Fentanyl preference among people who use opioids in New York City

Abstract

Background:

Several studies have pointed to a sub-sample of people who use drugs (PWUD) who prefer the use of non-pharmaceutical fentanyl (NPF) and, as such, pose a unique challenge for public health initiatives amidst the continued rise in overdose mortality. However, matters of drug preference and autonomy of choice remain under-studied and often misunderstood. This paper examined the experiences of PWUD reporting a preference for NPF or an NPF-heroin mixture, specifically how they navigate the perceived benefits of NPF and its established risks.

Methods:

22 in-depth interviews were conducted in New York City between March 2018 and August 2019 with PWUD who self-reported a preference for NPF or an NPF-heroin mix. Interviews were audio-recorded, and the resulting transcripts analyzed using a thematic approach.

Results:

Participants highlighted various factors that contributed to expressed preference for NPF or an NPF-heroin mix, including a desire to feel good, financial resources, drug availability, decreased consumer autonomy, and physiological demand. Participants reported practicing several risk reduction strategies; however, they highlighted that many, particularly carrying naloxone and always using with someone else, were difficult to implement in the context of illicit drug use.

Conclusions:

Our results demonstrate participants’ decreased consumer agency and greater exposure to systemic factors in the illicit markets, highlighting the need for expansion of various services, including drug checking resources and systems of outreach for PWUD who do not use intravenously. To promote tailored interventions, continued efforts in overdose prevention ought to more thoughtfully consider the context, perceptions, preferences, and behaviors of PWUD.

1. Introduction

In the 12 months ending in April 2021, there were an estimated 100,306 drug overdose deaths in the United States, the largest number of overdoses for a 12-month period ever recorded (Centers for Disease Control and Prevention, 2021). New York City (NYC) experienced similar trends in overdoses, with 2062 overdose deaths recorded in 2020, 77 % of which involved fentanyl (Nolan et al., 2021).

Fentanyl, an analgesic pharmaceutically utilized for pain management, is approximately 50–100 times more potent than morphine (National Institute on Drug Abuse, 2018). While fentanyl can be diverted for misuse, most fentanyl-related overdoses in the U.S. are linked to illegally manufactured, non-pharmaceutical fentanyl and its analogs (hereafter referred to as NPF) (U.S. DEA, 2017). Cheaper than heroin, NPF enters the illicit drug supply primarily as a heroin adulterant (Ciccarone, 2017), though it has been increasingly found in pressed counterfeit opioid pills (U.S. Drug Enforcement Administration, 2021), cocaine (Marinetti and Ehlers, 2014), and benzodiazepines (Arens et al., 2016). The presence, type, and amount of NPF in drug markets is largely unknown to consumers, placing individuals at increased overdose risk (Amlani et al., 2015; Ciccarone et al., 2017).

A growing body of literature has focused largely on identifying fentanyl “landscapes” (e.g., levels of exposure), overdose risk factors and prevention strategies, as well as people who use drugs (PWUD)’s experiences with NPF and responses to the expansion of NPF in their communities (e.g., Daniulaityte et al., 2019; Harris et al., 2021; McKnight and Des Jarlais, 2018; Silverstein et al., 2019). However, recent studies suggest a heightened consumer demand, with PWUD reporting preference and intentional use of NPF.

