Achieving improved health and human rights for street-based sex workers - an ethnographic exploration of factors that drive policing of cisgender female sex workers in a U.S. setting

Background: Building on a broader sociological discourse around approaches towards vulnerable populations, increasing public health evidence as a key health determinant, particularly among street-based sex workers. Despite the importance of policing as a structural health determinant, few studies have sought to understand the factors that underlie and shape policing practices towards sex workers. This study therefore aimed to explore the drivers for policing attitudes and practices towards street-based cisgender female sex workers. Methods: Drawing on ethnographic methods, 280 hours of observations with police patrol and 10 stakeholder interviews with senior police leadership in Baltimore City, USA were carried out to better understand the drivers for policing strategies towards cisgender female sex workers. Analysis was data- and theory-driven, drawing on the concepts of police culture and complementary criminological and sociological literature exploring the more explicitly spatial exercise of police discretion and structural forces operating on police decision making. Results: Ecological factors at the structural (e.g., criminalization), organizational (e.g., violent crime control), community and individual level (e.g., stigmatizing attitudes) emerged as key to shaping individual police practices and attitudes towards cisgender female sex workers in this setting. Findings indicate senior police support for increased alignment with public health goals. However, the study highlights that interventions need to move beyond individual officer training and address the broader structural and organizational setting in which police practices towards sex work operate. Conclusions: A more in-depth understanding of the circumstances that drive law enforcement approaches to street-based sex work is critical to the collaborative design of interventions with police in different settings. In considering public health-police partnerships to address the needs of sex worker populations in the U.S. and elsewhere, this study supports existing calls for decriminalization of sex work, supported by institutional and policy reforms, neighborhood-level dialogues that shift the cultural landscape around sex work within both the police and larger community, and innovative individual-level police trainings.


Introduction
3 The criminal justice system has been identified as a prominent apparatus for reinforcing stigmatization amongst vulnerable groups (Krüsi et al. 2016a). Police interactions are an omnipresent feature in the lives of many marginalized communities, shaping their everyday existence (Stuart et al. 2015). Extensive criminological and sociological literature has explored the role of policing in the social control of urban spaces and the communities within them (e.g., homeless persons, black youth, people who use drugs) (Stuart 2016;Goffman 2015;Bittner 1967;Moskos 2008). These works have included an examination of how policing drives socio economic exclusion and 'system avoidance' whereby marginalized groups fail to access healthcare and other forms of social support that represent fundamental human rights (Stuart et al. 2015). Street-based cisgender female sex workers (FSW) represent one such important marginalized group, whose working conditions (e.g., criminalization of sex work, community stigma, client violence, frequent drug use) (Sanders & Campbell 2007) place them in frequent contact with the police. Criminalization and police enforcement of sex work has increasingly been associated with negative health-related outcomes and human rights abuses among FSW including, increased risk of experiencing different forms of violence and sexually transmitted diseases (Platt et al. 2018).
FSW have been identified as a population who face a complex constellation of structural vulnerabilities (e.g., poverty, low health care access, exposure to violence, drug use) (Shannon et al. 2014). Within this broader portrait of social and material adversity, research has begun to map the pathways by which policing of FSW influences individual (e.g., condomless sex) and interpersonal (e.g., client violence) HIV related risk (Rhodes et al. 2012;Shannon et al. 2015). Laws relating to sex work furnish police with considerable discretion, which in turn has been shown to lead to exploitation in many contexts (Erausquin et al. 2015). Across different global settings research has documented human rights abuses by police towards FSW, including harassment, sexual and physical violence, and coerced sex under the threat of arrest (Decker et al. 2015). These types of violations impact women's safety (undermining FSWs' ability to seek police protection) and HIV risk (police-perpetrated sexual violence is often unprotected and associated with STI/HIV infection) (Decker et al. 2015). Studies have additionally measured enforcement strategies (e.g., police crackdowns, arrest, condom confiscation) 4 which can similarly impact women's safety and harm reduction around STI and HIV transmission and infection (Footer et al. 2016). For instance, frequent arrest, crackdowns, and move along-tactics that displace FSW to unfamiliar areas can increase vulnerability to violence and move women further away from outreach and HIV prevention services (Shannon et al. 2009) and are associated with client violence (Footer 2018). Condom confiscation can prompt unprotected sex and has been associated with STI infection (Erausquin et al. 2011;Wurth et al. 2013). A large proportion of street-based FSW in settings, including the U.S., are also illicit drug users and regularly face the dual risk of negative police interactions and arrest associated with drug use (Beletsky et al. 2012;Werb et al. 2011;Odinokova et al. 2014). Studies have shown the adverse impacts of drug-related enforcement practices on HIV, including a positive association between syringe confiscation and HIV infection (Strathdee et al. 2011;Flath et al. 2017). Law enforcement practices such as enhanced police surveillance represent a major impediment to utilization of syringe services programs and other harm reduction approaches (Davis et al. 2005;Shannon et al. 2008) that represent one of the most effective ways to protect rights and limit the human suffering associated with addiction.
