one night because you had nowhere else to stay?”; and “Have you injected any drugs in the last six months?” 15 additional participants were recruited through the purposive sampling strategy. Recruitment continued until the study team assessed that saturation was reached. Individuals received $25 for their participation in the study. All study procedures were approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board.

The semi-structured, in-depth interviews were conducted by two research assistants (A.K.W. and J.C.) and lasted between 20 and 90 minutes. Interview questions were informed by a socioecological model and designed to assess the ways in which societal disruption caused by the COVID-19 pandemic impacted the individual experiences of PWID through interpersonal, community, and societal levels (Bronfenbrenner, 1979). Questions prompted participants to identify pandemic-related changes in areas including housing, interpersonal relationships, drug use practices, and drug treatment engagement. To assess impacts on housing status, participants were asked: “Has your primary place of residence changed recently or since the start of the COVID-19 pandemic? If so, how has it changed?” All interviews were audio recorded with participant consent and transcribed verbatim by an external, professional transcription service. Transcripts were uploaded into Atlas.ti version 9 for analysis (ATLAS.Ti GmbH, 2022). A thematic analysis approach was used for data analysis, involving familiarization with the data, generation of an initial codebook, extraction of themes, review and discussion, definition of emergent themes, and subsequent analysis (Braun & Clarke, 2006). The coding process occurred over several consecutive months, and thematic analysis activities continued over a subsequent 12-month period. The research assistants (A.K.W and J.C.) and a senior qualitative researcher (S.M.G) each read two selected interviews and developed the initial inductive codebook. The research assistants then independently coded two more transcripts, meeting as a team to review coded transcripts for discrepancies and adapt the codebook as appropriate. The senior researcher oversaw the coding process, and subsequent transcripts were independently coded by the research assistants and reviewed collectively until a Krippendorff’s c-a-binary score of 0.70 was calculated, demonstrating internal consistency among coders (Krippendorff, 2011). The two research assistants then independently coded half of the remaining transcripts. Throughout the coding process, the research assistants met regularly with the senior researcher to discuss findings, review codebook application, and ensure that the codebook adequately fit the data. New codes were created as a team when the coders noted insights not captured sufficiently within the existing codebook.

While the CHANGES study aimed to explore the impact of the COVID-19 pandemic on people with a recent history of injection drug use broadly, the research team identified unique experiences among those who reported periods of homelessness in the analysis phase. These individuals shared accounts of unique challenges related to increased stigmatization and loss of economic opportunities, as well as shifts in substance use patterns. Given that people experiencing homelessness with a history of injection drug use are a particularly structurally vulnerable and thus difficult to access population, we chose to characterize the unique ways in which they were impacted by the COVID-19 pandemic. Thus, this analysis focuses upon the subset of the study sample who reported experiencing homelessness at any point during the pandemic. To determine those who fell into this category, the researchers exported all segments coded with “Housing” across all 28 transcripts, noting all accounts of homelessness, defined by any reports of individuals sleeping for one or more nights in a shelter, abandoned building, tent, outdoors, or other public space. The final sample included in this analysis was comprised of ten participants who reported experiences of homelessness occurring at any time during the pandemic. Among these ten participants, codes were aggregated into thematic categories through research team reflection and discussion. The researchers explored themes related to participants’ general accounts of homelessness, access to housing/resources, and changes in drug use during the pandemic.

Researcher positionality

Both research assistants (A.K.W. and J.C.) who led the interviews and coding had formal training in public health and qualitative methods. Both had prior experience conducting in-depth interviews with priority populations. A.K.W. is a white female and at the time of the study was a PhD student in public health. J.C. is an Asian female who at the time of this study was a research program coordinator for the ALIVE study. Data collection and analyses were supervised by two white, female senior researchers—a qualitative researcher and anthropologist (S.M.G.) and a mixed-methods, health services researcher and epidemiologist (B.L.G.). Both have over a decade of experience conducting qualitative research with priority populations in Baltimore, Maryland. S.M.G. and B.L.G. did not interact with the study participants; nor did other members of the research team.

Results

Among the subset of ten participants experiencing homelessness included in this study, 3 were female, 5 identified as non-Hispanic Black, 4 as non-Hispanic White, and 1 as non-Hispanic other race, and participants ranged in age from 38–68 years (Table I). The housing details of each participant are reported in Table I.

