High risk of suicide after discharge from involuntary psychiatric admission

Research presented at the World Congress of Psychiatry in Stockholm documents a markedly elevated suicide rate after discharge from involuntary psychiatric admission. The research found a nearly 200 times higher suicide rate compared to the total population in the first month after discharge. The nationwide study covered a full decade of every involuntary admission in Sweden. But its lead author cautions against simple conclusions. The researchers are now investigating further into this and aim to identify reasons why.

STOCKHOLM – Involuntary admission and treatment are among the most far-reaching interventions in European health care. A person is detained and treated against their stated wishes which raises serious concerns about autonomy and disability rights. Across Europe involuntary psychiatric admission and treatment is widespread, yet the evidence on what happens to people afterwards is surprisingly sparse.

At the World Congress of Psychiatry in Stockholm, Leoni Grossmann of the Karolinska Institutet presented work that addresses this gap. One study describes the suicide risk after discharge. A second sets out how the effects of involuntary measures could be studied in a way that is scientifically sound and ethically defensible.

A group that has rarely been studied

Grossmann, a doctoral researcher, says the starting point was that this group of persons with severe mental health problems is at extremely high risk and was almost absent from the research literature. “I think partially because they’re such a complex and vulnerable patient group, research on their suicide risk is particularly challenging,” she told the European Times.

The group is also hard to analyse, “since you don’t only have one specific psychiatric diagnosis in the patient group but you have a range of diagnostic profiles” she explains. The fact that some have a suicide risk as an indication at admission may also account for a difficulty of studying this group.

Her work therefore began with description: “as a first step, I think in any research it’s good to start with exploratory research and just report what do we see.”

What the data show

The first study was published in January 2026 in The Lancet Regional Health – Europe. It is described by its authors as the first nationwide, population-based analysis of suicide following involuntary admission. The researchers linked Swedish national registers and followed everyone discharged from involuntary admission between 2010 and 2020.

The cohort comprised 72,275 people and 134,514 involuntary admissions. Of these people, 2,104 (2.9 percent) died by suicide over a median follow-up of 4.4 years. On average there was one suicide for every 64 discharges.

The risk was greatest immediately after discharge. The suicide rate in the first month was 2,941 per 100,000 person-years. It fell to 1,321 over the first year and 738 over five years, but never approached the level of the general population.

The study compared this group with three others:

• All psychiatric inpatients: the rate after involuntary admission was 1.57 times higher over five years. In the first month the rates were similar.
• Psychiatric outpatients: the rate was 3.77 times higher over five years.
• The general population: the rate was about 198 times higher in the first month after discharge and 55.5 times higher over five years.

The excess compared with other inpatients grew with longer follow-up. The authors found that a “substantial part of their excess risk is thus long-term.”

Within the group, risk varied. People who died by suicide were younger, more often male and single, and more likely to have a history of self-harm and earlier involuntary admissions. Personality and substance use disorders carried higher risk than schizophrenia-spectrum disorders. The highest rate was among men aged 25 to 29. Among women the absolute risk was lower, but their risk relative to other women was higher, which led the authors to describe involuntary admission as “a particularly strong risk indicator for women.”

A companion analysis, presented at the World Congress of Psychiatry, followed 70,249 people for one year after discharge. In that shorter window, 901 (1.2 per cent) died by suicide. Chronic fatigue syndrome and previous self-harm were among the strongest risk indicators, followed by somatic conditions such as type 1 diabetes and fibromyalgia. A model combining the risk indicators produced risk-groups whose risk differed 13-fold between the top and bottom fifths.

The study notes that at the nationwide level, the findings confirm established clinical risk indicators, such as mental health problem and socioeconomic disadvantage, and importantly that this highlights “somatic conditions as previously unexplored markers for improved risk stratification and targeted care potential”.

Asked about the persistence of risk, Grossmann is plain: “We observe this extremely high risk that is worth following up on and trying to understand better who is at high risk and eventually maybe also why are they at high risk.”

What the numbers do not say

The published study does not show that involuntary detainment or coercive measures are actual causes of the excess suicide risk in this patient group, and the researchers state this. The paper says its design “precludes causal investigation.” It treats involuntary admission as three things at once: a legal category, a proxy for illness severity, and a coercive experience, including seclusion, restraint and forced medication, that “may itself influence outcomes.” The register data cannot separate these.

Grossmann stresses that a number of persons within the group was at risk before admission. She offers an example: “Someone may be admitted to involuntary care during an acute crisis, with concrete plans to end their life and an urgent need for help. Their suicide risk may still be high after discharge, but the care might have helped them through that immediate crisis and create an opportunity for continued voluntary treatment and recovery.”

She adds: “Simultaneously, there is likely to be some individuals who suffer from the involuntary care itself.” Working out who are at risk of suffering harm or will not be helped by the treatment, “will need a lot of follow-up studies.”

An intervention without an established evidence base

The central point of Grossmann’s second presentation may be the most urgent. Involuntary admission is widely used, yet its clinical effectiveness has not been established scientifically. Her abstract describes it as “a major medical intervention for severely ill psychiatric patients in crisis” and says “there is no established evidence base for its clinical effectiveness.” A survey of the literature on causal effects “reveals general a lack of robust evidence and consensus in the field.” The study concludes: “Without a strong and patient-centred evidence base, the ethical and legal justification for [involuntary admission] remains uncertain.”

