Lessons in “Slow” Engagement From Staff and Administrators at a Prebooking Jail Diversion Program

Some individuals with a serious mental illness experience an insidious cycle between homelessness and jail (1–3). Individuals experiencing “chronic homelessness” are homeless for at least 1 year, have a serious disability, and frequently rely on crisis services in nonideal settings (4–6). In the United States, 19.8% of individuals experiencing chronic homelessness also had a diagnosis of serious mental illness (7); moreover, they were often in jail or overpoliced (7–11). Individuals with serious mental illness were at least twice as likely to be arrested for the same offenses as people in the general population (12). Efforts toward “decriminalizing” homelessness and mental illness have not reduced recidivism (13). Criminal justice systems using a newer “managerial model” seem to inadvertently harm populations with mental illness by asking them to demonstrate rule compliance and prove their moral worth over time (14, 15).

Nationwide investment in prebooking jail diversion programs that redirect people from jail to mental health treatment has the potential to save upward of $1 billion annually (16). Engaging people in mental health treatment can reduce negative outcomes such as rearrest for minor infractions (17). In one review, key outcomes of prebooking jail diversion programs were decreased arrest rates and increased referrals to mental health services (18), often assessed over time as an indicator of a person’s change in trajectory (19). One program using pre- and postbooking diversion decreased rates of chronic homelessness, reduced the demand for acute care in jails and recidivism, decreased local crime, and improved public health and safety (20, 21). Another study found that jail diversion in three major cities (including Houston, where the present study took place) did not jeopardize public safety (22). Additional crisis response sites, such as jail diversion sites, can help end the cycle between chronic homelessness and jail (23). Although preventing pretrial incarceration alone is not enough to support people experiencing mental illness (24), mental health service providers can step up to help their clients remain free of further criminal involvement (15).

As part of a multisite project exploring innovative strategies for serving individuals with chronic homelessness and serious mental illness, our research team identified the prebooking jail diversion program, the Judge Ed Emmett Mental Health Diversion Center (JEEDC), as offering an unusual option for clients at this critical point of care. (A flow diagram describing the intake process is available as an online supplement to this article.) In 2020, our research team invited stakeholders and JEEDC clients to explore lessons learned from their attempts to offer services at the complex intersection of the criminal justice and mental health systems, which often hold different philosophies about approaches and goals (15, 25).

Methods

This study was approved by the Harris Center Institutional Review Board, and all participants gave informed consent. Participants were recruited via “purposive sampling” of mental health stakeholders (MHSs) and criminal justice stakeholders (CJSs) involved with the JEEDC (26). Our team also made site visits and attended administrative meetings to take fieldnotes on their observations (27). Audio-recorded, semistructured, hour-long interviews were transcribed and uploaded into Dedoose, a team-based, mixed-methods coding package. Using grounded theory methods, our team identified iterative patterns across the data to generate hypotheses about barriers and facilitators to client engagement (26). (For additional details on these methods, see the online supplement.)

Our team also drew on four sources of administrative data: a program evaluation (28), a Powerpoint presentation, a July newsletter, and an e-mail (W.Y., personal observations, September 2021), reporting data from January to July 2020. Demographic characteristics showed that since September 2018, in a sample of 1,170 individuals who were diverted, most were diagnosed as having a serious mental illness (N=879, 75%) (28). During our study period from January to July 2020, of 1,147 unique diverted individuals sampled, 78% (N=898) were male, 58% (N=665) were Black, 87% (N=998) had been picked up for trespassing, and 78% (N=895) reported being homeless; 51% (N=585) stayed in the program for <6 hours, and 33% (N=379) stayed >24 hours (W.Y., personal observations, September 2021). From September 2018 to July 2020, 25% (N=472) of 1,888 clients sampled were repeat clients (W.Y., personal observations, September 2021).

Results

From January to March 2020 (before COVID-19 lockdowns), our team spent 60 hours engaged in 10 site visits, attended five administrative meetings with 10 stakeholders (six CJSs and four MHSs), and recorded 70 pages of deidentified observational fieldnotes. From January to July 2020, we also audio-recorded up to two 45-minute interviews with 19 participants: 14 MHSs and five CJSs. Interview participants (Table 1) self-identified as somewhat ethnically diverse (32% Black and 5% Hispanic), and 47% were male.

