Experiences with the Philadelphia Police Assisted Diversion Program: A Qualitative Study

Abstract

Background:

The U.S. relies far too heavily on punitive criminal justice strategies to address problems that are better prevented and resolved through social and health services. The resulting harms are especially large and longstanding for people trapped in the failed War on Drugs. Philadelphia launched a Police Assisted Diversion (PAD) program to address the highest rates of poverty, incarceration, and fatal overdose among large cities in the U.S. PAD enables police officers to connect people with supportive services in many instances that would otherwise result in arrest or through outreach when no crime is suspected.

Methods:

We conducted semi-structured interviews with 30 clients, 15 police officers, and 12 other personnel involved with the program. Data were gathered in 2019 and 2020 and analyzed using standard qualitative methods.

Results:

PAD represents a new less-punitive model for responding to illegal purchasing of drugs, possession of drugs, prostitution and retail theft in Philadelphia. Clients reported mostly positive experiences with the program, identifying the primary benefits as avoided arrest and relational support from affiliated service providers. Police officers expressed support for the program in principle but skepticism about its effects in practice, questioning the quality of available services. Program personnel and police officers described multi-sectoral collaboration as essential to addressing frequent and diverse logistical challenges, including overly restrictive eligibility criteria, mistrust between police and service providers, and coordination across different neighborhoods. Finally, all three groups suggested that people can only benefit from service linkages when they are ready to engage and that inadequate access to resources like housing limits program effectiveness.

Conclusions:

The PAD program is a promising public health intervention for diverting people away from punishment and towards services to address unmet social and health needs. But increasing investments in training and in other supportive services is essential to sustainable and transformational change.

Introduction

Over a million people are arrested for drug-related offenses each year in the U.S. (FBI 2020). These arrests create immediate trauma and longstanding barriers to employment (Sheely & Kneipp 2015), housing (Geller & Curtis 2011), social support (Pettus-Davis et al., 2017), and healthcare (Brinkley-Rubinstein, 2013). There are few if any benefits to this punitive model. After fifty years of the War on Drugs, substance use is equally common (NIDA, 2015) but more harmful in the U.S. (Beletsky & Davis, 2017), with annual fatal overdoses exceeding 90,000 last year (CDC 2021). Laws criminalizing drug possession are primarily enforced against populations with the greatest unmet needs (Covin, 2012; Western, 2018), producing a counter-productive cycle of repeated arrests that has been lamented as the “revolving door” of the criminal justice system (Kushel et al., 2005; Helfgott, Parkin, & Fisher 2019).

One promising strategy for breaking this pattern is enabling police officers to divert people to immediate services in instances that would otherwise result in an arrest. Such programs have operated in various forms for decades outside of the U.S. When Portugal decriminalized drug possession in 2000, police instead began referring people to multidisciplinary panels charged with encouraging less and safer substance use through persuasion, and, as a last resort, minor non-criminal sanctions (Hughes & Stevens, 2010). This arrangement has resulted in dramatically more uptake of supportive services, including drug treatment (Pombo & da Costa 2016), and substantially fewer population-level harms like HIV transmission and fatal overdose (Hughes & Stevens, 2015). In Australia, police diversion programs have flowered in different forms through local legislation and administrative policy (Hughes & Ritter, 2008). Depending on the substance and the program, people are referred either to simple education or more intensive counselling. Police diversion of cannabis possession in Australia has multiple documented benefits including better social outcomes and fewer justice system costs; the effect of diversion programs for people possessing other substances is more limited, in part due to prohibitively tight eligibility criteria (Hughes et al., 2019).

The first large-scale adoption of police diversion in the U.S. was the launch of the Law Enforcement Assisted Diversion (LEAD) program in 2011. LEAD allows police officers to refer individuals who would otherwise be booked for drug possession and prostitution offenses to a case manager who links them with low-barrier, harm-reduction case management. LEAD began as a pilot in the Belltown neighborhood of Seattle, a dense economic corridor, and gradually expanded to other areas of the city in subsequent years. Evaluative research documented considerable benefits for people who enrolled, such as increased housing, employment, and income (Clifasefi et al., 2017) and decreased jail bookings, jail days, and incarcerations (Collins et al., 2019). LEAD also demonstrated some capacity to mediate community discord related to public intoxication, homelessness, and minor crime (Morrissey et al., 2019).

