Outpatient commitment: When it improves patient outcomes 1 year after Virginia Tech, many states’ statutes remain ambiguous and ineff ective n April 16, 2007, Seung Hui Cho shot® andDowden killed 32 Health Media students and faculty on the Virginia Tech campus Oand wounded 25 others before killing himself. A judge had declared Cho Copyrightmentally ill inFor 2005 andpersonal placed use only him on involuntary outpatient commitment (OPC). Cho apparently never sought treatment, and no one made sure that he did (Box 1, page 26).1 Much second-guessing has occurred about whether enforcing Cho’s OPC could have prevented the Vir- ginia Tech tragedy. Most states authorize OPC, but few PSYCHIATRY make much use of OPC statues that require patients to adhere to prescribed treatment in the community. Vir- CURRENT FOR ginia was typical; an OPC statute was on the books but rarely enforced. PARKER This article discusses the evidence on OPC laws’ ef- CURTIS fectiveness and offers recommendations on how to use these tools in psychiatric practice. Marvin S. Swartz, MD Professor and head Division of social and community psychiatry Jeff rey W. Swanson, PhD Mandating needed treatment Professor OPC—also called “assisted outpatient treatment” or “mandated outpatient treatment”—is a civil court Department of psychiatry and behavioral sciences procedure whereby a judge can order a noncompli- Duke University Medical Center ant mentally ill patient to adhere to needed treat- Durham, NC ment. OPC statutes exist in 42 states and the District of Columbia, although judges use these powers erratically.2,3 Most states have set identical thresholds for inpa- tient and outpatient commitment, such as when the patient is considered dangerous to self or others or (in Current Psychiatry some statutes) so gravely impaired that he is unable to Vol. 7, No. 4 25 For mass reproduction, content licensing and permissions contact Dowden Health Media. 025_CPSY0408 025 3/14/08 3:43:35 PM Box 1 Virginia’s OPC statute: Inpatient criteria for outpatient cases irginia’s commitment laws are in review Clinicians are uncomfortable using Vbecause of the Virginia Tech shootings inpatient criteria for outpatient commitment. in April 2007. The state’s existing OPC Labeling a patient as “ill enough” to be provisions are embedded in involuntary confi ned and then recommending that he or Outpatient commitment law, which is mainly directed she be released to outpatient treatment feels commitment toward inpatients. As a result, Virginia’s law: like a liability risk. • merely permits mandated outpatient Because the legislative intent in most treatment state statutes was to set criteria and • duplicates inpatient criteria for procedures for inpatient commitment, steps “imminent dangerousness” (although for implementing outpatient commitment are the state legislature is considering often ill-defi ned. An outpatient commitment relaxing this criterion) process requires: • provides no guidance on enforcement in • notifi cation to the responsible outpatient the event of treatment nonadherence local mental health authority, clinicians, Clinical Point • provides no administrative infrastructure and local courts that the order is in place Labeling a patient to make the law work. • expectations regarding the order Virginia’s statute is typical. It lacks • steps required to renew the order, if as ‘ill enough’ to a separate threshold for outpatient indicated. be confi ned, then commitment, using the same high threshold States such as New York with fully recommending his of imminent threat to self or others that is operational outpatient commitment statutes required for inpatient commitment. have clear implementation processes. release to outpatient treatment feels like safely care for himself in the community. Enforcement. Courts typically can request a liability risk These high thresholds have dramatically that law offi cers transport patients who fail reduced inpatient commitment eligibility to comply with OPC to a treatment facil- and yet may fl ag the patient as too danger- ity. There, patients will be encouraged to ous for outpatient commitment. comply with treatment or evaluated for in- OPC orders usually cannot force medi- patient commitment.2 This relatively weak cation. Periods of initial and subsequent enforcement authority has led some to ar- commitment vary across states but not dra- gue that OPC has no teeth. matically. In North Carolina, for example, Without clearly defi ned steps for im- initial OPC may be ≤90 days, after which plementation, an outpatient commitment a hearing must be held to renew the order order can be likened to a message in a bot- for ≤180 days. Depending on individual tle—a cry for help at risk for nondelivery. states’ statutes, OPC can be used as: In