Yes, asylums in the historic sense (large, isolated, long-term custodial institutions that warehouse people with mental illness) are no longer used in the United States and most of the developed world. They were largely dismantled between 1955 and 1994, when the US inpatient psychiatric population fell from about 559,000 to roughly 70,000. But the functions those institutions served did not disappear. They migrated into state psychiatric hospitals, acute inpatient units, forensic facilities, nursing homes, jails, and homeless shelters. The asylum as a building is mostly gone. The asylum as a set of unmet needs persists, scattered across systems that were never designed to hold it.
What an Asylum Actually Was
The word "asylum" originally meant a place of refuge, not a place of confinement. The first US institution built on that promise was the Eastern State Hospital in Williamsburg, Virginia, which opened in 1773. It was a moral-treatment facility, meaning patients were housed in calm surroundings and encouraged toward recovery rather than chained in cells.
The model shifted in the 19th century. Dorothea Dix, a teacher who documented the treatment of people with mental illness in jails and almshouses, lobbied state legislatures to build public asylums across the country. Between 1840 and 1880, the number of US mental hospitals grew from roughly a dozen to more than 130. Each was typically built on hundreds of acres outside a town, with the explicit design goal of removing patients from urban stress.
The mechanism that made asylums fail was structural, not moral. These institutions were funded per bed by state legislatures. That funding formula rewarded occupancy, not discharge. As patient populations grew from hundreds to thousands per campus, staffing ratios collapsed. By the early 20th century, many asylums had one attendant for every 30 or more patients. The refuge became a warehouse because the money followed beds, not outcomes.
In plain terms: the building was never the problem. The incentive to fill it was.
The Numbers Behind Deinstitutionalization
Three forces converged to empty the asylums, and the timeline matters because the popular story usually credits only one of them.
| Period | Event | US state hospital population |
|---|---|---|
| 1949 | Peak census before decline | About 559,000 |
| 1954 | Chlorpromazine (Thorazine) introduced | About 560,000 |
| 1963 | Community Mental Health Act signed | About 505,000 |
| 1965 | Medicaid and Medicare established | About 475,000 |
| 1972 | Supplemental Security Income replaces state grants | About 330,000 |
| 1981 | Omnibus Budget Reconciliation Act shifts funding | About 154,000 |
| 1994 | End of large-scale state hospital era | About 70,000 |
The drug arrived before the policy. Chlorpromazine was introduced in 1954, and the state hospital population had already begun a slow decline before the 1963 Community Mental Health Act. The larger accelerant was money. When Medicaid launched in 1965, federal funds could pay for nursing home care but generally not for care in institutions classified as mental hospitals for adults under 65. States discovered they could shift patients to nursing homes and capture federal reimbursement.
The 1972 creation of Supplemental Security Income added a second incentive. Patients became eligible for federal disability payments if they lived independently, which made discharge financially attractive to states. By 1981, the Omnibus Budget Reconciliation Act folded federal community mental health funding into block grants, and the community clinics that were supposed to replace the asylums never reached the scale the original 1963 law envisioned. The law authorized construction of community mental health centers but never funded enough of them.
Where the Asylum Function Moved
The function moved into six distinct systems, each with different rules and different capacity limits.
State psychiatric hospitals still exist. As of 2024, roughly 190 state-operated psychiatric hospitals in the US hold about 33,000 beds on any given day. That is down from about 200 facilities and near 500,000 beds in 1955, a bed reduction of more than 93 percent.
Acute inpatient psychiatric units in general hospitals are now the main entry point for crisis care. These units typically hold patients for 5 to 7 days. The national average length of stay in acute psychiatric units runs about 6 days, which is short enough that the work is stabilization rather than treatment.
Forensic psychiatric facilities hold people who entered through the criminal justice system. In 2020, about 28,000 people were in state forensic psychiatric beds, and that number has grown as waitlists for competency restoration have lengthened. In some states, people wait in jail for months for an open forensic bed.
Nursing homes absorbed a large share of the old asylum population, particularly people over 65 with dementia or serious mental illness. By the 1990s, an estimated 1.5 million nursing home residents had a diagnosed mental disorder. Federal law now bars nursing homes from admitting residents whose primary need is mental illness treatment, a rule designed to prevent a second wave of custodial warehousing.
Jails and prisons became the de facto asylums of last resort. The Cook County Jail in Chicago and the Los Angeles County Jail have each been described as the largest mental health facilities in the United States at different points, depending on the count. A 2017 Bureau of Justice Statistics report found that about 44 percent of people in jail and 37 percent in prison reported a history of mental health problems.
Homeless shelters and streets hold the remainder. Roughly 20 to 25 percent of the US homeless population has a serious mental illness, according to federal estimates, which works out to more than 150,000 people on any given night.
Asylums Compared to Modern Alternatives
The trade-offs are real, and the table below is the core of why the debate never fully closed.
| Model | Typical stay | Cost per person per year | Main strength | Main weakness |
|---|---|---|---|---|
| Historic asylum | Years to lifetime | Low per bed, high per campus | Continuous supervision, shelter | Loss of autonomy, poor outcomes |
| State psychiatric hospital | 3 to 12 months | USD 200,000 to 300,000 | 24-hour care for severe illness | Bed shortage, long waitlists |
| Acute inpatient unit | 5 to 7 days | USD 1,500 to 2,500 per day | Fast crisis stabilization | Too short for durable recovery |
| Assertive Community Treatment | Ongoing, lifetime | USD 10,000 to 20,000 | Housing plus treatment in place | Limited to people who engage |
| Assisted outpatient treatment | 6 to 12 months court order | USD 5,000 to 15,000 | Enforced engagement | Legal, ethical disputes |
| Jail or prison | Weeks to years | USD 40,000 to 60,000 | Availability | No treatment, high harm |
The comparison makes a point that is easy to miss. The modern community model is far cheaper per person than a state hospital bed. The historic asylum was cheaper per person than either, but only because it provided almost nothing per person. The current system is not failing because it is expensive. It is failing because the middle tier between crisis stabilization and total independence barely exists.