At least forty cities in the U.S. have attempted to replicate the LEAD model over the last few years (Lattimore et al., 2020) with mixed success. Programs in Albany, Atlanta, New Haven and San Francisco struggled getting police officers to make referrals instead of arrests (Worden & McLean, 2018; Jefferson & Jamision, 2019), with the latter two ending as failed pilots (Joudrey et al., 2020; Magana et al., 2021). Programs have sometimes struggled with narrow eligibility criteria (Worden & McLean 2018), poor access to services, or adversarial relationships between police and service providers (Magana et al., 2021; Joudrey et al., 2020). The Seattle LEAD program only overcame poor officer buy-in, and strained relationships between key stakeholders through regular work group meetings (Beckett, 2014). Narrow diversion eligibility criteria, lack of access to treatment and other resources, and changes in policing trends have complicated implementation of police diversion in other countries too (Hughes et al., 2019).

Philadelphia created a Police Assisted Diversion (PAD) pilot program in 2017 as part of a multi-pronged effort to address the highest rates of poverty, incarceration, and fatal overdose among large cities in the U.S. (The Pew Charitable Trusts 2019). The program launched in a “North” pilot area with about 90,000 residents, who are overwhelmingly Black and who experience some of the highest and most concentrated rates of deep poverty in the country. Unlike many police diversion programs, which focus on public disorder offenses in commercial areas, this residential area has high rates of serious violent crime. Initially, only people stopped for drug possession were eligible for PAD. However, in early 2018 a “social referral” pathway opened so that mere encounters with police could precipitate a warm handoff to social service providers in real time (with client consent), without requiring justification for an arrest. At that time, eligibility criteria also expanded to authorize officers to use “stop referrals” for people committing retail theft. These changes resulted from police officer statements that it was strange to only offer services to people once they committed an offense and that many retail thefts involve people with chaotic substance use. Later that year, PAD expanded to a second “East” pilot zone encompassing a more racially and socioeconomically diverse area with about 240,000 residents and one of the largest open air drug markets in the U.S. (Mars et al., 2014). The city-wide police assignment enforcing prostitution prohibitions started making PAD stop referrals the next year, and there has been expansion into other geographic areas and assignments since.

PAD staff work out of the local police buildings, and meet with potential clients immediately after qualifying arrests or referrals. If eligible and interested in the program, the client is immediately connected with one of three organizations, each of which provide services directly or through referrals as part of ongoing case management. This case management is grounded in a harm reduction approach of “meeting people where they are” that does not require sobriety as a precondition for receiving social and health services. All three organizations are based in the community and employ peer specialists to facilitate engagement. Unlike drug treatment programs which rely on the threat of criminal sanctions to enforce treatment compliance, charges are not reimposed on people who receive stop referrals but fail to engage with case management or follow suggested treatment plans. PAD case management services are available and free of charge for as long as clients want to remain engaged.

Since the program was launched in late 2017, there have been 1192 stop referrals and 822 social referrals. Stop referrals have mostly involved drug possession and drug purchasing (73%) rather than prostitution (13%) and retail theft (14%). According to administrative program data, 45% of clients are homeless, 63% are living in poverty and 69% have some problematic involvement with substance use. More detailed information about the operation of the PAD program and its effect on rearrest and other outcomes is discussed in a companion piece currently under review (Anderson et al. under review). Key features of the programs are provided in Table 1. This is the first study to explore the experiences of clients in a police-assisted diversion program in the US and one of only a few to describe the experiences of participating police officers and personnel with key operational roles.

Table 1 -.

Key history and characteristics of the Philadelphia PAD Program

History	Launched in PAD North (approx. population 80,000) in December of 2017; expanded to PAD East (approx. population 160,000) in early-2018 and the City-Wide assignment enforcing prostitution prohibitions in mid-2019. In mid-2019, a co-responder model, with behavioral specialists accompanying foot patrol officers was deployed in PAD East. In November 2020, this co-responder model was expanded to other districts.
Participating entities	Philadelphia Police Department, Managing Director’s Office, Public Defenders, District Attorney’s Office, Department of Behavioral Health, Pennsylvania Recovery Organization - Achieving Community Together (a grassroots advocacy and recovery support initiative), Prevention Point Philadelphia (a healthcare clinic offering harm reduction services including syringe exchange), The Salvation Army of Greater Philadelphia (a provider of trauma-informed and client-centered case management for people involved in sex work).
Stop referral process	When someone is stopped for a qualifying offense in a pilot area, the officer completes a screening form assessing the presence of disqualifying criteria. If none are met and the individual chooses to enroll, a brief intake process is completed by PAD staff who address any immediate needs and arrange an in-depth intake at an affiliated service provider within two days. Failure to remain engaged with supportive services does not result in further prosecution although it may limit future participation. Certified Recovery Specialists assist with intake and case management.
Social referral process	Social referrals are initiated through officer outreach, walk-ins at PAD satellite offices of affiliated providers, walk-ins at police districts, and engagement by behavioral specialists working in co-responder teams.
Eligible offensesa	Illicit purchase or possession of drugs, retail theft, prostitution
Disqualifying criteriab,c	Current intoxication; any prior convictions for homicide, rape, or arson; prior convictions within the last five year for robbery, aggravated assault, burglary, and firearm-related violations; open arrest warrants, participation in another diversionary program, or age less than 18
# of stop referrals	1192 (as of May 2021)
# of social referrals	822 (as of May 2021)

