We conducted in-depth, one-on-one qualitative interviews with ED providers, ED patients who had received Relay services, and Relay peer “Wellness Advocates” (WAs). The interviews aimed to examine barriers and facilitators to implementation of Relay from multiple perspectives as part of a larger clinical trial that has been described previously (Doran et al., 2023). The study was approved by the NYU Grossman School of Medicine Institutional Review Board. Methods and results are presented in accordance with the Consolidated Criteria for Reporting Qualitative Research (COREQ) (Tong et al., 2007).

The NYC Health Department’s Relay initiative currently operates in 15 NYC EDs (13 at the start of the study) and has been described previously (Welch et al., 2019). In brief, Relay sends trained peer Wellness Advocates (WAs) to engage with patients in the ED who have presented after a nonfatal opioid-involved overdose. WAs are generally required to have recovery coach certification; to be a Certified Recovery Peer Advocate (CRPA/CRPA-P), a Certified Addiction Recovery Coach (CARC), a Credentialed Alcoholism and Substance Abuse Counselor (CASAC/CASAC-T); or have 2 years of commensurate peer experience. All WAs complete a 4-day new hire orientation, and receive ongoing trainings (e.g., in specific topics such as motivational interviewing) and support through regular team and individual supervision.

ED providers call a Relay “hotline” to activate a WA, who generally arrives at the ED within one hour. Relay operates 24 hours per day, 365 days per year. Relay is grounded in the principles of harm reduction (National Harm Reduction Coalition, 2024). All patients engaged by WAs in the ED are offered immediate peer support, a naloxone kit and training, personalized overdose risk reduction education, and referrals to treatment and social services. For patients who consent to receive post-ED follow-up services, WAs provide ongoing support for up to 90 days after the ED visit. WAs offer support and resources across multiple realms, depending on the needs and interests of the patient, including harm reduction; substance use treatment; peer/social support; connections to social services (e.g., housing, education, employment); medical, mental health, and holistic care navigation; and assistance meeting material needs (e.g., obtaining a telephone or transportation assistance). While the Relay program model encourages in-person engagement when possible, most post-ED follow-up visits occur by telephone; this was particularly the case during the study period, which overlapped with the COVID-19 pandemic. Follow-up contact plans are intended to follow a Critical Time Intervention (Herman et al., 2011) model, with an initial high intensity of contacts tapering to fewer contacts as the follow-up period concludes (goal of at least 5 contacts in the first 45 days including a first contact ideally within 24–48 hours of ED discharge, at least 3 contacts in the second phase (days 46–75), and at least 2 contacts in the final phase (days 75–90). Notably, Relay WAs face many challenges reaching patients for follow-up, especially given the large proportion of patients experiencing homelessness. These challenges were exacerbated by the COVID-19 pandemic which limited WAs’ ability to meet patients in-person and conduct in-person outreach for patients without access to a telephone. As such, the exact number of Relay WA follow-up contacts varies, with many patients receiving no or few post-ED contacts.

For this study, we included participants from the four NYC EDs that were participating in a randomized controlled trial of Relay (Doran et al., 2023). These four EDs span three different boroughs and health systems in NYC and had participated in the Relay initiative for at least a year prior to the start of the study. We recruited ED providers and ED patients who had interacted with Relay from the study EDs, as well as Relay WAs and WA supervisors (together referred to in this paper as “WAs”) working with the study EDs. All participants had to be ≥18 years old and speak English. ED providers had to have engaged with the Relay initiative at least once in the past year (n=196 providers across the 4 EDs met this criteria). ED patients were selected from Relay intervention arm participants enrolled in the aforementioned RCT (there were n=112 Relay RCT arm patients who spoke English). Patients recruited for qualitative interviews had to have enrolled in Relay within the past 6 months and be reachable by phone. Also, after the initial round of interviews we added a selection criterion that patients should have had at least 2 contacts (either by telephone or in person) with a Relay WA, because we wanted to interview those patients who could speak about their experience with the full Relay model (including the ED visit interaction and post-ED interactions). WA participants had to have worked with Relay for at least 6 months and have experience responding to one of the study EDs for at least three consecutive months (n=26 WAs met this criteria). Lists of all individuals meeting these criteria for each group were compiled or obtained by the study team. Random sampling was then used to select individuals from these lists to invite for qualitative interview participation. By design, random sampling was later supplemented with purposeful sampling to strategically fill any gaps in key perspectives as needed (for example, in participant gender or ED provider role).