Three major themes about lessons learned emerged from these data: MHSs and CJSs needed to work flexibly toward shared philosophies and expectations of “success” of a jail diversion program, MHSs and CJSs had to work together to address barriers to clients’ engagement in the program, and well-trained staff needed to be prepared and supported to practice “slow” engagement that gave clients time to choose what the JEEDC offered.

Theme 1: Negotiating Success

Expectations of success varied between the CJSs and MHSs, and they held regular meetings to collaboratively calibrate expectations.

CJSs’ perspectives.

One CJS described success as “reduc[ing] recidivism . . . rearrest . . . contact with law enforcement . . . [being] connected with services.” Another saw it as “the reduction or termination” of calls for police assistance. Some expressed frustration with “familiar faces,” or individuals diverted multiple times, although administrative data suggested that this frustration applied to only one-fourth of those diverted (28). One CJS worried about diverting “someone that poses a public safety threat.” A few wanted the JEEDC to “hold” people involuntarily for a few hours to give CJSs a break and give MHSs more time to engage a client. One CJS talked about a client who gradually became more receptive to help during a longer stay at the JEEDC. CJSs often thought that having clients spend more time at the JEEDC led to success. The CJSs understood that their perspective sometimes differed from that of the MHSs: “When you’re talking about a police officer versus a social worker, totally different frames of reference, right?” They also appreciated that mental health services played a critical role: “My deputies . . . put up with a lot . . . but really the [MHS] . . . got [a client] into a facility, and he was . . . healthier . . . he had dignity . . . someone there to help care for him.”

MHSs’ perspectives.

The MHSs appreciated the CJSs’ political and logistic support, such as elected officials, law enforcement, and “the citizens [of Houston] . . . [who] want some return on their investment.” MHSs focused on slow, minor individual improvements that could amount to major positive outcomes. Most MHSs (N=12, 86%) talked about keeping clients at the JEEDC until they were housed but understood that engaging people immediately for a longer stay was unlikely. The more time they had with a client, the more they could help. However, “more time” could occur during one long visit or over multiple visits: “Some . . . want to stay 24 hours . . . next time, they stay 2 days . . . it’s . . . starting to work.” MHSs appreciated small triumphs such as “moments of lucidity” and “a conversation.”

Theme 2: Engagement Was Challenging

CJSs and MHSs worked together to better understand why about half of the clients left in the first 6 hours after their arrival. Most stakeholders (N=16, 84%) thought that many clients were sent to other facilities (e.g., medical detox centers and psychiatric emergency services) and then never returned. Administrative data showed that of a sample of 633 individuals, 15% (N=95) were triaged out for medical care, and 10% (N=65) were sent to the psychiatric emergency department (25).

Others thought that client-level factors such as “street culture” (N=9, 47%), which bred mistrust, played a role: “We’ve had people that have been on the streets since they were kiddos . . . that’s their family . . . [They are] a fish out of water . . . They don’t trust you . . . But that’s the engagement.” Some of the stakeholders thought that clients did not want any responsibilities. Another shared, “They’ve tried hard to not be engaged . . . they’ve had experiences that they don’t perceive as positive . . . They don’t trust organized or formal systems.”

Even after law enforcement officers dropped off a diverted individual at the JEEDC, one armed officer typically stood in the lobby. MHSs thought armed police presence enabled them to welcome more “high-risk” people for diversion. However, 47% (N=9) of all interviewees thought armed police presence made clients uncomfortable. As one observed, “Some [officers] are carrying a gun . . . that’s triggering.” Another described offering support: “[I] remind them, ‘hey, you’re not under arrest.’”

Nearly all stakeholders (N=17, 89%) thought “the rules” made engagement difficult: “clients don’t like a structured setting . . . [so it’s better to] make it a relaxed structure setting.” One of the rules mentioned most was the onsite smoking ban. Despite nicotine replacement options, clients “will leave . . . just [to] get a cigarette.’” Another stakeholder cited smoking as “the reason why clients are leaving.”

Most stakeholders (N=17, 89%) mentioned that they needed to enhance clients’ motivation to engage in a jail diversion program. One observed, “A fierce independent streak with sometimes a limited ability to understand how their mental illness might be impacting their circumstances . . . [and] . . . avoiding withdrawal or [sobriety] . . . combine into ‘I don’t want to stay here.’” About half of the stakeholders (N=11, 58%) wanted more training: “Have [staff] trained, because the people that I see leave here—it’s due to reaction [to staff] . . . how they were talked to or presented to by staff.”