^aPenalties for these offenses have all been substantially reduced in recent years. Starting in 2018, prosecutors stopped charging for marijuana possession or purchase, paraphernalia possession, and prostitution cases against sex workers with two or fewer previous prostitution convictions. At this time, retail theft of less than $500 worth of goods was charged as a summary offense. Policy also recommends diverting all drug possession and drug possession with the intent to distribute into diversionary court programs.

^bThese were the disqualifying criteria during the time data were collected. In January 2020, both sets of criminal history criteria were removed.

^cPresence of these disqualifying criteria are not assessed during social referral intake.

Methods

Study Overview

This study was approved by the institutional review boards at the University of Pennsylvania and the City of Philadelphia and followed the Consolidated Criteria for Reporting Qualitative Research (COREQ) (Tong et al., 2007). The qualitative findings reported in this paper are part of a multi-year mixed methods evaluation of the PAD program. The quantitative components of the evaluation, under review elsewhere, explore whether the program achieves hypothesized outcomes, including reductions in rearrest and improvements in health outcomes. This qualitative component explores the experiences and perspectives of people who designed, implement, and enroll in the program. There was particular emphasis on the “front-end” interactions between police and clients, given its novelty, rather than the “back-end” delivery of social and health services. Key informants and police officers were also asked to express their understanding about the purpose and aims of PAD.

A semi-structured interview format was selected to preserve confidentiality of participants (in comparison to a focus group style), and to allow for conversational flexibility between subject and interviewer, while maintaining some level of consistency between interviews. Initial interview guides were developed based on a review of relevant literature and an examination of evaluations of similar programs. Interview guides were enhanced throughout the data collection phase through iterative analysis of early transcripts (Kvale & Brinkmann, 2015). That is, early transcripts were reviewed before later interviews were conducted, allowing the research team to explore concepts which arose in early interviews with later interviewees, and across the three stakeholder groups (Saunders et al., 2018). The final interview guides are available in Appendices A, B and C.

Research Team

The study was designed and conducted by a team consisting of a public health law researcher with >10 years of experience evaluating the health effects of policing and criminal justice system practices and policies (E.D.A), two masters level social workers whose practice and research focuses on supporting people with complex unmet social and health needs and with criminal justice system involvement (R.K, R.S.), a research assistant with an MPH (C.M.), and a qualitative research scientist with >15 years of experience (R.F.).

Data Collection

Key informants were identified by their role in the design or implementation of PAD. This included people in leadership and supervisory roles from all city agencies participating in PAD as well as people in leadership, supervisory and direct service roles from affiliated non-government services providers. Clients were recruited through affiliated service providers. Recruitment flyers were provided to coordinators who were asked to distribute the flyers until there was an approximate balance between social and stop referrals from each pilot zone. Police officers were recruited from police districts and police assignments participating in PAD through word of mouth.

The one-hour interviews, which were conducted in private offices or in private meeting rooms in public libraries, were audio-recorded, transcribed verbatim, and de-identified by an independent third party. All participants provided written consent. Only the clients were compensated ($50) for their time. E.D.A conducted all the interviews with key informants and police officers; R.K conducted all the interviews with the clients.

Analysis

Using a directed content analysis approach, codebook development was informed by the literature and through preliminary review of a subsample of transcripts (Hsieh and Shannon 2005). Codes were developed collaboratively, and were agreed upon by each member of the research team. For each code, the study team provided an explicit definition and precise instructions to ensure coding accuracy and consistency. Two team members independently coded each transcript. The entire study team finalized the codebook, refined the coding process, resolved coding inconsistencies, and organized final codes into thematic categories. Transcripts were stored and analysis was conducted using NVivo12 software (QSR International).

