The quadrupling of opioid overdose mortality in the US, and in Baltimore, during the past two decades highlights a pressing need to prevent both fatal and non-fatal overdose (National Institute on Drug Abuse, 2018). The urgent need to prevent overdoses has become only more evident alongside the recent surge in fentanyl-related deaths (Seth, Scholl, Rudd, & Bacon, 2018). Fentanyl overdose deaths in Baltimore City, where the data for the present study were collected, increased 36.8% from 2016 to 2017 from 419 to 573 deaths (MDHMH, 2017, 2018). Overdose mortality may be reduced by diffusing overdose prevention and response skills and training people who use drugs (PWUD) and other community members likely to witness an overdose to administer naloxone, a lifesaving overdose reversal antidote.

Preventing overdose fatalities requires identifying individuals who are at risk for opioid overdose and reducing their risk through harm reduction and other types of interventions (Hawk, Vaca, & D’Onofrio, 2015; Lagisetty, Bohnert, & Fendrick, 2018). Throughout the US, including in Baltimore, local programs have implemented overdose response training and distributed naloxone through existing syringe exchange programs (Green, Heimer, & Grau, 2008; K. E. Tobin, Sherman, Beilenson, Welsh, & Latkin, 2009) with apparent success in increasing participants’ knowledge and skills to respond to an overdose event (Seal et al., 2005; Tobin et al., 2009). However, many individuals at risk of overdose do not frequent syringe exchange services, highlighting the need to diversify the locations providing harm reduction services (Beletsky et al., 2014; Gindi, Rucker, Serio- Chapman, & Sherman, 2009; Treloar & Cao, 2005). The current study sought to explore correlates of self-reported, non-fatal overdoses, including stigma and drug use settings, among PWUD in Baltimore City to inform future overdose prevention efforts.

The self-stigma and discrimination experienced by PWUD because of their drug use (i.e., stigma related to drug use) has not been adequately considered as a risk factor for opioid overdose despite increasing evidence linking stigma with health outcomes, injection-related risk behaviors, and utilization of harm reduction services (Cama, Brener, Wilson, & von Hippel, 2016; Couto E Cruz, Salom, Dietze, et al., 2018; Kulesza, Larimer, & Rao, 2013; Latkin et al., 2010; Rivera, DeCuir, Crawford, Amesty, & Lewis, 2014; von Hippel, Brener, & Horwitz, 2018; Wilson, Brener, Mao, & Treloar, 2014). Internalized stigma (i.e. self-stigma), the negative feelings PWUD have about themselves because of their drug use, has been previously linked to depression and decreased psychological well-being (Cama et al., 2016; Kulesza et al., 2013; von Hippel et al., 2018), which are known to increase overdose risk (Tobin & Latkin, 2003). Rivera and colleagues (2014) also found that among people who inject drugs, higher internalized stigma was associated with decreased recent use of syringe exchange programs. Additionally, enacted stigma (i.e. experiences of discrimination or being stereotyped because of drug use) has similarly been associated with depression and poor mental and physical well-being (Couto E Cruz, Salom, Dietze, et al., 2018; Kulesza et al., 2013), as well as with injection behaviors that increase overdose risk, such as being injected by someone else (Wilson et al., 2014). To our knowledge, only one study has examined the association of discrimination with overdose and found that experiencing discrimination because of drug use on a weekly or more frequent basis was associated with 60% higher odds of overdosing (Couto E Cruz, Salom, Dietze, et al., 2018).

In addition to individual-level factors, the micro-level risk environment, specifically the settings in which people use drugs, may influence the outcome of an overdose. Using drugs in public spaces, such as abandoned buildings or outside, has been associated with both calling for medical help and fatal overdose compared to use in private spaces (e.g., a participant’s home) (Bohnert, Tracy, & Galea, 2009; Tracy et al., 2005). In addition, setting may influence bloodborne virus risk through its influence on injection norms and practices (Rhodes et al., 2006). However, it is unknown whether using drugs in public versus private settings influences whether an overdose occurs. Setting-specific drug use norms may influence overdose risk. People who use drugs in settings other than their private residences may have less control over what transpires, may rush to use, and/or may fear police or others interrupting their drug use. Additionally, the number of settings where drugs are used may be indicative of drug use frequency and addiction severity.