Theme 3: Engaging Slowly

Engaging clients in the program was a top priority and could take a long time. As one MHS said, “they just keep coming back . . . [until] they finally make up their mind . . . ‘I’m gonna stay.’” Another suggested, “For some, we are playing the long game. . . . They have been arrested 80 times? Give me at least 80 shots at connecting with them.”

Slowly building relationships was key (N=15, 79%). Staff had to “actually care and listen.” Clients “stay sometimes because the right staff member talked to them at the right time.” Peer specialists were often the first people to meet a client, and they often described using microengagements: “I try to get to them on their level . . . play dominoes and crack jokes . . . lift them up . . . pull them to the side, ‘Hey, you need to talk? What’s going on?’” Another stakeholder said, “Engagement can happen the whole time . . . small little things . . . like just respect.”

Many stakeholders (N=13, 68%) also advocated for longer-term, caring relationships. The “aftercare team” continued to work with clients well after they were housed. One noted, “As long as I get them in an apartment, everybody’s happy. . . . But that doesn’t make them a whole person.” Many stakeholders moved clients toward stability by engaging them over time, every time; however, some feared this harder-to-quantify caring work was not always valued.

Discussion

Prebooking jail diversion programs show promise (18, 21, 23); however, ethnographic data from this setting illustrate that complications arise when the criminal justice and mental health systems intersect. Both CJSs and MHSs emphasized addressing challenges to engagement of clients in jail diversion programs as a top priority. This goal required ongoing negotiations between the mental health and criminal justice systems, which had different notions of program success. CJSs wanted the MHSs to engage clients quickly to rapidly change their circumstance and reduce contact with the police. MHSs aimed to divert people from future police contacts by engaging people into treatment and housing quickly; however, they endorsed slow engagement, or repeated microengagements with the client over time when needed, which did not always result in an immediate reduction in contacts with the police.

The CJSs and MHSs therefore regularly met together to negotiate goals, identify barriers to engagement, and work together to address these barriers. Research suggests that work on “philosophy” is crucial for success, especially to reduce the criminal justice system’s orientation toward punishment and to encourage recognition that people with mental illness are not “bad” but just need treatment (29). Most of the stakeholders we interviewed promoted a slow series of microengagements with clients at this intersection of care, an approach that resisted cookie-cutter concepts of success, such as efficiency and quantity, which are often used in value-focused systems (30). On the basis of the interviews, we hypothesize that accountability mechanisms focused on tenure in the jail diversion program may not capture what is successful. Focusing instead on the quality of the contact (i.e., trust building, harm reduction focused, trauma informed, and peer supported) and whether it produces improvements in material well-being (e.g., access to food, an identification card, wound care, or a filled prescription) might better represent and assess client needs.

Some CJSs had a high bar for success (e.g., reduced police contact with individuals) and wanted the MHSs to place “familiar faces,” that is, clients who were diverted repeatedly, on involuntary hold as an engagement strategy. MHSs valued the CJSs’ perspectives and efforts but wanted to keep services voluntary. During our study period, 67% (N=768) of clients left in the first 24 hours; although staff thought many of these clients were triaged to other services and therefore did not return to the program, 75% (N=1,416 of 1,888) were found to have left voluntarily (W.Y., personal observations, September 2021).

When discussing why people left, which most stakeholders called “not engaging,” many of the stakeholders thought that the armed police presence was a possible issue. Interactions with the police can negatively affect the mental health of Black Americans, who represented most of the clients served at the JEEDC (28). For example, Black Americans reporting a police interaction were twice as likely to report experiencing poor mental health as Black Americans with no such interaction (31). Reducing police involvement in jail diversion programs for clients with mental health problems could involve use of a co-responder model (e.g., a mental health–trained team) to improve deescalation, reduce stigma, and improve speed and clarity of pathways to mental health treatment (32), a model that the JEEDC does offer when possible. Mental health professionals trained in deescalation can limit police interactions with clients experiencing mental health problems by diverting between 5% and 17% of phone calls to mental health providers that otherwise would have gone to the police (33, 34). In any model, trauma-informed approaches that prioritize safety are ideal.

