High willingness to use overdose prevention sites among suburban people who use drugs who do not inject

Introduction Overdose prevention sites (OPS) are evidence-based interventions to improve public health, yet implementation has been limited in the USA due to a variety of legal impediments. Studies in various US settings have shown a high willingness to use OPS among urban and rural people who inject drugs, but data among people who use drugs (PWUD) via non-injection routes of administration in suburban areas are lacking. Methods We utilized cross-sectional data from a sample of suburban PWUD who have not injected drugs in the past 3 months (N = 126) in Anne Arundel County, Maryland. We assessed PWUDs’ likelihood of using a hypothetical OPS and perceived potential barriers to accessing OPS. We tested for associations between sociodemographic characteristics, drug use, service access, and overdose experiences with willingness to utilize OPS. Findings Participants’ median age was 42, and the majority were men (67%) and non-Hispanic Black (79%). Sixty-six percent reported willingness to use an OPS. Concerns about confidentiality (29%), arrest (20%), and transportation costs (22%) were the most anticipated barriers to using OPS. Men (75% vs 55%, p = 0.015), participants who used heroin (53% vs 32%, p = 0.017), and participants who used multiple overdose prevention behaviors (e.g., using fentanyl test strips) (36% vs 19%, p = 0.006) were more likely to report willingness to use OPS. Conclusion Most suburban non-injecting PWUD in the sample were willing to use an OPS. OPS implementation strategies in suburban settings should be tailored to reach PWUD via non-injection routes of administration while meeting the unique needs of suburban contexts.


Introduction
Overdose rates surged during the COVID-19 pandemic; a record-breaking 100,000 people died from drug overdoses between April 2020 and April 2021 in the USA, a nearly 30% increase from the prior 12-month period [1].This persisting crisis is shaped by a number of intersecting socio-structural factors that increase likelihood of overdose including changes in the drug markets (e.g., the pervasiveness of fentanyl) [2,3], impacts of the COVID-19 pandemic (e.g., harm reduction and drug treatment services interruptions), structural vulnerabilities (e.g., housing and food insecurity) experienced by people who use drugs (PWUD) [4], and drug criminalization [5,6].PWUD lacking safe, private locations may rush drug use to avoid police encounters, increasing the likelihood of overdose.Others hesitate to or completely avoid contacting emergency services during overdoses for fear of arrest, making some overdoses more likely to be fatal [7,8].Long-term reductions in overdoses require harm reduction interventions, developed in partnership with PWUD, that address social and structural drivers of overdose mortality [9].
Overdose prevention sites (OPS), sometimes referred to as supervised injection facilities or safe consumption sites, are evidence-based interventions implemented throughout Europe, Australia, and Canada that can fundamentally alter this risk environment [10].OPS allow individuals to use previously acquired drugs under the supervision of medical personnel trained to reverse overdoses, provide harm reduction counseling, case management, and link PWUD to resources, such as primary care and drug treatment [10,11].Studies of OPS demonstrate reductions in overdose, infectious disease incidence, improper syringe disposal, and crime [12][13][14][15][16][17].Studies have documented high willingness to use OPS among PWUD in several locations throughout the U.S [18][19][20][21][22][23].Existing evidence on OPS willingness in the USA has focused on urban settings and people who inject drugs rather than people who use drugs via other routes of administration (e.g., smoking, snorting).Little is known about willingness to use OPS among suburban PWUD and among PWUD who do not inject.In this study, we asked a sample of PWUD who do not inject to about their willingness and potential barriers to utilizing an OPS in Anne Arundel County, Maryland.

Study design
We used data from the Peer harm Reduction Of Maryland Outreach Tiered Evaluation (PROMOTE), a mixedmethods, cross-sectional study of people who use drugs (PWUD) in Baltimore City and Anne Arundel County (AAC), Maryland.Adjacent to Baltimore City, AAC has the third-most unintentional overdose deaths in the state and is largely constituted of suburban (more sprawling communities with limited access to public transportation) communities with limited harm reduction service coverage [24].In AAC, PWUD were recruited from seven street-based locations between November 2019 and March 2020; data collection ended prematurely due to the COVID-19 pandemic.We created heat maps of areas with high drug activity using drug arrest and overdose data from the Anne Arundel County and Annapolis City police departments.We extracted time signatures within these areas to develop a time-day-location sampling frame.Based on the time signatures, the study team created monthly recruitment schedules and parked the study van in seven recruitment zones.Once the study team reached a recruitment zone, they recruited participants on foot or by talking to interested people who came by a table with harm reduction supplies setup outside the van.All study procedures occurred inside the study van to ensure confidentiality and privacy.Eligibility criteria for the AAC surveys required participants to report being at least 18 years old and non-prescription use of any opioid in the previous 6 months.Eligible participants provided informed consent and completed a 30-min Audio Computer-Assisted Self-Interview (ACASI) on a tablet with a trained interviewer present for technical assistance.Following the survey, respondents were compensated with a $25 USD VISA gift card.In total, 173 participants were recruited.This analysis is restricted to participants who did not report injection drug use in the past 3 months (n = 132) with complete data for the outcome measure (N = 126).This study was approved by the Johns Hopkins Bloomberg School of Public Health and the Maryland Department of Health Institutional Review Boards.