Examining contextual and structural risk factors for experiencing an overdose, such as stigma and drug use settings, could inform overdose prevention programming. The primary aim of this analysis was to examine how stigma was associated with overdose history. We anticipated that higher levels of internalized and enacted stigma would be positively associated with experiencing a drug overdose. The secondary aim was to assess how drug use settings relate to overdose history. We hypothesized that using drugs in more settings, and especially in public settings, would be positively associated with experiencing a drug overdose. Finally, we examined how other factors, including drug use, syringe exchange use, selling drugs, and demographic factors related to overdose history.

We found that experiencing discrimination from others or stereotyping (i.e., enacted stigma) related to using drugs and using drugs in public settings were associated with recent and lifetime overdose history. These findings highlight the need to identify the mechanisms by which enacted stigma increases overdose risk so that overdose prevention strategies in Baltimore City and beyond can address the excess risk of overdose among stigmatized individuals. In addition, our finding of an association of overdose with injection at multiple venues suggests that PWUD should be recruited from multiple types of public settings for overdose education and response training and naloxone distribution.

Our findings around enacted stigma agreed with a study by Couto e Cruz et al. (2018), who found a 60% higher odds of experiencing an overdose among Australians who experienced persistent discrimination for injecting drugs. Surprisingly, internalized stigma was only marginally associated with overdose in our study, and the relationship was weaker in magnitude than that for enacted overdose. The differing associations of overdose with internalized and enacted stigma suggest that overlapping mechanisms by which both types of stigma increase overdose risk, such as through their positive associations with depression, do not fully explain the relationship of enacted stigma with overdose (Cama et al., 2016; Couto E Cruz, Salom, Maravilla, & Alati, 2018; Kulesza et al., 2013; von Hippel et al., 2018). One plausible mechanism by which enacted stigma could increase overdose risk is by encouraging dangerous injection behaviors, such as rushing drug use, to avoid being observed by bystanders. While there is limited prior work to inform this hypothesis, studies on police presence, which has been cited as a reason for rushed injecting, may be relevant, as a motivation for avoiding arrest by police is to avoid being identified as a person using drugs by bystanders (Cooper, Moore, Gruskin, & Krieger, 2005; Ti et al., 2015). In addition, Wilson et al. found an association between perceived discrimination with being injected by someone else, another injection risk behavior that could lead to an overdose (Wilson et al., 2014). Hesitancy to attend overdose prevention trainings or carry naloxone, using alone, hiding drug use, and discrimination from addiction treatment or other healthcare providers could also contribute to the increased risk of overdose that is associated with enacted stigma. It is also possible that greater addiction severity may lead to more drug buying, selling, consumption, and other consequences of drug use (e.g., incarceration) that increase stigma and overdose risk concurrently (Couto E Cruz, Salom, Maravilla, et al., 2018; Kulesza et al., 2013).

Further research on the longitudinal and contextual relationships between enacted and internalized stigma and overdose may aid in developing overdose prevention interventions. These two dimensions of stigma may have different roles, with interventions for PWUD having a greater focus on internalized stigma and interventions for health care providers, family members, and community settings focusing on enacted stigma. Social marketing campaigns have successfully been used to reduce HIV stigma, and such approaches could also be utilized to address stigma towards PWUD (Blankenship, Bray, & Merson, 2000; Mahajan et al., 2008; Rimal & Creel, 2008). Online trainings may be effective to reduce negative and discriminatory attitudes among healthcare workers against PWUD (Brener, Cama, Hull, & Treloar, 2017).

The finding that a greater number of public drug use settings are associated with overdose suggests that expanding overdose prevention and naloxone training into public settings frequented by PWUD may help reach populations who do not access training through syringe exchange services. While just over a quarter of participants reported using the syringe exchange program in the past six months, half reported using outside, and many reported using in other locations as described above. These results highlight the potential for diversification of settings for naloxone training and distribution as well as the potential to involve community members through a peer outreach approach to diffuse overdose prevention information in settings of drug consumption that may be difficult to reach by health professionals (e.g., shooting galleries and crack houses). Further, these results complement previous findings that supported expanding settings for overdose interventions based on the presence of overdose witnesses (Latkin, Edwards, Davey-Rothwell, Yang, & Tobin, 2018).

