Client Outreach in Los Angeles County’s Assisted Outpatient Treatment Program: Strategies and Barriers to Engagement

Abstract

Purpose:

Assisted Outpatient Treatment (AOT) programs can compel treatment-refusing individuals to participate in mental health treatment via civil court order. In California’s AOT programs, individuals first must be offered 30 days of outreach services and can accept services voluntarily. This study examines the use of outreach strategies in an AOT program with the potential for voluntary or involuntary enrollment.

Methods:

Outreach staff completed a survey in which they reported and rated outreach strategies and barriers to treatment for 487 AOT-referred individuals.

Results:

Outreach staff reported using a broad array of strategies to persuade and engage clients. Supportive and persuasive strategies were most common. More coercive strategies, including court order, were used when needed. More clients enrolled voluntarily (39.4%) than involuntarily (7.2%).

Conclusions:

Outreach, coupled with the strategic used of potential court involvement, can lead to voluntary enrollment of treatment-refusing individuals with many, often severe, barriers to engaging in outpatient treatment.

Treatment non-adherence is a critical issue for the care of individuals with serious mental illness (SMI). Non-adherence can refer to refusal of all services or to inconsistently following treatment and/or medication regimens (Haddad, Brain, & Scott, 2014; Kessler et al., 2001). Non-adherence rates among those with SMI are high. In a national survey conducted by the Substance Use and Mental Health Services Administration (SAMHSA, 2018), 35.9% of those with SMI did not receive any treatment for their mental health in the prior year, and medication non-adherence among those with schizophrenia is around 50% (Haddad et al., 2014; Lacro, Dunn, Dolder, Leckband, & Jeste, 2002; Tessier et al., 2017).

Individuals may refuse treatment for many reasons. In a systematic review of medication non-adherence among those with SMI, the most common reason (55.9%) was lack of insight (Velligan, Sajatovic, Hatch, Kramata, & Docherty, 2017). Other factors associated with treatment non-adherence include poor therapeutic relationships with providers, trauma from prior psychiatric treatment, desire to maintain autonomy, substance use disorder, cognitive impairment, depression, shorter illness duration, negative attitudes about medication, and poor system coordination (Haddad et al., 2014; Lacro et al., 2002; Lecomte et al., 2008; Priebe, Watts, Chase, & Matanov, 2005; Sajatovic et al., 2009; Tessier et al., 2017; Velligan et al., 2017). Adverse outcomes due to non-adherence can include clinical deterioration, self-harm, arrests, violence to others, homelessness, hospitalization, and early morbidity and mortality (Colom et al., 2000; Gilbert et al., 2010; Monahan, Vesselinov, Robbins, & Appelbaum, 2017; Swartz & Swanson, 2004; Walker, McGee, & Druss, 2015).

Assisted Outpatient Treatment.

Assisted outpatient treatment (AOT)—also referred to as involuntary outpatient commitment or community treatment orders—seeks to provide outpatient services to treatment-refusing individuals with SMI who have a recent history of arrests or psychiatric hospitalizations and/or are at risk of clinical deterioration, disability, or harm to self or others. Though AOT programs vary, a key feature is a civil court process in which a judge orders the individual to adhere to a treatment plan within the community. Since there are no legal consequences for non-adherence, it is hoped that individuals will respond to the “black robe” effect of being ordered into treatment by a judge (Kisely, Campbell, & O’Reilly, 2017). When individuals continue to refuse services, AOT staff also may take them to a hospital for evaluation and/or initiate additional efforts to persuade them to accept services (Swartz, Bhattacharya, Robertson, & Swanson, 2017).

As of 2019, AOT statutes have been approved in 47 states (all but Connecticut, Maryland, and Massachusetts), with varying designs and degrees of implementation. AOT programs feature three main pathways for entry into treatment: a community-entry model (for those not in treatment), a hospital or jail transition preventative model (for those exiting an institutional setting), and a safety-net monitoring model (identifying those at highest risk of engaging in violence to others) (Meldrum, Kelly, Calderon, Brekke, & Braslow, 2016). Most research has focused on hospital or jail transition models, with little research on the effectiveness of community-entry models. Though research methods and findings are varied and highly dependent on the particular program, there are some encouraging signs regarding reduction of adverse outcomes, such as arrest, victimization, hospitalization, and minor acts of violence, and increased use of medication (Cripps & Swartz, 2018; Gilbert et al., 2010; Hiday, Swartz, Swanson, Borum, & Wagner, 2002; Phelan et al., 2010; Schneeberger et al., 2017; Swartz et al., 2017; Swartz & Swanson, 2004; Swartz et al., 2010; Van Dorn et al., 2010).

In California, AOT programs have been implemented in several counties and follow a community-entry model, with some referrals made from jails and hospitals. Due to civil liberties concerns by mental health advocates, AOT in California differs in two key ways from other AOT programs, most of which are strictly involuntary. First, California’s statute requires a minimum of 30 days of outreach and engagement services. Second, individuals can agree to services voluntarily; if they remain unwilling, a civil order for treatment can be pursued after the 30 days of outreach (“Laura’s Law,” 2002). California’s AOT programs offer a unique opportunity to understand the role of outreach services in a program that encourages voluntary treatment while maintaining the potential for involuntary treatment.

Outreach and engagement services.

Outreach has become an essential component of mental health services (Burns & Firn, 2017). “Difficult to engage” populations include those who are experiencing their first psychotic episode, are homeless, and/or have co-occurring substance use issues (Dixon, Holoshitz, & Nossel, 2016).

Outreach services need to be flexible and adapted to the target population to be effective (Dixon et al., 2016; Olivet, Bassuk, Elstad, Kenney, & Jassil, 2010). They are designed to “meet people where they are” and take place in the community, such as on the streets or in shelters, homes, jails, or hospitals (Olivet et al., 2010). A review of the evidence on engagement strategies for difficult-to-engage individuals with SMI identified a number of components of successful outreach with individuals who are homeless (strong staff-client therapeutic alliance, persistent and consistent providers, emphasis on basic supports and not just medication, clients made to feel accepted, flexible strategies, and team-based outreach) and who have co-occurring disorders (shared goals, focus on treatment priorities beyond medication, psychoeducation, team-based care, and community-based outreach) (Dixon et al., 2016).