Table I:

Participant characteristics and housing details

Participant ID number	Sex	Age, years	Race	Drug use in past 6m	Injection drug use in past 6m	Housing Experience
4	Female	55	Non-Hispanic Other	Yes	Yes	Participant was experiencing homelessness before the pandemic Reported that she had been paying $20 a day for a room and a meal at an acquaintance’s house, but lost her capacity to earn this money during the pandemic Began sleeping in bus stops and abandoned buildings, until her son had to find new housing, when she was able to move into a motel with him
13	Male	68	Black	Yes	Yes	Participant was experiencing homelessness before the pandemic Reported that he had been consistently living in a shelter for over two years, which he had accessed before the start of the pandemic
17	Male	57	White	Yes	Yes	Participant was experiencing homelessness for 25 years at the start of the pandemic. Reported that he was placed in his own apartment through a non-profit behavioral healthcare corporation after the start of the pandemic, where he had been for nearly a year at the time of the interview
18	Male	50	White	Yes	Yes	Participant was experiencing homelessness before the pandemic Reported that he used to stay in abandoned buildings but had been sleeping in a tent in the woods for the past six months at the time of the interview
19	Female	38	Black	Yes	Yes	Participant was experiencing homelessness before the pandemic Reported staying in a women’s shelter when space is available, and otherwise sleeping in an open space under the porch of an abandoned building with a male partner
20	Male	49	Black	Yes	Yes	Participant was experiencing homelessness before the pandemic Reported trying to stay in men’s shelters when possible but often faced limited capacity Frequently slept in a cubby under a porch, within abandoned buildings, in parks, or under bridges when he could not access a shelter bed
24	Male	38	White	Yes	Yes	Participant was experiencing homelessness for ten years before start of pandemic Had been staying in shelters, within a homeless encampment, and on the street Was placed in a hotel during COVID-19 which he accessed through Healthcare for the Homeless Then linked to a treatment program with housing, where he had been for the past seven months at time of interview
25	Female	45	Black	Yes	No	Participant was experiencing homelessness for the past nine years off and on at the time of the interview Reported that she most often finds vacant and abandoned buildings to sleep in
27	Male	45	White	Yes	Yes	Participant was experiencing homelessness for seven years at the time of the interview Was residing in a shelter for two years before the pandemic, which shut down because of COVID-19 Participant was provided his own room at a local motel, where he remained throughout the pandemic
28	Female	53	Black	Yes	Yes	Participant became homeless during the start of the pandemic because she lost SSI benefits and was unable to pay her bills, which she was told was due to a clerical error Reported that she is currently staying in an abandoned house

The following sections detail the ways in which systemic social and economic changes caused by the pandemic impacted participants’ experiences with finances, resource access, housing status, and drug use. We first describe how participants experienced structural changes due to COVID-19, including changes in housing policy, and their impact on socioeconomic context and housing stability. We then examine how these changes related to participants’ drug use behaviors.

I. Impact of the shifting structural environment on individual socioeconomic context

a. Structural changes intensified by compounding community-level stigma

Economic and community-level shifts caused by the pandemic led to challenges earning sufficient money to meet basic survival needs and contributing to ongoing housing instability. In addition, a perceived increase in stigmatization and discrimination compared to pre-pandemic treatment compounded structural changes and further contributed to the loss of economic stability. Most participants were not formally employed at the start of the pandemic, but rather relied on informal sources of income, such as finding odd jobs or untaxed, “under-the-table” work. These participants faced the consequences of infrastructure changes, such as business closures, which limited their informal economic opportunities. Several other participants had relied on panhandling – a North American term used for begging – as a main source of financial support before the pandemic. Among these, many struggled to earn sufficient quantities to get by. With fewer people moving around due to lockdowns, business closures, and social distancing guidelines, the pool of people from which to beg for money was substantially reduced. Further compounding this reduction was a perception that community members were more likely to avoid or fear people experiencing homelessness, perceiving that they were viewed as more likely to be infected with COVID-19.

It’s like nobody—because I panhandle to support myself, and people won’t even roll their window down now hardly to help you, and then it’s like all the programs and everything—it’s like nobody wants to—especially being homeless, they just think you’re infected with COVID or something… Its just people used to give us bags of food and clothes, and nobody does anything anymore. Nothing. <laughs> I haven’t gotten any hygiene stuff. Nobody owes me nothing, but it was nice that people used to help, but they don’t do anything anymore.

– 50-year-old White male

Another participant had been staying with a woman in the community before the pandemic, who for $20 per day offered her dinner, a shower, and a safe place to sleep. She had easily been able to earn this money panhandling before the pandemic but reported being unable to continue doing so following its onset.

Because at the time when COVID hit I was homeless, and I panhandled to survive. And it was just really hard, people didn’t want to wind their—the people that did come by where I held my sign—if it wasn’t somebody basically that knew me or had already dealt with me in the past, they weren’t winding windows down… And I was barely making 20 dollars, where before COVID hit I literally could get 20 dollars within two hours, an hour sometimes.

