Commitment of Persons with Psychiatric Disorders

Abstract

Commitment of persons with psychiatric disorders is a form of forensic psychiatry: the involuntary legal detention of a person with a mental disorder for evaluation or treatment. Two justifications underwrite it — the state's parens patriae duty to care for those who cannot care for themselves, and its police power to protect the public from harm — and modern statutes fuse them into a dangerousness standard. The field's central difficulty is predictive: forecasting which patients will act violently is unreliable at the low base rates involved, so most people detained as dangerous would not, in fact, have harmed anyone. This article examines the legal standards, the arithmetic of risk prediction, the mixed evidence on whether outpatient commitment reduces rehospitalization and violence, and the enduring civil-liberties critique.

Keywords: involuntary commitment, dangerousness standard, outpatient commitment

What Commitment Is

Commitment of persons with psychiatric disorders — often called civil or involuntary commitment — is the legal mechanism by which the state authorizes the detention and, in most jurisdictions, the treatment of a person with a mental disorder without that person's consent. It is a form of forensic psychiatry, the branch where psychiatric judgment meets legal authority, and it is among the few circumstances in which a person who has committed no crime may nonetheless be deprived of liberty. MeSH files it under forensic psychiatry precisely because the act is at once clinical and juridical: a physician's assessment supplies the finding, but a legal standard and, usually, a court supply the power (#ref-testa-2010).

The commitment decision turns on three elements that statutes across jurisdictions share in some form: the person must have a mental disorder, that disorder must produce a specified consequence — dangerousness to self or others, or an inability to meet basic needs — and less restrictive alternatives must be inadequate. What varies, and what the rest of this article turns on, is how each element is defined and, above all, how the second is proven. The move from a purely clinical need for treatment standard to a dangerousness standard, cemented in United States law through the 1970s, narrowed commitment to those posing a demonstrable risk, and in doing so made the accuracy of risk prediction the hinge on which the whole apparatus rests.

The Two Legal Justifications

Commitment draws its authority from two distinct state powers, and keeping them apart clarifies almost every dispute about it. Under parens patriae — literally the state as parent — government acts benevolently on behalf of a person who, by reason of mental disorder, cannot make a competent decision to accept care. Under the police power, government acts to protect others from the harm a person might do. The first justifies detaining someone gravely disabled and unable to feed or shelter themselves even if they threaten no one; the second justifies detaining someone dangerous even if they would refuse help they are competent to refuse.

Through the twentieth century the balance shifted decisively toward the police-power rationale expressed as a dangerousness standard. A person could be committed not merely because treatment might help — the older, broad need for treatment test — but because they posed a substantial risk of harm. Courts required that the risk be shown by clear evidence rather than clinical intuition, and legislatures wrote the dangerousness requirement into statute. The reform was a civil-liberties victory: it forbade indefinite detention on a bare diagnosis. But it also loaded onto psychiatry a task it is poorly equipped for — the prediction of future violence — and made a genuinely hard statistical problem the legal centre of gravity.

Commitment rests on two distinct state powers. Parens patriae lets the state care for a person who cannot competently care for themselves; police power lets it protect others from harm. Classify each case by the power that best justifies detention.

The two powers can overlap in a single case, but they answer different questions: is the person unable to decide for themselves, or a danger to others? Dangerousness reforms shifted most modern commitments toward the second, protective rationale.

Inpatient and Outpatient Forms

Commitment is not a single act but a family of them, differing in setting and duration. Emergency detention holds a person for a short, statutorily fixed period — often 72 hours — on the certification of a physician or, in some jurisdictions, a police officer, to permit evaluation. Inpatient civil commitment follows, after a hearing, when continued involuntary hospitalization is judged necessary, and it must be periodically reviewed. The most contested modern form is involuntary outpatient commitment — known in England and much of the Commonwealth as the community treatment order — under which a person lives in the community but is legally obliged to accept treatment, typically medication and clinical contact, on pain of recall to hospital.