Outreach services can include a great deal of treatment coercion (Burns & Firn, 2017). Even among voluntary mental health treatment program participants, reports of feeling coerced or pressured are common. A majority (44%–55% across states) of adults with SMI in public mental health services in five states reported at least one form of leverage was used to induce them to accept treatment (Monahan et al., 2005). In involuntary community mental health programs, there is mixed evidence that compulsory treatment is associated with increased perceived coercion relative to control groups, with some studies finding greater coercion and others finding no differences (Pridham et al., 2016).

Persuasive strategies have been described using a five-level hierarchy of pressure (Szmukler & Appelbaum, 2008) that highlights the ubiquity of various levels of coercion even in the context of voluntary outpatient services. At the persuasion level, a clinician presents information that the client is free to accept or reject regarding the benefits of treatment. Interpersonal leverage refers to a clinician’s use of their relationship with the client (through expression of approval or disapproval) to influence the client’s decision. Inducement occurs when a clinician suggests that the client could receive additional support or services (e.g., assistance with housing or financial benefits (Monahan, 2008) if they participate in treatment. Threats involve warnings regarding potential loss of existing supports or services or the possibility of court-ordered treatment, hospitalization, or jail (through revocation of parole or lack of assistance with diversion from criminal proceedings). Lastly, compulsion is when a clinician obtains a legal order that requires adherence to treatment in a hospital or in the community.

In addition to persuading clients to enter mental health services, outreach services include efforts to engage participants. This relationship-building and social contact can itself be therapeutic. Burns and Firn (2017) created a classification of approaches to engage those refusing care: constructive, monitoring, and restrictive. Constructive approaches, common in recovery-oriented services, are client-directed, strengths-based, empowering, field-based, and help with resource linkage and community reintegration. These approaches overlap with the outreach strategies of interpersonal leverage and inducement. Monitoring approaches are less collaborative with clients and can involve intensive efforts to reach a client who is or may become avoidant (showing up their home; frequent texts) and use of key informants (e.g., family and others) to gather information about the client. Restrictive approaches are the last resort and correspond to the compulsion level of outreach strategies. They involve relationship-building during involuntarily treatment and can involve visiting the client in a hospital or jail.

Outreach and engagement strategies are common elements of community mental health services, but no studies that we are aware of have examined their use within involuntary program that allows for avoidance of court involvement through voluntary acceptance of services. The goal of this study is to explore the use and perceived effectiveness of outreach strategies in Los Angeles County’s AOT program and to understand the barriers to engagement present for the population it serves.

AOT in Los Angeles County.

On July 15th, 2014, the Los Angeles County Board of Supervisors voted unanimously to implement AOT (Sewell, 2014). On May 15, 2015, the Los Angeles County Department of Mental Health (LACDMH) launched its AOT program, which is based on a community-entry model. Flow through the program (Figure 1) begins with a referral, typically from mental health providers, family, or police, after which a committee reviews whether the individual meets AOT eligibility criteria. Eligible individuals receive a minimum of 30 days of outreach and engagement services. These are provided by the 15 members of the LACDMH Emergency Outreach and Triage Division’s (EOTD) two AOT Outreach and Engagement (O&E) teams, which between them cover all of Los Angeles County. The teams are composed of licensed clinical social workers, case managers, nurses, peer specialists, a nurse aide, and a psychiatrist, and include individuals who have lived experience or a family member with mental illness.

Figure 1.

Program Flow for Individuals from Referral to Possible Enrollment (Voluntarily or Under Court-Order) in Assisted Outpatient Treatment in Los Angeles

The O&E teams encourage individuals to enroll voluntarily, but outreach efforts may include mention of the potential for a court order. After 30 days of outreach a court order may be pursued if the client will not agree to services and is deteriorating. It is important to note that medication is not a component of the court order in California, and compulsory medication requires a separate hearing. When a client is either willing or compelled to enroll, the O&E team and the treatment team arrange a “warm handoff”: a facilitated meeting to help transition the client into services, using the trust and rapport with the O&E team to help the new treatment team build a relationship with the client.

Clients enrolled in AOT, voluntarily or involuntarily, receive services from the same array of AOT-contracted Full Service Partnership (FSP) or Enriched Residential Service (ERS; formerly Institution for Mental Disease Step-Down) providers. FSP services are a modified version of Assertive Community Treatment (Starks et al., 2017) in which interdisciplinary teams provide 24–7 field-based care. FSP programs normally have a 1:15 staff-to-client ratio, but the ratio for AOT FSP is 1:10 to allow for more intensive services. ERS facilities provide a more intensive level of care for clients who need housing and mental health services in a more structured but still unlocked facility. Single-site ERS programs operate both the housing and treatment services, while multi-site ERS programs provide only treatment services and contract with private housing operators. LACDMH has 300 FSP program slots and 60 ERS slots dedicated to AOT participants. Court orders for treatment are for six months and renewable if the client cannot be graduated to less intensive outpatient treatment.

In addition to the court order, the AOT statute includes two additional legal mechanisms for involuntary treatment, both of which are referred to as 5346 holds and are variations of a 72-hour involuntary hospitalization. These are in addition to the existing option in California of pursuing a 5150 hold, a 72-hour involuntary psychiatric hold for individuals who meet stricter criteria for grave disability or danger to self or others. Under 5346(d)(3), when an individual is the subject of a petition for a possible AOT court order and has refused to be examined by a licensed mental health treatment provider, the judge can order them transported to a hospital for evaluation to determine whether they meet clinical criteria for AOT, after which they will be transported back to court. Under 5346(f), when an individual has refused to comply with AOT treatment ordered by the court and is deteriorating, the treatment provider may have them transported to the hospital and held for up to 72 hours for observation and evaluation to determine if they meet criteria for a 5150 hold (“Laura’s Law,” 2002).

LACDMH chose to have dedicated O&E teams, rather than AOT treatment providers, perform outreach during the initial outreach period. First, LACDMH anticipated that highly experienced personnel with specific training in O&E would be required to successfully engage this population. Second, Los Angeles County is large. Locating and providing outreach to individuals in the field requires a great deal of transportation time that could pose a burden for treatment providers. Third, contracted FSP and ERS providers are only able to bill for services if they make contact with a client. LACDMH has funds to provide outreach services regardless of whether contact with a client is made. LACDMH currently has the capacity to provide AOT outreach services to at least 500 individuals annually. Though outreach strategies also are used by the treatment providers to engage clients after they enroll, this study focuses on the use of outreach strategies by LACDMH’s dedicated O&E teams.