– 55-year-old White female

The same participant had also maintained an alternative source of income, working part time in a carry-out restaurant several mornings per week. Yet when the pandemic forced the restaurant to close, she lost this source of income as well.

I actually had a little part time job. Even though I was using then, I had a part time job. I worked in this little carryout spot, and I would do their prep work… So it was great, and they paid me cash and that took care of my habit and my room rent. When COVID hit, the restaurant shut down. I had no job and had to struggle… Yeah, when COVID hit it was really hard. People just didn’t want to trust you. And I understood all that- they didn’t know if you were sick, especially being homeless. But it got to the point I just couldn’t get the help, because it wasn’t there, it just wasn’t there, and I really struggled. I ended up in the hospital a couple of times because I wasn’t able to eat. I just didn’t have money.

– 55-year-old White female

Beyond exemplifying limitations on informal work opportunities, these quotations demonstrate the perception of heightened stigma and discrimination towards people experiencing homelessness, grounded in the underlying belief that members of this population were more likely to be “sick” or “infected.”

b. Resource impact: amplified exclusion of people experiencing poverty

Structural changes caused by the pandemic created barriers to participation in the economy and access to resources. Public health COVID-19 prevention measures such as office closures, physical distancing, and restrictions on numbers of people congregating led to constraints on participants’ ability to access services that supported their well-being, leading in some cases to severe consequences such as job loss.

Well, when it was really, really bad, they couldn’t have as many people on the bus. So, some busses would ride right by me, even though I would be on the bus stop an hour-and-a-half earlier to make sure I was trying to get to work on time. But I think they was only allowing like 10 or 15 people on the bus, if that. So, that kind of caused me to be late a few times to work, which eventually I ended up losing my job due to lateness, and I thought they would be more understanding.

– 38-year-old Black female

In addition, the COVID-19 pandemic further perpetuated disparities in access to technology and its associated benefits. Specifically, limited access to the digital economy put some participants at a greater risk of adverse socioeconomic pandemic-related consequences. These technology barriers were particularly impactful among individuals who did not have their own cell phones or who had limited/no data plans and had therefore relied on publicly available internet sources. Specifically, a lack of access to free public internet sources led to challenges completing important tasks such as paying bills or applying for social services.

And if you didn’t have access to an internet, it was even more difficult, because the library was not open. Since the pandemic, I think it just made things additionally more harder to get resources that I needed, that was accessible, extremely more accessible before the pandemic… Resources in reference to like, you know, utilities and gas, electric and things like that. Trying to apply for like emergency food stamps and things like that. Before, you could walk in and get a ticket and wait for them to call your number on the ticket. As opposed to you have to be able to have access to an internet or something to apply for the application, you know? So, it was just the main thing about me, the internet access, which I was getting from the library before, or going to some stores, you know, that offer free wi-fi, like Dunkin Donuts or Subway or things like that.

– 38-year-old Black female

Limited access to technology also impacted treatment access, particularly when many in-person programs moved to virtual platforms. One 68-year-old Black participant was no longer able to access treatment and support groups once they became virtual, stating: “Just them not being available in person. They were available online, but I didn’t take advantage of that, because I didn’t have 100 percent access to the internet or a computer.”

c. Inconsistent impacts of shelter and housing policy changes

In response to the COVID-19 pandemic, many programs in Baltimore changed their policies around congregate living in attempt to reduce the number of people cohabitating near one another. Some shelters limited their capacity, while others closed altogether, which led to restrictions in the number of beds available to people experiencing homelessness. For those who had already struggled with the city’s limited service capacity before the pandemic, this led to further barriers to shelter access. The experiences of two participants (who were in a romantic partnership) exemplify the ways in which these structural changes further excluded people experiencing homelessness from accessing housing support. The couple frequently tried to stay in homeless shelters but were often unable to because of the limited capacity. While this problem pre-dated the pandemic, the pandemic further perpetuated the scarcity of available shelter beds, often forcing them to sleep outdoors or in abandoned buildings. The 38-year-old Black female participant was often turned away from a local women’s shelter, noting “If you don’t get there at a certain time, the beds get filled up,” in which case she slept in a small hole under the porch of an abandoned building. Her male partner was also frequently turned away from local shelters, often having to stay in outdoor areas or abandoned buildings. Speaking about the couple collectively, the male partner described:

Because we’re basically, like, homeless, so I’ve been staying where we can stay at because I’m just too poor. We’ve been trying to stay wherever we can stay to stay out of the cold. We stay at little abandos sometimes. Sometimes we’ll be down—you know, out downtown under the bridge. Then we’ll try to get up early to go to the shelter. Try to leave to get—try to be one of the first ones to the shelter, because, you know the shelters, they fill up real quick.

