Circumstances of overdose among suburban women who use opioids: Extending an urban analysis informed by drug, set, and setting

Abstract

Background:

Opioid overdoses are primarily discussed by the pharmacological properties of the drugs used. Research shows that other factors such as the social/physical environment and the mental/emotional states can have an impact on overdose events. Ataiants and colleagues (2020) used Zinberg’s “drug, set, and setting” framework to identify circumstances surrounding overdose experiences of street-involved women in Philadelphia. The aim of this paper is to extend their analysis to a diverse sample of suburban women who experienced overdoses.

Methods:

The mixed-methods design consisted of ethnographic fieldwork, in-depth interviews, and brief surveys with 32 suburban women who use opioids. Inductive theoretical reasoning and constant comparative analysis facilitated themes emerging within the “drug, set, and setting” framework.

Results:

Eighteen out of 32 women identified “drug” as the primary factor involved in their overdose events. Major themes were an inability to identify the synthetic opioid fentanyl, lack of knowledge or control over how much to use, poly-substance use, and an insufficient understanding of risks. Eleven out of 32 women linked “set” to their overdose experiences. Themes included emotional trauma, such as death of a child, child custody issues, and mental health conditions, such as depression. Six out of 32 women associated “setting” with one of their overdose experiences. Themes were related to being with friends or partners that used, and having recently been released from treatment or incarceration.

Conclusion:

Findings show similar themes found among an urban sample, adding insight on the need for effective overdose interventions targeted for suburban populations. The opioid crisis is not confined to the cities, and neither should services aimed at addressing opioid overdose. The knowledge provided here can help policy makers support female-centered harm reduction services not only in urban areas but also in the suburbs.

Introduction

Drug overdose is the leading cause of unintentional deaths in the United States (U.S.). Between 2010 and 2017, heroin related overdose deaths increased five-fold (Hedegaard, Miniño, & Warner, 2018). As death rates for prescription drugs and heroin began to decrease overall, fentanyl involved overdose deaths increased (Wilson, Kariisa, Seth, Smith, & Davis, 2020). Fentanyl is a Schedule II prescription drug with a potency that is up to 100 times higher than morphine. This synthetic opioid is sometimes prescribed for post-surgery or chronic pain management (NIDA, 2019). While it can be safely used under a medical provider’s supervision when the exact dosage is known, fentanyl is often mixed with other drugs such as heroin or cocaine in unknown quantities which can present catastrophic results (NIDA, 2019) Between 2016 and 2017, overdose deaths involving synthetic opioids increased by 45.2% (Scholl, Seth, Kariisa, Wilson, & Baldwin, 2019). In 2017, fentanyl accounted for 38.9 % of drug overdose deaths. Heroin accounted for 22.8%, and other opioids (oxycodone, morphine, methadone, hydrocodone) accounted for 24.6% combined (Hedegaard, Bastian, Trinidad, Spencer, & Warner, 2019).

Opioid involved overdose deaths are increasing at greater rates for women than for men (Mazure & Fiellin, 2018; Wilson et al., 2020). Studies show that women become dependent on drugs more quickly than men, have greater stigmatization, trauma, and stressors in their lives, and have more barriers to participation in treatment (Frazer, McConnell, & Jansson, 2019; Howard, 2015; Stone, 2015). Women are also more likely to be given an opioid prescription by health providers and more likely to hide their dependence from others, creating situations in which drug use becomes socially isolating (Goodman, Whalen, & Hodder, 2019; McHugh et al., 2018; Paltrow & Flavin, 2013; NIDA, 2018). Social factors such as homelessness, unemployment, marginalization, and social isolation increase risks for overdose (Dasgupta, Beletsky, & Ciccarone, 2018; Rhodes, 2009; Zoorob & Salemi, 2017); yet studies on overdose risks tend to focus on drug-related factors (Bode et al., 2017; Dilokthornsakul et al., 2016; Olfson et al., 2018).

A recent study on how social circumstances of urban women impacted their overdose applied a classic sociological theory to focus beyond the pharmacological and biological aspects of overdose risks (Ataiants, Roth, Mazzella, & Lankenau, 2020). Using Zinberg’s (1984) “drug, set, and setting” framework, the authors found that circumstances surrounding the women’s physiological, emotional, physical, and social environment often triggered the overdose event. The Ataiants et al. (2020) study was based on a street-involved sample from a large urban center. In this paper, we replicate their analysis using a “drug, set, and setting” framework with a suburban sample of women who experienced opioid overdose.

Replication studies are critical for validating findings in health research, and while more difficult to replicate social science studies, employing similar analysis to a different sample can validate as well as strengthen and extend findings from previous studies (Amir & Sharon, 1990; Atkins, Lewin, Smith, et al., 2008; Connelly, 1986; Hirsch, Conforti, & Graney, 1990; Lamal, 1990; Nakagawa & Parker, 2015). The purpose of this analysis is to broaden understanding of contextual and socio-psychological factors contributing to opioid overdose by examining patterns in circumstances leading to overdose among women using opioids in suburban communities. Knowledge of similarities and differences between urban and suburban overdose circumstances among women who use opioids will help target medical services and public health messages, and inform prevention strategies on “set” and “setting” patterns that differ by cultural and geographic contexts.