In response to this accumulating evidence, attempts to forge partnerships between sex worker organizations and the police across diverse settings (including India, Nepal, Thailand, Australia) have been documented. Interventions have included improving communication between the local sex worker community and law enforcers, police trainings, and introduction of new policies to improve sex worker safety (Tenni et al. 2015). In the U.S. there is a dearth of intervention work directly addressing the policing of sex workers. However, in the broader context of the failed 'war on drugs' (UNODC 2016), the U.S. is now mainstreaming public health-oriented harm reduction interventions aimed at illicit drug use (Flath et al. 2017;Cusick 2006;Jardine et al. 2012;Beyrer 2012) which has implications for many FSW who also use drugs. Studies have evaluated the effectiveness of police trainings and policy changes (e.g., diversion programmes) aimed at shifting police attitudes and practices, while securing police buy-in through a lens of occupational health and reducing rates of recidivism (Beletsky et al. 2011;Silverman et al. 2012;Collins et al. 2017).
Despite some examples of positive steps to address policing's contribution to the HIV risk 5 environment of two overlapping vulnerable populations, evidence suggests that shifting police culture and gearing day to day patrol practices towards public health orientated goals is both complex and challenging. In Krusi et al. (2016), a qualitative exploration of FSWs' policing experiences under a new Canadian policing policy intended to prioritize sex worker safety, the authors found that stigmatizing attitudes overshadowed and undermined any improvements to police protection (Krüsi et al. 2016b). More studies are needed that focus on understanding the mechanisms and drivers that underlie police behavior and decision making. Such insights are critical to the design of interventions and police public health partnerships that address the health and rights of FSW and other marginalized populations. Current research is heavily based on quantitative epidemiological surveys and/or self-report qualitative interviews with FSW, but not police. Burris (2004) has suggested that research has missed the opportunity to focus on looking at the behavior of the police towards vulnerable populations and the drivers for it, thereby creating avenues from which to design interventions that address the underlying causes of the negative health and human rights impacts of policing.
Ethnography engenders analysis that attends to the variance in and contextual specificity of policing taken towards marginalized groups. This approach allows for an examination of the gap between 'laws on the books' (Burris et al. 2004) (i.e., criminalization of sex work that demands arrest) and the realities of patrol officers' day to day practices. Bittner's (1967) formative work on Skid Row and many subsequent ethnographic studies have emphasized the discretion and autonomy with which street-level officers deploy the law and legal enforcement policies, with law often providing the legitimacy and authority to impose a variety of unchecked practices (Bittner 1967;Manning 2014). To understand what shapes police behavior towards vulnerable groups, the concept of police culture, defined as the shared norms, beliefs and values that shape police behavior (Loftus 2010), is important. Traditional accounts of police culture suggest a homogenous police mentality whose values can be highly problematic with respect to vulnerable groups and human rights compliant policing. Cultural racism, for instance, has manifested in over-policing and abusive practices towards black and other minority groups (Chan 1997;Loftus 2009). In the field of domestic violence, elements 6 of traditional police culture including notions of what amounts to 'real policing' and a masculine ethos have been shown to influence officers' exercise of discretion and dismissive attitudes to domestic violence work (Myhill & Johnson 2016). More contemporary accounts of police culture have moved away from a 'monolithic' interpretation of police culture, instead exploring the influence of broader social and political conditions, organizational considerations (e.g., leadership, supervisory styles) and individual officer occupational styles and agency in shaping behavior (Loftus 2009). Neighborhood context has also been shown to influence police behavior. Klinger (1997) in his theory of policing explores the connection between police decision making and their ecological environments, putting forward the proposition that officers police less in high crime areas (e.g., viewing victims as less deserving, and crime as more normalized.) (Klinger 1997). More recent work by Chan (2004) exemplifies contemporary efforts to understand the impact of the broader 'policing environment' in shaping officers' working culture (Chan 2004).
This study seeks to identify some of the underlying drivers shaping police practices that have been previously examined in terms of their impact on FSWs' health and HIV risk. In particular, the present study aims to explore what factors influence: a) police practices already identified in the literature as adversely affecting FSWs' HIV risk and human rights; and b) police officers' ability and willingness to adopt a more public health and rights orientated approach to street-based sex work. Through the observation of patrol officers in a major U.S. city, we explore these factors with particular attention to the broader ecological environment in which policing occurs. By ecological environment/context we are referring to economic, social, political and economic factors, in addition to broader physical environmental influences (Burris et al. 2004). Our methods were designed to produce a nuanced description of the policing of street-based sex work to better inform possible future health and human rights and public safety interventions that mitigate the established negative health impacts of policing on sex workers.