Outpatient commitment was designed for the revolving-door patient: someone who stabilizes in hospital, is discharged, stops treatment, relapses, and returns — repeatedly. Its promoters argued it could break that cycle at less cost to liberty than repeated inpatient detention; its critics answered that it extends coercion into the home and community, medicalizing supervision without clear benefit. Because outpatient commitment can be delivered under a written protocol and, in some places, allocated experimentally, it has been studied more rigorously than any other form — and the evidence it produced, examined below, is the field's most instructive and most disputed.

Table 1

Principal Forms of Commitment, by Setting, Duration, and Decision-Maker

Note. The forms differ in setting and coerciveness but share the same triple test — disorder, a specified consequence, and the inadequacy of a less restrictive alternative. Outpatient commitment and the community treatment order are two names for the same instrument in different jurisdictions; the randomized evidence on it is examined below (Burns et al., 2013).

The Prediction Problem

If commitment rests on dangerousness, everything depends on how well dangerousness can be foreseen — and the honest answer is: poorly. John Monahan's 1984 analysis marked the turning point, conceding that first-generation clinical predictions of violence were wrong far more often than right and calling for a second generation of actuarial, evidence-based risk assessment to replace unaided clinical judgment (#ref-monahan-1984). The empirical studies that followed confirmed the difficulty. Charles Lidz and colleagues, testing clinicians' predictions against later violence, found that while clinical judgment performed above chance, its accuracy was modest and heavily skewed by false positives — patients predicted to be violent who were not (#ref-lidz-1993).

The deeper obstacle is arithmetic, not clinical skill. Serious violence is, over any given follow-up, a low-base-rate event, and predicting a rare event generates false positives faster than true ones no matter how good the instrument. Matthew Large and colleagues drove the point home for schizophrenia, showing that because the base rate of serious violence is low, even a risk-categorization scheme with respectable sensitivity and specificity yields a positive predictive value so poor that most people classified as high-risk will never be violent (#ref-large-2011). The consequence for commitment is stark and unavoidable: a dangerousness standard, applied honestly, detains many people who would have harmed no one, because the mathematics of low base rates guarantees it. The worked example below makes the size of that effect concrete.

Of 1,000 people, some fixed fraction will truly become violent (the base rate). A risk procedure with a given sensitivity and specificity flags some as dangerous. Watch how few of those flagged would truly have been violent.

At the default 5% base rate with 50% sensitivity and 80% specificity, only about 12% of those flagged as dangerous would truly have been violent. Raising the base rate improves the predictive value sharply; lowering it drowns the true cases in false positives — the reason a dangerousness standard, applied honestly, detains many who pose no threat.

Does Outpatient Commitment Work?

Outpatient commitment was tested where commitment law rarely is: in randomized and controlled studies. The Duke group ran the pivotal randomized trial in North Carolina. Marvin Swartz, Jeffrey Swanson and colleagues reported that outpatient commitment reduced hospital readmission — but only when it was sustained beyond six months and combined with frequent outpatient services, not from the court order alone (#ref-swartz-1999). The same cohort showed a reduction in violent behavior among those under sustained, service-rich outpatient commitment (#ref-swanson-2000). Paul Appelbaum's assessment of New York's Kendra's Law, five years on, likewise found improved service engagement and reduced hospitalization, while cautioning that the program bundled the legal order with a large infusion of case-management resources, making the order's independent contribution hard to isolate (#ref-appelbaum-2005).