According to the “drug, set, and setting” framework, settings are drug use environments that are fundamental to drug use risk behaviors (Zinberg, 1984); yet, we know little about risk environments that influence drug use practices of women who use opioids in suburban settings. Epidemiological indicators show the demography and geography of people who use opioids have shifted away from urban centers (Cicero et al., 2014; Kuehn, 2014; Lippold et al., 2019; Okie, 2010). Rural towns and suburban communities surrounding large metro areas show the greatest increase in opioid deaths (Rigg, Monnat, & Chavez, 2018; Wilson et al., 2020). Communities outside large metropolitan centers often lack the medical facilities, harm reduction services, and social safety networks found in the cities (Boeri, 2013; Boeri & Lamonica, 2020; Woodall & Boeri, 2014; Haffajee, Lin, Bohnert, & Goldstick, 2019; Rosenblatt, Andrilla, Catlin, & Larson, 2015).

Zinberg’s (1984) framework also calls attention to psychological and emotional factors that increase risks for overdose, such as anxiety, trauma, loss of relationships, or stress of personal situations, such as threats to parental rights (Angelotta et al., 2016; Howard, 2015; Fraser, McConnell & Jansson, 2019; Lamonica & Boeri, 2020; Paltrow & Flavin, 2013; Nichols & Love, 2019; Stone, 2015). Like street-involved women in cities, women living in suburban areas who are using opioids are impacted by the pharmacological aspects of potentially lethal drugs, but while urban and suburban women might have similar emotional stressors due to gendered differences of drug use cultures, the distinctive overdose settings and contextual influences of suburban living conditions, and the strategies suburban women use for avoiding overdose death are unknown. There remains a gap in understanding the social determinants of contextual circumstances surrounding opioid overdose in suburban areas (Cicero et al., 2014; Marmot, 2005; Singer, Bulled, Ostrach, & Mendenhall, 2017; Wilkinson & Marmot, 2003). We seek to broaden overdose “drug, set, and setting” understanding by replicating the analysis used by Ataiants and colleagues (2020) in their urban sample of women and by applying it to a diverse sample of women living in the suburbs of Atlanta, Georgia, Boston, Massachusetts, and New Haven, Connecticut.

Methods

Local Context

The sample of women used in this analysis is drawn from a larger study examining opioid use in three suburban sites. The broad goal of the Suburban Opioid Study (SOS) was to gain a better understanding of opioid use patterns in suburban communities where opioid use is epidemic. The suburban settings provide contrast for comparison to the sample drawn from an urban setting in the Ataiants et al. study.

According to the National Center for Health Statistics urban–rural classification, Philadelphia, the study site in the comparison sample, is a “large central metro” area; whereas the suburban field sites around Atlanta, GA, and Boston, MA are categorized as “large fringe metro” areas, and the suburbs around New Haven, CT is a “medium metro” area (Ingram & Franco, 2014).¹ Between 1999 and 2016, overdose rates rose by 507% in large fringe metro counties and 388% in medium metro countries compared to 158% in large central metro areas (CDC, 2017; Rigg, Monnat, & Chavez, 2018). By 2018, deaths involving prescription opioids and heroin began to decrease in large central and fringe metro areas but continued to increase in medium metro counties; deaths involving synthetic opioids (e.g., fentanyl) increased in all areas (Wilson et al., 2020). Opioid overdose mortality rates were increasing in the suburbs of Atlanta and New Haven when data collection began in 2017 (Georgia Department of Public Health, 2019; Kaiser Family Foundation, 2019; Scholl et al., 2019), and while death rates decreased in Massachusetts in 2018, opioid-related overdose deaths were highest in the suburban counties surrounding Boston (Massachusetts Department of Public Health, 2019; Massachusetts Department of Public Health, 2020).

Suburban people who use opioids live in middle-class and upper-middle-class neighborhoods composed of single homes with lawns and well-kept gardens and smaller rental properties, as well as in densely populated depressed communities outside the city lines. These poorer communities include neglected trailer parks and deteriorating rental properties. There are no “open drug markets” for heroin, but opioids are easily obtained by a phone call and meeting in a designated area or by courier; drug exchanges are observed in parks, streets, malls, grocery stores, and abandoned buildings. Overdoses occur in private homes, hotel rooms, cars, parks, streets, the methadone clinic, and restrooms in fast food restaurants and gas stations (Boeri & Lamonica, 2020).

Sampling

The larger study used mixed-methods consisting of ethnographic fieldwork combined with in-depth interviews and brief surveys. Fieldwork was conducted by the principal investigators (first and second authors of this paper), who were familiar with suburban areas around Atlanta, Boston, and New Haven. Field coordinators and research assistants experienced with ethnographic studies helped with recruitment and interviews. All were trained in overdose prevention. The investigators were aided by “community consultants” who had knowledge of use patterns and settings of opioid use, typically individuals with opioid use experience and connected to networks of people who use opioids in the community (Boeri & Shukla, 2019; Page & Singer, 2010). Recruiting from different communities and from diverse networks helps to avoid data that reflect “social desirability” bias, such as local harm reduction messages that may not be practiced in reality (Fessel, Mars, Bourgois, & Ciccarone, 2019). In contrast to a focus on “homogenous sampling” (Ataiants et al., 2020), we employed theoretical sampling to maximize the differences of the sample (Glaser & Strauss, 1967; Van Maanen, 1988).

During fieldwork, the investigators talked with people on the streets, in parks, and other public spaces, and they left fliers or business cards with the study contact information. Recruitment for participation involved spending time establishing rapport by “hanging out,” providing people rides to needed services, offering a coffee or hot meal in a warm food establishment, and other caring activities that engendered trust. Criteria for participation include having used opioids in the last month, being 18 years old or older, and residing in a suburban location. Targeted and purposive sampling methods were employed to broaden the diversity of the sample for a better representation of the suburban population of people using opioids (Bluthenthal & Watters, 1995; Watters & Biernacki, 1989).² Participants were drawn from small and medium sized suburban towns, as well as semi-rural suburban areas. Some lived in exclusive communities and others lived in the street, under a bridge, or at a shelter, representing a range of women who overdosed on opioids.