As drug markets shift toward more novel and lethal substances, PWUD inevitably adapt; however, matters of drug preference and autonomy of choice remain under-studied and often misunderstood. As early as Preble and Casey (1969) argued that some PWUD prefer obtaining the strongest and most potent available heroin. Even before NPF infiltrated the illicit markets, research indicated the desirability of pharmaceutical-grade fentanyl (Comer et al., 2007; Firestone et al., 2009) and underscored the pleasant characteristics of fentanyl intoxication, which is marked by a quicker onset and a powerful high (Suzuki and El-Haddad, 2017). At the same time, in describing a concept they coined “connoisseurship of potency,” Mars et al. (2016) elaborated that while some individuals strive to obtain the strongest drug available, their decisions are shaped by an intersection of factors, including physical dependence, lack of resources and precarious living situations, as well as drug use stigma and the constraints of the U.S. heroin market. Similar arguments have been made by others, highlighting that it is both supply and demand that underlie the current proliferation of NPF (see Ciccarone, 2019; Dasgupta et al., 2018; Mars et al., 2018). That is, PWUD may develop a preference for NPF’s intense effects or lower cost, but as fentanyl availability grows, it may lead to a desire for stronger opioids or more frequent use due to changes in opioid tolerance and NPF’s shorter duration (Bluthenthal et al., 2020; Comer and Cahill, 2019).

Carroll et al. (2017), for example, found that while known or suspected NPF exposure was common in their Rhode Island sample, nearly every participant reported a strong dislike for its effects. Several others, however, consistently described a percentage of their sample expressing a preference for NPF. Estimates range widely, from 18 % to 44.7 % in Baltimore (Buresh et al., 2019; Gryczynski et al., 2019), 26 % across Baltimore, Boston, and Providence (Sherman et al., 2018), 31 % in North Carolina (Peiper et al., 2019), 29.9 % in New Jersey (Kline et al., 2021), and 40 % in Ohio (Daniulaityte et al., 2019).

Finally, Kenney et al. (2018) reported most of their Massachusetts sample (70.4 %) intentionally used fentanyl, though the connection between intentionality and preference is unclear. Those who elaborated on these sub-samples of PWUD indicated that preference was due to NPF’s “better high” (Foglia et al., 2021; Macmadu et al., 2017), easy availability (Daniulaityte et al., 2019; Mars et al., 2017) and inaccessibility of pure heroin (Duhart Clarke et al., 2022), higher and more predictable potency, as well as perceived effectiveness in delaying the onset of withdrawal (Mclean et al., 2019), particularly in the context of increased opioid tolerance (Cicero et al., 2017; Peiper et al., 2019). Other studies have found an association between preference for fentanyl, non-Hispanic White identity, more frequent drug-use (Morales et al., 2019) and younger age, homelessness, and daily intravenous drug use (Buresh et al., 2019).

While those who prefer fentanyl may represent a minority of PWUD, given the potency and role of NPF in overdose, these folks pose a unique challenge for public health initiatives and service provision. For example, Duhart Clarke et al. (2022), who reported that 25 % of their North Carolina sample preferred NPF and 28 % knew people who preferred NPF, found that those who preferred NPF also increased their use of stimulants (e.g., cocaine/crack and methamphetamine). Kline et al. (2021), in turn, found a positive correlation between fentanyl preference and persistent (i.e., multiple) overdoses. As such, it is critical to develop a better understanding of not only the extent of PWUD demand for NPF, but also the rationale behind and behaviors associated with its use, in order to develop more effective overdose prevention strategies.

This paper aims to examine the experiences of people who reported a preference for NPF or an NPF-heroin mixture, specifically how they navigate both the perceived benefits of NPF and its established risks. Findings will inform the nature and scope of responses, services needed, and public health messaging for this subgroup, as well as for PWUD at large.

2. Methods

This paper focuses on a subsample of 22 (27.5 %) participants from the Fentanyl Adaptation Study (FASt; N = 80), a mixed-methods, cross-sectional study of people who use opioids (PWUO) in NYC, who expressed a preference for NPF or a mix of NPF and heroin. Between March 2018 and August 2019, participants were recruited through convenience sampling from five syringe service programs (SSPs), street-based recruitment (e.g., approaching individuals engaging in public drug use), and snowball sampling. Potential participants were provided a brief study description and a contact number to determine eligibility which included: being aged 18 years or older; a resident of NYC or primarily using/buying drugs in NYC; awareness of NPF in illicit drugs; and use of “street” opioids (e.g., heroin or non-prescribed pain pills) at least twice in the past month. 51 potential participants did not qualify for the study, with the two most common reasons being having never heard of NPF, which is a significant finding in and of itself, and not having used opioids in the last 30 days.