Then came the trials that complicated the story. The English OCTET randomized trial, led by Tom Burns, compared community treatment orders against brief leave and found no difference in the rate of readmission — the orders did not deliver the hospital-reduction benefit their statutory rationale promised (#ref-burns-2013). Systematic syntheses have struggled to reconcile the evidence. Steve Kisely's updated meta-analysis of the randomized evidence found no clear advantage for outpatient commitment on health-service use, social functioning, or quality of life relative to voluntary or less coercive care (#ref-kisely-2014), and his Cochrane review concluded that it takes a large number of orders to prevent a single readmission (#ref-kisely-2017). Jorun Rugkåsa's review of the international evidence reached the same guarded verdict, noting the gulf between strong policy enthusiasm and weak trial support (#ref-rugkasa-2016). Phoebe Barnett and colleagues, pooling the controlled evidence, found compulsory community treatment did not reduce readmission or improve most outcomes (#ref-barnett-2018). The pattern is that the observational and non-randomized studies favor outpatient commitment while the randomized trials do not — the classic signature of confounding by the intensive services that accompany, but are not caused by, the order.

Each bar is one line of evidence on whether outpatient commitment cuts hospital readmission. Bars to the left mean fewer readmissions. Toggle to the randomized trials only and watch the apparent benefit shrink.

The observational studies (gold) suggest a large benefit; the randomized trials (navy) cluster near no effect. The gap is the fingerprint of confounding: in routine practice, orders go with intensive services, and it is the services — not the coercion — that most of the apparent benefit tracks. Illustrative values, drawn to reflect the direction of the cited findings rather than exact pooled estimates.

Worked Example

Suppose a clinician must decide, for a population being considered for commitment, who is dangerous. Let the true base rate of serious violence over the follow-up be 5%, and suppose the risk-assessment procedure has a sensitivity of 50% (it flags half of those who will actually be violent) and a specificity of 80% (it clears 80% of those who will not).

Take 1,000 people. The truly violent number 5% × 1,000 = 50; the remaining 950 will not be violent. Of the 50, the procedure flags 50 × 0.50 = 25 true positives. Of the 950, it wrongly flags 950 × 0.20 = 190 false positives. So the total flagged as dangerous is 25 + 190 = 215.

The positive predictive value — the chance that a person flagged as dangerous truly would have been violent — is:

PPV = 25 / 215 ≈ 11.6%.

Nearly nine in ten people identified as dangerous would not, in fact, have been violent. Put in commitment terms: if detention perfectly prevented violence, preventing one violent act would require detaining 215 / 25 ≈ 8.6 flagged people — of whom about 7.6 would have harmed no one. Raise the base rate and the picture improves; lower it and it worsens. This is the same low-base-rate arithmetic that governs preventive screening everywhere, and it is why a dangerousness standard cannot be applied without committing a large majority of false positives. The number is not a failure of any particular instrument; it is a property of predicting rare events, and no refinement of clinical skill repeals it.

Discussion

Commitment sits on an unresolved tension between two goods that cannot both be maximized: the liberty of the individual and the protection of that individual and others from the consequences of untreated disorder. The dangerousness reforms resolved the tension in liberty's favor, and the prediction arithmetic is the price. A standard honest enough to demand evidence of risk is a standard that, applied to a low-base-rate event, will detain many who pose none. There is no threshold that escapes this: tighten it and dangerous people are released, loosen it and the false-positive detentions multiply. The choice is genuinely value-laden, not technical, and pretending otherwise — treating commitment as a matter of getting the prediction right — obscures what is really a decision about how to distribute the risk of two different kinds of error.

The critique of commitment is older and more radical than the prediction problem. Thomas Szasz argued that involuntary psychiatric detention is not medicine at all but social control wearing a medical mask, and that committing a competent person against their will is a deprivation of liberty that no diagnosis can justify. Whatever one makes of his stronger claims — that mental illness is itself a myth — his challenge forced the field to defend commitment on explicit grounds rather than treating it as self-evidently benevolent, and David Rissmiller and colleagues trace how that antipsychiatry critique matured into the modern mental-health consumer movement, which reframed patients as rights-bearing users of services rather than objects of care (#ref-rissmiller-2006). The reforms that narrowed commitment, and the insistence on less-restrictive alternatives, are in part that critique's institutional legacy.