Target sample sizes were set for each subgroup, with flexibility allowed as needed to ensure theoretical saturation (the point at which no new significant concepts were arising from subsequent interviews) was met (Patton, 2002).

Study staff reached out to potentially eligible individuals by email, text, or phone call using the sampling scheme described above. Eligible individuals were given information about the study and invited to participate. Verbal informed consent was conducted immediately before the interview for ED providers and WAs. ED patients had provided written consent for qualitative interviews at the time of their enrollment in the RCT, but their consent was confirmed at the time of the interview. All participants received $50 cash gift card compensation for completing the interview. All participants completed a single interview (i.e., no repeat interviews were conducted).

All interviews were conducted either using the audio function of a secure web conferencing program or by telephone. Interviews were one-to-one (one interviewer and a single participant). All interviews were conducted by RF, a trained qualitative researcher with extensive experience conducting qualitative interviews and topical expertise related to substance use and harm reduction. The interviewer introduced themselves as a member of the university-based study team, specifying that they were not affiliated with Relay.

All interviews were digitally audio recorded and transcribed by a professional transcription company, with transcriptions additionally checked for accuracy by a member of the study team.

The interviewer used a semi-structured interview guide, which ensured that key concepts were covered in all interviews yet provided flexibility to allow for additional probing questions as needed and to maximize interview flow. Interview guide questions included Consolidated Framework for Implementation Research (CFIR) (Damschroder et al., 2009) domains of individual characteristics, intervention/innovation characteristics, inner setting, and outer setting as relevant to the implementation and effectiveness of the Relay intervention and the specific interviewee type. Separate interview guides were made for each of the three types of interviewees (ED patients, ED providers, and WAs) given their unique experiences and points of reference from which they could offer insight about Relay. Draft interview guides were workshopped with Relay WAs, Relay program leadership, and ED providers; they were edited based on this feedback, and additional small edits were made for clarity after initial interviews. Final interview guide details are provided in Table 1. At the conclusion of each interview, the interviewer recorded participant self-reported demographics and “field notes” with any additional pertinent observations.

Early analyses used rapid turn-around qualitative methods including creating templated summaries of interviews and developing a summary matrix table (Hamilton, 2013; Hamilton & Maietta, 2016) focused on key domains organized by CFIR construct to identify barriers and facilitators to implementation. Initial template summaries were created independently by RF and TC (the study’s project coordinator, who was trained in qualitative research techniques), and cross-checked to ensure accuracy and consistency; subsequent template summaries were completed independently by TC. A summary matrix, which distilled information from template summaries to facilitate comparisons across interviews, was drafted by TC and checked by LG (a medical student with experience working in EDs who was trained in qualitative research techniques) and KMD (an emergency physician and researcher with significant training and experience using qualitative research methods).

Additionally, line-by-line coding was conducted for all interview transcripts. All interview transcripts were coded independently by three researchers (TC, LG, RF, and/or KMD), with coding discussed and reconciled in group coding meetings. Coding initially used an a priori code list developed based on study goals, salient concepts, and CFIR constructs (i.e., inductive coding). New codes were also allowed to emerge organically from the text (i.e., deductive coding) in the grounded theory tradition (Patton, 2002); a final code list emerged after three coding meetings. Dedoose qualitative research software was used to assist with data organization. Participants were not re-approached for their feedback on the codes or themes, but the authorship team for this paper included physicians who work in EDs that offer Relay services (EC, IW) and NYC Health Department Relay initiative employees (AJ, AEW, DC, KLK) who were able to evaluate the face validity of the findings.