Results

Semi-structured interviews were conducted with 42 individuals involved with the PAD program. This included 12 key informants with operational involvement in PAD working in municipal leadership, the police department, the public defender’s office, the prosecuting attorney’s office, and the department of behavioral health, as well as directors, coordinators, and case managers at community based non-governmental service providers affiliated with PAD. Thirty semi-structured interviews also were conducted with clients, of whom eighteen clients were male and twelve were female. Seventeen clients were Black; twelve clients were white, two were Hispanic, and one selected “other” for their race and ethnicity. The age of participating clients ranged from 22 to 64 years, with an average of 42.8 years. There was an approximate balance between clients who enrolled in PAD through stop (n=14) and social (n=16) referrals and in PAD North (n=16) and in PAD East (n=14). Finally, semi-structured interviews were conducted with 15 police officers, with seven from PAD North and eight from PAD East. Length of tenure ranged from 1 to 35 years, and previous and current assignments varied widely. Demographic characteristics of key informants and police are not provided to preserve their anonymity. Analysis of transcripts resulted in data being organized into 5 thematic categories, each described below and supported by participant quotations.

Key Informant and Police Officer Perspectives on the Goals and Aims of PAD

Key informants and police officer participants were asked to define the aims and goals of the program. Key informants described an ambitious and multi-faceted program with goals of reducing arrests, addressing unmet needs of people in the community, and improving relationships between police and the community in which they serve. To achieve these goals, the respondents identified six aims of the PAD program: non-punitiveness, root-caused-ness, readiness, person-centeredness, adaptability, and institutional support.

The traditional punitive approach to minor criminal offenses was widely acknowledged as intensifying challenges that many people in the community face by undermining access to employment, housing, and supportive services with little or no durable benefit in terms of individual or community health. It was acknowledged that this non-punitive approach would need to extend to individuals who sometimes have lengthy and ongoing criminal activity. As one government official remarked,

we’re trying to intercede . . . from our perspective, I don’t really care if it’s your first or 10th [arrest]-our job is to try to connect you with the services so we can intervene and break the cycle.

Respondents suggested that PAD avoids punitive means in responding to public safety challenges by attempting to address the root causes of those challenges such as unstable housing and other unmet social and health needs. Respondents noted that this approach is important because, for many potential PAD clients, these challenges are often numerous, overlapping, and exacerbated by previous involvement in the criminal justice system.

Readiness refers to the overall aim of the PAD program to provide services to people at a moment when they are ready to receive them, with the underlying idea that readiness to engage is fluid and might not endure over the time periods spanning referrals in typical outpatient services. One key informant respondent likened this strategy to being ready for a “lightning strike. You find the person who’s ready to change, and you’re there for them.” Practically speaking, this underscores the aspiration for PAD to operate in the community to “offer people opportunities to get connected to meaningful help right there on the spot.”

An embrace of “person-centeredness” in the design and delivery of PAD was identified by key informant respondents as “meeting people where they are” meaning privileging the client’s assessment of their own needs. According to respondents, the deployment of PAD personnel with longstanding ties to local communities or with similar life experiences was important in helping clients identify their needs. Key informant respondents also described person-centeredness in the sense that the affiliated services were flexible and could be rolled out in a way that responds to each client’s unique needs, aims and strengths. As one client-facing key informant suggested

[T]he missing piece has been social services, but not just traditional social services . . . we flip the switch and we say, “Okay, you know what’s worked, you’re the expert of your own story. You need to be connected with social services but what does that look like to you? What does that mean to you?”

Key informants described the PAD program as adapting in the face of new opportunities and challenges. This included tweaking eligibility criteria, changing practices and adopting entirely new referral pathways like a co-responder model. The pace of this evolution was viewed differently, but respondents acknowledged that ongoing change is important and inevitable.

Police officer respondents described the program as a general opportunity to provide community members with connections to services, noting that that was difficult before PAD existed. Some placed emphasis on the benefit of deflecting people out of the criminal justice system whereas others stressed the importance of decreasing the burden on the community from repeated reoffending. There were diverse perspectives on whether the program was primarily a way to reduce substance use or a mechanism for addressing a broader set of unmet needs. One officer described PAD as “basically it’s a try to help those that are looking for help. . . [PAD has the] resource available. . . at the beginning we didn’t.” Another described the logic of PAD as

It’s a way to try to steer people into rehabilitative programs rather than taking them straight into the criminal justice system. It’s geared towards getting people help rather than punishing them . . . if you fix those problems, you won’t be locking up the same person again and again and again. And if you do that enough times then hopefully it will [reduce] . . . crimes motivated by drugs.