Method

This study is a cross-sectional examination of the outreach phase of LACDMH’s AOT program. Data are primarily from a cross-sectional survey completed by AOT O&E staff about each eligible AOT-referred client at the end of outreach, providing their perspective of the client and the outreach services they received. This survey is one part of a larger evaluation of LACDMH’s AOT program, and was developed to provide information on all individuals who received outreach services, including those who do not ultimately enroll in treatment. Because O&E staff are responsible for this outreach, and often have extensive contact with referred individuals and their families, they are uniquely able to provide data on this broad sample of AOT-eligible referred clients. Surveying clients directly at the outreach phase was not feasible due to LACDMH concerns about interfering with a delicate outreach process and the expectation that clients willing to complete a survey at that stage would not be a representative sample. Secondary data from program administrative and claims datasets supplement the survey data.

Participants.

Post-outreach surveys were completed by O&E staff at the end of outreach for all AOT-eligible clients who received outreach services. Data presented are from surveys completed during an 18-month period between July 25, 2017 (survey rollout) and January 22, 2019, with a total sample of 487 clients about whom staff completed surveys (Table 1).

Table 1.

Client Characteristics (n = 487)

	N	%
Gender
Female	174	35.7%
Male	312	64.1%
Other/Unknown/Does not identify	1	0.2%
Age (2018 minus birth year)
18–30	173	35.5%
31–40	142	29.2%
41–50	87	17.9%
51–60	65	13.4%
61–71	20	4.1%
Race and Ethnicity
Asian	44	9.0%
Black	103	21.1%
Hispanic	163	33.5%
White	149	30.6%
Multiple or Other	23	4.7%
Unknown or Not Reported	5	1.0%
English Speaking
Yes	478	98.2%
No	6	1.2%
Unknown	3	0.6%
Diagnosis
Schizophrenia	223	45.8%
Schizoaffective	84	17.3%
Psychotic Disorder	74	15.2%
Bipolar	68	14.0%
Mood Disorder	32	6.6%
Other	4	0.8%
Missing	2	0.4%
Referral Source
Clinician/hospital	186	38.2%
Family member	119	24.4%
Psychiatric mobile response or mental health/law enforcement team	108	22.2%
Social service agency	39	8.1%
Law enforcement/probation officer	32	6.6%
Unknown	2	0.4%

Data Collection.

The post-outreach survey was developed with input from O&E staff and programmed into REDCap’s secure web-based electronic data capture software hosted at UCLA (UCLA CTSI; UL1TR001881 (Harris et al., 2019; Harris et al., 2009)). UCLA REDCap accounts and project access were requested for all O&E staff by project PI JTB. Modifications to the survey were made based on feedback provided during an in-person training on how to complete the survey, and final instructions were sent to O&E staff on July 13, 2017 asking them to begin completing surveys about any clients whose outreach had recently ended as well as at the end of outreach for all other clients going forward. Data used for imputing missing demographic data and cross-checking the reason outreach ended are from AOT-LA program administrative data provided by EOTD to UCLA on May 20, 2019 as well as demographic, outreach, and claims data provided by LACDMH Clinical Informatics on May 28, 2019. Data were provided to UCLA identified by a common study ID that facilitates linkage of de-identified data across data sources, in accordance with protocols approved by UCLA’s Institutional Review Board and the LACDMH Human Subjects Research Committee.

Measures.

The survey asks O&E staff about the client’s circumstances, where outreach occurred, barriers to engagement, outreach strategies, why outreach ended, questions about family involvement (to be reported elsewhere), and a number of open-ended questions.

Outreach Services Locations and Outcomes.

Locations in which outreach services were delivered and reasons for outreach ending were selected by staff from a list developed with O&E staff or entered as free text if not pre-listed. Staff could select more than one location in which services were delivered but were asked to select a single reason that outreach ended. Staff reports of the reason outreach ended were cross-checked with program administrative data and reconciled as needed.

If staff reported the client was enrolling in AOT but administrative data showed that the client did not enroll, we used claims data to confirm that the client did not initiate services and recoded to “Expected Client to Enroll in AOT, but Did Not.” When administrative data noted that the client did not meet criteria, we used claims data to confirm that the client did not initiate services and recoded to “Did Not Meet Criteria.” When staff reported a voluntary AOT enrollment but administrative data listed a court order or settlement agreement, we confirmed that these took place after the client enrolled and noted those cases in the text or, if they took place prior to enrollment, recoded to “AOT, Involuntary.” If staff reported an involuntary AOT enrollment but administrative data listed no court order or settlement agreement, we recoded to “AOT, Voluntary.” If staff reported that outreach ended due to inability to find the client but administrative data showed an enrollment, we checked the timing of the enrollment. We recoded to “AOT, Voluntary” a client who continued to receive outreach services throughout their enrollment to assist with engagement, and whose post-enrollment outreach ultimately ended because they could not be found. Finally, if staff reported that outreach ended due to enrollment in Community-Based Restoration to Competency but administrative data showed an AOT enrollment, we checked the timing of the enrollments and services and recoded to “AOT, Voluntary” or “AOT, Involuntary” if these were the enrollments associated with the end of outreach.

Because of the complex nature of the AOT program, reasons reported for the end of outreach are not necessarily static or mutually exclusive. As mentioned above, some voluntarily-enrolled clients continued to require outreach services and eventually were court-ordered during their enrollment. AOT-enrolled clients also could be conserved, lost to follow-up, or linked to non-AOT mental health services after graduating from or being prematurely discharged from AOT. Additionally, enrollment in either AOT or other mental health services does not guarantee that clients will engage with or receive services. Here we focus on the outreach process, but in future articles we will use claims data to analyze treatment services received after referral to AOT, overall and among those who enroll in AOT services.

Barriers to Engagement.

Staff reported whether they thought an issue was present for each client, selecting from a list of issues developed with O&E staff or entering free text for issues that were not pre-listed. They rated each selected or entered issue based on their perception of whether the issue was a barrier to treatment for that client by selecting from the response options serious barrier, moderate barrier, minor barrier, or not a barrier.

Outreach and Engagement Strategies.