Interviews highlighted the multiple competing demands emblematic of the ED environment (inner setting), which both underscored the utility of Relay and led to some practical challenges for its delivery. Primary among competing demands was the lack of time ED providers had to spend with their patients which limited their capacity to have in-depth discussions or provide interventions for those presenting after an overdose. As a result, ED providers overwhelmingly felt that Relay was useful because it addressed this need by employing a dedicated person to interact with patients.

I think they [Relay WAs] help a gap ‘cause they’re willing to sit and talk to the patient. They’re there for a long time. …I don’t have that luxury to sit and talk to the patient like that.

While limitations in ED provider time underscored the usefulness of the Relay intervention design (which uses dedicated NYC Health Department staff WAs rather than relying on existing ED providers to offer an intervention), this inner setting factor also sometimes presented practical barriers to Relay’s implementation. Though Relay is designed to require as little ED provider time as possible, even the few minutes it took for providers to call the Relay hotline was sometimes noted as an obstacle to fully utilizing the program:

I think the whole process of activating Relay probably takes, if I’m being honest about it, somewhere between three and four minutes. Which seems like nothing. You’re like, ‘Oh, what’s three or four minutes to help somebody?’ But in reality, three or four minutes is a really long time in the Emergency Department when you’re that busy.

Patients themselves were also aware of the competing demands for time and attention in the ED, noting that they felt like less of a priority amidst these other demands. For example, one patient recalled:

They looked at me like, you’re just another drug addict trying to hustle the system, and we don’t have time for that. This is an emergency room, and this is for people that chop their arms off or break their legs. That’s what this is. That’s what this facility is.

At the same time, some ED patients had their own competing demands that interfered with participation in the Relay intervention, such as being in a state of withdrawal (for example, if they were given a large dose of naloxone after their overdose), scared, or wanting to leave the ED for other reasons. As summarized by one WA:

You have people who are still under the influence…or the symptoms of coming out of their overdose, whether it’s they’re irritable, they’re tired, they’re frustrated, they’re scared. It’s really the full gamut of human emotion. Being in a hospital or emergency room alone is just a harrowing experience.

One beneficial element of the Relay intervention design noted by interviewees was that the WA could wait in or near the ED for a period of time until patients were alert and comfortable enough to talk. For example, one ED provider reflected:

I think that that’s part of the beauty of Relay is that the program allows for them to be where the patient is and just hang out until they’re ready to talk, to say somethin’, but to be able to jump in there. Whereas I can’t jump in there ‘cause I’m jumpin’ all over the place.

Finally, demands related to the ED physical environment presented a few practical challenges for Relay WAs including limited physical space and shared patient rooms; one WA commented, “You wanna zero in on that patient, that participant, and get their story, and sometimes it’s hard to get that story with all that happening” (ID27). Due to competing space needs, generally there was not a dedicated area in the ED for WAs to use.

Sometimes I feel like I am in the way. Usually, those situations are where the person, maybe they’re asleep or they’re not ready to be interviewed. I can’t really hang out there [in the ED]. I gotta go hang out somewhere else, and come back, and check. … I better get out of their way ‘cause it’s a busy place.

Overall, the implementation barriers presented by competing demands for time and space were seen as surmountable. WAs described using a variety of strategies to adapt to the competing demands in the ED, for example, working to develop relationships with ED providers to smooth operations and facilitate information exchange. WAs also described being mindful of the ED environment and “staying out of the way” while waiting for a patient to be ready to engage.

Participants commonly described the salience of WAs being peers with lived experience related to drug use and overdose in terms of their approach to and ability to connect with ED patients. This status of being a peer was commonly described as being distinct from and in contrast to the non-peer status of ED providers. The fact that WAs were peers was almost universally described in a positive light, and encompassed their ability to understand and develop rapport with patients, and the credibility they brought to discussions about overdose. One patient summarized, “She has a good understanding, like I’m talkin’ to a addict myself” (ID38). Another patient remarked:

She [the WA] had some experience herself. There was a connection. You know what I’m sayin’?…Somebody who hasn’t been there just doesn’t get it. …that’s like Nancy Reagan tellin’ the kids, “Just say no.”

