Implementation Status of Assisted Outpatient Treatment Programs: A National Survey

Statutes for involuntary community mental health treatment have been broadly adopted in the United States; in fact, 45 states have statutes authorizing assisted outpatient treatment (AOT). AOT refers to court-mandated treatment in community mental health settings for individuals who meet stringent criteria. Advocates for AOT consider it an important mechanism to improve the lives of individuals who lack insight into their need for treatment. Numerous studies have found that persons with mental illness benefit from AOT participation; these studies have defined the major criteria of effectiveness as reduction in rehospitalization and recidivism, reduction in violence and victimization, and improvement in quality of life.

Researchers have found that participants in AOT have lower odds of arrest (1,2); reduced risk of victimization (3,4); lower risk of harm to self or others (5,6); reductions in hospital admissions and recidivism, length of hospitalization (7–10), and emergency visits (7); improved engagement in services (9); and higher quality of life (11). However, nearly all reports were from one state only or from outside the United States; most failed to consider the patient’s perspective (12); and nearly all had methodological flaws, ranging from use of retrospective study designs (1,9,10) to absence of a control group (7), nonrandom assignment to AOT (3,6,9), nonrandom extension of AOT orders (8,11), and exclusion of persons with a history of violence (12). Attempts to produce definitive evidence of the effectiveness of AOT may be “a quixotic quest” (13), given how many questions remain about the use of AOT (4,14). In fact, mandated service programs remain highly controversial because of concerns that they require coercion to achieve improvement in treatment engagement and have the potential to divert resources from programs for clients seeking voluntary mental health services (5,11).

Engagement of the “revolving door” individual, someone who has repeatedly avoided long-term treatment despite recurring crises, is a major objective of the AOT strategy. Thus most AOT programs and, consequently, many of the studies evaluating AOT have focused on patients who were referred as they transitioned into the community from inpatient mental health facilities or, in a few cases, prisons or jails. Yet many families and advocates also see AOT as filling a need for a treatment “gateway” for members of the community who may be severely ill but who are not engaged in treatment and do not meet the criteria for hospitalization. A secondary objective that often drives the passage of AOT legislation is the hope that treatment will forestall violence by individuals with mental illness; yet such violence is rare and notoriously difficult to predict (15). Thus the literature appears to indicate that AOT is best defined as a service, involving the collaboration of courts and treatment providers, to recruit individuals with mental illness who may be at risk of grave disability, of deterioration, of self-harm, or of violence to others and engage them in outpatient treatment (16). The literature further suggests that AOT is considered appropriate both for individuals who are leaving institutions and for those with untreated mental illness who are living in the community.

Despite growing interest in AOT and the existence of AOT statutes in 45 states, little is known about the ways in which individual states have utilized this mechanism and the ways in which utilization varies from state to state. The most extensive evaluation of the use of AOT to date was a qualitative review carried out by researchers at RAND Health in 2001 (16). The researchers reviewed all the available AOT statutes and conducted detailed informant interviews in eight states. Their major finding was that AOT had had limited and inconsistent implementation. Many states used AOT primarily in hospital discharge planning, and informants from the selected eight states reported a lack of adequate treatment resources, lack of “teeth” for enforcement, and inadequate monitoring and data tracking capability. The conclusion of the RAND researchers was that “outpatient commitment is neither as effective a solution to the problem of compliance as its advocates claim nor, in its practical application, as repressive a law as consumer/survivors fear” (16). The report and other recent research suggest that many states have AOT statutes that are unused or used sporadically (16,17). For an accurate assessment of AOT programs that can guide future policy, it is important to examine how these programs are currently implemented, how patients are enrolled in AOT and through what pathways, and the types of common problems that arise in practice.