Staff reported whether a particular strategy was used with each client, selecting from a list of strategies that had been developed with O&E staff or entering free text for strategies that were not pre-listed. Strategies fell into three categories: services provided to the client during outreach (inducement and interpersonal leverage); services advertised to the client as benefits of treatment (persuasion and inducement); and legal strategies (threats or compulsion). For each strategy selected or entered, they rated the extent to which they perceived it as contributing to the engagement of that client by selecting from the response options very effective, somewhat effective, not effective, or counterproductive.

Open-Ended Qualitative Responses.

Staff completed free-response questions about the outreach process, including why they thought outreach had been successful or unsuccessful with the client; how confident they were that the treatment provider would be able to engage the client in treatment after enrollment; and any other comments they wanted to provide.

Outreach Services Contacts.

Analyses of the number of outreach services are from Community Outreach Services data delivered to UCLA by LACDMH Clinical Informatics. These data capture non-billable outreach services by LACDMH providers, including the AOT O&E teams.

Analysis.

Data management and analyses were performed using Stata 14 statistical software. All survey variables were categorical or free-text responses. Variables that allowed free-text responses, including mental health diagnoses, race/ethnicity, and “other” responses to categorical questions, were reviewed and coded into categories by authors SLS and ELK to allow inclusion in data tables or figures.

Descriptive analyses of participants’ clinical and personal characteristics (number and percentage) were performed. For treatment barriers we present the percentage of individuals for whom a barrier was reported, overall and within each staff-rated response category describing the degree to which staff perceived the issue to have been a treatment barrier for the client (serious barrier, moderate barrier, minor barrier, or not a barrier). Similarly, for outreach strategies we present the percentage of individuals for whom a strategy was reported to have been used, overall and within each staff-rated response category indicating how effective staff perceived it to be for the client (very effective, somewhat effective, not effective, or counterproductive). Strategies are described in the text as having usually been perceived to be effective when O&E staff indicated that they perceived them either as very effective or somewhat effective for the majority of clients for whom the strategy was used.

Responses to open-ended questions were reviewed and coded into thematic categories by SLS, with review by ELK and other co-authors, to deepen our understanding of the client population and outreach process and to facilitate the presentation of client examples.

Administrative data on the number of outreach services provided for clients in the survey sample were merged with administrative program referral and eligibility data. The dataset was restricted to outreach services that were provided by the AOT O&E teams within 183 days (6 months) of the client’s referral to AOT or, if referred more than once, the first referral in which the client met AOT eligibility criteria. The mean number of services (and standard deviation) is reported across all clients in the outreach sample, as well as for the conditional sample of clients with any services of each type.

Results

Sample characteristics.

Table 1 summarizes characteristics of the client sample for whom outreach services were completed. The sample was racially and ethnically diverse, predominantly male, and relatively young. Most clients had a diagnosis of SMI. Typically, clients were referred to the AOT program by a clinician or hospital, family member, or a mobile mental health team, such as a psychiatric mobile response team or a joint mental health-law enforcement team. Less typically, a social services agency or a law enforcement/probation officer referred clients.

Table 2 shows the reported reasons that O&E staff ended outreach for individual clients. AOT enrollment was reported as the reason for just under half of clients (46.6%). For 39.4%, outreach led to voluntary enrollment in AOT services, though ten of these clients were eventually court-ordered while enrolled. For the other 7.2%, enrollment was involuntary. For other clients outreach was reported to have ended due to enrollment in treatment services other than AOT; inability of the O&E staff to locate or meet the client (often reflecting the difficulty of engaging individuals who are homeless); client refusal (including instances in which the client was not deteriorating); ineligibility; or death.

Table 2.

Reasons for Outreach and Engagement Ending for Each Client (n=487)

	N	%
AOT Treatment
AOT, Voluntary	192	39.4%
AOT, Involuntary via Court Order	17	3.5%
AOT, Involuntary via Settlement Agreement	18	3.7%
Other Treatment Services
Conservatorship	31	6.4%
Community-Based Restoration to Competency*	18	3.7%
Other Treatment	42	8.6%
Unable to Find or Meet Client
Can’t Find Client	80	16.4%
Unable to Meet Client	6	1.2%
Client Refusal
Not Deteriorating	16	3.3%
Refused Available Providers	1	0.2%
Expected Client to Enroll in AOT, but Did Not	20	4.1%
Not Eligible
Long-term Incarceration	14	2.9%
Criminal Charges; Judge Cancelled Petition	1	0.2%
Out of County	7	1.4%
Did Not Meet Criteria	10	2.1%
Safety Risk	1	0.2%
Private Insurance	1	0.2%
Referral Withdrawn	4	0.8%
Client Deceased
Deceased	8	1.6%

^*Note: Los Angeles County has two community-based restoration programs, Misdemeanor Incompetent to Stand Trial (MIST) and Felony Incompetent to Stand Trial (FIST), where individuals charged with a crime but found incompetent to stand trial can be treated for restoration to competency in the community rather than jail.

Barriers to Engagement in Treatment.

Figure 2 shows how frequently various potential barriers to engagement were reported for these clients, and, when reported, whether O&E staff viewed them as a barrier for the client. Client mental health was frequently reported as an issue, and usually rated as a serious or moderate barrier; in particular, clients exhibited lack of insight (“Client simply does not believe she has a mental illness. Client extremely psychotic”), paranoia (“Outreach was not successful due to client’s mistrust/paranoia and refusal to engage with team”), anger issues, substance abuse (“Client continues to use heroin and meth, which coupled with his mental illness impairs his ability to engage in services”), and lack of motivation. Threatening words or behaviors also could pose a serious barrier (“He has no insight and also very aggressive during O&E”), sometimes to the point that O&E requested assistance from law enforcement (“Client is a very angry individual and the police had to accompany AOT in all our O&E efforts”).

Figure 2.

Barriers to engagement: staff-reported frequency and perception of severity (n=487).

Treatment barriers related to attitudes about mental health treatment were common. Distrust of mental health providers and fear of treatment due to trauma from past psychiatric hospitalizations were significant barriers: “Client eloped from mental hospital demonstrating a strong… distrust for mental health services,” and “It was difficult to engage client because he had a bad experience with past mental health providers, and therefore did not trust AOT team.”

Although California’s AOT program does not include an involuntary medication component, and individuals are able to enroll in AOT services without agreeing to see a psychiatrist or take medication, O&E staff often rated resistance to medication as a serious or moderate barrier for clients: “The client would not agree to meet with AOT team… Client was only interested in receiving some form of financial support and strongly opposed medication therapy.” At times resistance was explicitly due to previous adverse experiences with medication: “The client thought of mental health services as medication. The client had a bad reaction to medication [and] seemed reluctant to give up the notion that mental health treatment meant forced medication.”