– 49-year-old Black male

For some individuals in Baltimore, however, the pandemic served as an opportunity to move from a congregate living space to a private room through the city’s pandemic mitigation strategy. For example, one participant had consistently been living in a shelter for two years before the pandemic. When the shelter closed, the participant and other shelter residents were moved into rooms within a nearby hotel as part of a Baltimore City policy to reduce the risk of COVID-19 transmission within congregate shelter programs. After two years in his own room at a hotel, he had found a part-time job and was awaiting placement in an apartment. In fact, he was one of several participants for whom pandemic-related policy changes led to housing placement and subsequent impacts they perceived to be positive.

Well, it’s changed by actually a better place to stay, because they actually shut down the shelter I was in and moved us into the Holiday Inn. So I actually have my own room and everything now, and I’ve been there for two years, so that was good. I mean I’m actually waiting on an apartment now. Let me see, what else has changed? I was able to get a job because of my situation now… As soon as we got there, I went into a room and I have been there two years. I pretty much treat it like an apartment basically to where I come and go as I please… We do have a curfew, but other than that, I do everything like it’s basically an apartment. I go out, I buy groceries and bring them back and just set them in my room and I watch TV… before you had to be out at 7:30 in the morning, you had to be put out into the street, and you wasn’t allowed to come back until I think two o’clock. Now like I said, we go wherever we want, we do whatever we want to do. I actually got a job now, it’s only a couple of days a week, but that’s actually changed too.

– 45-year-old White male

II. Impact of shifting socioeconomic conditions on drug use

The shifting socioeconomic conditions experienced by participants because of the COVID-19 pandemic had important impacts on their drug use. All ten of the participants reported having used drugs in the past 12 months at the time of their interview, with nine of the ten reporting having injected drugs within the past six months. However, the pandemic led to significant changes in drug use behaviors for all participants. In one case, the emotional toll of the pandemic led to feelings of depression and a subsequent lack of desire to stop using:

Participant: It’s made me not even think about—it’s kind of like not even made me feel like even wanting to quit or get clean anymore. I always used to have a little bit of just this thing inside me that would want me to always feel like I could finally beat this thing one day, but now it’s like I don’t even feel like it anymore.

Interviewer: Why do you think that is?

Participant: I don’t know. I feel like there’s no use anymore. It just gets you so down. Yeah, it’s kind of weird to explain. <laughs> Lately I’ve been feeling pretty down, I guess because I haven’t been feeling well, and I’m worried that I’ve got hep-C again, and it’s just really—I guess I’ve gotten kind of depressed. So, I mean, I’ve noticed it’s hard for me to even get up in the morning, and I always used to jump right up out of bed and go about my business, but now it’s like it’s so hard to get out of bed. – 50-year-old White male

The remaining participants who reported substantial pandemic-related drug use changes experienced decreases in their substance use. The financial impact of the pandemic was one cause of this shift, as less income led to less money for drugs.

Actually, I’ve slowed down a lot, because the money not plentiful like it used to be, so I was forced to slow down on my drug use, and I’m okay with that… Like I said, since I don’t have the money for it, I had to slow down. I haven’t lost no friends. I haven’t, you know what I’m saying, burnt no bridges with people, because when I don’t have it, I don’t have it. I don’t go out here robbing people. I don’t go out here trying to see who I can steal from. I don’t do none of that. When I don’t have it, I just don’t have it.

– 45-year-old Black female

Importantly, among those who secured housing, a drive to sustain improved housing conditions was a consistent motivator to reduce the frequency of substance use. There were several mechanisms that demonstrated how housing motivated reductions in drug use. In one case, the desire to use less was linked to respect for the needs of family. When the pandemic interfered with her ability to make enough money through informal sources to afford the room she had been renting, one participant moved in with her son, a shift which motivated her to reduce her use out of respect for his wishes.

But I mean I was using really heavy, real heavy and then when I moved in with my son it wasn’t like that. I couldn’t just go into my room and start shooting dope, I couldn’t do that, I had to respect him and respect his roommate and respect his home and respect myself basically.

– 55-year-old White female

Another mechanism which contributed to reductions in drug use was the role of physical comfort, which for one participant who had been experiencing homelessness for ten years before the pandemic, made the physically and emotionally taxing act of going out to purchase drugs less desirable. After years of moving between encampments and short-term shelters, he had finally been able to access longer-term housing during the pandemic, where he had been residing for seven months at the time of the interview.