In general, people living in the Atlanta suburbs had less access to social and medical services than people living in the suburban areas around Boston and New Haven. Whereas harm reduction services were available in many of the towns in the suburbs of Boston, there were no harm reduction services in the Atlanta suburbs at the time of data collection. Harm reduction services via a mobile van were available on an irregular basis in the suburbs of New Haven with limited outreach. Although most people who used opioids illicitly bought and sold different types of opioids (e.g., prescription pills, heroin, methadone) in the suburban communities where they lived, a few went into the city to obtain heroin. Of the 173 interviews collected in the SOS study, females represented 43.9% of the larger sample. The 32 women who discussed their overdose experiences were included in the analysis for this paper (see Table 1).

Table 1:

Social, demographic, and drug use characteristics of interviewees (N = 32)

Characteristic	Total, % (n)
Age
Mean (range)	40.0 (28–63)
Race
White	71.9 (23)
African-American/Black	18.8 (6)
Hispanic/Latina	9.4 (3)
Education
High School Graduate /GED or above	75.0 (24)
Sexual Orientation
Sexual minority (homosexual or bisexual)	15.6 (5)
Family situation (current) 1
Living with a partner	65.6 (21)
Having children	75.0 (24)
Unstable Housing
Current1	34.4 (11)
Ever2	71.9 (23)
Incarceration (prison, jail)
Current1	18.8 (6)
Ever2	68.8 (22)
Employed (current) 1	59.4 (19)
Drug use profile
Currently injecting1	43.8 (14)
Ever injected2†	78.1 (25)
Heroin (past 30 days)	75.0 (24)
Prescription Opioids (past 30 days)	65.6 (21)
Cocaine or Crack (past 12 months)	65.6 (21)
Methamphetamine (past 12 months)	12.5 (4)
Cannabis (past 12 months)	31.3 (10)
Treatment (including medication assisted treatment)
Current1	68.8 (22)
Ever2	90.6 (29)

¹During year of interview

²Since first use of opioids

^†a few women injected before first use of opioids

Data collection

Data were collected by the two principal investigators, field coordinator, and research assistant involved in ethnographic fieldwork. Interviews were conducted in homes, private offices, harm reduction centers, shelters, library rooms, investigator’s cars, parks, and on the street. The Institutional Review Board (IRB) from the investigators’ academic institutions approved the study, and a “Certificate of Confidentiality” was obtained from a federal agency to protect study data from subpoena. Participants received $40 for their time at the end of the interview, which is consistent with ethical reimbursement for this type of study (Ritter, Fry, & Swan, 2003).

A typical interview started with the completion of a structured survey on social demographics and past 30-day drug use, and the collection of social history and drug history data using a color-coded “yearly trajectory.” This trajectory provided a visualized overview of sociodemographic activities and events since a year before first use of opioids. Both survey and yearly trajectory data were collected using an iPad. This was followed by an audio-recorded in-depth interview that focused on drug transitions, risk behaviors, and important life transitions shown on the yearly trajectories. Interviews generally lasted between two and four hours with snack breaks. The semi-structured interview guide revolved around context of drug use, setting of use, and interaction with others, but participants were free to discuss areas of interest that emerged in their life stories.

The three sources of data collected allowed “participant verification and informant feedback” during the interview process (Harper & Cole, 2012, p. 511). When interviewers noticed a discrepancy between the survey data and the qualitative interview, they were able to discuss the different answers with respondents and modify the survey data or collect more detailed information during the qualitative interview. Additionally, participants could see the visualizations themselves and were able to amend answers that were discrepant with details of their recalled events (Fendrich, Mackesy-Amiti, Wislar, & Goldstein, 1997). Member checking was also used as a tool to increase internal validity (Birt, Scott, Cavers, Campbell, & Walter, 2016). This technique allowed us to explore the credibility of our collected data. Sources of data collected were also triangulated with field notes and interviewer observations.

Theoretical framework

The concept of combining “drug, set, and setting” properties was first introduced by Zinberg (1979) in a chapter on “nonaddictive opiate use” revealing the different impact of interactional effects. In a later book, Zinberg’s (1984) theory of “drug, set, and setting” is more thoroughly examined with support from studies on soldiers returning from Vietnam with opiate addiction (Robins, Davis, & Goodwin, 1974). In Zinberg’s terms, “drug” refers to the “pharmacology of the drug,” set” refers to the “ the personality of the user,” and “setting” to social situations where use occurs (1984, p. 2), including cultural and political environment of drug use in time and place (p. 10). Zinberg argues that the “interaction of set and setting” is more influential on continued drug use than “addictive personality” psychoanalytic theories or pharmacological effects (p. 104). While Zinberg’s findings on “controlled users” never gained substantial consideration in research, his concept of “drug, set and setting” achieved some popularity as a theoretical framework (Perrone, 2011; Singer & Schensul, 2011; Golub & Bennett, 2013; Hartogsohn, 2017). Until the Ataiants et al. (2020) article, there have been no studies analyzing overdose with a “drug, set and setting” framework. In this paper, we extend their formative work with a more diverse sample, addressing some of the limitations reported in their article.