Police officer respondents stressed that PAD was meant to provide access to rapid and high quality services. In the words of one officer, PAD services were not “going to scam you. They’re actually helping you . . .It’s not, you pay for this and we’re going to take your money and then we don’t care about what happens to you.”

Overall positive experiences with PAD among clients

The majority of client respondents reported positive experiences with the PAD program, often describing it as life-altering if not life-saving. Several clients identified the primary benefits as avoiding arrest and gaining access to resources. The latter emerged repeatedly despite the focus in the interview guide questions on the initial referral steps involving police. Most respondents mentioned receiving instrumental assistance obtaining healthcare, personal identification (which is essential to accessing public services (Sanders et al., 2020)), and other tangible resources (e.g., food). But the most consistently expressed benefit was emotional support from case managers. As one person who enrolled through a social referral noted

[T]hey’re walking with me, I don’t have to do this alone . . . Even if I just want to vent or scream or somebody pisses me off, instead of picking up a drug, I pick up the phone and tell them what’s going on…I’m so used to figuring shit down on my own . . . sometimes I [don’t] ask for help until I’m backed into a corner. Then, I have to fight my way out, and usually by then it’s treatment, jail or end up hurting myself and ending up in a hospital.

These sentiments were widely shared among clients enrolled through stop referrals too, exemplified by the remarks from another respondent that

[My case manager is] willing to go step by step with me. . . I’m just learning how to be an adult. . . and having somebody there holding my hand … it comforts things a little bit. . . their case managers do walk with you side by side.

Many respondents also suggested that the program provided respite spaces that were safe and inviting. When pressed on what they would change about PAD, some clients suggested that advertised resources like housing were lower quality than expected or unavailable.

In reflecting on their experiences, multiple respondents were emphatic that services provided though PAD only benefit people who were ready to engage. Respondents often described their own decision-making related to PAD as fluid. For some, readiness existed in a particular moment when police were present and acting with genuine compassion. Some of these instances coincided with the potential for arrest and others occurred through foot patrol outreach. One respondent recounted

[The officers] came walking through on their beat. [One officer] was talking to a few of us, and was like, “Hey, do you want to get clean? . . . come on. Let’s go. Let me make a call right now, and I’ll take you right now.” . . . [the officer] knew that I wasn’t ready. What [the officer] said to me, though, was, “When you’re ready, [name removed], I’ll be around.” . . . [then a few days later] I had overdosed . . .when I got up [to use the] bathroom, and they were pulling around the corner. [The same officer] was like, “Hey, [name removed]. You all right?” And I said, “No.” [the officer] said, “You want to go to detox?” I said, “Yeah.” I forgot all about going to the bathroom. I couldn’t get in the back of that police car fast enough.

Police support the program in principle but some question it in practice

The police officers who participated in interviews broadly supported the aims of PAD, with many officers characterizing it as a useful “tool” and “fantastic option” to connect community members with services. “[O]ur traditional approach . . . isn’t working,” one officer stated, “and I think this is great [be]cause we’ve never been able to link people with services before, and we’ve always known they needed it.” Some officers described the program as a potentially effective way to break a cycle of punishment, with benefits to individuals and their communities:

It’s geared towards getting people help rather than punishing them… if you fix those problems, you won’t be locking up the same person again and again and again. And if you do that enough times then hopefully it will make a dent in . . . crimes motivated by drugs and things like that.

These endorsements were often qualified by statements that people could only benefit from available services when ready to do so and that readiness is sometimes fleeting.

Conceptual support for PAD was tempered by practical concerns. Although perspectives on the issue were diverse, some officers suggested that there should be penalties for failing to remain engaged in services after a stop referral. These officers recognized that prosecution for qualifying offenses was unlikely (see Table 1) but worried that not enforcing program “compliance” sends a problematic message to the community. Some officers also stressed that coercion was necessary to change behaviors like substance use. Other officers, however, expressed very different views including that PAD referrals are worthwhile even if many attempts fail to result in uptake of services:

Police officer: If one person gets the help they need and actually sticks with it, then it’s a win for me.