Finally, clients often had challenging circumstances, such as legal issues, lack of resources, and complicated family situations. Lack of housing often was a serious barrier, because homeless clients were difficult to locate during outreach or after assignment to a provider: “It will be difficult for provider to provide consistent services given [that] client is homeless and does not frequent one specific location.” Frequent hospitalizations also could make clients challenging to locate, especially when discharge communications from hospitals were inadequate: “Outreach was unsuccessful because team was unable to make contact with client upon her discharge from the hospital.”

Additionally, clients who were chronically homeless often resisted the structure of a residential treatment or housing placement, especially when substance use also was a factor: “Client very hard to find since he lives on the street. Client also has significant anger issues and is easily triggered, which means he leaves arranged housing after a day or two. But the biggest problem is client’s use of meth.”

The barriers and ratings in Figure 2 and the open-response statements from O&E staff make it clear that AOT-eligible clients are a particularly challenging population to engage in treatment.

Outreach Service Contacts and Locations.

Table 3 shows the number of outreach services recorded in the LACDMH Community Outreach Services system in the 6 months following each client’s referral to AOT or, if referred more than once, after the first referral in which the client met AOT eligibility criteria. The mean number of services is reported across all clients, and for the conditional sample of clients with any services of each type. Nearly all clients in the sample (94.1%) had outreach services delivered to them, and 29.0% had outreach services delivered to their family members. O&E staff delivered an average of 10.6 client or family services for each client in the sample (SD=9.3), either in person or by phone.

Table 3.

Number of AOT Outreach Services in the 6 Months Following Referral (n = 487)

	Full Sample*		Conditional Sample*
	Mean	SD	N	%	Mean	SD
Any Outreach, Client or Family	10.6	9.3	460	94.5%	11.2	9.2
Outreach with Client	9.7	8.7	458	94.1%	10.3	8.6
In Person	6.4	5.7	443	91.0%	7.0	5.6
By Phone	3.3	5.5	279	57.3%	5.8	6.2
Outreach with Family Member	0.9	2.3	141	29.0%	3.1	3.4
In Person	0.2	0.8	58	11.9%	1.8	1.5
By Phone	0.7	1.9	115	23.6%	2.9	3.0

^*Note: The full sample includes all clients in the outreach sample (n=487), including those who did not receive any outreach services for the category in question. For each category of outreach services (e.g., in-person with the client), the conditional sample includes only the clients with a non-zero number of services in that category (n varies, and is presented in each row). For services that were received by most clients, such as in-person client services, the means differ only slightly between the full sample and the conditional sample. For less-common services, the conditional-sample mean is much greater than the full-sample mean. For example, in-person outreach with a family member was provided for only 12 percent of clients; therefore, while the mean number of such services was 1.8 for the conditional sample, the mean number for the full sample was only 0.2 such services. The full-sample means give a better sense of the volume of services the outreach teams are providing overall, while the conditional-sample means give a better sense of how intensively a given type of service was provided when it was part of the range of services for a client.

Table 4 shows the locations where O&E staff reported delivering outreach services to clients. Over a third of clients received outreach in their home; nearly a quarter on the street; nearly a quarter in a hospital; and over ten percent in jail. Clients also received outreach in living situations such as supported living facilities, motels, emergency shelters, and other temporary housing facilities; in the homes of family members; in cafés and restaurants; in parks and local establishments; and in service agencies including clinics and methadone clinics. This diversity of locations is indicative of persistent outreach across the spectrum of constructive, monitoring, and restrictive approaches to engagement.

Table 4.

Where Outreach Services were Delivered To Clients (n = 487)

	N	%
In the client’s home	177	36.3%
In a hospital	119	24.4%
On the street	114	23.4%
In jail	53	10.9%
At a family member’s home	33	6.8%
In a café/restaurant	18	3.7%
At a supported living facility	23	4.7%
In a hotel/motel	12	2.5%
At an emergency shelter	11	2.3%
In court (via Other)	11	2.3%
In a clinic, urgent care, or services agency (via Other)	8	1.6%
In a park (via Other)	6	1.2%
In a car, shed, or family member’s yard (via Other)	5	1.0%
In a library, church, or other establishment (via Other)	4	0.8%
At other temporary housing situation (via Other)	4	0.8%
At a methadone clinic (via Other)	2	0.4%
At a juvenile hall (via Other)	1	0.2%

Reflecting on the success or failure of outreach in engaging clients in treatment, O&E staff sometimes cited this flexible outreach: “O&E was successful because AOT team was able to engage client in various settings, i.e., hospitalization, home, streets, and Starbucks” and, similarly, “The outreach was successful because we met with client wherever she was that was considered safe. I believe she realized how AOT team members wanted to help her.” The willingness of staff to meet clients where they are, and the persistence that these multiple meetings reflect, were viewed by staff as critical to building rapport with clients.

Outreach and Engagement Strategies.

Figures 3–5 show the frequency with which O&E staff reported using various outreach and engagement strategies with clients and the extent to which they perceived these strategies as being correlated with engagement in services for each client. From these figures, it is evident that O&E staff reported drawing from a wide array of strategies and the effectiveness of a strategy was dependent on the individual client. A strategy could be rated by staff as very effective with one client and counterproductive with another. Given this reality, O&E staff often appear to have exhausted all options to find a way to connect with clients. The types of strategies they most frequently reported using were support to client and families (interpersonal leverage; Figure 3) and persuasion and inducements, such as telling clients about the benefits of treatment (Figure 4). When these strategies were not enough they turned to legal strategies (threats, compulsion), including discussion or use of court-ordered AOT, psychiatric holds, or mental health treatment for jail diversion (Figure 5).

Figure 3.

Staff-reported frequency and perceived effectiveness of outreach strategies: services provided to the client during outreach (n=487).

Figure 5.

Staff-reported frequency and perceived effectiveness of outreach strategies: Legal strategies (n=487).

Figure 4.

Staff-reported frequency and perceived effectiveness of outreach strategies: Services advertised to clients as benefits of treatment (n=487).

Services Provided to the Client During Outreach.