The purpose of this study was to describe the actual operation of AOT programs in practice nationally in 2014 rather than in terms of models discussed in the literature (18) or of statutory variance. Our approach used a qualitative survey design and an analysis that involved four steps. First, we reviewed and compared the statutory criteria and procedures for AOT in 45 states to inform our development of interview items. Second, we collected interview data to determine the extent to which AOT had been implemented in each state and the availability of data on implementation. Third, we developed a simple set of enrollment pathway descriptors that reflected our findings on how AOT had been implemented in practice. Finally, we compiled strengths and limitations of the programs most often identified by our informants.

Methods

Our methodology began with compilation and review of state statutes from 45 states, a process that served as the basis for our discussions with AOT program staff in each state. We then conducted phone interviews with informants in each state, using a semistructured script, and we supplemented the results with online research about the programs. Interviewers either audio-recorded their phone interviews for transcription by research staff or took handwritten notes at the time of the call, which they wrote up as a formal document immediately thereafter. We began by contacting the state mental or behavioral health agency and explained that our objective was to identify and talk to the person or persons with the most thorough knowledge of the AOT program, who could tell us how implementation occurred in practice. Our basic criteria for identifying an informant, in addition to referral from the responsible state agency, were length of experience with AOT and self-identified knowledgeability about the program.

Finding an informed contact took at least two calls and, in some cases, as many as eight; in some states, two or three people served as informants. In most cases, the informant was an administrator (N=25) or provider (N=7) in the state mental health services department or the state psychiatric hospital; other informants were judges or legal coordinators (N=11), staff members of the managed care plan that administered public mental health services (N=3), or officers of a mental health advocacy group (N=3). In several states, where implementation was limited to one or more counties, interviewers spoke to informants at the county level. As noted below, in several states, a good contact could not be identified. The majority of the research was completed from December 2013 to May 2014, with some follow-up in December 2014 and January 2015. The research was determined to be exempt from University of California, Los Angeles, institutional review board review and was conducted responsibly and ethically.

Results

We were able to contact informants and obtain information on AOT or comparable programs in 37 states. Of the remaining 13 states, five (Connecticut, Maryland, Massachusetts, New Mexico, and Tennessee) had no current AOT legislation. Our attempts to find an informed contact willing to discuss the statutory program were unsuccessful in eight states (Alabama, Alaska, Colorado, Georgia, Kansas, Minnesota, Virginia, and West Virginia). In 14 of the 37 states we were able to contact, implementation of the statutory program appeared to be limited or was not well documented. In three of the states with limited implementation (Missouri, South Dakota, and Wyoming), our informants stated that they preferred other options to the use of AOT; in the other cases, the reasons for limited implementation or documentation were not specifically identified. In three of the 37 states, the statutes were relatively new, and experience with the program was therefore minimal.

Twenty states had AOT programs that were active, operational, and documented to some extent. Table 1 summarizes statutory criteria for AOTs in the 20 states with the most active programs. Details about the active and relatively inactive state programs are summarized in Table 2 and Table 3, respectively.

Operational Pathways

Although the statutory criteria were quite similar across the 20 states (Table 1), the numbers of patients placed under AOT orders and the working approaches to implementation varied considerably. According to our informants, clients were most often referred to AOT via one of three pathways (of the available pathways described in the statutes). We developed the following set of descriptors for these pathways: a community gateway pathway, in which there was an emphasis on identifying unengaged or noncompliant individuals in the community with serious unmet mental health needs and getting them into treatment (eight states); a hospital transition pathway (sometimes combined with jail transition), in which patients were ordered into outpatient treatment after discharge from an inpatient commitment (ten states); and a surveillance, or safety net, pathway, in which the main emphasis appeared to be on monitoring, keeping tabs on, and ensuring some treatment for persons considered a danger to others (seven states). As Table 2 shows, these pathways often overlapped within a single state, and it would be inaccurate to say that any state used one pathway exclusively; rather, these descriptors characterize the different approaches taken by agencies and providers to apply the concept of AOT to their particular needs.