Verbal informed consent was obtained, and eligible individuals participated in an anonymous interview conducted in public or semi-public venues (e.g., fast food restaurant). Study participants were provided with a $50 honorarium and were offered a naloxone kit. Participants were asked to refer up to three individuals to facilitate snowball sampling and received an additional $20 for each person that enrolled in the study. All study procedures were approved by the Institutional Review Board at the New York City Department of Health and Mental Hygiene.

The interview guide consisted of qualitative open-ended questions and close-ended quantitative items which covered the participant’s: (1) demographic characteristics; (2) knowledge of and response to fentanyl (e.g., bought fentanyl deliberately, how they heard about fentanyl); (3) drug purchasing practices; (4) experience with fentanyl test strips; (5) experience with naloxone; (6) drug use during the past month; and (7) opinions of health messaging related to fentanyl. While the parent study focused on participants’ overall experiences with, opinions on, and adaptations as a result of the increasing presence of NPF in the drug supply, only data relevant to the focus of this manuscript are reported. All interviews were recorded and professionally transcribed. Interviews lasted for an average of 68.2 min.

Dedoose (2018), a web-based data analysis tool, was used to code and analyze the data employing a thematic analytic approach (Braun and Clarke, 2006; Guest et al., 2014). Four members of the research team independently coded four transcripts to develop a code structure which was revised until a consensus was reached on a final codebook. Interviews were independently coded by two members of the research team with any discrepancies reconciled. The research team served as an interpretive community to develop themes from the coded excerpts, and methodological rigor was maintained through an audit trail and analytic memos (Wolf, 2003). Pseudonyms chosen by participants are used in lieu of real names to protect confidentiality.

3. Results

Participants ranged in age from 19 to 62 (mean = 36.1); 15 identified as men and seven as women. Participants identified as Latinx (n = 9), non-Hispanic White (n = 8) and non-Hispanic Black (n = 4), with one participant identifying as another race/ethnicity. Four participants did not complete high school, eight completed high school or equivalent, six completed at least some tertiary education, and four had tertiary degrees. Almost two-thirds (n = 14) were living in their own home or that of a friend/family member, four were residing in a shelter, and three were “street homeless” (e.g., living on the street, in a subway station). One participant resided at a residential treatment facility.

Participants’ demographics were mostly comparable with the rest of the FASt sample, though this subsample included slightly younger (mean=36.1 compared to mean=43) and more non-Hispanic white (36.4 % compared to 25.8 %) than non-Hispanic Black (18.2 % compared to 25.9 %) and Native American (0 % compared 8.6 %) participants. Participants did not differ from the larger sample in gender or education level; however, fewer participants in the larger sample lived in their own home or that of a friend/family member (55.2 % compared to 63.4 %) and were residing in a shelter (29.3 % compared to 18.2 %). In this subsample, five participants were recruited using street-based recruitment; the rest were referred to the study via snowball sampling.

All participants in this study believed they had used heroin containing NPF and a subset (n = 16) had purposefully attempted to buy a product that contained NPF. Nine reported a preference for NPF and thirteen a preference for a mix of heroin and NPF. Participants reported using opioids an average of 23.7 days in the past month (range 4–30 days). The majority (n = 14; 63.6 %) also reported using other non-prescribed substances, most frequently cocaine or crack (n = 10; 45.5 %), alcohol (n = 8; 36.4 %), and/or benzodiazepines (n = 6; 27.3 %). Nine participants had used opioids intravenously in the past month and 13 had used them intranasally. Seven had experienced an overdose in the past year.