Among outreach strategies that fall under the category of services provided during outreach (Figure 3), O&E staff most frequently reported using motivational interviewing, psychoeducation to clients and their families, and various kinds of support, including emotional, informational, and crisis support and support during legal proceedings. They usually rated these strategies as effective in engaging the client and never rated them as counterproductive.

Reflective of the high rates of poverty among those with SMI, staff frequently reported providing basic resources as part of outreach. Purchasing food or coffee for clients was a common way to get them to spend time talking with O&E staff: “[C]lient would walk, talk and eat with AOT Team members and would admonish team if we were late to appointments” and “[B]ringing her food seemed to help rapport building tremendously.” A third of clients received assistance with housing. Staff also assisted with obtaining identification or benefits, accessing other programs such as substance abuse services, and accessing medical care. This assistance was an important way of developing rapport, improving clients’ well-being, and building positive associations with the mental health system. Development of rapport was essential, especially in instances where staff had to take actions that could threaten their relationship with the client: “We developed a trusting relationship, even though we hospitalized. Client found that the hospitalizations led to a dramatic improvement in her functionality.”

When asked why outreach was or was not successful, O&E staff rarely explicitly mention mental health services provided during outreach. However, the benefits of motivational interviewing and psychoeducation are visible between the lines. Staff frequently stated that outreach went well because the client was willing to engage or motivated to accept services due to an awareness of their illness or a desire for potential benefits of treatment: “Client didn’t feel he needed services but he knew he was about to be homeless and didn’t want to live on the street, so that gave us some leverage.” These successes highlight the importance of building connections with clients to understand their potential motivations. In some cases they show how barriers to treatment, such as history of traumatic hospitalization, could be transformed into motivators to treatment, as clients saw services as a means of addressing their problems: “O&E was successful because he was motivated to engage. Client repeatedly stated that he does not like to be hospitalized and does not want law enforcement contact.”

Services Advertised to the Client as Benefits of Treatment.

Outreach also makes use of persuasion or inducement strategies, such as telling clients about the services they can receive if they enroll (Figure 4). The mental health services O&E staff most frequently reported telling clients about were case management, medication, and therapy. Housing and benefits also were mentioned frequently, followed by assistance with employment, reintegration, or going back to school: “Outreach was successful because client expressed needing to be connected to a supportive team to assist with housing resources and other services.” Substance use treatment services were reported to have been mentioned to a fifth of clients, with lower rates of perceived effectiveness. Groups and socialization opportunities were mentioned least frequently. AOT services are usually provided by FSP programs, in which services are delivered to clients in their home or at another preferred location, and therefore are less likely to include clinic-based group services than the other services advertised.

Legal Strategies.

Use of legal strategies (Figure 5) was much less frequently reported than other types of strategies. The most frequently reported legal strategy, used with 30.2% of clients, was advising the client that the outreach team would or could pursue court-ordered AOT, though they reported actually seeking a court order with only 10.1%. Psychiatric holds were reported to have been mentioned as a possibility to 8.0% of clients and pursued for 9.9% of clients, and staff reported assisting in efforts to obtain a conservatorship for 4.1% of clients. For some clients, staff reported suggesting that participation might prevent future arrests, jail time, or hospitalizations; suggesting or pursuing mental health services as a diversion from jail; advising the client that they might face jail or a probation violation; and reporting clients to law enforcement or a probation officer.

Free-text responses from O&E staff show the challenges, pitfalls, and benefits of pursuing legal strategies. In some instances, staff highlighted the strategies’ limitations. With a client whose outreach ended in conservatorship, they described efforts that failed prior to conservatorship: “Law enforcement refused to lay hand[s] on client, and at least three 5150s had to be called off due to [inability] to get client on gurney.” With a client who enrolled involuntarily and was later conserved, they noted, “Client was not interested in services and only complied when court ordered… It was fairly obvious from the beginning that client would not be successful in an outpatient setting, even if it was court mandated.”

The legal strategies available through the AOT program, such as the 5346 involuntary holds, were at times instrumental in getting a client into court ordered treatment: “The 5346 order was critical in hospitalizing client so that client can be transported to AOT court hearing and also client can receive the long acting injectable.” Offering mental health treatment as an alternative to incarceration also worked for some: “Client did not want to face jail time so treatment was included as part of probation.”

Other staff statements hint at the complex interplay between outreach efforts and the potential involuntary nature of AOT. For example, the threat of a court order was sometimes necessary to tip a client over into accepting services: “We developed a decent rapport with client after many visits. When we said he would either have to sign and consent for services or be taken to appear before the judge, the client decided to accept services.” Similarly, with a client who ultimately accepted services voluntarily after around six outreach visits, they noted that “the development of a relationship… was helpful before we sprung the idea of court mandated treatment on the client. I think he was more open because of it.”

Finally, with other clients, the outreach team was able to offer AOT as an alternative to incarceration on criminal charges. In these cases, the legal threat was coming from the criminal court, not an AOT petition, and the O&E team’s assistance with diversion was a way of building a supportive relationship as well as a way of directing the client into treatment through his desire to stay out of jail:

With [the] AOT public defender, AOT [was] able to work out a deal with [the] criminal court public defender that client’s case will be continued, and client will be under AOT court supervision. Otherwise, considering the offense, he has to be in prison for at least three years…. [U]nder AOT court supervision instead of regular probation [if the] client fails to follow court-ordered mental health treatment, client will be hospitalized under AOT program, instead of being sent back to prison [for a probation violation].

In the cases of clients who ultimately enrolled involuntarily via court processes, clients’ initial responses to a mention of court petition were variable. One client initially responded with skepticism: “Client did not believe he could be brought before the court before he was finally, actually, brought before the court. He said we were lying about it.” Another reacted angrily to the prospect of attending court: “When AOT team informed client about AOT court hearing, client got very upset and angry and made verbal threats towards AOT O&E team.” However, warnings of court involvement to compel treatment engagement could be a positive as well: “Client began to trust the AOT team and somewhat knew things were serious when she received the court summons.” The court experience was a valuable tool to compel treatment in some cases: “[The] ‘judge effect’ really worked.”

