Outpatient Commitment and Its Alternatives: Questions Yet to Be Answered

Not to be overlooked in current commentaries and debates about outpatient commitment and its alternatives (1,2), such as conditional release and guardianship, is the fact that important questions remain to be answered. We have recently had the opportunity to review the literature on the effectiveness of outpatient commitment and its alternatives (3), updating previous evidence reviews (4,5) and adding our assessments of more recent research. Here we highlight some of the unanswered questions and methodological limitations identified in our review and suggest a general strategy for undertaking further research.

Outpatient commitment

Outpatient commitment is a civil law mandate ordering an individual to obtain psychiatric treatment against his or her will or risk sanctions up to and including forced hospitalization. Although there is an extensive research literature on outpatient commitment (6–8), it rests largely on uncontrolled studies. Consequently, any reported reductions in hospital use may be due to selection or other factors.

The three major controlled studies about the effectiveness of outpatient commitment were all conducted in the United States. The Bellevue Pilot Study (9) was a randomized controlled trial of an early version of assisted outpatient treatment (see below) conducted at the psychiatric service of a public general hospital in New York City. The study compared individuals randomly assigned to court-ordered treatment with enhanced services with those who received only the enhanced services. After a 12-month follow-up, no differences were found between groups on a number of outcomes, including hospitalizations and arrests. However, the validity of these null findings was undermined due to a number of methodological limitations, including lack of enforcement of court orders, small sample size, and nonequivalent comparison groups.

The other two studies were conducted by the same research team in North Carolina in the late 1990s and in New York State a decade later. The North Carolina study employed a randomized experimental design (10). The study enrolled involuntarily hospitalized patients who were randomly released to outpatient commitment status from a state psychiatric hospital or from one of three general hospitals. Patients in the control group remained hospitalized and were released under usual procedures. Both groups were followed into the community for a 12-month period. The New York State study was a retrospective, quasi-experimental evaluation of the state’s assisted outpatient treatment (AOT) law (an outpatient commitment statute) that used administrative data from 1999–2007 (11). Most participants had been placed on AOT while hospitalized in a state or local hospital; others were placed on AOT directly from the community. Comparison groups were constructed with propensity weighting procedures. Study participants were limited to those with Medicaid benefits so that claims data could be used to track their community service use before and after placement on AOT. Medicaid claims were linked to other administrative databases to assess rehospitalization and arrests.

Our literature review indicated that there is moderately strong evidence from both of these studies about the effectiveness of outpatient commitment, both with regard to reducing admissions to psychiatric hospitals and to engaging recipients in community-based services (3). However, because of methodological limitations, only weak support was found for conclusions that outpatient commitment orders of longer duration are more effective (10,11), that outpatient commitment reduces criminal justice involvement (12), or that it saves states money (13). These limitations are attributable to inadequate comparison groups or inadequate statistical controls (3).

Further, outpatient commitment appears to work only in settings where there are ample and intensive community mental health services. Whether a court order without intensive treatment has any effect cannot be answered from current research, because both the North Carolina and the New York studies compared groups exposed to service systems that were in some way enhanced, including through the provision of case management that was otherwise not available. No study has yet demonstrated what might happen when outpatient commitment is implemented in areas with more modest mental health systems, lacking the level and type of services that were available in the North Carolina and New York studies. In addition, there is scant literature examining other important outcomes, such as clinical improvement, safety, quality of life, or even societal participation through labor markets or other mechanisms. Admittedly, no single study can address all outcomes. Most person-centered outcomes are seldom available through administrative data, and thus they are challenging to investigate. However, with use of advanced methodologies, there is room for more important work on outpatient commitment.

Conditional release

Conditional release is a form of outpatient commitment in that individuals who are experiencing an inpatient commitment can be released from the hospital because they no longer require intensive inpatient care and are deemed by their clinician or hospital executive as capable of functioning in the community with the aid of outpatient services. Most of the evidence for conditional release comes from studies conducted in Australia (14–18) and the United Kingdom (19). These studies reported positive findings regarding reduced psychiatric rehospitalization.

The main difference between outpatient commitment and conditional release is that under conditional release, there is no separate court order mandating the outpatient phase of treatment. In principle, many outpatient commitment statutes in the United States are broader than conditional release. They allow for outpatient commitment directly from the community as well as from an inpatient hospitalization, whereas, by definition, individuals first have to be hospitalized before they can be conditionally released. Direct community commitments are allowed under so-called preventive outpatient commitment statutes (8) for individuals who currently do not meet inpatient commitment standards but are mandated to receive outpatient commitment as a way of avoiding the deterioration that would result in a future hospitalization.

In practice, however, outpatient commitment seems to be used less for prevention and more as a way to shorten a current hospitalization. For example, the studies in North Carolina (10) and New York State (11) were conducted under preventive outpatient commitment statutes, but a large majority of participants (100% in North Carolina and about 80% in New York State) were placed on outpatient commitment or AOT, respectively, as a step down from inpatient hospitalization. In that sense, except for the presence of a court order mandating outpatient treatment, these studies were actually conditional release studies.

Does this mean that outpatient commitment and conditional release are equally effective? The only research-based answer to this question comes from the Oxford Community Treatment Order Evaluation Trial (OCTET), a head-to-head comparison of preventive outpatient commitment and conditional release conducted in Oxford, England (19). This randomized controlled trial compared the British versions of the two practices and found that conditional release was as effective as outpatient commitment in avoiding hospitalization with less legal compulsion. The implication is that conditional release is the more user friendly of the two practices, with no loss of effectiveness. However, the relevance of the OCTET study for the United States has been questioned partly on the grounds that it focused on a group of patients who would not meet preventive outpatient commitment criteria in the United States (20).