Data analysis

The recorded interviews were transcribed word-by-word into Microsoft Word documents, and research assistants served as quality-controllers in order to ensure accuracy and anonymize the transcripts. The investigators triangulated various sources of data (surveys, yearly trajectories, transcripts) in order to increase validity and credibility of the findings and to gain a better understanding of complex social issues (Becker, 1998; Caudle, 1994; Denzin & Lincoln, 1994). Using the framework of “drug, set and setting,” inductive theoretical reasoning was followed to allow meaning to emerge from the data through constant comparative analyses between our sources of data and ongoing coding (Strauss & Corbin, 1998). The first coding was conducted by two of the authors (AK and JT) using the Word documents and coordinating the accuracy of their understanding of the data. In order to extend the findings of Ataiants and colleagues and compare our analysis with their results, we employed the same coding categories of “living situation,” “drug” type, “set” type, and “setting” type, adding sub-categories as they emerged from the data. Meanings and interpretations were discussed between the authors. All the transcripts were imported into NVivo, a qualitative data management program, by the second author (MB), who conducted a third coding of the data.

The sociodemographic and drug characteristics of the participants were tabulated using SPSS for Mac. While most of the characteristics presented for the sample in the Ataiants et al. (2020) article are included in our table, we added characteristics provided by our multiple sources of data. As shown in Table 1, drug use characteristics are reported for “past 30 days,” “past 12 months” and “ever” since initiation of opioid use. Heroin and other opioid use were asked in the survey for “past 30 days.” We had data on recent overdose and past experiences. We included data on treatment experiences since this impacted injection and overdose characteristics.

Results

Sociodemographics and living situation

The ages of the women in this sample ranged from 28–63 years old, and our participants were predominantly White. Twenty-four women (75%) had children. Of those, 15 (62.5%) had lost custody in the past to child protective services, spouses, or other family members. More than half (65.6%) were currently living with a spouse or steady partner. Twenty-two women (68.8%) had been arrested or spent time incarcerated. Over the course of their lives, 23 of our participants (71.9%) experienced unstable housing situations, and 11 (34.4%) were unstably housed during the past year. More than half of our participants (59.4%) were currently employed in some type of legal work. Twenty-five women (78.1%) had injected drugs in their lifetime, and slightly less than half (43.8%) were currently injecting opioids. Six women reported overdose experiences when not injecting: five reported “sniffing” and one said she swallowed heroin and pills. The majority of women used other illegal drugs in the past year. Most women (90.6%) had been in drug treatment at some time in their lives, and over half (68.8%) were in treatment during the past year. The women who were currently in medication-assisted treatment were also misusing or abusing other opioids. Many women (59.4%) spoke of a combination of physical, psychological, and sexual periods of abuse throughout their lives.

While the suburban women lived in different states in the U.S. and had diverse socioeconomic environments, distinct patterns were found between those with stable versus unstable housing, and those employed versus unemployed. Women who had recently been incarcerated emerged as most vulnerable to experiencing an overdose. These socioeconomic aspects of their lives often occurred concurrently, with one impacting another.

Housing was influential in the women’s lives. Women who could not afford their own place, lived with family members or roommates. The federal housing choice voucher program was utilized by five women who qualified as very low income or disabled. Two others were waitlisted. Women who experienced a period of homelessness in the past year, typically mentioned losing a relationship with a partner or family member, such as a woman in a Boston suburb who was living with her mother:

When my mom went to jail, and I lost the apartment, about eight months ago. If I sleep I either sleep on the stairs right there or at the shelter- back of the shelter, or the front of the shelter, or the parking garage. […] I lay a blanket down if I have one. (White, 30)

Work played a critical role in a little over half of the women’s lives. Although four women had worked in pink-collar semi-professional jobs at some time in their lives, drug use and incarceration interrupted their careers. During the past year, seven women worked in service jobs, such as in restaurant or cleaning businesses, or in elderly home care. One participant from the Atlanta suburbs described her work:

I was doing a lot of front of house, like serving and bartending, but I’m actually cooking right now in the kitchen, and then, in retail. So I’ll also work at [grocery store]. And you know, it’s, is it my choice of what I would want to do? No. But it’s like the only job that doesn’t do a background check or a drug test. I mean a lot of people smoke weed. Does that mean they’re undeserving of a job? No. Um, so…you know, and also I have no transportation right now. (White, 38)

Similar to these women, those who were unemployed at the time of the interview but actively sought work, said they were limited by mandatory drug tests or lack of transportation, and were also restricted by a criminal record. A 52-year-old White woman from a Boston suburb explained her work history after being incarcerated: “I went back to the book bindery business; I did construction...I did housekeeping. A lot of stuff was under the table.” Five women who were not working received disability or social security checks; three women relied on their spouses or steady partners to help support them financially, and one was retired. Eight women revealed that they had engaged in sex work in their lives.

The women in our sample all resided in the suburbs; however, this was not necessarily where they purchased their opioids. Women indicated that they would avoid buying where they lived to protect their privacy, because it was not available in their area, or it was more expensive. A 56-year-old Black woman from the New Haven suburbs said: “I don’t usually buy, you know, in the town that I live in ‘cause I know too many people, and they know my family.” A 44-year-old White woman from the New Haven suburbs added: “it’s not really available where I live, and the stuff that is available where I live, it’s…a lot smaller and more expensive.” In contrast, six women discussed that they did not leave their residences because they knew dealers who delivered to their homes. A woman from an Atlanta suburb described a typical scenario:

But when I got the Dilaudid, they would deliver them to my house because they went to the pharmacy right down the road every Monday, so they would just come by my house afterwards and, and drop them off. (White, 30)

Others called a dealer and met someone in a car, hotel, park, or on the streets; seven mentioned buying drugs directly from people on the street when desperate. Three women relied on their partners to provide them with drugs and did not engage in any purchasing habits.