The Non-Obvious Mechanism: Why Bed Counts Do Not Predict Outcomes
A common assumption is that more psychiatric beds would fix the system. The data on this is more complicated, and it cuts against intuition. Between 1955 and 1994, the US cut state hospital beds by more than 90 percent, and the suicide rate among people with serious mental illness did not rise in proportion. But other outcomes did.
A 2023 analysis of international data found that bed counts alone explain little of the variation in outcomes across countries. What predicts outcomes better is the availability of supported housing and the speed of access to outpatient care after a crisis. In plain terms: a bed is a container, not a treatment.
The counterintuitive finding is that the most effective intervention measured in randomized trials for people with serious mental illness is not a hospital service at all. It is supported housing combined with assertive community treatment, which reduced homelessness and psychiatric hospitalization more than standard care in multiple trials. The mechanism is that housing removes the daily crisis that drives repeat admission. You cannot stabilize someone who loses their shelter every 90 days.
There is also an edge case that rarely gets attention. The US still operates about 190 state psychiatric hospitals, and a small number of people with the most severe illness spend decades in them. This population is small, estimated at a few thousand nationally, but it is the closest thing to a surviving asylum cohort. Their existence is a reminder that to close all long-term institutions would require the community system to hold the people it currently cannot.
Misconceptions About Asylums Today
The first misconception is that asylums no longer exist anywhere. Several countries still run institutions that fit the classic definition. Russia, parts of Eastern Europe, and some Asian and Latin American countries maintain large long-stay psychiatric institutions with populations in the thousands per facility. The World Health Organization has documented these as a continuing human rights concern, and it has called for their replacement with community care since the 2001 World Health Report.
The second misconception is that deinstitutionalization caused homelessness. The causal chain is more indirect than that. Most people who became homeless after the 1980s did not come directly from state hospitals, which had largely closed by then. They were pushed out by a shortage of affordable housing and by the loss of low-skill jobs. Mental illness raises vulnerability to homelessness, but the housing market did more of the pushing.
The third misconception is that transinstitutionalization means the system simply moved people from one institution to another. That is partly true for jails and nursing homes, but it overstates the match. The vast majority of the 500,000 people discharged from state hospitals between 1955 and 1994 were not incarcerated or institutionalized elsewhere. They went to family homes, board-and-care facilities, or independent apartments with varying levels of support.
The fourth misconception is that long-term institutional care is a thing of the past. As of 2026, about 1.3 million people in the US live in nursing homes, and a meaningful fraction of them have serious mental illness as a secondary diagnosis. The institution did not vanish. It was renamed and reimbursed differently.
What to Actually Do If Someone Needs This Level of Care
The practical question most people face is what to do when a family member or neighbor needs more than outpatient care and less than long-term hospitalization. The path depends on the level of immediate danger.
If there is an immediate risk of harm to self or others, call emergency services and ask for a crisis team trained in mental health response where available. More than 200 US cities now have some form of mobile crisis unit.
If the situation is urgent but not immediate, contact the 988 Suicide and Crisis Lifeline, which routes to a local crisis center and can dispatch a mobile team in many areas.
If the person needs evaluation, go to the nearest emergency department. Most hospitals with psychiatric units can hold a patient for evaluation under a crisis hold. The duration varies by state, typically 72 hours to 5 days.
If the person is being discharged from a hospital, ask for a warm handoff to an assertive community treatment team. Ask specifically whether the team has capacity, and get the answer in writing if possible.
If a family member is refusing treatment and deteriorating, ask a local mental health court or county behavioral health office about assisted outpatient treatment eligibility. About 20 states and the District of Columbia have AOT laws on the books as of 2026.
If housing is the real blocker, contact the local Continuum of Care, which coordinates federal homelessness funding and often has priority access for people with serious mental illness.
The Limits of the Current Evidence
Honesty about what is known matters here. The evidence base for community treatment is stronger than it was in 1980, but it has clear gaps. Supported housing has good trial evidence. Assertive community treatment has good evidence for reducing hospitalization in people with high service use. But the evidence for many other services, including most peer-support programs and many crisis response models, is mixed or based on observational data rather than randomized trials.
The size of the surviving long-stay population is also not precisely counted. Federal surveys track state hospital admissions and discharges, not point-in-time census by length of stay. Estimates of the number of people who have been hospitalized for more than five years range from a few thousand to about 10,000 nationally, with wide uncertainty.
The most important limitation is definitional. The question "are asylums still used" depends on whether asylum means the building, the legal authority to confine, or the social function of separating people with mental illness from the rest of society. By building, no, not in the US. By legal authority, yes, in every state, through civil commitment and guardianship. By social function, the answer is uncomfortable: the separation still happens, but it now occurs in jails, shelters, and underfunded apartments rather than behind a single gate. The institution was dismantled more completely than the need it met.