Current Directions

Two strands dominate the contemporary literature. The first is comparative and descriptive: rates of involuntary hospitalization vary enormously between countries with broadly similar health systems, and the variation tracks legal framework, service organization, and social factors rather than any measured difference in illness. Luke Sheridan Rains and colleagues, comparing legal frameworks and detention patterns across many jurisdictions, found wide and poorly explained international variation, implying that much of who gets committed is a product of law and system design rather than clinical necessity (#ref-sheridan-rains-2019). The finding reframes commitment as, in part, a policy variable — something a society chooses the level of — rather than a fixed medical response to a fixed need.

The second strand presses the field to hold involuntary hospitalization to the evidentiary standard it demands of any other intervention. Nathaniel Morris and Robert Kleinman argue for an explicitly evidence-based approach to involuntary psychiatric hospitalization, observing that one of medicine's most coercive acts rests on a strikingly thin randomized evidence base, and calling for the trials, outcome measures, and reporting that would let the field learn whether, when, and for whom commitment actually helps (#ref-morris-2023). Between the comparative work showing how contingent commitment rates are and the methodological work showing how little is known about their effects, the direction of travel is toward treating commitment as a testable practice rather than an unexamined necessity.

Common Misconceptions

Commitment is a criminal punishment.

It is a civil proceeding, not a criminal one. A committed person has been convicted of nothing; commitment rests on the state's parens patriae and police powers, and its stated purpose is care or protection, not punishment (#ref-testa-2010).

A diagnosis of mental illness is enough to commit someone.

Modern statutes require more: a mental disorder plus a specified consequence — dangerousness to self or others, or grave inability to meet basic needs — and the inadequacy of less restrictive alternatives. Diagnosis alone will not sustain a commitment (#ref-monahan-1984).

Clinicians can reliably predict who will be violent.

They cannot, at the accuracy the standard implies. Because serious violence is a low-base-rate event, even a decent risk instrument produces a positive predictive value low enough that most people flagged as dangerous will never be violent (#ref-large-2011).

Outpatient commitment is proven to keep people out of hospital.

The randomized evidence does not support that. The order alone showed no readmission benefit in the English OCTET trial (#ref-burns-2013), and the Cochrane review finds no clear effect on service use; benefits in observational studies largely reflect the intensive services bundled with the order (#ref-kisely-2017).

Glossary

Base rate.

The underlying frequency of an outcome, such as serious violence, in a population over a defined period; low base rates make any prediction of the outcome dominated by false positives.

Community treatment order.

The term used in England and much of the Commonwealth for involuntary outpatient commitment: a legal obligation to accept treatment in the community, backed by the possibility of recall to hospital.

Dangerousness standard.

The legal test that permits commitment only when a person's mental disorder produces a substantial risk of harm to self or others, replacing the older need-for-treatment test.

Emergency detention.

Short-term involuntary hold, often 72 hours, on a physician's or officer's certification, to permit psychiatric evaluation before any longer commitment is decided.

Forensic psychiatry.

The branch of psychiatry at the interface of mental health and the law; the parent kind under which MeSH files commitment.

Involuntary outpatient commitment.

A court-ordered obligation to accept treatment while living in the community rather than in hospital, aimed chiefly at the revolving-door patient.

Least restrictive alternative.

The legal principle that commitment may be imposed only when no less coercive option would meet the need, requiring inpatient detention to be justified against outpatient and voluntary options.

Need-for-treatment standard.

The older, broader commitment test permitting detention whenever treatment might benefit the person, largely displaced by the narrower dangerousness standard in twentieth-century reforms.

Parens patriae.

The state's power to act as guardian for those who cannot care for themselves; the benevolent justification for commitment, distinct from the protective police power.

Police power.

The state's power to protect the public from harm; the protective justification for commitment, invoked when a person poses a danger to others.

Positive predictive value.

Among those a test flags as positive, the proportion who truly have the outcome; for violence prediction at low base rates it is small, so most people flagged as dangerous are false positives.