Willingness to use OPS
Each participant received the following description of OPS: "An overdose prevention site, also known as a safe consumption space, is a place where it would be legal for people to safely inject, snort, or smoke, or otherwise consume drugs that they buy somewhere else.You would not be arrested while in the site.There would be staff on site to respond to an overdose, and to provide basic medical care and referrals to health and social services upon request.While overdose prevention sites are not currently approved in the USA, they operate in several countries worldwide."Participants then indicated their likelihood of using OPS by selecting their response from a Likert scale (very likely, somewhat likely, somewhat unlikely, very unlikely), which was dichotomized for analysis (likely vs. unlikely), consistent with prior work [21].

Awareness of and barriers to using OPS
After receiving a description of OPS, participants shared whether they had previously heard of OPS (yes/no).
Participants were asked to select anticipated barriers to OPS access (yes/no) from a predetermined list of 10 options, including concerns about arrest, confidentiality, cost of transportation, and lack of interest.We summed endorsement of these 10 barriers to create a categorical variable for analysis (number of anticipated barriers: 0, 1, 2 +).

Sociodemographic characteristics
Sociodemographic characteristics included age (continuous), gender (man/woman; no participants reported non-binary gender identities), race/ethnicity (trichotomized: non-Hispanic white, non-Hispanic Black, other races (including Hispanic, Asian, Native American, and Multiracial), and educational attainment (dichotomized: high school diploma or G.E.D. vs. did not graduate high school).Structural vulnerabilities included currently experiencing homelessness (yes/no), recently experiencing food insecurity (dichotomized: weekly or more compared to less than weekly in the past 3 months) and arrest within the past year (yes/no).

Drug use characteristics
Participants were asked about their recent (past 3 months) use of eight substances (i.e., opioid pills, heroin, fentanyl, cocaine, crack, benzodiazepines/tranquilizers, synthetic cannabinoids, and PCP) and method of use (i.e., smoking, snorting, swallowing).We generated binary variables to indicate any recent non-injection use of each drug and created a dichotomous variable to identify participants who had recently used three or more substances vs. less.Lifetime history of any drug injection was captured as a binary variable (yes/no).Participants selected their typical locations for drug use from a predetermined list including: your or someone else's home, street/park, abandoned building, shooting gallery, car/ vehicle/bus/metro, stairwells, public bathroom, other.Other locations were then specified by participants using free text.We then created a categorical variable for public/semi-public (e.g., shooting galleries, abandoned buildings, streets, or parks) and private locations (e.g., at home or at someone else's home), consistent with our prior work [7].

Overdose experiences and prevention behaviors
Lifetime and past 6-month overdose experiences were assessed with binary variables (yes/no).Participants were also asked what they do to prevent overdose (yes/ no) from a list of six behaviors: buying from the same dealer; using fentanyl test strips; keeping naloxone with me; using with other people; calling or texting a friend to let them know what I'm doing; and using in a place where someone will see me if I overdose.We summed endorsement of these items to generate a categorical variable (number of overdose prevention behaviors: 0, 1, 2 +).

Analysis
Covariates of interest were selected a priori based on prior work and hypothesized associations between covariates and willingness to use OPS.We compared differences in covariates of interest by OPS likelihood using Chi-squared tests for categorical variables with cell sizes > 10, Fisher's exact tests for categorical variables with cell sizes ≤ 10, and two-sample Mann-Whitney tests for continuous variables.Descriptive analyses were conducted in Stata SE 15.1.