When discussing why they thought outreach was successful for a client or whether they were confident that the client could be engaged, O&E staff frequently mentioned stabilizing factors such as a consistent living arrangement, even when homeless (“I think [O&E] activities were successful because the client frequents the same area”) or supportive family members (“Client family members provided additional support to reinforce the client’s treatment plan and mental health goals”). These factors made them easier to locate, facilitated monitoring, and could be a source of additional interpersonal pressure to comply. In other cases, they mentioned restrictive approaches involving the justice system (“Able to outreach and engage client in jail; able to advocate to public defender to have client diverted to mental health treatment; client consented to treatment.”) or hospitalizations during which a client received medication (“Successful due [to] the following reasons: 5150 hospitalization to help client be compliant on psychotropic medication and client started eating; Provided psychoeducation to family; Set firm limit with father and mother”). These examples again highlight the interplay of voluntary and involuntary strategies and the complex role of coercion in engaging clients.

Client Likelihood of Engaging with Provider after Enrollment.

Once a client consents to or is court-ordered to services, the O&E team performs a “warm handoff” to a contracted treatment provider. O&E staff also may be called upon to provide additional post-enrollment outreach if the client is insufficiently engaging with the provider. For some clients, O&E staff report high confidence that the provider will be able to engage the client, but for others they have low confidence, often due to the same barriers to engagement observed during outreach.

When O&E are confident in the likelihood of success, they often cite willingness to engage, usually with insight: “Because client is eager and opened to receiving services. Client has some insight into her illness and is opened to medication support.” However, lack of insight was not always perceived to be meaningful barrier: “I am very confident that the provider would be able to engage client, due to her willingness to engage in spite of her lack of insight.”

Sometimes a client’s willingness was based on specific concerns or goals, with staff reporting that they were very confident because “Client was interested in the services and wanted the support,” “client is willing to get help for the sake of her [young] daughter,” or “client is motivated; doesn’t want to be hospitalized.” In another instance, they noted that outreach was “very successful since client’s legal case is in jeopardy and client runs the risk of going to jail for criminal charges if not compliant with mental health services.” These motivations hint at successful use of various outreach strategies and highlight the uniqueness of each client’s situation.

The role and limitations of legal strategies and other inducements remains apparent at the handoff stage. With some clients, O&E staff were confident about engagement due to the legal threats available through the AOT statute (“Client does not want to be court-ordered and will more than likely follow through with treatment to avoid the courts”) or legal strategies in which O&E staff advocated for diversion (“Because of the possibility of jail time hanging over his head, I think he will readily engage”). With others, they expressed concern that they were enrolling only because of the court order or other pressures and were unlikely to engage in treatment. Clients’ reasons for enrolling, but not necessarily engaging, include avoiding hospitalization (“Team may have a difficult time engaging client due to her lack of insight. Client more than likely agreed to be linked because she saw it as her ticket out of the hospital”), obtaining housing (“Moderately confident; client wanted housing, not treatment”), pressure from family, which may have been leveraged in O&E (“Client is engaging in treatment only because parents are forcing him to. So it’s unclear how effective treatment will be”), or compulsion by O&E staff and the court (“Client simply refuses to talk with mental health. Client was finally 5150ed, and taken to court, where the warm hand off occurred. Client signed releases in court, but it’s unclear if he actually understood what he was signing”). O&E expressed optimism that one client would engage with the treatment team, along with doubts that treatment could be successful: “Very confident he will engage. Successful treatment is another story.” These assessments emphasize that the barriers to engagement O&E staff faced during outreach will continue to be faced by treatment providers after clients enroll.

When O&E staff observed connections forming between client and provider at the handoff, or cultural concordance between the client and provider, these factors were noted as promising: “AOT requested a FSP provider that has the same cultural background as client, and speaks the same language with client’s mother, who provides support and collateral information,” and “Client seem to respond well to the therapist. The client and therapist had a lot of things in common.”

Pairing the client with a provider whose skills and experience were well-matched to seriously ill, treatment-disengaged clients also was important. O&E staff developed extensive experience with this population, and sometimes had concerns about whether the assigned providers had the same understanding: “The therapist kept referring to doing ‘therapy’ with the client, which is not a smart thing to do with treatment resistant individuals because of the stigma associated with the word.”

In many cases, however, O&E staff noted positive provider characteristics that made them feel optimistic that the client would engage. Clinical experience and skill in working with higher-risk individuals with SMI were essential: “AOT O&E team matched a clinician who has more clinical experience and able to set firm boundaries and limits with client but at the same time able to build rapport and engage client in treatment,” and “Treatment [provider] seemed able to handle the client’s angry outburst and redirect the client behaviors.” Provider openness to client perspectives also was seen by staff as useful in engaging clients: “During the warm-hand off it seemed like the provider was really good at letting the client express his true feeling about mandated mental health treatment.” After extensive efforts by the O&E teams to encourage clients to agree to treatment or respond to a court mandate, the warm handoff and treatment enrollment represent the beginning of a second, also potentially difficult, period of engagement, this time by the assigned treatment provider.

Discussion

Provision of care to individuals with SMI in the least restrictive setting possible requires a difficult balance between fulfilling an imperative to care for those most vulnerable in our society while respecting their autonomy. California’s AOT model attempts to accommodate both concerns by maintaining the possibility of a court order but allowing individuals the opportunity to accept services voluntarily in response to outreach. This study shows how intensive this outreach is in LACDMH’s AOT program, with staff delivering an average of 10.6 outreach service contacts to clients and their families, across multiple locations and using strategies that span the hierarchy of coercive outreach pressures and engagement approaches. Although just under half of clients enrolled in AOT treatment services, this is comparable with a non-AOT outreach program in San Diego, California that targeted individuals who are unengaged but require mental health services, 22.3% of whom met California’s AOT criteria (San Diego Behavioral Health Services, 2014). In some cases, LACDMH’s AOT clients did not enroll in AOT services because they needed a different level of treatment, such as conservatorship and inpatient treatment. In other cases, particularly if they were homeless, they were able to avoid outreach by disappearing.

By definition, the AOT-referred population is extremely challenging to engage. Many barriers to engagement were reported and often rated as severe. Similar to other studies with individuals with SMI (Kessler, Chiu, Demler, & Walters, 2005), lack of insight was the most frequently reported barrier, followed by other clinical symptoms. It is important to note that the O&E team reported trauma from prior psychiatric hospitalizations for over a third of clients, and distrust of mental health providers and resistance to medication for similar numbers of clients. They also noted that some clients resisted treatment because they equated mental health treatment with medication, which may reflect on past experiences with services that were narrow in scope rather than supportive and holistic. These observations underline the need to consider the potential costs of poorly delivered services and the use of restraints in psychiatric hospitalizations. When restraints are used on individuals with prior trauma history, they can be even more deeply traumatizing (Priebe, Bröker, & Gunkel, 1998). The use of coercion should always be approached thoughtfully by clinicians and policymakers, not only out of concern for civil liberties but also as a factor that could impact clients’ future willingness to engage with the mental health system, with resultant costs and risks to the individuals and to society.