The suburban women used drugs in a variety of settings. For example, seven of the women indicated that they preferred the privacy of their homes, while others preferred using at other people’s homes or at hotels. A woman from the Atlanta suburbs explained:

I would get high when I’d go to the people house that I knew, the people that sold. I would go to their house, and they would let me get high...and then I would go home. (Black, 60)

Women who bought at a dealer’s house reported drugs were either consumed at the dealer’s home or in the car depending on the social circumstances. A woman from the Atlanta suburbs recounted:

You go get it, and you go get in your car, or if the people that you’re buying it from are cool enough, you go to their bathroom, and you do it...in some of the social settings because there were so many of us that were doing it, you know, we would all ride down together, all sit in the car and do a shot, and then go back to the house and chill, play video games, listen to music, go out to eat. (White, 38)

The majority of suburban women used multiple types of opioids including methadone and fentanyl. Six women used their opioids in combination with other drugs, such as cocaine and alcohol. Women in all three suburban areas described using heroin while on medication-assisted treatment. One woman explained how this increased the risk of overdose:

If you’re on a low dose it’s easy to cut through; if you’re on a high dose it’s harder. You have to use more. You become a risk for an overdose ‘cause you risk doing too much, so you already risk an overdose just trying to cut through it. (White, 36)

Most of their overdose concerns, however, were not knowing how much fentanyl was in the heroin they bought. By the end of data collection in 2019, most women said they could not find heroin without fentanyl.

Because our women came from three separate suburban areas, they had a range of experiences related to “drug, set, and setting.” Women shared common attributes of their collective overdose experience related to “drug.” Their “set” was affected by struggles to procure and use drugs while also fulfilling other social roles as mothers, daughters, friends, or employees. The women’s “set” was also affected by the participants’ psychological issues and diagnosed mental illnesses. Half of our participants indicated that they experienced depression, while 11 struggled with anxiety, and nine with post-traumatic stress disorder. The impact of their “setting” was characterized by whether the women had a stable environment or not, and if they were recently in treatment or incarcerated. Women in northern U.S. states, where healthcare and treatment services were more accessible, typically fared better.

“Drug” type

In Zinberg’s theory (1984), “drug” refers to the “pharmacology of the drug” or its physiological effect on the person using it (p. 2). Analysis of the women’s experiences showed the “drug” effects were primary factors of their overdoses. Themes related to the effects of the drug included the inability to gauge the presence or amount of fentanyl and its analogs (e.g., carfentanil), lack of knowledge or control over how much to use, poly-substance use, and a false perception of risks related to routes of administration.

Over half the sample, 18 women, indicated that at least one overdose event was connected to the pharmacological effects of the “drug.” Purchasing heroin adulterated with the potent synthetic opioid fentanyl was the most common reason given for an overdose event. Women said they ingested unknown quantities of fentanyl unintentionally and later identified fentanyl as the catalyst of their overdose. One woman from the Boston suburbs, who was informed by her primary care physician that she had been using fentanyl, explained:

I was surprised because I thought it wasn’t fentanyl. People had been putting it, and I wouldn’t know it. That’s why I’d been – I OD’d five times, because I didn’t know. So I thought it was like straight up.” (Latina, 32)

Two women suspected that the product they were about to purchase did not look like their regular dope. A woman from the Boston suburbs described how she was set up by her own dealer:

A dealer gave me a bag—okay, I know heroin’s brown, right? And sometimes I’ve seen it gray, I’ve seen it yellow, okay? This was like a gray, odd, like a odd color but—no, no, no! It was white. I said I don’t want coke. I don’t want coke. He’s like that’s not coke, its dope. I said that’s not dope. So I did a frickin’ line of fentanyl this big. I was fuckin’ dead standing up. Yeah, dead. (White, 42)

She suspected that something was not right with her heroin based on her prior experience with the drug, but she used it in her desperation to avoid withdrawal and overdosed as a result.

At least five women said they favored fentanyl-laced heroin although they also said they were aware of the potential consequences. A woman from the New Haven suburbs preferred the potent synthetic opioid over pure heroin. In her own words:

But surely everybody would want [heroin laced with fentanyl] because it’s a better high. If it puts you out, you’re like, “Damn. I’d rather have that.” You’re takin’ a big chance. (Black, 56)

Three women in our sample indicated that their overdoses were tied to using multiple drugs at one time. The use of multiple drugs was reported as desiring a greater “high” or increased tolerance, as a woman from the New Haven suburbs explained:

I was just using more and more tryin’ to chase that high. And then that’s when it became dangerous ‘cause then I wasn’t feelin’ it anymore. ‘Cause then, even though I’d get high ‘cause I’d add the benzo’s to it, but then sometimes I would overdose. ‘Cause I would take too much, and I have low blood pressure, and I would overdose. (White, 34)

Other women were combining heroin with prescription opioids underestimating the combined effects of both drugs, such as a woman from the New Haven suburbs who used heroin on top of her prescription drugs:

The first time I did it, I fell asleep for eight hours, and he [husband] was so scared. He had to get me in the shower and everything. It was awful, and that was only a little matchhead; tiny little matchhead thing. ‘cause I had already take my Percocet’s. (White, 55)

Being unfamiliar with an injecting route of administration, particularly by a novice user, also affected women’s overdose events. A woman from the Atlanta suburbs blamed her overdose on the potency of the product due to her lack of experience with injecting heroin:

It was when I first started using, and I had done some, and I guess it was just a lot better than the rest or something. That I had first started injecting, which had a lot to do with it, and before I know it, I’m on the floor and he’s giving me CPR. (White, 45)