3.1. “Preference” for NPF is multifaceted

3.1.1. Effects and functions of NPF use

Participants reported varied reasons for designating a preference for NPF and NPF-heroin mixtures, although “preference” was often a complicated concept. Many referenced NPF’s qualitatively different effects compared to heroin, including a more potent and “cleaner” high and superior analgesic effect, often accompanied by a deep “nod” and time loss.

“It’s definitely a different high, depending on how much you do. But for me, they both give me that downer, you know what I mean? That down feeling. So, yeah. But fentanyl, like I said, when you – not blackout, but I would nod off and then I’d wake up and a whole fucking hour went by. I’m like, ‘What the fuck?’ You know what I mean? So, yeah. It’s a deeper – It’ll put you in a deeper trance, you know what I mean, as opposed to heroin.” (Trump, 45).

Despite its documented shorter duration of action, NPF may result in a prolonged elimination and respiratory depressant effect (e.g., Suzuki and El-Haddad, 2017), and some participants perceived NPF to have “longer legs” (i.e., longer lasting effects).

For others, NPF seemed to also serve a different function than heroin, which informed when and how it ought to be used. Daniel, for example, conveyed a general nostalgia for heroin while still underscoring his use of NPF in moments of emotional pain.

“Fentanyl is kind of like getting decked in the face and heroin is like, is kind of like getting put out in the chokehold. Does that make sense? Fentanyl – it’s going to hit you and you’re going to knock out. You know, you’re going to hit the table. You know, hit your head on the table. Dope, though, heroin is, it’s li-, when you get, when you get good heroin, it, it’s a very subtle, euphoric layers that kind of slowly – You know, like that poem. It takes you slowly off into the good night, you know, and it’s kind of like being wrapped in a blanket…. You know, it [fentanyl], it’s always, it’s a super ugly high. It, it’s super ugly, it’s super like – you’re nodding – like, if you’re going for the nod out, then like – that’s what you want. So, at the moment I’m in emotional pain, so I want the nod out, but the feeling of euphoria, bliss and love that you get from heroin is unparalleled to just about anything else.” (Daniel, 22).

Others also reported tailoring their substance use to personal needs such as using NPF to aid sleep or address chronic pain. Interestingly, one participant, referring to NPF’s potency, recognized that she uses heroin to curb opioid withdrawal and uses NPF when she wants “a treat.”

“Like I said, like, dope, like,…it just keeps you well enough to function, like, it’s not like getting fucked up, but maybe to help sleep or something like —— it sounds sick but [I use fentanyl] like as a treat at the end of the night or like something like that.” (Erin, 30).

3.1.2. Secondary gains of NPF use

Additional reasons for participants’ preference were adjacent to the physiological effects of NPF. For example, one participant, who commented on NPF’s “longer legs” stated that this provided an opportunity to curb withdrawal and use less product. He noted that NPF enabled him to sleep through the night and not worry about waking up in withdrawal and having to go buy more, something he had struggled with when using heroin.

Similarly, others noted how using NPF, which is generally cheaper than heroin, was a financially conscientious decision. Daniel, who spoke about the difference between heroin and fentanyl above, also underscored the role of financial sensibility in drug buying behaviors.

“[Fentanyl] lasts longer because it’s so much stronger, and when you’re buying the same amount of weight compared to other kind of drugs, why aren’t you going to get something that’s going to be more bang for your buck?”.

Similarly, for participants like French, the ability to reduce expenditure was an advantage that even helped suppress concerns about overdose.

“I heard about a year and a half [ago] that [fentanyl] was introduced into heroin, I’m not sure if it’s been around for longer than that but my personal experience in New York, about a year and a half ago that I heard it was getting – and everyone was like, ‘Stay away, stay away, ODs, ODs, everyone’s overdosing,’ and then however we ended up getting it and we scared to do it, we did a little and we were like, ‘Oh my god, this lasts longer,’ so you have to buy less heroin. And then that’s it, then fentanyl basically took over heroin, it lasts longer, less money to spend.” (French, 38).