Interviewer: Okay. Even if, let’s say, two other people go in and walk [right] out? Police officer: Yeah. I’d rather do that. I rather see someone who gets some help, and then have two people walk out.

One officer suggested that a coercive approach would decrease the number of clients choosing to enroll in PAD through stop and social referrals:

My feelings . . . is, if a person wants to be helped, they’ll stay. If you start putting repercussions on people, they’ll be afraid. And trust me, words get by quick here. . . . [w]hen you first come in contact with them, your trying percentage is going to go lower.

A second source of concern was skepticism about affiliated service providers. Some officers described instances when, in their perspective, an attempted referral failed because the service provider did not emphasize availability of services like detoxification, but instead only provided information about harm reduction. This was especially frustrating when, in the respondent’s view, that outcome did not align with what the person expressed to the officer moments earlier:

We have transported people [to PAD] and we have gone around the corner and . . . the person we just dropped off 5 seconds ago, is walking around the corner. . .if we’re bringing the people there and they want the help, why are they walking [right] out.

Skepticism and mistrust of service providers was often qualified by statements that the officers did not know much about how service providers operate. Some officers judged a service provider based on its inability, in their view, to stem a tide of public disorder in the community including, in a few instances, around the provider’s building:

[I]f I’m sending this person to get help, do I really have confidence that this person is going to get help because I see what [the service provider] does on the outside

These negative views of service providers contrasted with frequently positive views of PAD staff embedded within police districts. “[T]he PAD personnel are good people and they work their asses off back here,” one officer stated, “and, if you can’t work with [them], you seriously have a problem. . .[They] will go out of his way to help somebody.”

Operational challenges require and stimulate multisectoral collaboration

Mistrust between police officers and affiliated providers was identified as one of three important challenges in implementing PAD. That mistrust, according to police, providers and city personnel, originated in historic tensions between law enforcement and harm reduction practitioners and necessitates enhanced training. Improved information sharing was noted also as one way to increase a sense of shared purpose and faith in the value of services, which was sometimes low. One police officer respondent shared that

[Y]ou see a lot of negative stuff . . . It’s very heartening for us to find a story to see when something works out. It’s just wow, finally something worked. Actually, they got better, a situation got better.

The second logistical challenge was overly restrictive eligibility criteria for stop referrals. When the program began, potential clients were ineligible if they had previous convictions for specified crimes or if they had an open criminal case or arrest warrant (Table 1). Service providers, city personnel and officers often lamented that too many potential clients were excluded who could have benefited from the program. As one officer noted,

I understand there has to be some criteria regarding people who are getting enrolled into the program, but at the same time you’re missing a large population of people who could benefit.

Rigid eligibility criteria, another officer mentioned, seemed to frustrate linking people with services to address the underlying causes of their behavior:

Like shouldn’t there be a little bit of flexibility there? Like I get it, okay they’re repeat [offenders], but why are they stealing? That’s what I keep going back to. Like I can’t say that that person’s a criminal. What they did, stealing is a criminal act. But is that person a criminal? Like normally if they weren’t high would they do this?”

Based in part on preliminary presentation of these findings, eligibility restrictions based on prior criminal convictions were removed in January of 2020 (Table 1). However, efforts to modify other eligibility criteria - like restrictions due to warrants for failure to appear in court - are ongoing because they involve yet another stakeholder, individual judges in the municipal court system.

Coordinating PAD services across very different pilot zones frequently emerged as a third important challenge. The pilot areas varied tremendously, according to many respondents, on a number of characteristics including demographics, patterns in housing stability, drug use, and crime, and community relationships with police and service providers. It was essential, one key informant respondent stressed, to “tailor PAD. . . the model of . . . North wasn’t ideal for . . . East, [so we] created a new model.” The expansion into the second area was viewed by some as too rapid, especially given the different and perhaps greater needs of the second area. Although concerns about coordination and capacity persist, key informants highlighted that expansion stimulated creation of a co-responder model there, which relied on the ongoing multi-sector collaboration forged through PAD implementation (Table 1).

PAD benefits are contingent on readiness and resources

Members of all three groups spoke explicitly about the idea that some potential clients were not ready to engage with services because their lives are so chaotic. These groups of people, according to one respondent, included “people who are using multiple times a day, who are street homeless, who are living in deep poverty, who have no income” and who were often too focused on impending withdrawal to contemplate behavior changes. Helping this population, another respondent stated, requires getting “out of this crisis management stage of service provision [and] . . . into the goal being stability.” However, such a transformation is complicated, other respondents cautioned, by the complexity of unmet needs and the limits of available resources. It was broadly acknowledged, including among many police officers, that providing access to substance use and other treatment is sometimes necessary but often insufficient to enduring wellness. As one police officer responded

You can give these people their treatment and help them kick their habit, but now they’ve got to work, they need a place to live… that’s got to be a full spectrum of things.