the Virginia Tech case, the judge issued • an alternative to hospitalization for pa- an outpatient commitment order for Cho, tients who meet inpatient commitment but how the local clinic understood its re- criteria sponsibilities and what resources and en- • a form of conditional release for pa- forcement power it had were unclear. tients completing an involuntary inpa- tient commitment • an alternative to hospitalization for Noncontrolled studies noncompliant patients at risk for re- Evidence from noncontrolled OPC studies lapse and involuntary inpatient com- is diffi cult to interpret because of: mitment.2 • lack of comparable committed and Few states have lowered the threshold noncommitted groups to the last variant, allowing OPC use to • diffi culty in comparing treatment avert relapse and hospitalization. Newer across comparison groups statutes in New York, North Carolina, and • selection effects, whereby clinicians elsewhere have incorporated these preven- and courts select patients for a pre- Current Psychiatry 26 April 2008 tative outpatient commitment criteria.4,5 dicted good outcome.2 026_CPSY0408 026 3/13/08 4:16:52 PM Most noncontrolled studies have con- Patients ordered to OPC had fewer hos- cluded that OPC improves treatment out- pital readmissions and spent fewer days comes and decreases hospital readmission in the hospital only if they received OPC rates and lengths of stay under some cir- plus consistent community services for ≥6 cumstances.6-12 The largest study reported months.14 Patients who received this model on New York’s initial 5 years’ experience of care were: with more than 3,000 patients under its • less likely to be homeless,15 criminally OPC statute, known as “Kendra’s Law“ victimized,16 arrested if they had past (Box 2, page 28).12 Under this law—the arrests,17 or violent18 most intensively implemented OPC stat- • more likely than the control group ute in the United States—the court’s order to comply with recommended specifi es a detailed plan of medications treatment.19 and psychosocial treatment. Patients received no benefi t from OPC Most of New York’s OPC recipients <6 months—even if combined with consis- stayed in assisted outpatient treatment tent, frequent mental health services—or longer than the court-mandated 6 months OPC of any length without consistent, fre- (average 16 months). The incidence of hos- quent mental health services. Clinical Point pitalizations, homelessness, arrests, and Study limitations. Length of time on Hospitalizations, incarcerations was far lower while patients OPC could not be randomly assigned, homelessness, and participated in OPC, compared with the even though this was a key variable in previous 3 years of their lives (Table, page the intervention. If lower-risk subjects arrests were far 34). Medication adherence improved from had been selected for longer periods of lower while patients 34% before OPC to 69% after commitment, commitment, positive fi ndings could participated in OPC and engagement with treatment improved have been overstated. Legal criteria for than in the previous from 41% to 62%, respectively.12 renewing OPC also prevented us from selecting lower-risk subjects for longer 3 years of their lives exposure to court-ordered treatment. Confl icting controlled trials Higher-risk subjects appeared in prelim- Duke Mental Health Study. In the fi rst inary analyses to have received longer controlled study of OPC, the Duke Men- periods of commitment, but unknown tal Health Study (DMHS) enrolled 331 selection factors could have affected seriously mentally ill inpatients being dis- OPC duration. charged from involuntarily hospitaliza- Outpatient service intensity was not tion to court-ordered outpatient treatment controlled but varied according to clini- between 1993 and 1996. Patients with a cal need and other unknown factors. As a history of violent behavior in the previ- result, selectively providing services could ous year were placed in a nonrandomized have infl uenced outcomes, although other comparison group and remained on OPC analyses argue that this factor was not for at least 90 days. The remaining 264 pa- important.4 tients were randomly assigned to: • an experimental group that received New York. In 1994, the state legislature OPC for ≤90 days (could be renewed for established a 3-year pilot program to ≤180 days) plus consistent community men- evaluate OPC in New York City’s Bel- tal health services levue Hospital as a fi rst step toward con- • a control group that was released from sidering permanent OPC legislation.20 The OPC but received the same community randomized, controlled study compared a mental health services as the experimental court-ordered group (N=78) and a control group.13 group (N=64) during 1 year after hospital