Policing the drug trade is a deeply entrenched goal of many law enforcement agencies. Police practices, whether sanctioned or unsanctioned, often have a detrimental impact on the health of people involved in the drug trade, including people who inject drugs (PWID), Such practices range from high-frequency harassment and arrest, to police violence and killings (Beletsky, Grau, White, Bowman, & Heimer, 2011; Cooper, Moore, Gruskin, & Krieger, 2004; Csete et al., 2016; Hayashi, Ti, et al., 2013; Human Rights Watch, 2017; Kutsa et al., 2016; Landsberg et al., 2016; Sarang, Rhodes, Sheon, & Page, 2010). In the U.S. setting, the “war on drugs” has relied on racialized policies incentivizing the number of arrests at the expense of constitutional rights and justice, particularly among low-income Black and Latino communities (Alexander, 2012; American Public Health Association, 2016; Cooper et al., 2004; Lunze, Lunze, Raj, & Samet, 2015). Intensified police interactions include the “militarization” of policing, which can often result in violent encounters (Dansky, Bunting, Solon, & Bohm, 2014). Drug-related law enforcement activities have paradoxically been shown to increase community-level violence in the U.S. context (Werb et al., 2011).

Law enforcement practices constitute a key aspect of the “risk environment” of PWID, influencing their health and safety (Burris et al., 2004; Rhodes, 2002). In the U.S., almost a third of syringe services program (SSP) attendees experienced recent syringe confiscation by police (Beletsky et al., 2011), which is of concern given that fear of syringe confiscation by police is associated with higher likelihood of receptive syringe sharing (RSS), a major risk factor for HIV and HCV acquisition (Bluthenthal, Lorvick, Kral, Erringer, & Kahn, 1999; Maher & Dixon, 1999; Small, Rhodes, Wood, & Kerr, 2007). Police activity (including “crackdowns”) that target people in the drug trade are also associated with reductions in access to SSPs, drug treatment and HIV testing programs (Davis, Burris, Kraut-Becher, Lynch, & Metzger, 2005; Gu et al., 2014; Rhodes et al., 2003; Ti et al., 2013); some of the practices may be coersive or violent in nature, and influenced by stigma and prejudice against drug addiction (Cooper, Moore, Gruskin, & Krieger, 2005; Hayashi, Small, Csete, Hattirat, & Kerr, 2013; Lunze et al., 2015; Maher & Dixon, 1999; Wood et al., 2017; Cooper et al., 2004; Wood, Taylor, Groff, & Ratcliffe, 2013). Specific correlates of exposure to physical police violence among PWID include male gender, younger age, homelessness, incarceration, compulsory drug detention, cocaine injection and syringe sharing (Hayashi, Ti, et al., 2013; Kutsa et al., 2016; Landsberg et al., 2016), however, no quantitative studies on this topic among PWID have been conducted in the U.S.

Baltimore city, with a population of about 620,000 faces many chronic social and economic and challenges that impact the drug epidemic and health of its residents. A quarter (24%) of the city’s residents live at or below the poverty line, the city has over 20,000 vacant houses and over 10,000 people seek access to homeless shelters annually (Maryland’s Interagency Council on Homelessness, 2016; U.S. Census Bureau, 2010). Policing is a challenging task in a city where crime is rampant with over 9,500 violent crimes recorded in 2016, of which 318 were homicides (Baltimore Sun, 2017). A large and established drug market persists and subsequent encounters between police and people in the drug trade are frequent (Baltimore City, 2017; Beletsky et al., 2015). Frequent encounters between police and participants of the street drug economy are often manifestations of “zero tolerance” policing strategies (Collins, 2007), with officers “often resorting to force—with minimal training and insufficient oversight from supervisors or through other accountability structures” (pg. 5) (U.S. Department of Justice, 2016). Such frequent and severe constitutional violations when experienced, witnessed or discussed may shape community norms and expectations around police interactions, increase risk of psychological distress from secondary trauma, and fuel mistrust of the criminal justice system, particularly among racial/ethnic minorities (Alang et al., 2017; Collins, 2007; Cooper, 2015; U.S. Department of Justice, 2016), impairing effective policing and safety promotion (International Association of Chiefs of Police, 2006).

In order to address the current gaps in the literature, particularly on exposures to police violence among U.S. PWID, and the virtual absence of quantitative research on the phenomenon of knowing someone who has experienced police violence among PWID, this study assessed the magnitude and correlates of physical police violence among PWID recruited from Baltimore city. We modeled direct experiences of police violence as well as knowledge of violence directed towards other PWID peers, the latter is of specific interest given the influence of peers in shaping norms and health behaviors (Davey-Rothwell, Latkin & Tobin, 2010).

Data were from the National HIV Behavioral Surveillance (NHBS) system, which recruited PWID from August to December 2015. The procedures are described in detail elsewhere (German, Linton, Cassidy-Stewart, & Flynn, 2014; Villanti, German, Sifakis, Flynn, & Holtgrave, 2012). In short, PWID were recruited using respondent driven sampling (RDS), a chain referral sampling technique that is suitable for recruiting “hard-to-reach” populations (Heckathorn, 1997). Nine seeds diverse in age, sex, race/ethnicity, and area of residence were selected through random street encounters and outreach at local organizations. Upon study enrollment, all seeds were given five coupons to refer other participants. Subsequent eligible recruits were provided 3–5 coupons.