Taken together, both the primary and secondary effects of NPF reinforced it to be the best option available. As Angel (38) summarized, “It’s a quick cheap alternative to heroin…The fentanyl produces a way stronger result and way faster, and it doesn’t take as much.”

3.1.3. Impact of available drug markets

The appeal of NPF was also tied to the perceived decline in the availability of “good” heroin. For many, their preference for NPF was market-driven, both in terms of drugs available to them and their desire for a potent high. An example of this is Kurt, who while stating a preference for an NPF-heroin mix disclosed, “Technically, I would prefer heroin because of not wanting the sickness but the euphoria.” He reported he became “used to [fentanyl],” due to the proliferation of NPF in the drug market. He was among several participants who described how their preference had shifted, not as a conscious choice, but because of the drug supply. As Joon further explained:

“The heroin, like I said, everybody’s putting stuff in it, and I wanted real heroin. I didn’t want no fentanyl inside and stuff like that. But I had to take what I could get, because that’s what was around. And, like, that or I have to go to different places to go find out and get the heroin and see what’s the good stuff. And then when I started using fentanyl, the regular heroin wasn’t good enough for me, it was too – it was not as potent as before.” (Joon, 43).

Similarly, several participants suggested that extended use of NPF resulted in an increased physical tolerance to opioids, indicating a cycle in which people buying heroin containing NPF quickly became accustomed to it and, as a result, were locked into continuing use.

“If there’s no fentanyl out here, you can’t go get other dope, and it’s going to make you feel well, you know… You got to have fentanyl, and that’s the only thing. And then when you’re withdrawing from it, there’s nothing that can help you with the withdrawals.” (Stephanie, 27).

Of note, one participant, who reported a preference for an NPF-heroin mix, emphasized that heroin is already frequently adulterated with other substances and stated, “If I’m not getting pure heroin, I would rather have heroin cut with another opioid than heroin cut with a non-opioid.”

3.2. Perception of overdose risk

3.2.1. Awareness of and attitudes toward overdose risk

Seven participants reported having overdosed in the last year, most of them suspecting it was due to NPF. Several others also recounted recent instances of friends and/or family overdosing. Thus, participants were largely aware that the presence of NPF in the drug market was driving increases in overdose deaths. At the same time, participants conveyed both trepidation and a sort of powerlessness when it came to overdose risk.

“Fentanyl is so strong that it might kill someone that’s just taking pills and they don’t know that. So, I mean, you come on the street and you buy something, you have no idea what– I have no idea what’s coming in this little bag. But just to– the reward is so much higher than the risk for me, because I’m just trying to function. I’m willing to take that risk every day because what’s my alternative?” (Billy D., 41).

Other participants similarly considered NPF in the context of opioid use more generally, highlighting the risk that has always accompanied illicit drug use and suggesting that fentanyl may not warrant additional concern.

Among some participants, attitudes toward overdose were more explicitly linked to their quality of life, mental health concerns, and broader ideas about life and death. Rick, who had overdosed in the last year and admitted that it had no impact on him “though it should have,” demonstrated how depressive thoughts can intertwine with drug use, resulting in a more fatalistic attitude toward overdose risk.

“I don’t wish to die, but there’s some times where I say ‘Whatever happens, it’s going to happen.’ I’m down and out. It’s the bottom, so I take a chance I guess, because there’s nothing to look good to or nothing to look forward to.” (Rick, 42).

3.2.2. Perceived protective factors

Despite complex positions on overdose risk, participants named several factors they perceived to be protective against overdose. Most participants (59 %) in this sample used intranasally and several, particularly those who had never experienced an overdose, were adamant that this route of administration was protective.