And the opportunity [for housing] came and it happened to be a program, but I said, ‘You know what? I kind of need that, too, so I’ll take it.’ So, I’ve been there ever since. So, like I said, being up at this program really slowed me down a lot. You know, I am still using, but it’s definitely slowed me down a whole lot because of how far away it is, and I’m always on my feet walking. It takes me about 40 minutes to walk down to my program, 40 minutes to walk back… So, when I get back, I don’t feel like going out… I realized all the work I got to do to get what I want to get. It’s not even worth putting time and effort in, so knowing I got a nice warm room and I don’t feel like going out in the cold, kind of a blessing, you know what I mean? So, I’m like, ‘Shit, I ain’t got to go nowhere and worry about freezing my ass off. I’m going to stay in today.’ You know what I mean? So, it definitely slowed me down a lot being up in this room.

–38-year-old White male

For others, secure housing fostered a sense of personal responsibility and value on priorities other than drug use. This was demonstrated by a participant who’d been placed in his own apartment through a local non-profit social service agency after 25 years of experiencing homelessness and living on the streets.

Participant: “I don’t do drugs like I used to. I do it once in a while, but not like that. Not like I used to.”

Interviewer: So how frequently were you using drugs before, and how frequently now?

Participant: Before, it was every day. Now, it’s maybe like once a week, maybe.

Interviewer: And can you tell me a bit about this? Why has your use decreased?

Participant: I got me a place to live now. I got more responsibility… make sure my rent is paid, food, and body washed, soap, laundry detergent, laundry machine, you know? Things like that. – 57-year-old White male

Finally, for some, the psychological pressure not to risk eviction from stable housing led to decreased use, as was the case for a participant who had been placed in a long-term shelter at the start of the pandemic.

Participant: I would say [I’m using] less than before the pandemic started. Yeah, I would say less smoking, and smoking and snorting was the methods that I was using. And as far as heroin, maybe 50 percent less and as far as cocaine, probably 30 to 40 percent less than before the pandemic.

Interviewer: And can you talk a little bit more about the change in frequency, like why you’re using less? I know you’ve shared it was hard to get out. Are there any other things that you think influenced how frequently you’ve been using?

Participant: No, like I said before, just not wanting to lose my place here in the shelter because I have a tendency that when I use, I’ll stay out for a couple of days or go on a binge and that would jeopardize my standing here in the shelter, so I think that was a motivating factor. –68-year-old Black male

Discussion

These data describe the ways in which the shifting social and economic context caused by the COVID-19 pandemic impacted the individual experiences of individuals who use drugs with a history of recent injection drug use, a uniquely vulnerable population of people experiencing homelessness. The structural changes brought about by the pandemic created challenges through diminishing resources, both in terms of income generation and access to services, which were amplified by community-level stigma. Further, changes to local housing policies and programs enacted during the pandemic to reduce congregate living had an inconsistent impact: while limits in shelter beds forced some to sleep outdoors, others accessed stable housing, which in turn contributed to decreases in substance use. The experiences of these ten participants can help expand our understanding of the unique needs of this vulnerable sub-population of people experiencing homelessness and can be used to guide programmatic and policy changes specifically designed to meet these needs.

In particular, evidence is mixed on the impact of housing interventions on people experiencing homelessness who have substance use disorders. While many housing advocates promote the HF model as an evidence-based approach to improve outcomes among individuals experiencing homelessness, evaluations of the program’s impact on substance use outcomes remains inconclusive (Aubry et al., 2020; Baxter et al., 2019; Collins et al., 2012; Padgett et al., 2011; Tsemberis et al., 2012). Given this mixed evidence, several evaluations of HF programs call for more nuance in understanding the housing and social service needs of unique sub-populations of people experiencing homelessness, suggesting that there may not be a ‘one-size fits all’ approach to housing interventions (Baxter et al., 2019; Woodhall-Melnik & Dunn, 2016). Importantly, many of the studies examining HF’s impact on substance use were not conducted among individuals recruited on the basis of “active” or “severe” addictions (Kertesz et al., 2009). This makes it difficult to draw explicit conclusions about the program’s impact on the broader population of people who use drugs and highlights the need to examine housing program impacts among more specific sub-populations.

In the case of our sample, many participants experienced changes in housing following local efforts to mitigate COVID-19 transmission by moving people into non-congregate shelters such as hotels, enabling physical distance and reducing COVID-19 transmission risk (Kim, 2022; Rosecrans et al., 2022). Studies examining the impact of similar programs among individuals experiencing homelessness in other US cities during the pandemic identify that participants report a sense of stability, privacy, autonomy, and control from access to their own rooms, all of which are associated with improved mental and physical outcomes among participants (Robinson et al., 2022). Our findings suggest that for people with more severe or active substance use disorders, housing access may engage these mechanisms to disrupt the pathways linking homelessness to substance use.