Context and research setting
Baltimore is an East Coast city that has been challenged by a sustained history of social and economic hardship, largely divided along racial lines, with white prospering neighborhoods sitting in stark contrast to impoverished and largely African American ones. A 2011-2015 U.S. census report highlighted that 23.7% of city residents live below the poverty line (American Community Survey, 2011. Deeply embedded structural inequities are rooted in the decline of the city's manufacturing industry which led to major demographic change within the city, often referred to as the 'white flight', but also marked by the later exodus of Baltimore's black middle class (Levine 2000).
Today Baltimore's low-density city is best described as enclaves of wealth nested within expanses of city terrain characterized by open-air drug markets and high rates of violence. Street-based sex work, all aspects of which are criminalized (buying and selling) in the state of Maryland, is predominantly located in spaces that are within or on the borders of Baltimore's surviving working class communities and the most hollowed-out neighborhoods. In the latter, boarded-up row houses often outnumber inhabited homes and the stoops are occupied by the runners of an illicit drug economy that perpetuates the city's violent crime. Street-based FSW in this setting are characterized by high rates of STI infection and co-morbidities related to drug use and violence (Davey-Rothwell & Latkin 2008;Footer et al. 2019;Sherman et al. 2019). In particular, studies in this setting highlight the magnitude of the overlap between street based sex work and injection drug use. Amongst a cohort of 250 cis-FSW recruited in Baltimore City, 70% reported daily heroin use and 60% reported ever having drugs or drug paraphernalia confiscated (Footer et al. 2019). Despite the city's predominant black population, a number of the main sex worker strolls in Baltimore are dominated by white drug-using women, a group who in many ways embody the city's growing urban poverty.
Dating back to the early 1990s, the police response was one of 'zero tolerance' policing, characterized by large number of stops, searches, and arrests (Collins 2007). During the time that this study was conducted, Baltimore city saw a surge in murder rates, with 344 murders recorded for the year of 2015 when data collection commenced (up from 211 murders in 2014) and a police department under unprecedented scrutiny and pressure (Lowery 2016). This scrutiny was occasioned by the death of Freddie Grey, a 25-year old black African American whilst in police custody. A subsequent report by the US Department of Justice concluded that Baltimore Police Department (BPD) requires widespread reform. Findings included unconstitutional stop and searches, and excessive use 8 of force by officers, further eroding already fragile community trust in law enforcement. Although sex workers were not considered as a distinct population by the Department of Justice investigations, the report suggested that there could be a gender bias affecting police treatment of vulnerable women in the reporting of sexual assault, including FSW (Department of Justice Report, 2016 pg. 11). Our ridealongs occurred in police districts characterized by these abuses and police deficiencies, as well as the socioeconomic disparities that foster the police department's ongoing major enforcement challenges.

Methods
Ethnographic methods (i.e., key informant interviews, observational ride-alongs with police officers during their shifts) were employed in four police districts in Baltimore. In selecting these police districts, we first looked at open-access Baltimore arrest data between 2013 and 2015 for prostitution and drug arrests, selecting arrest records for females only (Open Source Baltimore). Districts that had higher levels of prostitution or prostitution and drug arrest activity were selected. Secondly, we held meetings with the Commanders at identified districts and leadership at outreach organizations (e.g., Baltimore Needle Exchange Program) who validated our initial choice of districts based off the arrest data. This paper reports on observational fieldwork conducted in these districts between 2015 and 2016.
The research team consisted of six field observers all trained in ethnographic observation techniques including social epidemiologists (KF, SP, SL), a criminologist (BS), an anthropologist (VK) and a research assistant with extensive knowledge of Baltimore City and our target population (SH).
Observational ride-alongs with officers were conducted throughout the year during a variety of shift times to maximize variation in observations and officers. In addition to routine patrol officers, ridealongs were done with two other specialized units that focused on sex work and drug policing. A total of 54 ride-alongs (281 hours) were completed between July 16 th 2015 and February 25 th 2016 with 64 unique officers. The majority of ride-alongs (n=46) were with routine patrol, and a small number were with two specialized units, described below, n=2 with Vice and n=6 with District Operations.