A few participants suggested that there is a mismatch between the services that PAD advertises and the services that PAD provides. Some of these perceived deficits related specifically to PAD services while others, like inadequately diverse housing options, were features of the broader referral ecosystem in the city. One client stated

They don’t have no housing. . .Every single person in that building tries . . . in any way possible down to the water and coffee. If they don’t have resources for this, how can they help? . . . The only thing that they can honestly do is call.

Many respondents interpreted the volume of people proactively requesting PAD services through officer outreach or by walking into PAD satellite offices, or even police stations, as evidence of an inadequate supply of social and health professionals working in the community. One officer captured this common sentiment among colleague respondents by saying:

we need more . . . from the civilian end. Like we need boots on the ground. . . . they need to come out and engage them more.. . .Take us out of it a little bit more. Like we’re limited. Law enforcement in America is not set up for addiction.. . .We don’t have the resources.. . I’m not a counselor. I’m not a social worker. . . .I know like what’s going on, but I’m not a professional.

Discussion

Our qualitative exploration of the experiences of people implementing and enrolling in the Police Assisted Diversion (PAD) program in Philadelphia is subject to important limitations, including potential selection and social desirability bias. While we asked providers to recruit a wide sample of clients in order of response, it is possible that individuals with positive experiences in PAD and with affiliated services were more likely to participate. This limitation is notable notwithstanding that the emphasis of this study was on the police referral to services and not on the services themselves. There may also be social desirability bias in that key informants, clients, or police officers were more inclined to speak positively about this alternate public health strategy given their expectations about our preferences or in light of current discourse about aggressive policing. We attempted to blunt social desirability bias by underscoring the confidentiality of interviews.

Despite these limitations, our study provides some valuable insights about a large and ambitious police diversion program operating in the poorest cite in the U.S. with highest rates of fatal overdose. Similar to the LEAD program in Seattle (Beckett, 2014), the PAD program illustrates that sustained high-level collaboration between law enforcement and health partners can provide a platform for continuing reform. This evolution is essential because, as observed in replications of LEAD elsewhere, implementation challenges for such programs are often considerable (Jefferson & Jamision, 2019; Worden & McLean, 2018) if not insurmountable (Joudrey et al., 2020; Magana et al., 2021). These programs require trust between potential clients and service providers and between police and service providers. Providers build trust with clients through tailored service and embeddedness, suggesting the need for many community-based providers with different models of engagement. This complicates the task of strengthening fragile trust with police officers, who move between areas, but exist within a hierarchical and routinized institution.

The experiences reported by clients were overwhelmingly positive. That clients primarily focused their responses on backend case management services, despite redirection by the interviewer on the initial police referral, underscores the importance of social support for a portion of the people who are trapped in the revolving door or the criminal justice system. It also may reflect the reality of broad reductions in charging and prosecution for the qualifying offenses in the city over recent years (Table 1). Medications and other health services have important roles to play in addressing harmful substance use. However, our findings provide a reminder that frayed social connections often underlie high risk substance use (Dasgupta et al., 2018) and other public health harms.

In providing linkages to services without further involvement in the criminal justice system police diversion programs are an “attractive alternative” to court-based programs (Csete, 2019), which compel participation under threat of punishment (Belenko, 2019). These “pre-booking” programs may also capitalize on a moment when readiness to engage with services is greatest, in contrast with court programs, which often occur weeks after the precipitating event (Butters et al., 2020). Given historic problems with over-policing and growing support – even among law enforcement (Rouhani et al., 2019; Vermeer et al., 2020) – for public health approaches to substance use, these programs could improve how policing is experienced by officers and communities while promoting individual and community health.

Police assisted diversion embodies the interdisciplinary least harm approach championed by many reform campaigns including President Obama’s Commission on 21^st Century Policing. It is not, however, a no harm approach. Many laws that police are asked to enforce remain manifestly unhealthy (Burris et al., 2020) and enforcement, like all interventions, is risky (Anderson & Burris, 2017). PAD is a concession – probably necessary given sobering recent experiences with attempted reforms (Herndon, 2020; Goodman, 2021) – to an incremental approach to unwinding counterproductive systems of punishment. This muddling through makes sense to the extent that contingency and context dependence necessitate an iterative and adaptive process of change. But it is essential that this muddling is a march to a destination and not merely a meander along feelgood pathways. That destination must center the experiences of people in distress and not the police or providers serving them. Even this incrementalist approach relies on far greater investments in public health infrastructure like housing and services that meet people where they are.