Following the standardized national protocol, eligible participants were (1) aged 18 years and above; a Baltimore-Columbia-Towson metropolitan area resident (which includes Baltimore city and Baltimore county); and (3) reported injection drug use (Buckley, Calvert, Lapidus, & Morris) in the past 12 months. Seeds were eligible if they met the prior criteria and identified as male or female. Participants completed informed consent procedures and provided oral consent prior to study participation. The anonymous study visit included a 45-minute interviewer-administered computer assisted personal interview (CAPI) survey, which consisted of the 2015 NHBS core questionnaire and local survey items. Together, these incuded questions on socio-demographics, sources of income, area of residence (including ZIP code), housing, drug use behaviors, police interactions and violence, and network size. Participants were compensated 25 USD for completing the survey and 10 USD for each eligible recruit who completed study procedures. The study was approved by Institutional Review Boards at the Johns Hopkins Bloomberg School of Public Health and Maryland Department of Health.

A total of 592 participants were eligible; participants were excluded from the analytic sample if they were seeds (n=10); did not inject within the last year (n=5), had missing values for either outcome (n=7). The analytic sample included 570 participants.

This study is among the first to examine the correlates of physical police violence among U.S. PWID. The prevalence of experiencing physical police violence in the past year was 7% among this urban sample of PWID (and 4% after RDS-weighting), which is consistent with estimates observed internationally (3–14%) (Hayashi, Ti, et al., 2013; Kutsa et al., 2016; Landsberg et al., 2016). Moreover, one in four PWID were socially connected to someone who had recently experienced physical police violence, which may generate negative attitudes and mistrust of police officers even among those who do not directly experience police violence. Police violence was associated with male gender, homelessness, and arrest, as well as drug paraphernalia confiscation and RSS, placing PWID at higher risk of HIV and HCV acquisition. Addressing this complex public health and criminal justice issue will require strengthening monitoring systems and accountability mechanisms for illegal and unconstitutional policing. The implementation of evidence-based strategies targeting police officers, institutional policies and the law may be required.

Within a relatively short period of time, PWID were affected directly and indirectly by physical police violence, and the burden was not evenly distributed. The prevalence observed in the current study is similar to that observed in a previous study that examined the issue among a general urban population (DeVylder et al., 2017). Male PWID were more likely to experience physical police violence than female PWID, which has been shown among PWID in Canada, Ukraine and Thailand (Hayashi, Ti, et al., 2013; Kutsa et al., 2016; Landsberg et al., 2016). One previous study from Ukraine has shown that male PWID are more like to be stopped on the way to drug treatment, bribed, pressured into diclosing information, and detained without an official charge compared to female PWID (Kutsa et al., 2016). In our study, male gender was not correlated with our limited set of law enforcement variables (arrest and syringe confiscation). Future studies could examine the role of gender dynamics between PWID and police officers to help elucidate these findings. Other unmeasured forms of violence (e.g., sexual violence), may disproportionately impact female PWID and require separate investigation (Kutsa et al., 2016).

Syringe confiscation and arrest were prevalent and emerged as types of police encounters that were independently correlated with increased odds of police violence exposure. Results from a qualitative study from Thailand showed that police violence against PWID most often occurs during drug-related searches, arrest and interrogation (Hayashi 2013). More studies are needed to examine the contextual factors surrounding police violence experiences in U.S. settings. While not measured in this study, tactics such as “stop and frisk”, “warrant checks” and “racial profiling” have historically been a part of policing the drug trade in the U.S., which elevate the number of arrests and often result in the violation of constitutional rights (Sarang et al., 2010; U.S. Department of Justice, 2016). Some authors have suggested that decriminalizing drug possession would systemically alter PWID’s risk environment (Debeck et al., 2017). Further research is warranted to evaluate whether such changes could reduce risk of violent encounters between police and PWID.

We noted several direct and indirect implications of police violence on health; PWID who experienced police violence in our study had a twice higher odds of RSS, a key HIV and HCV risk factor, corroborating findings from previous comparable research (Hayashi, Ti, et al., 2013). Other studies examining relationships between police encounters and health have documented that fear of police and arrest can drive unwillingness to carry sterile syringes, rushed drug injection, RSS and HIV infection as well as posing a barrier to accessing SSPs (Beletsky et al., 2011; Bluthenthal et al., 1999; Booth et al., 2013; Maher & Dixon, 1999; Rhodes et al., 2003; Sarang et al., 2010; Werb et al., 2008). Police violence can also fuel institutional mistrust posing a further barrier to accessing services (Beletsky et al., 2015; Cooper, 2015; Hayashi, Small, et al., 2013; U.S. Department of Justice, 2016). There are substantial immediate and chronic impacts of physical trauma documented elsewhere, including the risk of death (Cooper et al., 2004; Cooper, 2015; Human Rights Watch, 2017; Krieger, Chen, Waterman, Kiang, & Feldman, 2015; Sarang et al., 2010). The potential for psychological harms from experiencing or witnessing violence, and being told about a traumatic experience, is concerning given that PWID commonly have histories of trauma (El-Bassel, Gilbert, Witte, Wu, & Chang, 2011; Shaw et al., 2016). A recent study of urban U.S. residents showed higher odds of suicide attempts among victims of police violence; the mental health ramifications of police violence should also be addressed (DeVylder et al., 2017).