Vice is a centralized unit that conducts undercover stings to arrest sex workers and clients, and conducts enforcement around sex trafficking. In addition, district-level operations units consist of plainclothes officers whose role includes conducting undercover drug purchases and street stops investigations to build cases against violent offenders and drug organizations. All officers provided oral informed consent and field observers accompanied them for a 2-10 hour period of a 10-hour shift, with an average ride-along lasting 5 hours. A field note aid and an encounter form were used to help field observers structure note taking (e.g., features of environment, policing priorities, officers' attitudes/views, interactions, routines and practices), capture specific elements of police-FSW encounters (e.g., location, nature of incident, actions, use of discretion), and ensure consistency across observations. At the same time fieldworkers were also encouraged to record observations, reflections, and even idiosyncratic details that enabled them to better contextualize their experience as it occurred in the moment, while preventing misinterpretations that might occur if quotes or observed details were decontextualized. Observers generally rode along with a single officer and sat in the front of the vehicle, which aided conversation and rapport building. During street stops and citizen encounters observers would, if safe and appropriate, leave the patrol vehicle and observe exchanges. Field-notes were written up within 48 hours by observers to minimize recall bias. Standalone quotes from field notes or quotes integrated within a field note are italicized to denote they are the direct quotes. All identifying information has been removed to ensure anonymity of the police districts and those observed. Field notes and interview extracts are labeled with the ride-along officer(s) type or rank, gender, ethnicity and age. In addition, any other individuals directly quoted in the field note are additionally identified (i.e., FSW, local resident).
Key informant interviews (n=10) were conducted with police leadership (e.g., District Commanders, Community Liasons, Vice). Sampling was purposive, based on knowledge of the research team and opportunistic, with interviews limited to those available and willing. Seven key informant interviews were audio recorded. Extensive notes were taken during the 3 remaining interviews. Notes were written up within 48 hours of an interview and audios transcribed verbatim. Key informant interviews explored organizational attitudes among senior officers around policing sex work, attitudes towards public health-orientated approaches to policing marginalized populations, opportunities and challenges to improve relationships, and alternative policing approaches towards the FSW community in Baltimore City. All key informant quotes are anonymized except for the type of officer interviewed (e.g., District Commander).
Ethnographic data was analyzed using a data-and theory-driven approach (DeCuir-Gunby et al. 2011) and was both inductive and deductive (Fereday & Muir-Cochrane 2006). The concept of police culture (Loftus 2009), including how broader ecological factors can shape police culture (Klinger 1997), and complementary criminological and sociological literature exploring the more explicitly spatial exercise of police discretion, guided our analysis (Herbert 1997;Hubbard 2004). Initial open-coding involved reading and line-by-line analysis of typed field notes from all 4 districts and key informant interviews by the first author to categorize and conceptualize the data by hand. This was followed by multiple rounds of revisions by the coding team using Nvivo Software. Team feedback on emergent categories built an audit trail to ensure the inclusion of multiple perspectives and enhance reliability (Maxwell 2012).
Approval for the study was obtained from the ethics committee at Johns Hopkins Bloomberg School of Public Health and conducted in accordance with a pre-obtained memorandum of understanding entered with Johns Hopkins Bloomberg School of Public Health and the Baltimore Police Department.

Results
Of the 64 officers in our study, 86 percent identified as male and 14 percent identified as female.
Approximately 63 percent of ride-along participants were white and 27 percent were black, the remaining were Asian or other. Thirteen percent of participating officers identified as Hispanic.
Officers' years of experience ranged from less than one year up to 34 years, with the average being seven years. Most officers reported that they did not live within Baltimore City.
Our observations with police in Baltimore City and key informant interviews with police leadership identified important themes that: 1) influence ongoing enforcement practices known to have a detrimental impact on FSWs' health and rights; 2) impact non-enforcement that have implications for FSWs' health and safety; 3) shape police's continuing failure to protect FSW, including addressing sexual violence and 4) impede a shift towards a more public health and human rights orientated approach to policing sex work. Our results point to ecological factors at the structural, organizational, and community levels that shape individual level police practices and underpin police attitudes.
Based on our findings, the discussion addresses the implications of this study for intervention and policy development to drive forward a public health and human rights centric policing of sex work.

Factors influencing 'police enforcement' practices towards FSW
Observations in this study highlight key ecological factors driving police enforcement approaches. Patrol officers, whilst rarely able to arrest sex workers for prostitution (due to the high evidential Consistent with other literature (Lyons et al. 2017;Hubbard et al. 2007), our findings also pointed to community association and homeowner complaints being a strong driving force behind FSW-targeted Vice 'sweeps.' Vice, unlike patrol, are responsible for conducting 'stings,' during which plainclothes officers solicit services from women to provide the necessary grounds for a prostitution arrest.
However, an additional ecological factor driving these 'sweeps,' alongside community pressure, is the periodic need to respond to upticks in violent crime, such as shootings.