Given that several study participants who had been engaged in active injection drug use reported using less or cessation of drug use in response to this housing access, it is crucial to examine the mechanisms through which these changes occurred to guide future policies and programs. Specifically, these participants highlighted physical and cognitive changes associated with housing access, such as comfort, warmth, dignity, responsibility, independence, and agency. These physical and mental conditions are associated with increased self-esteem, coping, resilience, and self-efficacy, all of which are psychological mechanisms associated with decreases in adverse mental health outcomes such as depression, anxiety, and Post Traumatic Stress Disorder symptomatology, which in turn are associated with decreased substance use (Gamarel et al., 2019; Jose & Novaco, 2016; McGarty et al., 2021). This suggests that housing access serves an important role in impacting the psychosocial conditions needed to support individuals to reduce substance use. Some studies have begun to characterize people experiencing homelessness into distinct well-being profiles, with experiences such as healthcare discrimination, comorbid mental health disorders, or histories of childhood adversity found to be associated with subsequent levels of well-being (Mejia-Lancheros et al., 2022). Other research has categorized individuals experiencing homelessness by their level of need for mental health support services (Stergiopoulos et al., 2019). It is suggested that the many complex and intertwined relationships between these and other mechanisms associated with well-being or unmet mental health needs may help to explain how HF and other interventions demonstrate differential impacts among sub-populations of people experiencing homelessness (Mejia-Lancheros et al., 2022). This concept merits further research attention, particularly given the inconclusive evidence on the relationship between HF interventions and certain psychosocial outcomes, including substance use indicators.

While low barrier programming is considered by many to be a best practice, one participant in this study indicated that the perceived risk of eviction once he was provided housing served as a motivator for him to cut back on his drug use. This suggests that some individuals may prefer the more structured or restrictive programming associate with Treatment First models. Accordingly, it is likely that there is not one single approach best suited to support people experiencing homelessness with active or severe addiction. Rather, we suggest a tailored, person-centered approach to housing which engages individuals in the decision-making process and offers programming designed to best meet each person’s unique psychosocial needs.

Alternative approaches are also merited to bolster the effects of housing interventions and support improved psychosocial outcomes among members of this population. Specifically, research shows promising effects of some psychosocial, non-pharmaceutical interventions in improving mental health outcomes among people experiencing homelessness (Hyun et al., 2020). Examples include cognitive behavioral therapy, motivational interviewing, and mindfulness-oriented recovery enhancement. One such study among individuals experiencing homelessness with a substance use disorder found that relaxation response training and acupuncture therapy were associated with decreased levels of anxiety and improved overall mental health (Chang & Sommers, 2014). Other studies have explored the role of community integration as a key outcome with the potential to improve the well-being of people experiencing homelessness, a construct comprised of three dimensions: physical, social, and psychological integration (Wong & Solomon, 2002), which have shown positive associations with housing stability, mental, and physical health outcomes among this population (Gaetz, Dej, Richter & Redman, 2016). A 2019 systematic review on a diverse range of housing interventions for individuals experiencing homelessness found mixed impacts of a range of housing programs on these elements of community integration, concluding that housing interventions should more directly focus on supporting the development of community integration among participants to best support people experiencing homelessness (Marshall et al., 2020).

Our findings also suggest the need for expanded economic policies and programs designed to support people who use drugs and are experiencing homelessness. We found that during the pandemic, informal channels of procuring money that had been reliable means of income generation became challenging and contributed to ongoing housing instability. One solution which may offer relief from economic distress among members of this population is cash assistance programs. In recent years, several pilot projects have explored the potential effectiveness of cash transfer programs to support people experiencing homelessness (Von Wachter, Bertrand, Pollack, Rountree & Blackwell, 2019). Such programs provide direct cash assistance to individuals, often free from conditions or restrictions on its use. Substantial evidence demonstrates that within low and middle-income countries, unconditional cash transfer programs have been associated with increased economic and psychological well-being, physical health, and household purchasing power (Homelessness Policy Research Institute, 2021). Yet engagement of this approach to reduce homelessness within the North American context remains limited. Currently, several pilot studies are underway (Sehgal & Ark, 2022; VandenEinde, 2022), and early research has demonstrated more rapid transitions into stable housing and fewer days spent unsheltered following one-time cash transfers to people experiencing homelessness (Dwyer et al., 2023; Foundations for Social Change, 2021; The New Leaf Project Society, 2019, Evans et al., 2016). However, some programs attach conditions to the funds or engage more restrictive eligibility criteria, potentially limiting individuals with substance use disorders from benefitting (Homelessness Policy Research Institute, 2021). Further, research has not yet explored potential adverse impacts of cash assistance programs on individuals experiencing homelessness: for example, that such cash infusions could be characterized as income and thereby make it difficult for some people to quality for income-based housing assistance. Given the promising evidence, though, additional programs of this nature and rigorous impact evaluations are merited to support future expansions of cash assistance programs to improve outcomes among people experiencing homelessness, particularly among those with substance use disorders. Yet further evidence is needed to determine the effects of this type of intervention in the face of disruptions faced by people experiencing homelessness during periods of social upheaval.