While harm reduction programs have trained people who use opioids to successfully administer naloxone, there remains a pressing need to train PWUD to educate their peers to diffuse overdose prevention and treatment training skills and encourage PWUD to carry naloxone (Sherman et al., 2009). Working with PWUD to identify contextual characteristics of drug use settings that influence overdose prevention and response may also be valuable to inform strategies for peer-driven intervention. For example, what are the range of norms and policies across different types of settings related to naloxone availability, expectations about peer monitoring, and overdose response procedures and what factors help to produce contexts that better facilitate safety? Whereas public bathrooms present challenges around identifying an overdose due to using alone (Fairbairn, Coffin, & Walley, 2017; Holloway, Hills, & May, 2018), shooting galleries have the opposite characteristic: an association with larger social networks and sharing injecting equipment (Tobin, Davey-Rothwell, & Latkin, 2010). The social relationships of PWUD in public settings, or lack thereof, may also guide expectations for responding to a witnessed overdose, calling for medical assistance, or monitoring overdose victims (Holloway et al., 2018). Future research should examine the mechanisms by which different drug use settings contribute to overdose so that the risks may be adequately addressed.

The association between syringe exchange utilization and experiencing an overdose was attenuated after adjusting for heroin and cocaine injection (i.e., speedball). This change is not surprising because the vast majority of individuals who attend syringe exchange programs inject drugs. Nonetheless, the fact that >40% of participants who overdosed attended the syringe exchange provides strong support for continued outreach to people who inject drugs at syringe exchanges for overdose education and naloxone distribution. As overdose from heroin and synthetic opioids continues to present a major risk to people who inject drugs, the BCNEP’s existing overdose prevention and naloxone program will be a critical resource in preventing overdose deaths.

This study has several limitations. The study’s findings are limited in generalizability to individuals meeting the parent clinical trial’s eligibility criteria (presence of HIV risk behaviors, aged 18–55, and willing to attend study visits). Additionally, we relied on self-reported data for overdose history. Overdose was described to participants as “a life-threatening condition that occurs after someone has taken drugs by injection or by any other route.” This leaves room for personal interpretation by the participant, and there is no way of determining whether overdose experiences were under- or over-reported. However, unlike medical records or hospital admissions, self-reported overdose data captures overdoses that are unreported (Darke, Mattick, & Degenhardt, 2003). Furthermore, it is unknown what percentage of reported overdoses were caused by drugs other than opioids and therefore would not respond to naloxone administration. However, 94% of study participants had a history of opioid use. Thus it is likely that most reported overdoses were caused by opioids or opioids in combination with other drugs. In addition, these data were cross-sectional and cannot establish the temporal sequence between stigma, drug use, and overdose. It is plausible that having an overdose, in and of itself, could increase stigma, and that this could in turn increase the risk of future overdose. Longitudinal studies to further characterize these relationships are warranted. We also did not have reports on where participants had injected more than 6 months ago. Participants in our study either never exchanged or exchanged syringes infrequently (less than once per month). We were therefore unable to examine the relationship of overdose and frequency of exchanging syringes or accessing other harm reduction services at the syringe exchange program. These data also do not capture the experiences of users who have succumbed to fatal overdose and therefore could not participate in the study. It is plausible that unique risk factors exist for fatal overdose and these should be addressed in future research.

It should also be noted that this study was conducted before the influx of synthetic opioids (i.e., fentanyl and fentanyl analogs) on the drug market and preceded recent local policies that increased the availability of naloxone. In October 2015, the Baltimore City Health Commissioner issued a standing order for naloxone (i.e., a blanket prescription covering anyone in the city who presents a certificate from a state-approved naloxone training program) (Baltimore City Health Department, 2015). Since then, more than 20,000 Baltimore residents have been trained through the Baltimore City Health Department, with >800 reported overdose reversals (Baltimore City Health Department, 2017a). In June 2017, an additional standing order removed the training requirement, making naloxone accessible to all residents (Baltimore City Health Department, 2017b). The impact of standing orders on stigma and overdose prevalence and mortality has not yet been determined. It is unlikely that changes in rates of overdose or standing orders have dramatically changed the level of stigma experienced by PWUD.

Our results suggest the potential utility of supplementing existing overdose education and naloxone distribution policies in Baltimore with measures that reduce stigma. For example, peer educator programs that connect PWUD who are disengaged with harm reduction services to peers engaged with harm reduction services could help minimize the impact of stigma associated with drug use, which may facilitate carrying naloxone, attending overdose trainings, or accessing syringe services programs. Moreover, providing PWUD with prosocial options for engaging in harm reduction in the community, such as becoming peer educators, may help to reduce self-stigma. Further, additional training to destigmatize drug use is warranted for health professionals, social service providers, families of PWUD, and the general public.