"There has been a crime initiative the last two weeks as the crime is out of control, so we put boots on the ground including Vice. The last two weeks we've just been out rounding the women up [FSW]." (Vice Command) While FSW themselves do not commit these violent crimes, they often observe crimes occurring on the street and can provide the police with valuable information. As this Vice officer explained during an observation of a sting operation: I took the opportunity to ask the officers whether they thought what they were doing was getting women (FSW) off the streets. The Vice officer responded as he let out a laugh: "Ha. Not at all. They'll be walking the streets again as soon as they get out (of jail). We're out here for information basically." (Vice Officer)

Factors influencing 'non-enforcement' approaches to sex work
Observations revealed that in marginalized and high-crime city locations, patrol officers do not police sex work as aggressively as they do in more residential and gentrified areas. Marginalized neighborhoods are characterized by few residential complaints and high levels of drug activity and violence. Observers' field notes consistently documented that police take no action to displace FSW in these areas, despite frequent sightings of sex workers actively working the strolls. Additional factors shape patrol officers' approach to enforcing or not enforcing drug laws. A neighborhood context in which drug use is ubiquitous and police prioritize calls for service associated with violent crime, is coupled with variations in individual officers' outlooks. Whilst 'rookies' seem more enthusiastic about making low-level drug arrests, more seasoned officers seem more cognizant of the bureaucracy entailed in arresting vulnerable groups such as FSW, and of the limitations on their time. As this patrol officer explained: 13 "She [FSW] probably had a warrant if I had run it." However, he doesn't want to find out because "if she goes to central booking, she won't pass medical and then I have to take her to the hospital and stay until she's cleared before taking her back to booking." He says they're already short on officers, "We try to avoid locking up certain groups …handicapped people," or other people who might not pass their medical. "Once you have a couple years on you, you know these things," he says. (Patrol, Male, Asian, 32) However, the mere potential to invoke drug related arrests, remakes FSW as important assets for intelligence. District commanders described FSW as "the Google search engine of the street." Although some FSW are paid informants, more commonly the exchange of information is informal and drives the frequency of interactions in marginalized spaces between police patrol and FSW. These exchanges are predominantly characterized as simply 'checking in' and being friendly to FSW, although as this quote illustrates there is always the underlying coercive threat of enforcement: The officer started talking about FSW as informants, "We will buy them food all the time, or now with it getting colder, we'll get them hot chocolate or coffee. They will literally tell you everything." Another officer chimes in, "The dealers and stuff don't treat them good, so they have no problem giving up information on them." He elaborated, "We have a good relationship with them. We leave them alone [i.e. don't arrest], and in return they give us information." (District Operations, Male, White, 24) It appeared that these exchanges primarily contribute to many officers' intimate knowledge of FSWs' lives and vulnerabilities (e.g. homelessness, lack of employment opportunities, hunger, need for wound care). However, officers rarely utilize this knowledge and daily contact with FSW to offer any bridge to health or social services. Instead, observations highlighted that help is ad hoc and predominantly reflects the moralistic and gendered attitudes of individual officers, with no guiding organizational norms around public health policing: "Sometimes I do run into women who are genuinely selling their pussy to make ends meet -feed their kids, pay rent, etc., and when that happens I try to connect them with some services." (Patrol Officer, Female, Black, 35)

Factors hindering police protection and assistance to FSW
Many officers view violence towards FSW as an inescapable part of the street existence, as opposed to crimes against vulnerable women that properly deserve police attention. Officers' attitudes towards FSW as less deserving of police assistance and protection emerged as closely linked to broader stigmatizing attitudes to FSW in this context, specifically shaped by their drug use. This was underscored by the frequent documentation during ride outs of dehumanizing police banter including phrases such as "pregnant prostitute junkie", "skanky", "disgusting" and descriptions of FSW drug related sores and poor physical condition. One officer referred to a FSW he knew on the strolls simply as "abscess". Some of the most dehumanizing language involved the use of animal imagery, including describing FSW as junkies waking up for "feeding." A small number of officers did describe FSW safety as more of a priority and provided examples of proactively addressing FSWs' needs. This emerged at the level of the individual officer and did not appear to translate into a more generalized organizational concern to address and prioritize FSW safety. Although, during key informant interviews police commanders would give anecdotal examples of their officers assisting women by buying food, bringing them warm clothes, and taking them to local service providers.