Many stakeholders also thought that the smoking ban prevented program engagement, and they advocated for a harm-reduction approach to nicotine use, such as having designated smoking areas, to better engage clients. Managed alcohol interventions and the Housing First program have shown positive results in addressing substance use disorder and chronic homelessness when harm reduction is fully implemented (35–38).

Limitations of this study included the size of the interviewed sample and dependence on administrative data. Although the interview sample was limited in size, it was adequate for achieving theoretical saturation for MHSs and CJSs through a purposive sample (39). We also contextualized the interviews through site visits and observational fieldnotes. Study attrition in this population has been shown to be high. However, this study marks a starting point for understanding prebooking engagement (40).

Clients experiencing chronic homelessness, who have serious mental illness, and who qualify for jail diversion are difficult to engage in services (41). We identified no studies focused on client engagement in prebooking or precharge jail diversion. Findings from engagement studies on postbooking jail diversion suggest that engagement is affected positively by female gender and a lack of substance use (41); engagement is affected negatively by a history of criminal justice involvement and threat to public safety (42). Research suggests that identifying people who are at higher risk for disengaging from services on arrival may be helpful for increasing engagement rates, possibly through the use of tools that assess such risk (43, 44) or engagement potential (45), both of which would require further research. Offering primary care and psychiatric services under one roof, which the JEEDC does, is also recommended (46). One study suggested placing high-risk clients in an involuntary hold on a psychiatric unit for 3–5 days before enrolling them in a diversion program (44), but the JEEDC remains committed to voluntary treatment to promote engagement. One review of 25 jail diversion studies recommended reduced pressure for quick fixes, collaboration between CJSs and MHSs, patience, humor, more time, and peer support (29). All of these practices were used by the JEEDC.

In addition, MHSs described “slow engagement,” or a gradual, gentle, and persistent microinteractional approach that they thought nudged clients toward greater stability. This slow approach, as we have called it, did not always align immediately with other goals, such as quickly getting a client into housing, out of jail, out of the public eye, and off the police’s radar. Instead, slow engagement achieved a balance between encouraging clients to stay voluntarily and allowing them to leave. Peer specialists greeted newly arriving clients right away to offer the presence of an ally and then reminded the clients consistently that they could leave at any time, thereby offering respect and accentuating choice (47). Staff gently eased clients’ fear of the police, noting that clients may have had very recent or past traumatic experiences with police. MHSs reported patiently supporting clients across multiple intakes, after hours, and long after they found more stable housing.

Thus, MHSs gently built respectful relationships with clients for whom a brief, casual social interaction might be enough (or even too much). Research suggests that “positive withdrawal,” engaging in brief microinteractions with others, is a strategy to avoid sensory overload that is used by individuals who have diagnoses of schizophrenia and experience homelessness (48, 49). This strategy also seemed ideal for the JEEDC’s clients.

MHSs thought slow engagement was important for clients who were not accustomed to kindness or care, had negative relationships with “the system,” and found the structure provided by a jail diversion program unappealing. Individuals who are chronically homeless can find a structured lifestyle undesirable (50). One study found that clients at times resist affiliation with mental health–related housing because being labeled as “crazy” made them more vulnerable to victimization on the street (51). Being aware of these concerns and addressing them over time in slow engagements seemed to help.

Our concept of “slow engagement” may be read as a fleshing out of the “engagement” component of the “relationship-based care” model, which foregrounds the importance of empathy, respect, and connectedness over time (44). Practices such as these have been flagged elsewhere as part of a new “slow” style of social work modeled after the ability to be present, intentional, and attentive and, as with the Slow Food movement, to “savor” the relationship needed to engage (30, 52, 53). The work of the MHSs at the JEEDC offers a practical example of slow engagement at this point in the care continuum and underscores its potential importance for effectively engaging people in prebooking jail diversion programs. However, more research is needed to investigate whether slow engagement is an effective approach in other criminal justice and mental health settings that work together at this point in the care continuum.

Conclusions

The challenges of diverting persons who are chronically homeless and are living with mental illness from jail and toward recovery are considerable. Prebooking jail diversion programs such as the JEEDC show promise, especially when CJSs and MHSs work together. The results of our study suggest that potential evidence-based practices for this point of care might include limited or no use of police intervention, the use of harm-reduction and trauma-informed approaches, and flexible metrics of success that incentivize slow engagements with clients.