Revolving-door patient.

A person who repeatedly cycles through hospitalization, discharge, treatment cessation, and relapse; the population outpatient commitment was designed to help.

Sensitivity.

The proportion of those who will experience an outcome that a test correctly flags; one of the two accuracy parameters that, with the base rate, fix the positive predictive value.

Specificity.

The proportion of those who will not experience an outcome that a test correctly clears; low specificity at a low base rate generates most of the false positives in violence prediction.

Key Researchers

Paul S. Appelbaum (living). Professor of psychiatry, medicine and law at Columbia University; a leading scholar of mental-health law whose five-year assessment of New York's Kendra's Law is among the most cited evaluations of outpatient commitment. ORCID - Wikipedia - Faculty Page

John Monahan (living). Professor at the University of Virginia School of Law and the founder of modern violence risk assessment; his 1984 call for a second generation of actuarial prediction reframed how commitment law thinks about dangerousness. ORCID - Wikipedia - Faculty Page

Alan A. Stone (1929-2022). Harvard professor of law and psychiatry and past president of the American Psychiatric Association; his Mental Health and Law: A System in Transition (1975) framed the debate over commitment standards for a generation. Wikipedia - Wikidata

Jeffrey W. Swanson (living). Medical sociologist at Duke University who led the violence-reduction analyses of the North Carolina outpatient-commitment trial, showing benefit only under sustained, service-rich orders. Wikipedia - Faculty Page

Marvin S. Swartz (living). Professor of psychiatry at Duke University and principal investigator of the randomized North Carolina outpatient-commitment trials, the pivotal experimental evidence on the practice. ORCID - Faculty Page

Thomas Szasz (1920-2012). Psychiatrist and author of The Myth of Mental Illness (1961); the most influential critic of involuntary commitment, which he argued is social control masquerading as medicine. Wikipedia - Wikidata

Frequently Asked Questions

What is civil commitment? It is the legal process by which the state detains and usually treats a person with a mental disorder without their consent, on the grounds that they are dangerous to themselves or others or unable to meet their basic needs. It is a civil, not a criminal, proceeding and it is a form of forensic psychiatry.

On what grounds can a person be committed? Most jurisdictions require three things together: a mental disorder, a resulting consequence such as danger to self or others or grave disability, and the inadequacy of any less restrictive alternative. A diagnosis by itself is never sufficient.

What is the difference between parens patriae and the police power? They are the two legal justifications for commitment. Parens patriae is the state acting benevolently for someone who cannot competently care for themselves; the police power is the state acting to protect others from harm. Most modern commitments invoke the protective, dangerousness-based rationale.

Why is predicting dangerousness so difficult? Serious violence is a rare event, and predicting rare events produces many false positives. Even a reasonably accurate risk instrument yields a low positive predictive value at a low base rate, so most people identified as dangerous would never have been violent.

What is involuntary outpatient commitment? It is a court order requiring a person to accept treatment while living in the community rather than in hospital, backed by the possibility of recall. Known as a community treatment order in England and much of the Commonwealth, it targets patients who repeatedly relapse after stopping treatment.

Does outpatient commitment reduce hospital readmission? The randomized evidence is largely negative. The English OCTET trial found no readmission benefit from the order itself, and Cochrane and other meta-analyses find no clear effect on service use. Apparent benefits in observational studies mostly reflect the intensive services provided alongside the order.

What was Thomas Szasz's objection to commitment? Szasz argued that involuntary psychiatric detention is not medicine but social control, and that depriving a competent person of liberty on the basis of a diagnosis is unjustifiable. His critique helped push the law toward narrower, dangerousness-based standards and toward the mental-health consumer movement.

How does commitment relate to forensic psychiatry? MeSH classifies commitment of persons with psychiatric disorders under forensic psychiatry, the field where clinical judgment meets legal authority. A physician supplies the finding of disorder and risk, but a legal standard and usually a court supply the power to detain.

References