Results
Participants' median age was 42 years and most reported being men (66%) and non-Hispanic Black (79%; Table 1).Sixty-five percent of participants had received a high school diploma or equivalent.Many participants reported experiencing structural vulnerabilities, as half (52%) were currently unhoused and 39% experienced recent food insecurity.One-third (31%) had been arrested in the past year.Most participants had never injected drugs (89%).The most prevalent non-injection drugs recently reported included opioid pills (66%), crack (53%), heroin (44%), fentanyl (42%), and cocaine (41%).Multiple substance use was common, with 64% reporting recent use of three or more drugs.Thirty-five percent of participants reported having ever overdosed, and 14% reported having an overdose in the previous 6 months.Participants reported implementing several overdose prevention behaviors.The most common strategies included buying from the same dealer (29%), using in a visible location (25%), and using fentanyl test strips (18%).Forty-four percent had heard of OPS before it was described to them, and 58% reported being likely to use OPS if available (37% very likely, 21% somewhat likely, 9% somewhat unlikely, 33% very unlikely).The most common perceived barriers to using OPS were concerns about confidentiality (29%), cost of transportation (22%), and concerns about arrest (20%).
Participants willing to use OPS were significantly more likely to be men (74% vs 55%, p = 0.024) and to have recently used heroin (53% vs 32%, p = 0.017) than those unwilling to use OPS.Those more willing to use OPS also reported engaging in a significantly greater (2 or more) number of overdose prevention behaviors compared to those who were unwilling (33% vs. 11%, p = 0.006).Overdose prevention behaviors were significantly more common among those willing to use OPS than those who were not, including buying from the same dealer (38% vs. 16%, p = 0.014) and using fentanyl test strips (25% vs. 8%, p = 0.018).Those willing to use OPS anticipated a significantly greater number of barriers (2 or more) to OPS utilization compared to those who were unwilling (36% vs. 19%, p = 0.048).

Discussion
We found high willingness to use OPS among a sample of suburban PWUD who did not inject.Our finding adds to the growing evidence showing high OPS acceptability among people who inject drugs in urban and rural areas [18,21,23].Individuals who use drugs but do not inject are seldom included in discussions of OPS.Our findings speak to the demand for OPS among suburban people who smoke, snort, or swallow their drugs, highlighting the importance of expanding OPS framing to consider diverse PWUD beyond urban areas.Two-thirds of participants had previously overdosed and most engaged in behaviors associated with increased overdose risk, including public/semi-public [7] and polysubstance drug use [2].Consistent with prior work, heroin use was associated with OPS willingness [20,21].We previously documented limited awareness of naloxone access points and limited knowledge of legal protections for help-seeking during an overdose in this population [8].These findings demonstrate how PWUD but do not inject are at an increased risk for overdose because of social and structural marginalization.Participants who were likely to use OPS reported implementing overdose prevention behaviors, including using fentanyl test strips and buying from the same dealer.Participants also reported engagement with social supports to reduce overdose risk, including using around others or while on the phone with a friend.These findings demonstrate how PWUD without access to OPS are mobilizing to prevent overdoses, and structural interventions, like OPS, are needed to ensure ongoing access to harm reduction services to mitigate overdose risk.
Many PWUD anticipated barriers to accessing an OPS, and anticipated barriers were more common among those willing to an OPS than those who were not.Commonly reported concerns included transportation cost, confidentiality, and arrest, consistent with previous research in urban settings [18,21].However, approaches to eliminating these barriers should be tailored to meet needs of PWUD in suburban environments, where there may be greater travel barriers (farther distances to navigate and lack of public transport), less population density, and variation in PWUD/law enforcement relationships.Overcoming transportation barriers for suburban and rural PWUD requires structural approaches that can benefit health service access beyond just OPS access.Vouchers for taxis, dedicated shuttle services, and improving public health infrastructure all have a role in overcoming this key barrier.Input from local constituents-including PWUD who do and do not inject-should be at the forefront in shaping decisions around OPS implementation.

Limitations
Data collection was stopped prematurely due to the COVID-19 pandemic, resulting in a modest sample size that limited our ability to explore additional correlates or perform adjusted analyses.The survey also did not capture methamphetamine use or differentiate between overdoses caused by opioids versus other drugs.It is also possible that participants did not fully understand the description of the OPS presented in the survey, which may have influenced their responses to questions about willingness to use an OPS or any barriers.It is also possible that relevant barriers to PWUD were not included in the survey question and are therefore missing from this analysis.Participants may have also had different levels of previous awareness of OPS that were not captured in our measures that could have influenced their perceptions of barriers or willingness to use an OPS.Future qualitative work is warranted to further explore and verify our interpretations around this topic.Additionally, our findings  reflect PWUD in suburban Maryland; demand for OPS and anticipated barriers to access may differ for people in other settings.

Conclusions
We found high OPS willingness among suburban PWUD who do not inject, extending the growing evidence of high demand for OPS throughout the USA among diverse populations of PWUD.Participants anticipated several barriers to OPS access.While many of these barriers are consistent with existing literature, experiences of these barriers may differ between PWUD who do and do not inject in urban, suburban, and rural settings.OPS expansion in the USA should facilitate access for diverse PWUD and tailor strategies to specific environments and structural contexts.

a
Two-sample Mann-Whitney test b Fisher's exact test -Significance tests are not presented due to zero cells

Table 1
Factors associated with OPS willingness among PWUD who do not inject in Anne Arundel County, Maryland, 2019-2020 (N = 126)