Figure 1 –

Emergent themes within and across interviewee groups

Highlights.

• Clients reported positive experiences with the program.

• Police officers support the program in principle but question its effectiveness in practice.

• Mistrust between police and service providers is a challenge to successful collaboration.

• All stakeholders agree that clients can only benefit from service linkages when they are ready to engage.

• The program is promising but investment in other services is essential to sustainable change.

Appendices

Key Informant Interview Guide.

1. Please describe how you define the PAD program.

2. What is/are the purpose(s) of PAD?

3. Is PAD currently serving this/these purpose(s)?

4. What are the keys – or facilitators – to current successes?

5. What are the challenges to success right now?

6. What are the challenges to sustained success going forward?

7. What do you think I should know as I explore the perceptions and experiences of officers and clients?

8. What are the key things you think I should measure about participants?

9. Are there factors that make it more likely that a PAD referral will succeed? Or falter?

10. Is PAD serving the right population?

11. Will it / can it / should it target those in the “revolving door” of the criminal justice system?

12. What inter-organizational dynamics facilitate the success or PAD?

13. What inter-organizational dynamics challenge the success or PAD?

Client Interview Guide.

1. How long have you lived in Philadelphia?

   Prompt: What part of the city? Where'd you come from? What brought you here?

2. Before we start talking about PAD, I'd like to learn about your experience with police before you started in the PAD program. So could we go back and can you tell me a bit about your experience with the police before PAD.

   Prompt: How does that make you feel when you see other police?

3. I've learned a lot about what happened before. Now we're going to move on to your experience in PAD. Can you tell me about the day you got involved in PAD - just walk me through that day.

   Prompt(s): What happened next? What was that like for you? Why did you choose to participate in PAD? Was it really a true choice? Had you heard about PAD before the encounter? Whaťd you hear?

4. You've been in the program for some time now so you're an expert on what it looks like and how it feels to go through it. We'd like to tell the people running the program how iťs going and what could be done to make it better. Leťs start with what do you like about PAD?

5. What do you not like about PAD?

6. What kinds of other services or help should also be offered as part of PAD?

7. Have your experiences with PAD changed your life? Tell me about it.

8. Earlier we talked about your experience with the police before PAD. Now that we've talked about your experience in PAD, I'm curious about How has PAD changed your perception of police?

9. Have you told anyone about the PAD program? Whaťd you tell them? Did you recommend it?

10. We're going to go back to the people running the program with the things we're hearing today and they really want to know how to make the program better. If you had to sum up your experiences in 1–2 sentences, what would you say?

11. I've learned so much from you and we've talked about so much today so I want to give you a minute to think about the conversation and tell me if there is anything else you want me to know?

1. Police Officer Interview Guide.

1. Tell me briefly about your experience as a police officer in Philly?

   Prompts: How long? What places? What shifts/units (strike force, vice, community relations etc.)

2. We are going to talk about the PAD program today. Suppose an officer from another city heard about the Philadelphia PAD program and asked you to describe it - what would you say?

3. Do you think PAD is accomplishing its goal or goals?

   Prompts: Why or why not?

4. Tell me about the last time you referred someone to PAD.

   Prompts: Is that typical? If no: How did it look different? If yes: Are there any other notable stories? If the last time was social referral, please tell me about the last stop referral (and vice versa)

5. What kind of person is most likely to benefit from PAD? What are their characteristics?

6. Do you know the eligibility criteria for PAD?

7. Are these the right eligibility criteria?

8. Do you feel emotionally invested in an individual’s treatment once you refer them to PAD?

9. You’ve gone through some training for PAD. Is there additional training that would help better participate?

10. I’d like you to think about your job before and after PAD. How are things different now that PAD is an option?

    Prompts: How does that make you feel about your job? How does that affect your job satisfaction?

11. Does PAD align with your values as police officer?

12. We understand that participation in treatment for people referred to PAD is not really enforced. What do you think about that?

13. This has been a great conversation and I have learned a lot about your participation in PAD and views of it. What else would you like us to know about being a police officer participating in the PAD program? Any questions?