Police violence was higher in the lives of PWID who sold drugs; they were twice more likely to experience police violence, and three times more likely to know a peer who had experienced police violence. Law enforcement often intervenes to disrupt the drug trade with the intent to reduce crime, however, a systematic review has suggested that such interventions may paradoxically escalate community violence (Werb et al., 2011). Further research will be required to better understand this dually criminalized and underrepresented subpopulation who may at higher risk of negative interactions with police. We also found that older PWID were less likely to experience police violence though this finding did not remain significant in multivariate modeling. It is plausible that older PWID, particularly lifelong residents, were more likely to underreport police violence due to having a higher perceived threshold for violence reflecting desensitization to institutional racism and chronic intensive policing strategies; this finding will require further exploration. We did not observe an association between police violence and place-based characteristics, i.e., ZIP code level racial composition and poverty. This may be due to lack of variability between the ZIP codes that PWID reside in, lack of statistical power, and the possibility that residents are searched or arrested outside of their ZIP code of residence. ZIP codes are administrative boundaries established for mail delivery and may not adequately reflect the activity spaces and environmental features that increase vulnerability to violent interactions with police. While the finding that White PWID were more likely to report recent police violence than PWID who are racial minorities is counterintuitive based on the literature, this finding may reflect a mechanism through which neighborhood racial segregation increases the likelihood of White PWID being perceived as outsiders in predominantly Black neighborhoods where drug-related law enforcement is frequently targeted as observed in Philadelphia (Rosenblum et al., 2014).

Reducing the number of negative encounters between police and PWID are achievable. Strategies recommended by the American Public Health Association include reversing the militarization of police and addressing the social and structural determinants of crime and police encounters, including homelessness, unemployment, racial discrimination and the criminalization of drug use (American Public Health Association, 2016). Our survey was conducted in 2015 following the death of Freddie Gray and the highly-publicized Baltimore City protests, and prior to the completion of the Department of Justice investigation of the Baltimore City Police Department (BCPD). Sweeping reforms built on the tenets of community-oriented policing are underway at BCPD, including the pilot of a Law-Enforcement Assisted Diversion (LEAD) program (Rouhani et al., under review); continued monitoring will be necessary to assess whether violent encounters are reduced as a result of these changes.

Decreasing police violence against PWID has been shown to be possible in other settings; the prevalence of past 6 month police violence against PWID in Vancouver decreased from 14% to 3% over the span of ten years following the implementation of a harm reduction informed policing strategy (Landsberg et al., 2016). Arrest as behavioral deterrence may not be effective for addiction (Wood et al., 2013) and the decriminalization of drug use and possession is likely to have a substantial impact on the likelihood of human rights abuses experienced by PWID, particularly at the hands of police (American Public Health Association, 2016; Csete et al., 2016). Currently, drug possession for personal use” and drug paraphernalia convictions can result in four years of incarceration and drug possession with “intent to distribute” carries a sentence of up to 20 years of incarceration (Md. Ann. Code § 5–619; Md. Ann. Code § 5–601, 5–402, et. seq.). Continued research (e.g., population-based surveys) will play an important role in monitoring the progress made through structural change.

This study is subject to some limitations. Our prevalence may be an underrepresentation of the true prevalence of police violence due to survivor and healthy participant bias. Given the low representation of transgender PWID in the sample, there is certainly a need for similar research among transgender communities where risk of negative police interactions is also high (U.S. District Court for the District of Maryland, 2017). Future studies are needed to assess better methods of reaching transgender PWID and examine whether the nature of police encounters are different among this dually stigmatized and marginalized community. Our analysis approach was somewhat limited by the small number reporting the main outcome (n=42). Studies of larger sample size may help to further elucidate the relationship between police violence and individual and contextual factors, as well as health outcomes. The definition of police violence used in this study should be distinguished from excessive or intentional force as further contextual data are necessary to deepen our understanding of the findings. We did not capture other forms of police violence including sexual and psychological violence, and witnessing violence; we do not know the circumstances of any reported violent encounters. The cross-sectional design does not allow assessment of temporality. Survey data may be subject to recall and social desirability bias. The strengths of this study are that this is the most representative sample of PWID available in the greater Baltimore area and that RDS permits us to engage highly marginalized and vulnerable population in research. Future epidemiologic studies could incorporate a more comprehensive measure of police violence and mental health effects that are measured longitudinally.

These findings add to the growing literature demonstrating police encounters as an integral component of the risk environment of PWID that may directly and indirectly influence public health goals. Alongside enhanced research, comprehensive social, cultural, structural and legal reforms will be necessary in reducing the burden of police violence among PWID.