There have been no careful prospective studies of the nature and amount of supervision that occur under conditional release, so the whys and hows of the favorable results achieved in these studies remain to be identified. On the practical side, there is also the issue of monitoring mechanisms and of how revocations would be enforced if the case mix of patients on conditional release was expanded to include individuals who otherwise would be candidates for outpatient commitment. New York State, for example, invested tens of millions of dollars annually to develop enrollment, service, and monitoring mechanisms to support its AOT program. Currently, most states do not have that type of infrastructure in place for either outpatient commitment or conditional release.

Guardianship

Guardianship and conservatorship enter discussions about outpatient commitment and its alternatives through considerations of decisional incapacity or incompetency as an alternative to dangerousness in civil commitment hearings and as a way to increase the likelihood of outpatient service utilization. These alternatives have been discussed in the United States (21–24) as well as the United Kingdom (25). Guardianship and conservatorship both represent a legal relationship between a protected person and one or more individuals appointed by the court to make decisions on his or her behalf. Guardianship invests the guardian with decision-making authority over all personal affairs (including treatment), whereas the decision-making authority of the conservator is limited to the management of property and financial affairs. Here we focus our comments on guardianship.

A handful of studies have examined the effectiveness of guardianship in reducing adverse outcomes among adults with mental illnesses in the United States (26,27) and the United Kingdom (28). Findings provide some evidence that guardianship may be effective in reducing hospitalization, although the results are mixed. In addition, there is some evidence that guardianship is associated with higher levels of psychosocial functioning compared with conditional release. However, many methodological issues limit the strength of these findings, including differences at baseline between intervention and control groups and small samples. Moreover, individuals who meet the grave-disability or incompetency criteria of guardianship likely differ on relevant clinical characteristics from those who meet criteria for involuntary outpatient commitment or for involuntary inpatient commitment, thereby potentially limiting the generalizability of findings to the target population for outpatient commitment.

Many issues that merit consideration may also arise in the application for and implementation of guardianship. The process of obtaining full guardianship can be quite slow and time consuming, involving different courts from those involved in civil commitments of persons with mental illness, and guardianship may therefore not be appropriate in urgent care situations. In addition, identification and appointment of an appropriate guardian may be challenging. Family members, for example, may be reluctant or unable to become guardians or may be unavailable. Guardians may be provided through public guardianship systems or social service agencies; however, these guardians or ombudsmen typically have very large caseloads and are overburdened. Moreover, although the court-appointed guardian is expected to act as an advocate for the protected person and to make decisions in his or her best interest, there is potential for abuse of the guardian role and exploitation of the protected person for personal gain. Finally, the appeal process is onerous, and case reviews are relatively infrequent. Consequently, the protected person may remain under guardianship long after competency has been restored.

Research strategy

Although the empirical literature is expanding, we have not yet reached a preponderance of evidence in favor of outpatient commitment or its alternatives, and many questions remain about their mechanisms and effectiveness. Much of the evidence for outpatient commitment rests on single-site studies or on research conducted in other national contexts, with attendant difficulties in extrapolating findings to the United States. The likely limits to generalizability are associated with differences in the availability and nature of community-based mental health services, local cultures, participant characteristics, social welfare entitlements, and policy contexts, among others. Moreover, the research as a whole typically neglects the consumer or even societal perspective; has largely focused on hospital and criminal justice outcomes, with notable exceptions; and may need to be rethought in the current U.S. context given expanded treatment options and greater system enhancements resulting from the Affordable Care Act (ACA).

A further question is also crucial: what is the best strategy to follow for future research in this area? We side with those who caution about the costs and limitations of additional randomized controlled trials. Instead, we believe that much greater payoff can be achieved with efficient, well-conducted observational studies that use administrative data to answer both clinical and policy questions. Advances in methodologies, such as instrumental variables and propensity weighting, can be used with administrative data to create constructed comparisons of service recipients and nonrecipients in ways that approximate experimental conditions by balancing groups on an array of observed characteristics. Precedents for the use of administrative data can be found in the New York AOT studies, the OCTET study in England, and the conditional release studies in Victoria, Australia. The key is to assemble a rich array of observable variables for large samples of individuals so that adequate statistical controls can be applied. Linkage to external data sets, such as Medicaid, employment records, and death certificates, can be used to assess outcomes beyond hospitalization and recidivism, such as participation in labor markets and mortality.

In principle, the ACA presents many new opportunities and resources to improve care in both community and medical settings. Its implementation in the United States reflects a new reality against which practices such as outpatient commitment, conditional release, and guardianship should be compared. In the post-ACA era, persons with serious mental illness will be more likely to be insured (29), possibly facilitating the receipt of services in the community, especially in the long run as service systems catch up to the initial increase in demand. The implication is that better treatment options should be available in the community and persons with serious mental illness will have greater opportunities for better health care outside hospital, jail, or prison walls (30). In practice, however, service expansion will depend on the prevalence of state Medicaid expansions, the adequacy of reimbursements for mental health care, and the availability of clinicians, because most U.S. counties face shortages of mental health professionals (31). Although the literature on outpatient commitment and alternative models is promising in redefining collaborations between the justice and mental health treatment sectors, the literature is nowhere near its apex. Thoughtful, efficient research in this area is still needed as newer health care opportunities unfold.