During the time that ride-alongs were conducted, a significant institutional shift towards the broadening of collaborations between public health and public safety was instigated by the BPD. This involves officers carrying Naloxolone (narcan) for overdose victims. However, as this vignette illustrates, patrol officers' attitudes embody a lack of sensitivity and embedded 'canteen (discriminatory) culture'(Van Hulst 2013) towards FSW who inject drugs: She [FSW] had nodded out from heroin and was laying on the steps but was still breathing. "Wake up, ma'am! Ma'am! Ma'am wake up!" the officer said to her, with no response. He then shouted, "Narcan!" and was about to use it but she shot up to her feet instantly and began walking away. She made it about a block and a half away when she hunched over and fell asleep standing up. He again yelled, "Narcan!" and she woke right back up and walked away. "Man! This Narcan stuff is great! It works so good you don't even have to use it!" he said. Everyone shared a laugh and the female officer showed us a video entitled 'Baltimore Gold' of a women nodding out from heroin in a parking lot and a male that was with her trying to keep her awake. (Patrol,Male,White,29)

Identifying barriers to more public health and human rights orientated policing of FSW
In interviews with district commanders the phrase "you can't arrest your way out of the problem," Command also recognized the realities of day-to-day patrol policing and organizational resource constraints (e.g., shortages of officers, other policing priorities, lack of specific training) considerably hamper a proactive response. This was backed up by observations with patrol where field observers noted understaffed shifts, backed up calls for service, and a focus on the surge in violent crime.
Interviews with District Command advocated for organizational level improvements to officer trainings, aimed at better equipping officers for the "social worker aspects of policing" (District Commander, Male, White, 45). However, patrol officers sense of low self-efficacy around helping FSW focused on the more structurally embedded lack of external support from other agencies, coupled with a fatalistic attitude towards FSWs' prospects of finding a way out.
"People expect too many things from us -we're not social workers." He says, "I feel sorry for them, same as for young guys on the corner selling drugs. But it's not our primary duty to deal with them directly. There are other agencies getting paid [to provide these sort of services], but where are they?
It takes an act of God, damn near, to get them to help them [the women] -they are failing all of us." (Patrol Officer,Male,White,35) The officer explains that in his experience there's not really much you can do -sex workers are stuck in what he calls a "revolving door." (Patrol, Male, White, 47)) Crucially, the broader structural context of criminalization and stigmatization emerged as hampering a shift in policing at an individual officer and organizational level towards public health and human rights approaches. Despite many patrol officers' negative attitudes to FSW, many advocated that a first step would be to move away from criminalization: "Don't tie my name to this but they just need to decriminalize it. From talking to these girls out here you see they have so many issues and they see drugs and prostitution as their only way out. If it was

decriminalized, and I don't know what kind of [employment] benefits they would have or what that
would look like, maybe you can figure that out, but then they could get the help they need." (District Operations, Male, White, 24)

Discussion
Our ethnography provides some of the first evidence generated from studying police that both elucidates and confirms factors shaping police practices towards FSW, largely absent from the public health and social science literature to date. The study seeks to complement and bridge the divide in the existing literature, which on the one hand has included prominent policing ethnographies (Stuart 2016;Weitzer 2017;Goffman 2015) investigating police practices towards other vulnerable populations, and on the other socio-epidemiological literature specifically focused on documenting police's contribution to FSWs' risk environment (Shannon et al. 2015;Platt et al. 2018). In seeking to identify the underlying factors that influence the policing of FSW, our findings provide a platform from which to discuss how to most effectively intervene to alleviate the negative health and human rights impacts surrounding the policing of FSW in both the U.S. and other contexts.
Social-spatial control through law enforcement tactics has been highlighted by criminologists and sociologists to aid the 'cleaning up' of urban spaces (Beckett & Herbert 2008;Yarwood 2007;Klinger 1997). In this study, "move along" tactics by patrol and 'sweeps' by Vice, were shaped by an ecological context in which more affluent neighborhoods are dominated by communities' moral concerns (e.g., children seeing sexual activity on the way to school, discarded condoms on porches).
Interestingly, many patrol officers expressed frustration toward community complaints, which they viewed as born of social anxieties rather than actual crimes, and disproportionately drawing on already overextended patrol resources. These findings confirm existing literature (Krüsi et al. 2016b;Hubbard et al. 2007) that point to the importance of gentrification and community level stigma in driving an ultimately ineffective police response to sex work that compromises FSWs' safety. An interesting additional contextual factor shaping displacement of FSW by the police, are the periodic upticks in violent crime that characterize cities such as Baltimore. Spikes in violent crime prompt crackdowns by Vice; these sweeps are not intended to address sex work, but utilize the threat of arrest to coerce FSW into providing criminal intelligence. These tactics displace FSW from their usual strolls, as they attempt to avoid such crackdowns, and highlight the coercive nature of police relations with FSW that can contribute to a climate of vulnerability in which FSW rarely view police as interested in their rights as individuals or collective wellbeing. Public health research additionally highlights the direct and indirect impact these different spatial displacement tactics can have on FSWs' safety and HIV risk factors (e.g., client violence, condom use, current STIs) (Platt et al. 2018).