ED providers noted that WAs were able to have higher quality interactions with patients than other ED staff could. For example, one ED social worker noted, “When we do have a peer advocate come in to meet with the patient I feel like they’re much more comfortable talking to someone who’s been there and not our attendings and not nurses” (ID1). This was confirmed in patient interviews, and as one patient reflected:

Actually, I felt quite comfortable with them, like kind of relieved, like it was someone to talk to and, like, to let the shit off my chest and just—I didn’t feel like I was by myself. I felt somebody was there for me.

As peers, WAs were also able to serve as role models and provide inspiration to patients based on their own lived experience, while meeting patients where they are and not being prescriptive about the exact path an individual should take.

It’s important to let them know…that you have a multilevel of experience to be able to empathize and sympathize with them, on both levels, and following up [with] them to move forward wherever they wanna go. If they don’t wanna move forward, they just wanna go side to side, well, how can we do that in a less harmful way.

The status of WAs as peers was described as especially important in the ED setting, where patients – particularly those with drug us and potentially frequent ED visits – were sometimes met with skepticism or stigma. As explained by one WA:

It could be… the person’s been there enough times they’re known in the emergency room. The doctors are tired of [them]. The nurses are tired of seeing them. I think as peers, we can relate. Yeah, it sucks to just be asked all these questions and still be told that you’re a liar, a junkie. …it’s a very cold response sometimes in the emergency room compared to, say, Relay, which can come with a lot of warmth or add a human approach to what could easily be an escalated experience for someone.

Overwhelmingly, ED providers reported that they valued the services provided by Relay in the ED. This was particularly true in the context of the competing demands noted earlier. For example, one ED provider described the value of having a dedicated person who could talk with patients:

For the most part, most people are so eager to have that human interaction with somebody, and here you have someone sitting down and talking to you who you could tell is there for you. … That person’s there for you, and that probably is nice to have that, someone to talk.

ED providers also described the high burden of need of their patient populations accompanied by relative resource scarcity (e.g., in terms of services available to patients), which further heightened their appreciation for Relay. At one ED in particular, where there was no overnight social work coverage, providers noted that Relay could fill in some gaps for patients after an overdose. As one ED provider reflected:

Especially in our patient population, especially, at our hospital, we’re one of the poorest zip codes in the state or we’re the poorest, I think. Whatever resources we can get, whatever help we can get in trying to educate our patients is very, very welcome.

The fact that ED providers valued Relay seemed critical for its successful implementation and ongoing operations because this motivated providers to call the Relay hotline and help facilitate patient interactions with WAs. For example, one WA noted:

A lot of times the doctors and the medical staff, they won’t discharge somebody until we’re able to get to them, you know? Because they realize that sometimes it’s so important to make sure that we engage with these folks….

Interestingly, while ED providers expressed their appreciation for Relay they also acknowledged they typically were not aware of patient outcomes. Further, some providers explained they did not know exactly what a WA-patient interaction involved and commented on the lack of information regarding services patients received from Relay following their ED encounters.

Regarding the ubiquitous lack of awareness of what happens to patients once they leave the ED – which is not unique to the Relay initiative – one provider explained:

I have no outcomes. The ED is the land of “we don’t know the outcome.” We never have any idea what happens with our patients. We do these things, and we’re like I hope they’re OK; they were OK when they left.

WAs also suspected that ED providers did not fully understand the extent of services provided by Relay after the ED visit. For example, one WA commented:

I don’t think that they [ED providers] understand the lengths that we go to. I think that when we’re there they respect us because of our personalities and how we respect the ED staff and they respect us at the time that we’re there. I don’t believe that they have any idea of what we—they [are] only aware of our initial engagement. …. After the ED, it’s just over as far as they’re concerned.

An interesting paradox emerged in the interviews: participants very commonly described how structural factors (e.g., homelessness, poor healthcare, and treatment access) were major challenges for patients who had experienced an overdose, yet at the same time their language often placed the locus of responsibility for “success” on the individual. First, related to the prominent role of structural factors, one ED provider noted:

We have a really large undomiciled population. … People don’t have working phone numbers. People don’t have addresses that are permanent. You know, it’s hard to say one way or another what the issues are, but those are definitely road blocks and barriers that we face in trying to connect these patients with the proper resources.