Seven women in the suburban field sites who did not inject, as a harm reduction strategy, were surprised by their overdose experiences. Five of these women never injected in their lifetime and only “sniffed” opioids; one “ate” heroin by the spoon with water, and one woman had recently switched from injection to “sniffing” to avoid overdosing. A 55-year-old Black woman from the New Haven suburbs remembered she was “sniffing a bag I don’t know what had happened but…I sniffed a bag, and then it was just like I went out.” Participants revealed different perspectives when asked why they overdosed while sniffing, often blaming it on using too much. A 39-year-old White woman from the Boston suburbs said she “felt like it. […] I just overdid it. I overdid it.” A woman from the New Haven suburbs, echoed this experience:

When the dude gave me the bag he said, “don’t sniff it all,” he said, “it’s powerful.” Me hardheaded…sniffed it all. Got home, I went upstairs to my neighbor to get a plate so I could fix the kids something to eat. When I turned around the next thing I know…I’m at the emergency room. (Black, 50)

Both women were found by their boyfriends and given naloxone to reverse the overdose, but a few of the “sniffers” ended up in the hospital. A woman from the Boston suburbs said she “snorted” a large amount of what she thought was heroin:

Somebody gave me a $60 bag of—so they said it was brown. It was fentanyl. I snorted the whole $60 bag which was way too much. I should have known better ‘cause I know what color it is. It’s kinda like a yellowish-white. So I snorted the whole thing. (White, 52)

She passed out in the streets before she was found and an ambulance was called. She was not the only woman to suspect a “yellowish” or “brownish” color indicated the presence of fentanyl.

When discussing overdoses in terms of fentanyl-laced heroin, combining drugs, desiring a greater “high,” and inexperience or false assumptions regarding route of administration, participants could not entirely separate these overdoses from “set” and “setting.” These examples illustrate that although strongly linked to the pharmacological properties of the “drug,” the women also connected “set” and “setting” to the overdose experience.

“Set” type

According to Zinberg (1984), “set” refers to the “attitude of the person at the time of use, including his personality structure” (p.8). Applications of “set” include life experiences, relationships, and emotional motives for using drugs (Ataiants, 2020; Reinarman & Levine, 1997). In total, 11 out of 32 participants connected “set” directly with their overdose experience. Women discussed the impact of “set” on their overdose experiences in terms of the emotional trauma related to losing their children, negative life events, and states of depression.

A pattern of accumulated physical, psychological, and sexual abuse emerged. For one woman from the Atlanta suburbs losing custody of her children was the “last straw.” Dealing emotionally with child abuse, trauma, and the loss of custody of her children played a significant role in her overdose:

[My mother] left me whenever I was little for dick and dope, and I followed in her footsteps. […]. And then I had also found out my children’s father was cheatin’ on me too then, so I had all of that. So yeah, I really did wanna die; I didn’t wanna live. I didn’t wanna deal with all that shit. I mean I didn’t know how to deal with it. I had nobody to help me, back me up, support system, nothing. All I had was me and my two kids. The straw that broke the camel’s back was my children’s father took the kids and told me I would never see my kids again. And that’s when I was like “fuck this shit,” and I went and got high. Never came down I felt like. (White, 36)

While this woman indicated she wanted to die in an overdose due to accumulated stress and trauma, two mothers claimed they intentionally wanted to escape the emotional misery of the death of a child. A woman from the New Haven suburbs had spent seven years caring for her son with a rare genetic disorder who ultimately died in her arms. During that time, she also lost an infant son to SIDS. When she was asked how she grieved for her children, she revealed:

I went and got high and overdosed. […] Like, that’s what I wanted to do; I didn’t wanna be here. I wanted to be with my kid, and that’s what I kept sayin’ to my family, I don’t wanna be here anymore. [...] I went and got two bundles of heroin, and I had Xanax, ‘cause I was prescribed them ‘cause my kids died so they were givin’ me whatever. Like, that’s crazy too. Like in a tragedy, they will prescribe whatever just to get you away from them. Like, go take these pills instead of crying in my office. It’s terrible. But anyway, I had my pills and my drugs […] and I sniffed one whole bundle. (White, 36)

Negative life experiences were associated with overdose events, whether intentional or not. To cope with trauma or other negative events, women increased their drug use, ignoring the potential of a fatal overdose. A 28-year-old White woman from the Atlanta suburbs had already overdosed upwards of ten times. She explained: “every time like something goes wrong in my life, what I usually always turn to is drugs.”

A few women linked their overdoses to depression, as illustrated by a 32-year-old Latina woman from the Boston suburbs describing her state of mind: “I was in deep depression...I stayed indoors all the time. Never wanted to go outside.” Her state of mind lead to several overdose events.

The women in our sample linked some of their overdose events with “set” even while acknowledging the “drug” effects of opioids. For some, their overdose was intentional due to amassing negative life events; for others it was unintentional but a desired escape from insufferable emotional pain. While “drugs” and “setting” cannot be ignored in these overdose circumstances, in these examples, the women’s “set” emerged as the predominant factor.

“Setting” type

The component of Zinberg’s theoretical framework that he thought was most ignored in research on heroin use was “setting,” defined as the social setting where use occurs as well as the cultural and political influences (p. 10). The “setting” was also the most hidden aspect of suburban women who overdosed while using opioids. Typically, women did not link their overdose to their social circumstances or acknowledged the cultural and political environment of suburban communities. Yet, analysis of their stories shows that their social setting (whom they are with and where) as well as the cultural stigmatization of women who use opioids, particularly in suburban places, and the political impact of incarceration were instrumental in their overdose experiences. Six out of 32 women associated the “setting” with one of their overdoses.