The AOT court order is expected to “exert its primary direct effect on the compliance behavior of the client through threat of force to be applied if the individual fails to comply with a regimen of outpatient treatment as mandated by the court” (Swanson et al., 1997). The results of the present study suggest that attentive, persistent, flexible outreach predominantly involving persuasion, leverage, or inducements can lead nearly half of those refusing treatment to enroll in services voluntarily. Stronger treatment pressures, including the possibility of involuntary treatment, appeared to be an important part of outreach for some clients—nearly a third were advised that a court order could be pursued if they did not voluntarily agree to treatment—and a small number needed to be compelled via court order. With some clients O&E staff explicitly mentioned the importance of building rapport through less coercive outreach strategies prior to mentioning or pursuing a court order. While the tailoring, blending, and balancing of more- and less-coercive outreach strategies is a common part of mental health outreach services, it is less commonly part of AOT programs because most are strictly involuntary and do not involve a period of outreach with the potential for voluntary enrollment. Purely involuntary AOT programs may represent a missed opportunity to engage a subset of eligible clients without resorting to a court order, as well as to develop more positive associations with the mental health system.

It is evident from their responses that LACDMH O&E staff consider voluntary outreach to be a critical part of the program, even when legal strategies are invoked; but also that they are willing to jeopardize their relationship with a client by pursuing a legal strategy to persuade the client to enroll. This willingness to use strategies that might be unpopular with the client is facilitated by the design of LACDMH’s AOT program, with dedicated O&E teams that are separate from the providers that treat clients once they enroll in services and are ultimately responsible for keeping them engaged.

According to O&E staff ratings, a strategy that works well with one client will not work for everyone. Strategies rated as effective with clients represent every level on the hierarchy of persuasion (Szmukler & Appelbaum, 2008). Future analyses will examine whether particular strategies were more frequently reported or rated as effective at engaging particular subpopulations.

Outreach setting also may be important. Outreach was reported to have been delivered to almost a quarter of participants while hospitalized and 10% while in jail. This suggests that outreach in restrictive settings, such as those found in discharge-transition model AOT programs, was used with a substantial but non-majority proportion of clients. Clients’ experiences with incarceration and hospitalization—both of which are among the eligibility criteria for AOT—also highlight the complexity of the debate about coercion in outpatient mental health treatment for individuals who, in the absence of care, may have potentially more coercive experiences with the criminal justice system or involuntary hospitalization. Future analyses will examine the relationship between staff reports of whether a client received outreach in restrictive settings, strategies used, and perceived effectiveness. Mixed-methods analyses incorporating ethnographic data will further explore AOT clients’ experiences with law enforcement, courts, jail, and hospitals.

Because outreach is by necessity flexible and targeted to the individual, our data and analyses are observational and descriptive, and the strategies O&E staff use reflect what they believe is necessary and feasible for a given client. For example, more coercive strategies may be reserved for the most challenging clients, who also may have less successful outcomes. Nonetheless, the data can provide some insight into for whom different strategies may have been seen as necessary and as either effective or ineffective. For example, we see indications that stabilizing factors such as residence with family may affect O&E staff’s choice of strategies and perceptions of successful engagement. Future analyses will look at outreach and enrollment among key subpopulations including those who are homeless, those who reside with families, and those who have criminal justice involvement. A forthcoming manuscript will incorporate ethnographic and interview data with data from this survey to better understand the role of families in outreach and treatment for AOT-eligible individuals.

Limitations.

The strategies described here reflect the responses of two teams of outreach and engagement providers delivering services to a small proportion of individuals with SMI in Los Angeles County. Reports and ratings are from the perspective of O&E staff, who are only able to identify and rate potential barriers that they become aware of in their interactions with clients, family members, and referring parties. O&E staff also may not have identified every strategy used with each client; nonetheless, given the high frequencies with which many strategies were identified, and the requirement that they rate the effectiveness of every strategy selected, it is likely that they were attentive to their ratings. Free-text responses similarly show thoughtful consideration of each client. Nonetheless, these are solely from the perspectives of the O&E team. In future analyses we will examine field notes from ethnographic observations of outreach services, to provide another perspective on how they operate and their effectiveness. Additionally, we will combine data from this survey with administrative and service utilization data to examine the relationship between client characteristics, barriers, outreach strategies, and whether clients engage successfully in treatment after enrollment, as measured by enrollment duration as well as service volume while enrolled.

Conclusions.

Over the past several decades the mental health field has oscillated about the necessity and value of coercion in mental health services. The enactment of AOT statutes and implementation of AOT programs represents a shift toward the use of compulsory treatment to address concerns about treatment-disengaged individuals with SMI. California’s AOT model differs from those in many other states through its use of a community-entry model as well as its requirement that clients be provided with outreach services and given an opportunity to enroll voluntarily.

The experiences of Los Angeles County’s AOT O&E staff suggest that persistent and creative outreach and engagement, coupled with the strategic use of potential court involvement, can lead to voluntary enrollment of treatment-refusing clients who have many, often severe, barriers to engaging in outpatient treatment. For policymakers and public mental health administrators considering the implementation of an involuntary outpatient program, this argues in favor of including a voluntary, outreach-based component to such programs. For AOT programs and for outreach and engagement programs more generally, this study also suggests the importance of trainings and system-level supports to enable creative, flexible, and individually tailored engagement strategies.

Future research is needed to develop such trainings, evaluate the effects of specific engagement strategies on short- and long-term mental health outcomes and service use, and identify sub-groups referred to AOT programs that would benefit most from specific engagement strategies. Understanding when and how to apply varying levels of persuasive and coercive pressures is key to striking an appropriate balance between respecting individuals’ autonomy and fulfilling the need for care. By studying the use of a broad range of outreach strategies in a program that holds the potential for court-ordered treatment, the present study contributes incrementally toward a better understanding of these factors and highlights areas for additional research, some of which is planned in future analyses of the quantitative and ethnographic data from our LACDMH AOT evaluation.