“I guess it’s just, like, maybe the chances [of overdosing] are way less, you know. You know, it’s way higher for somebody who’s injecting than somebody who’s, like– And people who, like, sniff it, I guess the process is different too. It’s just like I guess you start off small or whatever. You test whatever you get. It’s just the process is completely different as opposed to putting something directly into your bloodstream, you don’t know what it is, where you got it for, you just as soon as you get it, all of it is going right into your bloodstream and you don’t know what it is.” (Lee, 30).

Moreover, many reflected on their opioid tolerance, levels of experience with NPF and opioid use more broadly, combined with the perceptions of their own bodies and personal “limits” as a vessel of knowledge that served to mitigate overdose risk.

“And I know like myself and a bunch of other users kind of get upset when we get regular heroin because it’s not as good. So whereas, yes, it’s dangerous – at this point, my tolerance is so high, and I’m so used to it…. That it doesn’t scare me as much as it should. Especially with like how many friends I’ve lost due to fentanyl.” (Andrew, 29).

“I feel like my body is like a tank, like… I know, I know that sounds stupid, but, like, there’s been points where years ago I tried to kill myself and stuff like that, and it never worked out.” (Erin, 30).

Interestingly, for some, the preference for NPF itself seemed to be perceived as a protective factor. Angel (38), reflecting on her experience of overdosing after she had used what she thought was heroin (i.e., without altering her approach or using smaller amounts), highlighted the risks of not knowing what is in the drug supply – “I didn’t have bad experiences until I was unaware,” she stated. She clarified that she began to consistently expect NPF and use smaller amounts at a slower pace.

Several participants suggested that an NPF-heroin mix may pose a better and safer alternative to using “straight” NPF. For one, a mix provides “the best of both worlds” by combining the effects of fentanyl and heroin in one product. One participant made this point clearly, commenting on NPF lacking the “drip”— mucus or saliva that mixes with the substance and drips down the back of the throat after inhalation.

“You’re not going to get a drip, a dope drip, unless you have dope involved. And the sniffers, the people that sniff, they don’t like anything that they can’t taste. So that’s why people started cutting the fentanyl with dope.” (Juan, 62).

Others suggested there may even be a “right” way to mix heroin and NPF to produce the desired effects with minimal negative consequences, and several believed the risks associated with NPF were predominantly the result of inappropriate heroin to NPF ratios and/or careless or partial mixing.

“Whenever you can find heroin and fentanyl together, it multiplies it –– it makes it even stronger than just fentanyl or heroin when they’re separate… A lotta times with fentanyl, a lotta times, it’s not mixed properly. And you know you can have one bag that’s, you know, not that strong, as the other bag. So you really have to be careful.” (Marvin, 27).

3.2.3. Impact of overdose and adverse experiences

Those who had previously overdosed often referenced the event as having led to a change in drug use behavior such as using less or buying from a more reliable source. Even those who had not experienced (or were not aware of experiencing) an overdose referenced changing their drug use in response to a “bad” high that was accompanied by significant time loss or other adverse effects. As Greg described:

“I know my safe limit. If there’s something new, I’ll do less, so I know I’ll be safe or okay. You know what I mean, if I do a little bit and it doesn’t do anything, I’ll do the rest, or do half of another one, until I get to my limit. But sometimes it works instantly, sometimes it works after. Because the one time it worked instantly, another time I waited like an hour and then I started feeling the effects, and I was like, wow, it’s all starting to hit me all at once, everything I just did. So then you just got to be cautious.” (Greg, 31).

While for most, overdoses and adverse effects were suspected to be caused by NPF, the finding points to the potential differences in credibility and “weight” given to personal experiences versus more abstract ideas of overdose.

3.3. Overdose risk reduction strategies

3.3.1. Fentanyl test strips

Aside from the few participants who had tested positive for fentanyl during urinalyses at emergency rooms or detox facilities, most had no objective way of confirming whether they were using NPF. Only eight participants had used fentanyl test strips, two of whom expressed using them only to confirm the presence of NPF in their product. The majority highlighted limitations of test strips (e.g., limited accessibility, inability to detect specific analogs) and the lack of test strip utility in a market flooded with NPF. Stephanie, for example, stated, “There’s just no point… Because I know there’s a lot of fentanyl in every dope out there.”