Being with a drug-using friend or partner prompted three of our participants to either use more than they would usually ingest or change the route of administration. A woman from the New Haven suburbs explained:

I was with somebody that did—that used to shoot it. I don’t know. I wanted to try it... I fell out in the backseat of the car though. I couldn’t wake up or nothing. I really thought I was—I could hear what was going on, but I couldn’t open my eyes. It was the most scariest thing in my world. (Latina, 45)

Other participants had similar experiences and discussed how friends and acquaintances injected them with heroin before they overdosed. These women indicated that the presence of other people who inject drugs was associated with the overdose event and normalized the route of administration.

Leaving treatment triggered overdose events for two participants. One woman who was recently discharged from a rehabilitation facility explained:

This was after treatment. I overdosed; I relapsed and overdosed. When I was in the inpatient and got out and stayed clean for a while. I met my husband in rehab. We stayed clean for a while and then relapsed. I overdosed, or I relapsed and did heroin for the first time and I, I did too much, I guess, and it made me overdose (White, 30)

Others said their release from jail or prison led to an overdose. Returning to familiar neighborhoods with drug-using acquaintances after incarceration can result in strong cravings that participants said are difficult to ward off. A woman from the New Haven suburbs indicated her overdose was a direct impact of a change in setting:

When I overdosed––when I got outta jail my body was clean. […] And I’m thinking when I left all—what I got high off that I can go back to that. ‘Cause say if I did nine bags before I went to jail, then when I came home, I thought I still could do them eight or nine bags. […] And that’s when I went out. (Black, 56)

While “setting” was not always linked to the overdose by the women, having a drug-using companion, returning to an old neighborhood, or being released from jail/prison and treatment contributed to their overdose. In some incidences, the “setting” emerged as the strongest link, although the interaction between “drug, set, and setting” was difficult to disentangle.

Discussion

In this paper, we applied the “drug, set, and setting” framework (Zinberg, 1984) to a suburban sample of women who experienced opioid overdose to extend the findings of previous analysis of recent opioid overdose experiences among “street-involved” urban women (Ataiants et al., 2020). Our findings strengthen the importance of analyzing the interactional effects of “drug, set, and setting” beyond the effect of the drug alone. A comparison of the street-involved sample (N=29) with the suburban sample (N=32) indicates that the women in the suburban study were on average older, and more likely to be living with a partner and in stable housing, and less likely to be incarcerated during the past 12 months. There was little difference between the sample populations in educational attainment, having children, lifetime incarceration, and racial/ethnic diversity. The diversity found in the suburban sample included women who owned their own homes, were employed, and never lost custody of their children, as well as women who were chronically homeless, unemployed, and had lost or given up custody of their children. Regarding past 12-months drug use, women in both samples used multiple drugs, but there were more street-involved women who used cocaine or crack, and more suburban women who used prescription opioids. All women in the urban sample injected opioids recently; in contrast, some of the women in our sample did not inject drugs in the past year or had never injected, adding to the literature knowledge on non-injection opioid overdose. Accounts of the women’s overdose events support the finding that variances in their mindset and/or social situations can “trigger” an overdose, providing more insights to develop strategies that address the recent increase of overdose resulting from an influx of fentanyl in the U.S. illegal market (Comera & Cahill, 2019; Dai et al., 2019).

Ciccarone (2019) describes the recent decades of increasing opioid-related overdose mortality rates as a “triple wave epidemic” caused primarily by prescription opioid pills in the first wave, heroin in the second wave, and synthetic opioids such as fentanyl in the third and current wave. By 2017, unintentional deaths by opioids exceeded those caused by motor vehicle deaths or gun violence (Hedegaard, Miniño, & Warner, 2018; Katz, 2017). This coincided with the increase of fentanyl sold as heroin. Eighteen of the suburban women discussed their overdose in terms of the drug itself, often linking the potency of the drug to their overdose. Similar to the urban women, the suburban women were aware of fentanyl-laced heroin being sold on the streets, but some suburban women indicated they sought the more potent concoction. Other suburban women said they used strategies to avoid overdose, including routes of administration that were considered “safer.” Unlike the street-involved women who bought drugs in the “open-air drug market” (Ataiants et al., 2020, p. 6), most of suburban women said they called their dealers, and sometimes the drugs were delivered, or they met dealers at private homes or establishments. Quite a few of the women said they knew their dealers well and often trusted them to reveal if the heroin they sold contained a high proportion of fentanyl. A few of the women relied on their partners to procure opioids, as well as to administer naloxone if they suspected an overdose. Those women who had stable housing were less at risk for public exposure of drug use than street-involved women in the suburbs, but still at risk for overdose. As the women indicated, fentanyl was in all the heroin bought from known or unknown sources.

The dramatic increase in overdose death rates over the past decades is one of the “deaths of despair” linked to increased economic distress and deteriorating social support systems in the US; as well as increased availability of opioids (Botelho et al., 2017; Monnat, 2016; Scutchfield & Keck, 2017). Although unintentional deaths caused by opioids are not always linked to emotional and psychological mental health status, the interaction of access to potent drugs with circumstances of social isolation or emotional trauma has been shown to have an impact on overdose events (Dasgupta et al., 2018; Olfson et al., 2018; Zoorob & Salemi, 2017). The emotional state of mind defined by Zinberg (1984) as the concept of “set” was discussed by 11 suburban women in relation to their overdose attempts. Typically, this was related to depression or negative experiences throughout life, most notably, the loss of their children. Trauma, physical, sexual and emotional abuse, stigma, social isolation, and violence accumulated in the lives of the women, leading to depression, despair and suicidal ideation, often hidden behind suburban facades.