Our results indicate that neither patrol nor command view legal enforcement tactics as effectively addressing sex work. Instead officers are engaged in reproducing socially dominant ideas from local communities about what is criminal conduct, regardless of actual evidentiary standards for those crimes and despite its potential to undermine FSW safety. Interventions need to focus on educating police departments and the communities they serve about the public health implications of these policing tactics. Through public safety and health partnerships and officer training, police departments could work in these communities to highlight the ineffectiveness of existing policing approaches and question communities' categorization of FSW as a public nuisance rather than a public health issue. Such steps could do much to alleviate the presently damaging spatial control of 18 FSW in this study's context, and others. It is potentially harder to address organizational police policies that utilize prostitution enforcement as a tool for gathering criminal intelligence where there is such obvious police capital. Instead this example of police misuse of power serves to support and emphasize the importance of decriminalization in dismantling the legal apparatus under which police discretion is used to exploit sex workers in ways that not only undermine their health and human rights, but limit any trust that could be fostered between this population and the police.
In line with other policing studies, enforcement practices in 'marginalized spaces', such as open-air drug markets, contrasted enforcement in 'prime spaces' such as more affluent residential neighborhoods discussed above. As Stuart (2014) explains, 'police discretion is, first and foremost, contingent upon how officers conceive the neighborhoods they patrol' (Stuart 2014). Within the milieu of the open-air drug trade, FSW were not prioritized as targets of sex work-related patrol enforcement at all, instead it was their identity as drug users that shaped patrol interactions. Observations revealed considerable variation in drug enforcement based on patrol officers' discretion. Less experienced patrol officers were characterized as being more willing to conduct drug arrests, while more experienced officers recognized the bureaucracy entailed in low level arrests and lack of real police capital. In contrast, all officers utilized the underlying threat of drug arrests to extract information on violent crime. Many police officers tended to achieve this by building relationships with FSW, developing considerable knowledge of women's vulnerabilities along the way. Interventions that seek to shift police's role towards helping women (e.g., on the spot referrals for HIV testing and counselling), could consider how such frequent day-to-day interactions with FSW and knowledge can be utilized to bolster FSW' rights to access social and health related services. However, it must be considered that whilst during observations such interactions appeared friendly and good-natured, they are underscored by unequal power dynamics, shaped by broader structural forces of criminalization of sex work and drug use.
Consistent with findings in other settings (Deering et al. 2014;Klambauer 2017) police confirmed high levels of violence experienced by FSW, which has been linked in the public health and human rights literature to FSWs' risk of HIV acquisition. In addition, studies, most recently in the U.S., have found that FSW who have experienced abusive police practices are more likely to experience client violence (Footer et al. 2019). Officers acceptance of FSWs' violence as inevitable was reinforced by the violent context in which FSWs' lives are situated, and appeared to act as a normalizing factor underscoring officer attitudes. In addition, gendered assumptions and a collective masculine police culture, amongst both male and female officers, influenced the ranking of sexual violence as a low policing priority, compared to other violent crime. Only a small number of officers' patrol interactions with FSW reflected a more guardian (i.e., protective policing) style. Furthermore, observations in this setting revealed little evidence of police proactively helping women (e.g., with social service, drug program referrals), coupled with evidence of deeply embedded stigmatizing attitudes. Studies have highlighted the dehumanization of vulnerable populations by the police, often connected to the spaces they occupy (Stuart 2016;Goffman 2015;Manning 2014). In this study, both male and female officers' language reflected highly stigmatizing normative attitudes toward FSWs' drug use. There was an extraordinary tension between many officers' intimate knowledge of FSW lives and moments of empathy, and their frequent dialogue of 'othering'. Legal literature has highlighted the role of legal controls in reinforcing negative stereotypes and perpetuating social exclusion and vulnerability (Wakefield & Uggen 2010). Here observations suggested that the relentless exhaustion and sense of futility embodied in police attitudes and practices is amplified by the structural environment of poverty and drug use in which they encounter FSW, in turn reinforcing officers' stigmatizing attitudes.
Collectively our findings highlight the need for more far reaching structural changes to address the broader environmental context in which the policing of sex work operates. In particular, patrol officers cited the need to remove criminalization of sex work and ensure there is adequate and integrated external support from health and social services, as key to better human rights orientated public health and safety partnering. Police command also pointed to the very real organizational limitations to prioritizing public health partnering around vulnerable populations, such as sex workers.