While economic and housing interventions can be tailored to support individuals experiencing homelessness who use drugs, their impact may be impeded by ongoing societal-level stigma and discrimination. Evidence suggests that people who hold stigmatizing beliefs about people who use drugs or other vulnerable populations are less likely to support programmatic or political investments designed to help stigmatized groups (Clair et al., 2016; Kennedy-Hendricks et al., 2017; Tsai et al., 2019). In this study, perceived stigma and discrimination towards people experiencing homelessness with a history of recent injection drug use amplified the economic challenges experienced during the pandemic. Study participants perceived that they were viewed by others as “dangerous” or “spreaders of disease” and therefore more likely to carry and transmit COVID-19. This aligns with prior research among PWID in Toronto, Canada where perceived stigma was believed to contribute to bystanders’ refusal to help in an overdose event and increases in crime perpetrated towards PWID (Bowles et al., 2023). Taken together, this research demonstrates the ways in which stigma and discrimination, particularly when amplified by external societal factors, contribute to resistance to help people experiencing homelessness who use drugs, and may therefore impede community-level efforts or political support for programs and policies designed to support members of this population.

Accordingly, programs and interventions designed to reduce stigmatization and discrimination towards people experiencing homelessness and PWID are urgently needed. Our findings support a recent call on the importance of shifting away from the view of both homelessness and substance use as individual-level problems, moving instead to a focus on the wider social, structural, and contextual factors that contribute to their prevalence (Thomas & Menih, 2022). Group-level stigma reduction interventions could be designed to reduce stigmatizing attitudes towards those who use drugs or experience homelessness held by service providers, law enforcement officials, and the general community (Heijnders & Van Der Meij, 2006). For example, stigma-reduction programs within healthcare settings often involve efforts to challenge and reduce stigmatizing attitudes and norms at an institutional level, engaging practitioners to identify ways in which their own knowledge and behaviors can reduce negative treatment towards clients or patients and create a culture of acceptance (Corrigan & Matthews, 2003). Others include workshops to raise awareness about stigma and discrimination and identify an organization-level responsibility to protect people impacted from a stigmatized condition or behavior from discrimination (Busza, 2001). Research further suggests that building individual connections between members of marginalized groups and those in dominant social roles is a means by which to dismantle stigmatizing societal beliefs (Canham et al., 2021; Clair et al., 2016). Reducing stigmatizing attitudes at an individual level is a key mechanism to facilitate the broader policy and programmatic changes needed to support vulnerable communities, including PWID and people experiencing homelessness (Tsai et al., 2019).

These results must be interpreted in the context of the study’s limitations. Most notably, the CHANGES study was not designed to specifically explore experiences of homelessness during the pandemic. Therefore, the analytic sample for this analysis only comprised ten participants. While this is a small sample, it consists of a highly structurally marginalized sub-population of people experiencing homelessness who are considered especially hard to reach with both research activities and direct services. It is therefore crucial to enumerate their unique experiences, particularly given the dearth of evidence on effective housing interventions specifically designed to support this population. Further, while the concept of validity is not considered compatible with qualitative research methods, we assert that these interviews demonstrate both credibility (the extent to which our interviews capture the phenomenon of interest) and transferability (the ability to apply the knowledge we derived to other situations – in this case, to other people experiencing homelessness who inject drugs in urban, US areas) (Lincoln & Guba, 1985). Nonetheless, larger qualitative studies are merited to better understand the complex interactions between homelessness, resource access, housing, and substance use in the context of COVID-19. In addition, Baltimore, Maryland is an urban environment in which multiple social service agencies have been established specifically to support people experiencing homelessness. It is therefore likely that the experiences of individuals experiencing homelessness in more rural US areas differ.