3.3.2. Naloxone

Although all participants knew of naloxone, and many recognized its utility, few reported carrying it on their person, although some did keep a kit at their home or in their car. Since most participants used drugs alone at least some of the time, they commented that even if they carried naloxone, they would not be able to self-administer it in the event of an overdose. Mostly, participants expressed concern that having naloxone may draw further attention in the event they are stopped by law enforcement.

“Sometimes I feel a little bit weird about having it in the glove box, just in case I get like stopped by a cop and searched, and they’re like why the hell do you have this, and then they’re gonna search harder.” (Andrew, 29).

3.3.3. Other risk reduction strategies

Many participants reported that their previous experiences with NPF had tempered their drug use behavior. Strategies to reduce risk often involved using less product at a time, including doing a “test shot” or not being “greedy” (i.e., using a large amount in a short period of time).

“I learned respect for it when I started hearing overdoses from it. I know to respect – I didn’t take it for granted. Then I started paying attention to how to use it. Before, I didn’t care, I’d just put it in the needle and just do whatever I have to do.” (Joon, 43).

Other risk reduction strategies were less often cited, and few participants reported consistently using with another person who could intervene in the event of an overdose. Those who predominantly used alone rarely described having a check-in plan (i.e., letting another person know that they are using), describing various barriers including wanting to keep their substance use private and not wanting to experience stigma or shame from others. As one participant explicitly stated, “[I’m] too macho [for check-in plans].”

Moreover, the tension between the aspirational risk reduction messages and the complex realities of illicit opioid use meant that some participants had adopted a fatalistic attitude toward their ability to stay safe, and as one participant commented on overdose risk, “It is what it is.”

4. Discussion

This study suggests that while there are several reasons for using NPF, the concept of preference is multifaceted and largely market driven. Participants highlighted various factors that contributed to their expressed preference for fentanyl or a heroin/fentanyl mix, including not only a desire to feel good and opportunity to cope with emotional pain, but also financial resources, drug availability, decreased consumer autonomy, and physiological demand. We found support for existing evidence that PWUD still value product consistency and swift alleviation of withdrawal symptoms (see e.g., Bancroft and Scott Reid, 2016). Morales et al. (2019) cautioned about the detrimental political and media discourse placing the onus of fentanyl use on consumers, stating “PWUD are not [emphasis added] inflexible consumers driven chiefly by addiction’s compulsions” (p. 7). Similarly, our findings highlight PWUD’s decreased consumer agency, exposure to systemic factors, and the difficulties of negotiating these factors while mitigating overdose risk. Moreover, the range of functions that opioids served for participants in this study points to the continued need to identify potential underlying causes for substance use and misuse, including psychological disorders and/or trauma and suicidality (see Cleland et al., 2020).

Given the presence of NPF in the U.S. (U.S. Drug Enforcement Administration, 2021), regardless of preference, avoiding fentanyl in the drug supply is likely more difficult than finding it. Thus, continued public health messaging about the presence of fentanyl in the illicit market and its lethality may be insufficient in reducing PWUD’s risk of overdose. The call to provide a safer and regulated opioid supply has become a prominent theme among advocates and drug user groups, who argue that safe supply initiatives directly address the toxic drug supply while reducing the need to acquire money and drugs through the informal economy (Tyndall, 2020).

Comparably, the expansion of reliable and accessible drug checking resources, such as fentanyl test strips and Fourier Transform Infrared technology (FTIR), is another important strategy to curb overdose fatalities and promote agency among PWUD (Peiper et al., 2019; Sherman et al., 2019; Weicker et al., 2020). Participants in this sample were confident they were using NPF and, as such, did not find test strips particularly useful. The finding is consistent with other studies, suggesting that these services may be more critical for people using drugs other than heroin, those who are trying to avoid fentanyl, or those whose lack of awareness of NPF in their product puts them at greater overdose risk (Ciccarone et al., 2017; Firestone et al., 2009).