Limitations include a lack of dedicated financial resources and considerable constraints on police time, with drastically understaffed shifts being the norm. It is these bigger shifts in the structural landscape that will ultimately pave the way to a new era of public health informed policing towards 20 sex work, and could ensure similar systemic changes to those that occurred around the policing of mental health (Paterson & Pollock 2016). Additionally, addressing the prevalent gendered and stigmatizing assumptions of police officers towards FSW in this setting and others has been tackled through police sensitivity trainings in other settings (Erausquin et al. 2015), including peer education of officers in collaboration with sex worker organizations (Tenni et al. 2015), and should be explored in this context. Trainings, however are insufficient without an organizational shift towards prioritizing FSWs' human rights and health, in particular, introducing policies and systems of accountability requiring police to respond equitably to FSWs' experiences of violent crime. Another important dimension of officers' embedded stigmatization towards FSW surrounds their drug use. In line with ongoing political and policy shifts at a national and state level to introduce pre-booking diversion for low level offences in the U.S., Baltimore City has recently joined the 'LEAD' program (Collins et al. 2017). LEAD introduces diversion to drug and social services at the point of arrest, albeit currently targeted specifically at drug using populations, rather than FSW specifically. Despite this, it presents an opportunity to better institutionalize harm reduction and public health thinking into policing practices around vulnerable populations drug use. It also has the potential to ensure a more uniform approach to police officers current highly discretionary interactions with low level drug users, including FSW.
The study was characterized by several limitations. In setting out to conduct this ethnography with the Baltimore Police Department, the scrutiny around police practices may have led police to be less frank in their opinions, or for districts to pick 'by the book' officers for field observers to ride-along with. The nature of observations also meant that there were no expectations around witnessing the most egregious police behaviors towards FSW during ride-alongs, despite these having been documented in work with FSW in this setting [Blinded ref]. Despite these limitations, the study was able to establish an open and frank dialogue with the police department around the aims of the study through initial key informant interviews, allowing us to capture an honest portrayal of policing approaches and priorities towards FSW. Although the presence of observers in the field can influence officers' actions, field notes revealed the extent to which officers let down their guard and shared 21 frank opinions during observations. In turn it is acknowledged that the ethnographers involved in this study carry with them their own sources of bias, particularly working with stigmatized populations like sex worker populations. Diversity in background and experience (including gender, time in the city, previous work with the police) was intended to help mitigate the effect. A further limitation of the data relates to the lack of focus on race as a socio-structural category that may have had a bearing on police-sex worker interactions. FSW that were observed and encountered during ride-alongs tended to be white. It is beyond the scope of this paper to unpack the reasons for encountering a predominantly white street-based FSW population in the context of city where economically deprived neighborhoods are disproportionately black. The predominance of white FSW could be tied to the relationship between street-based sex work and the drug economy, or a preference among other FSW populations to utilize indoor venues.
In conclusion, much can already be garnered from current public health and human rights literature on the types of police practices that compromise FSWs' health and safety, particularly as it relates to HIV and violence. Still, a more in-depth understanding from the policing side of the circumstances that drive law enforcement approaches to street-based sex work has been lacking. Such an understanding is critical to the collaborative design of interventions with police in different settings. In considering public health-police partnerships to address the needs of FSW populations in the U.S. and elsewhere, this study supports existing calls for decriminalization of sex work, supported by institutional and policy reforms (e.g., around coercive police practices, FSW street safety), neighborhood-level dialogues that shift the cultural landscape around sex work within both the police and larger community and innovative individual-level police trainings (e.g., to address sex work and drug related stigma). Points of leverage include the general acceptance of officers around the need for change and a frustration with ineffective policing strategies towards sex work. A shift in police culture that questions communities' moral sensibilities around sex work and categorization of FSW as a public nuisance, rather than a vulnerable group in need of health and social support could help alleviate the damaging spatial control of FSW that have been shown to promote HIV risk behaviors (e.g., rushing negotiations, screening clients, and condom use). Many officers of all ranks favored some form of legalization or decriminalization, which is a view is heavily supported by the public health literature that shows both the enormous negative health implications for sex workers of criminalization and the impact that decriminalization could have on the course of the HIV epidemic among sex worker populations (Shannon et al. 2015;Platt et al. 2018). This study highlights police support for this type of legal change in the State of Maryland, and suggests common ground on which police and sex worker activist could unify to instigate change. Other achievable policy reforms include a shift at the institutional level in the cultural landscape of policing sex work (e.g., introducing a sex worker liaison officer to facilitate follow up around FSW experiences of violence.) However, for such reforms to be effective, the study also highlights the importance of fostering extensive structural support for FSW and service referral options in collaboration with criminal justice reforms. Shifting police culture away from viewing the vulnerability of FSW as a resource to be exploited and towards viewing it as a harm to be reduced would require a seismic shift away from the current crime control climate, which is tightly bound to the city's ongoing epidemic of violence and to the broader socio economic problems of Baltimore city. Finally, FSW must themselves be empowered in their legal and social rights, thereby holding police and civil society accountable for promoting health and equity amongst this population.

Declarations
Ethics Approval to participate: Ethical approval was obtained from the Johns Hopkins Human Subjects Research and the Institutional Review Board. All participants provided informed consent. Consent for Publication: Not applicable.