Finally, while we began to characterize experiences based on the overlapping marginalized identities of persons who inject drugs and those experiencing homelessness, we were unable to explore differences based upon other characteristics, including sex, gender, age, or race/ethnicity in our analytic sample. We recognize that individuals possessing additional minoritized statuses, such as racial/ethnic minorities, older adults, or sexual/gender minorities face additional compounding stressors. Future studies should thus engage an intersectional stigma lens to understand the experiences of people experiencing homelessness with a history of drug use who face additional compounding minoritized social identities (Thomas & Menih, 2022). We further call for future research which specifically centers individuals with lived experience of homelessness and substance use, as well as other stigmatized social identities, to help inform the study’s underlying research questions, interview guides, and analysis process, in order to engage principles of community-based participatory research and ensure research priorities are aligned with the perceived needs of members of the community they are seeking to serve.

Conclusions

These findings provide important insights into the impact of COVID-19 on people experiencing homelessness with a history of recent injection drug use, some of society’s most marginalized. We highlight the ways in which shifting socioeconomic conditions impacted the daily lives and experiences of members of this population, limiting resource access and opportunities to generate income and, in some cases, facilitating secure housing. These changes influenced substance use behaviors and, notably, among those who procured stable housing during the pandemic, decreases in drug use were explicitly linked to these housing changes. We suggest that there may not be one “best practice” to support people experiencing homelessness who have substance use disorders, and support calls for more research on the impact of HF programs among unique sub-populations. However, these findings highlight the pathways through which low barrier housing access can lead to substance use reduction among this sub-population of people experiencing homelessness, suggesting the potential benefit of this program among people with an active addiction. We also demonstrate the perceived amplification of stigma and discrimination experienced by members of this population during the pandemic. Reducing stigmatizing attitudes at both a community and individual level will be critical to facilitate the broader policy and programmatic changes needed to support vulnerable communities, including PWID and other vulnerable sub-populations of people experiencing homelessness.

Biographies

Abigail K. Winiker is a PhD candidate in the Department of Health, Behavior and Society at the Johns Hopkins Bloomberg School of Public Health, where she also completed an MSPH. Her work focuses on the design and implementation of harm reduction-based behavioral interventions targeting vulnerable populations, including people who inject drugs. Her current research explores the intersections of trauma and violence with mental health and substance use outcomes, as well as the moderating role of resilience.

Eshan U. Patel, MPH is a PhD candidate in the Department of Epidemiology at the Johns Hopkins Bloomberg School of Public Health. He is an epidemiologist who primarily conducts research on the epidemiology, detection, prevention and control of infectious diseases, including HIV, viral hepatitis, other sexually transmitted infections, and COVID-19. Eshan is particularly interested in using research to understand and address social determinants of infectious disease transmission and prevention among key populations such as people who inject drugs.

Becky L. Genberg, PhD, MPH is an Associate Professor of Epidemiology at the Johns Hopkins Bloomberg School of Public Health. She is an epidemiologist who uses research to understand and intervene on the social determinants of health behaviors and outcomes related to infectious disease and substance use.

Jennifer Ching, MSPH completed her MSPH in the Johns Hopkins Bloomberg School of Public Health in the Department of International Health. She is currently serving in the Peace Corps in Botswana.

Catherine Schluth, BS is currently pursuing a Master of Science (ScM) in Infectious Disease Epidemiology at Johns Hopkins Bloomberg School of Public Health. Her intended focus for her graduate thesis and subsequent research career is HIV/AIDS among key populations such as adolescent girls and young women, sex workers, transgender people, and people who inject drugs. Catherine has engaged in several years of research in infectious disease epidemiology.

Shruti H. Mehta, PhD, MPH is a Professor of Epidemiology and Vice Chair for Research and Administration in the Department of Epidemiology at the Johns Hopkins Bloomberg School of Public Health. She is an epidemiologist whose primary research interests include working with hard-to-reach populations to understand the epidemiology, natural and treated history of HIV, hepatitis C virus (HCV) and HIV/HCV co-infection. She has a special interest in identifying and overcoming barriers to care and treatment of HIV and hepatitis C virus among people who inject drugs and men who have sex with men as well as their sexual partners in both Baltimore and international settings.

Gregory D. Kirk, MD, PhD, MPH is a Professor of Epidemiology and Vice Dean for Research at the Johns Hopkins Bloomberg School of Public Health. He is a physician-scientist and epidemiologist whose primary research focuses on understanding the natural history of viral infections, particularly HIV and the hepatitis viruses, in both domestic and international settings.

Suzanne M. Grieb, PhD is an Assistant Professor in the Center for Child and Community Health Research (CCHR) in the Johns Hopkins School of Medicine and holds a joint faculty appointment in the Department of Health, Behavior and Society at the Johns Hopkins Bloomberg School of Public Health (JHSPH). Her research focuses on utilizing community engaged research to develop and implement interventions that address social determinants of sexually transmitted infections (STIs) among minority and underserved populations.