Variations, Similarities and Problem Areas

Active state programs had many statutory similarities. All were found to be quite similar in the formal statutory criteria; in the referral, evaluation, and hearing procedures for AOT; and in the initial duration of outpatient commitment (usually 90 to 180 days). All but five states explicitly included inability to care for oneself and the consequent risk to one’s welfare as criteria. All states cited danger to self or others as a criterion; in most states, the danger was explicitly defined as a “real and present threat” or a “substantial risk of immediate harm.” In a few states, however, such as California, the criterion focused on the prevention of harm to individuals or by individuals at risk of becoming dangerous at some future time. Differences in the statutes’ definition of danger to self or others are described in Table 1.

All states had statutory protections for patient rights. The individual had the right to an attorney in all programs (provided by the state if an individual cannot afford one), as well as other clearly defined rights, allowing him or her the opportunity to participate in the process to the extent that he or she is able to do so. In states with a statute allowing involuntary medication, the procedure nearly always required the consent of a guardian (if a patient lacks capacity to consent) or a court hearing or other formal procedure to determine necessity.

Variations among programs in states with active programs fell into four primary categories, as reported by our informants. First, although each state statute listed several criteria for referral of patients to AOT, the criterion used most often by particular states varied. Some states used the criterion of anticipated dangerousness of the individual, and others used an individual’s disability and deterioration. A few states made frequent use of the criterion of lack of compliance with treatment. Second, the agency responsible for overseeing AOT was not always the state mental health department or division but could be a regional or local county agency or a contract agency. Third, in some states, the AOT order was formalized in a written treatment plan or as a set of defined services approved by the court and included in the court order, whereas in others, AOT treatment services appeared to be left to the discretion of the treatment provider. Fourth, states varied in the effectiveness of their participant monitoring and data tracking.

There were several major problem areas in AOT implementation, as defined by informants from states with active programs. First and foremost, informants from ten states cited inadequate resources. Although the courts in almost all states were required to consider availability of services and to issue treatment orders on the basis of advice from approved treatment teams, several respondents noted that the AOT order committed the providers, as well as the participants, to the designated treatment plan; and respondents expressed concerns that voluntary mental health clients might face cutbacks as a result. Our informants stated that they were able to maintain adequate service levels, but that limited resources were an ongoing problem. Only four state legislatures had authorized any designated funding for AOT.

The second set of major problems, identified by ten informants, was lack of interagency collaboration; the informants described resistance among local providers to accepting AOT patients; resistance by local courts to accepting AOT cases because of overloaded dockets or a lack of understanding; inadequate monitoring of participants; or a basic lack of communication and coordination.

Third, although three informants attributed the success of AOT to the “black robe” effect, stating that participants were more strongly motivated to comply with the authority of a court, four others thought that their program lacked any real power of enforcement. If a participant was noncompliant, the court could stay the AOT order; the participant could be put under a short-term hold and reexamined, and the case would be reviewed again. Staying or revoking an AOT order in itself was often a problem—informants in three states described difficulties in finding an individual or in determining which law enforcement agency was required to find and transport him or her for reexamination. What options would a court then have to enforce compliance? The individual usually had not committed a criminal offense, nor did he or she necessarily meet the criteria for inpatient commitment. Furthermore, most states had inadequate psychiatric bed capacity and preferred that patients participate in less expensive outpatient treatment, if possible. Involuntary medication without hospitalization was generally not a legal option. The usual action taken by the court in cases of noncompliance was to revise or extend the original order, placing the responsibility back on the treatment provider to engage the participant.

Finally, nearly all the informants from active programs judged their state’s AOT implementation as successful and reported that many patients complied appropriately with the ordered treatment. However, slightly less than half of these states had strong data tracking (9). Although the programs were often able to document reductions in hospital census and admissions, this decrease was usually systemwide and AOT accounted for only a part of it. Several informants believed that there had been a decrease in violence involving persons with mental illness, but very little hard data were available to support this statement.