Interviews revealed that these structural factors were often challenging to address solely through an individually-focused intervention like Relay. One WA explained, “we only have 90 days, so that’s hard to do, to be a case manager and get somebody housing. It’s just not going to happen” (ID3). Additionally, structural factors made it more challenging for WAs to fully implement the Relay intervention as designed. WAs commonly reflected on difficulties in providing post-ED follow-up when patients were homeless and did not have working telephones. WAs described creative solutions to try to work through these challenges, such as trying to find patients at their shelters or common “hang out” places. However, sometimes housing instability and the lack of reliable telephone numbers made it impossible to engage Relay participants during the follow-up period.

While interview participants acknowledged the deleterious impact of structural factors such as widespread homelessness, they still commonly placed responsibility on the individual, sometimes even within the same segment of an interview. For example, one ED provider stated:

I think the difficulties just come down to the socioeconomics of the patients or support system outside of the hospital. At the end of the day, you can lead a horse to water; you can’t make it drink.

Patients also reflected on their drug use in ways that on the one hand acknowledged structural factors, and on the other hand attributed their substance use to individual traits. For example, one patient described his living environment:

It’s the supportive housin’… I don’t like the neighborhood. The neighborhood is drug-infested, and it’s crime-infested too… I’m livin’ in a buildin’ where they sell and use drugs. I don’t like walkin’ out my apartment building and then my apartment and smelling all these drugs and leavin’ beer bottles in front of my door….

In addition to the four described primary themes, we also observed an unsurprising meta-theme of stigma toward people who use drugs which manifested directly in interview content (e.g., WAs or patients describing poor treatment received in EDs) and indirectly in the language used across all participant types.

One patient described watching ED staff provide food to other patients while themselves being told there was no food available; the patient suspected this occurred “because I was a drug user. That’s what I’m thinkin’. Maybe” (ID18). Another patient remarked:

When I got to that hospital, it was like a bad dream. You think you’re gonna go to a hospital and you’re gonna get help. … They couldn’t do nothing for me because they looked at me as, I’m a drug addict. “Get the fuck out of our hospital,” and that’s it. “This ain’t a hotel.”

Another way that stigma manifested in ED provider interviews was in the feeling of pessimism about the potential for patients to “change,” and some WAs described attempting to confront this stigma among ED staff. As one WA recounted:

One nurse was like, “Oh, they never change. They da, da, da, da, da.” She was just ranting, so I turned to her, and I say, “Yes, but you coming at them like that is not gonna help them. You gotta treat them with dignity, respect, and compassion. People are gonna still do what they’re gonna do, but we just gotta keep educating them how to be safe so they don’t die. ‘Cause you’re here to save lives. I’m here to save lives. We’re in the same kinda work. We just do it differently.”

We did note, however, that occasionally WA interviewees and patients themselves used stigmatizing language when talking about people who used drugs (e.g., “addict,” “nasty.”) It is important to consider that Relay is being implemented within a broader environment in which attitudes (i.e., outer context) toward people who use drugs (i.e., the intervention recipients for Relay) remain marked by stigma.

In qualitative interviews with ED providers, patients, and Relay WAs, we found that: 1) EDs are characterized by multiple competing demands, underscoring the utility of Relay and leading to some practical challenges for its delivery; 2) There is strong role distinction of WAs as peers with lived experience; 3) ED providers value Relay, even though they have a limited understanding of its full scope and outcomes; and 4) While the role of structural factors is recognized, responsibility is often placed on patients for controlling their own success. We plan to use these findings to contextualize and interpret results of a recently completed randomized controlled trial of Relay. These findings can also help inform implementation of similar interventions in EDs outside NYC. As EDs and other settings nationwide seek new interventions to reduce the toll of overdose fatalities, implementation research is needed alongside effectiveness research to learn how to rapidly and successfully bring effective programs to scale.