Moreover, while participants in this study were practicing some well-versed risk reduction strategies (NYC DOHMH, 2019), our findings suggest that these recommendations are not always feasible, particularly those of carrying naloxone and not using alone. While the distribution of naloxone has been a key national strategy in overdose prevention, barriers, such as apprehension about increased attention from police, still prevent PWUD from carrying naloxone (see Dayton et al., 2019; Khatiwoda et al., 2018). While decriminalization of possession of opioid antagonists has been a priority in some jurisdictions, additional policies that prohibit the presence of these substances being used as a pretext for further search may be needed. Ongoing naloxone distribution efforts ought to consider addressing these concerns during overdose prevention trainings. Similarly, Papamihali et al. (2020) have identified multiple reasons for PWUD using drugs alone, including convenience and comfort, stigma, having no one to use with, and not wanting to share drugs. Existing literature has consistently emphasized the importance of considering other risks around health, safety, and use of state services that PWUD face when using illicit substances, especially as those may supersede the risk of overdose (see Ciccarone, 2017; Harris et al., 2021).

Lastly, participants in this study, the majority of whom used opioids intranasally, believed this route of administration to be protective against overdose. While there is some evidence suggesting an association between injection and increased overdose risk (e.g., Carlson et al., 2020), it is possible that the perception of intranasal ingestion as less risky is an artifact of risk reduction messaging concerning blood-borne infections, such as hepatitis C (e.g., Des Jarlais et al., 2011). Findings from Relay, an NYC-based non-fatal overdose response system, showed that 77 % of individuals who participated in the program reported a route of administration other than injecting prior to their overdose event (Welch et al., 2019). Syringe service programs are efficacious in overdose prevention efforts (CDC, 2019), but there are fewer systems of outreach for PWUD who do not use intravenously. In addition to investigating the role of route of administration in overdose risk, we need to better engage people who do not inject and increase the utilization of alternative risk reduction strategies, such as using smaller amounts of the drug.

This study had some limitations. Data were collected in a large U.S. metropolitan area and may not be generalizable to people in other communities. Relatedly, our findings center only people who reported a preference for NPF or an NPF-heroin mix; future research would benefit from contextualizing experiences of preference within broader populations of PWUD. Data were based on self-report, which can be affected by social desirability bias, recall bias, and general reporting errors. As with most cross-sectional studies, the chronology of exposure to NPF, development of preference, and changes in drug use behaviors is not entirely clear, though some participants reported more explicit causal links. Due to the qualitative nature of the study, we were unable to evaluate correlations between NPF preference and participant factors (e. g., demographics, drug use patterns), though it may be significant that participants in the subsample were slightly younger and included a larger number of non-Hispanic White folks than in the larger FASt sample, which is consistent with previous findings (Morales et al., 2019; Buresh et al., 2019); moreover, the majority of participants reported use of other non-prescribed substances, primarily amphetamines, suggesting the need for continued exploration of preference within polysubstance use (see Duhart Clarke et al., 2022). Moreover, participants’ use of NPF was not confirmed via urinalysis. Daniulaityte et al. (2019) reported that 97.7 % of their sample who self-reported suspected street fentanyl use tested positive for NPF; given the prevalence of NPF, self-report data is likely more unreliable when PWUD report not having used fentanyl (Martinez et al., 2020).

In conclusion, among participants in this study, preference for NPF, or an NPF-heroin mix was often influenced by market forces including drug availability, price, and quality. Despite the increased risk of overdose from NPF, participants often struggled to implement risk reduction strategies. Future overdose prevention efforts should consider the context, perceptions, preferences, and behaviors of people who use drugs themselves and promote tailored interventions that engage PWUD more effectively.