Discussion

States with active AOT programs used the program to maintain engagement of both discharged inpatients and treatment-recalcitrant individuals. Nevertheless, they faced several logistical and resource difficulties in implementation and were often unable to offer good documentation of program effectiveness. Moreover, the community gateway pathway, which is often discussed by advocates hoping to improve engagement of individuals with mental illness who are not currently in treatment (19), was used in a small number of states for relatively few participants, and its implementation has not been well studied. The majority of participants appeared to enter AOT following hospital discharge.

Our data suggested some general observations. First, AOT programs did not operate in a vacuum. An effective AOT program requires a well-developed mental health system with adequate funding—not only to ensure coverage for the actual services but also to support staff training; provide resources to meet additional needs, such as housing and treatment of co-occurring disorders; ensure inpatient bed capacity as backup when AOT doesn’t work; and facilitate active collaboration among mental health agencies, the courts, and law enforcement personnel on the basis of mutual agreement on eligibility criteria, compliance goals, and monitoring. In the states that had the most active programs, our informants reported recognition of these needs and a commitment by mental health leaders to work toward the most effective program within existing constraints. However, their reports also made clear that an AOT program is only as effective as the system in which it is based.

Second, we found that AOT was implemented across the country less widely than might be inferred from the prevalence of statutes and extent of the literature (20). Where it has been implemented, its effectiveness in engaging participants in long-term treatment was not always well documented. AOT laws are based on the concept that if participants are compelled to accept services for a defined period, they will recognize the benefits and continue treatment. Several informants from states with long experience reported that although many individuals transitioned successfully after receiving mandated services, others stayed with treatment only as long as the court order was in effect and then returned to their former “revolving door” habits. For the former group of patients, the real issue is how best to establish effective engagement and reeducation through access to effective services. For the latter group of patients, existing AOT programs appeared to provide only temporary benefits. Most of the extant research has focused on individuals making hospital and jail transitions, and there has been little research on programs using community gateway pathways. Future research may evaluate the effectiveness of community gateway programs; however, these efforts will be limited until states track data consistently, using a common set of outcomes.

Finally, the stringency of the statutory framework may not allow for the flexibility and insight necessary to reduce the incidence of violence by, or against, persons with mental illness or the likelihood of deterioration. We found that AOT statutes were usually written in an attempt to balance patient rights and welfare with public safety; assessment criteria and enrollment procedures were defined at length. No state statute allowed for patient commitment without the application of stringent criteria. Given that the intent of AOT is to create a legal mandate for care, courts are required to meet state evidentiary standards before issuing an AOT order. This stringency may militate against the ability of a program to assess and deal with individuals living in the community who have mental health issues but who are not seeking care and whose behavior has not yet suggested a serious threat or risk to the individual or to others.

There were several limitations to this study. In each state, the research team used its best efforts to establish contact with the person or persons who were most knowledgeable about AOT and its current operation. It is possible that we missed speaking to some informants who may have had the most current awareness of how AOT was used within each state or who could contribute additional insights or that we missed identifying some programs active at the county level. However, all our contacts were identified by members of the state behavioral health agency as key personnel knowledgeable about AOT, and all were willing to participate. Limited data availability within several states also affected our ability to draw conclusions. Many state agencies faced resource and practical limitations in tracking outcomes for individuals in AOT programs. Finally, this was a cross-sectional survey during a defined time period, and several respondents reported that the AOT program in their state or county was in a state of change.

Conclusions

Nationally, the implementation and use of AOT programs are growing. However, there was little evidence about the effectiveness of programs that recruit participants directly from the community. Furthermore, many states have implemented AOT on a limited scale and have made minimal evaluation efforts. Evaluations of AOT performance must take into account differences in state statutes, such as whether criteria for inpatient or outpatient commitment are similar or distinct. Variations in implementation and inadequate resources have limited the development of a strong evidence base. Nevertheless, the use of AOT has important implications for the well-being of individuals with serious mental health issues and for policies guiding the provision of treatment resources. Therefore, it is critical to develop and examine rigorous evidence on the operation and the impact of AOT programs in practice.