According to Zinberg (1984), the most important factors promoting control of substance use were the social sanctions and structural forces in society, which are associated with his concept of “setting.” Society’s response to illicit drug use, such as incarceration and increased governmental surveillance, has social and economic repercussions that lead to disadvantaged situations. Settings of opioid use are difficult to disentangle from the structural forces in society that lead to situations where overdose occurs. While research has established the link between settings and drug type or frequency of use (Boeri, Sterk, & Elifson, 2004; Hellman, 2017; Hussong, 2000), there is little literature linking settings to overdose other than the marginalized circumstances of homelessness and low-income housing (Bauer, Brody, León, & Baggett, 2016; Rowe et al., 2019). The effects of “setting” are perhaps the most critical contributions illustrated in the women’s stories, but mostly unidentified in their own understanding of overdose experiences. Although the suburban settings of this study were diverse, ranging from affluent and majority White neighborhoods to economically-depressed towns in the outer suburbs, and our sample provided a diversity of racial and ethnic women who discussed their overdose experiences, few women connected their overdose to a setting in which they were using. Six women discussed their overdose experience in relation to their social environment. Most often, this was linked to being in the company of other people using drugs or injecting with a partner or friend. Being released from jail/prison or treatment was also linked to overdose events.

Emphasizing set and setting interaction, Zinberg (1984) argues that the interaction effects of “drug, set and setting” are more influential on continued drug use than medical and psychoanalytic properties alone:

All three variables––drug, set, and setting–– must be included in any valid theory of drug use. It is necessary to understand in every case how the specific characteristics of the drug and the personality of the user interact and are modified by the social setting and its controls. (p. 12–13)

Our study adds further insights on the interaction between these components when looking at the impact on opioid overdose, underscoring the centrality of “set and setting”.

Until the Ataiants et al. (2020) article, there have been no studies analyzing overdose with a “drug, set and setting” framework. In this paper, we extend their formative work with a more diverse sample, addressing a few of the limitations reported in their article. We contribute to their seminal work in two critical areas. First, our findings show that harm reduction services should not stop at city limits. Harm reduction strategies must be communicated to more than the most visible populations using potentially fatal drugs. While street-involved women are more noticeable than women living in the suburbs who are using opioids, suburban women have less access to harm reduction services and public health messaging regarding drug use. Harm reduction centers and mobile units typically start in dense urban areas near open drug markets, and only during our current opioid epidemic are harm reduction services becoming more acceptable in suburban communities. This often leaves women who need these services to fend for themselves when attempting to avoid overdose, or they must rely on their friends and partners. Second, the realities faced by street-involved women are not the same realities faced by women living in suburban settings. Although street-involved women are exposed to more violence and are under greater surveillance that leads to incarceration, suburban women who use opioids suffer from similar but often hidden traumatic experiences, including loss of their children. Our analysis shows that suburban women were less often part of a community of people who use opioids; therefore, they were reliant on male partners or family for protection during an overdose. Suburban women had less access to social services that could help with legal matters, including finding housing, treatment, and regaining child custody.

As common in qualitative research, our suburban study has some limitations similar to the urban study (Ataiants et al., 2020). Findings from a convenience sample cannot be generalized to other populations. While our participants were recruited from different suburbs in three northeastern U.S. states, adding to the urban findings in terms of diversity of experiences and backgrounds, similar studies need to be conducted in rural locations and different countries with diverse cultural and political environments. Another limitation of this study is that we did not specifically ask women about their overdose experiences, although this was typically discussed during the lengthy life-history interviews. While we cannot claim to have inter-rater reliability for the drug, set, and setting codes used in the analysis of this paper, all three authors coded the interviews independently and later discussed the codes until agreement was reached. This approach strengthens the validity of the analysis. Recall bias and social desirability bias are common limitations in all retrospectively collected survey and interview data, but can be reduced using the strategies we employed, including rapport-building, using timelines and historical cues, and member-checking themes that emerged (Becker, 1998; Boeri, 2017; Harper & Cole, 2012; Denzin & Lincoln, 1994; Shaw, 2005).

While we answered Ataiants and colleagues’ call to investigate whether other populations who use opioids are more likely to link their overdose events with their “set” than “drug” or “setting,” there is much work to be done in this area. In contrast to the urban sample discussed in Ataiants and colleagues’ work, our suburban sample of women who use opioids was more likely to attribute their overdose events to pharmacologic effects of the drug itself rather than the “set” or “setting.” Future research in this area can focus on replicating this finding in a similar sample and also examine mixed gender samples or male-only samples. Further, research should focus on how route of administration affects overdose events and is linked to drug, set, and setting in countries with diverse political and cultural environments.

In conclusion, this study supports the need for the structural interventions suggested by Ataiants and her colleagues (2020), specifically calling for policy that legalizes and supports supervised injection facilities, heroin-assisted treatment, and drug-checking services (Boeri & Tyndall, 2012; Broadhead, Kerr, Grund, & Altice, 2002; Kerr et al., 2005; Kolla & Strike, 2019; Oviedo-Joekes et al., 2008; Rhodes, 2009; Tupper et al., 2018). We add to this by calling for these services to be extended beyond the city. Much like the implementation of syringe exchange programs, these life-saving strategies typically start in urban areas and are implemented outside cities when overdose incidents become more prevalent (Nadelmann & LaSalle, 2017; Showalter, 2018). Our study adds to the literature showing that the opioid crisis is not confined to the city, and the structural factors that impact risk of overdose are similar in urban and suburban locations. While some suburban women benefitted in other ways from structural factors in the suburbs, such as delivery of drugs to their homes, they were at the same risk for overdose and had little harm reduction education or supplies. Having harm reduction services primarily in urban areas means that suburban people are left with less access to these resources.