The difficulty to ascertain causal effects of involuntary care is fundamental. People are usually admitted because they are in a personal crisis or at acute risk, and those same factors shape what follows. “If people who receive involuntary care have worse outcomes, we cannot conclude from that association alone that involuntary care caused those outcomes,” Grossmann says. “Conversely, improvement following admission does not by itself show that involuntary care caused the improvement.”

It is also partly a matter of ethics. Experiments with randomisation to involuntary measures would raise concerns, thus a robust evidence base will most likely be achieved through triangulation of evidence across different methods with trade-offs of each study design.

The psychiatric use of involuntary measures varies

Grossmann points to a review by Sheridan Rains and colleagues that compares admission rates across Europe. “Within Europe you would assume more similar legal criteria, similar patient mix, similar population. And yet there are admission rates that differ by a factor of 100 for some countries.”

She declines to speculate on the reasons. “It would be wrong to say we know what it is. I think we need to study it and we need to let the data talk.”

The same holds for what happens during admission. In Sweden, she notes, coercive measures such as belt restraint may be used only under the compulsory care act. What is permitted varies widely across countries. “How come that Denmark uses certain methods, Sweden not, whereas most likely we have similar people that we try to treat?” she asks. “And also there, I think we heavily call for more evidence.”

Sweden’s law requires a severe mental disorder, an indispensable need for care, and refusal or inability to consent. Risk of harm is not one of the three criteria. Grossmann nevertheless notes, “it is however an aspect that clinicians take into consideration,” for example it is an additional requirement for the conversion from voluntary to involuntary inpatient care in Sweden.

Combining methods

Grossmann argues that no single study can settle the question. “Qualitative studies can help us understand the experiences and perspectives of patients, clinicians, and others involved in coercive care. Large observational studies can show us patterns and outcomes across entire patient populations. And experimental or quasi-experimental approaches can help us get closer to answering causal questions.”

The research group she works in is now moving from description toward causal designs that try to isolate the effect of admission itself. She wants to separate how much of the excess risk is explained by involuntary mental health service users being in a worse condition to begin with, and how much by the coercion. “Of course, the part that is explained by the coercion is a negative effect that we don’t want,” she says. “We all can agree that no one wants to have this effect, neither the clinician nor the patient or people who make the policy development.”

She also wants outcomes measured beyond mortality: “It would be readmission, self-harm, but also social characteristics are important to look at, for example going back to work, having a social support network.” Register data cannot capture how care felt to the person receiving it, so she says mental health service user interviews are essential. Qualitative research shows that “many patients experience aspects of coercion as distressing, harmful, or even traumatic, while others describe feeling safe, receiving help they felt they needed.”

She rejects a simple verdict either way, and wants to know “what potential benefits and harms may arise from coercion, for whom, under what circumstances, and at what point in time.” The shared goal, she says, is “to use as little coercion as possible while achieving the best possible outcomes for patients.”

The human rights question

This research arrives in a debate that has shifted since 2006. The UN Convention on the Rights of Persons with Disabilities (CRPD) recognises people with psychosocial disabilities as full rights-holders. Article 14 states that “the existence of a disability shall in no case justify a deprivation of liberty.” Article 12 affirms equal recognition before the law and the right to exercise legal capacity.

The UN Committee on the Rights of Persons with Disabilities reads these provisions strictly. In its guidelines on Article 14 and in General Comment No. 1 on Article 12, it rejects detention and substituted decision-making based on impairment. It calls instead for supported decision-making and for services based on free and informed consent.

This pressing question was also discussed at the World Congress of Psychiatry. In the Presidential Panel “Coercion in psychiatric treatment,” dr. Anis Ahmed a presenter from the UK noted that psychiatry is “the only medical specialty which forces treatment on people and we do it with a good heart, but obviously the effect on the patients and the carers can be very devastating.”

Research needed

The research of the Swedish research team does not provide the answer to the ethical dispute. But it is clear though, that if involuntary measures are to be carried out under state sanctioned laws and practices, the case for them needs evidence built around the experience of mental health service users. Where evidence is lacking, uncertainty should be acknowledged. Where harms are identified, they must be weighed against potential benefits and the consequences of withholding care.

Such harms can take many forms, and in the extreme case death or the lack of preventing suicide. Dr. Marianne Engberg from the Institute of Psychiatric Demography in Denmark, conducted a nationwide study on mortality and suicide of the entirety of persons receiving psychiatric services in Denmark similar to the new study from Sweden. The Danish nationwide study covered two periods of four years in the 1970s and 1980s and similarly found that “committed patients, in spite of the commitment, constitute a high-risk group for suicide, also in relation to psychiatric patients in general.”

Dr. Engberg discussed the difficulty of predicting who will die by suicide, even within this high risk group and argued that applying some involuntary preventive interventions to everyone in these groups would expose too many people who would not die by suicide to that intervention. She in consequence concluded “involuntary measures cannot be used as prevention, even if the suicide risk is high.”

The new Swedish findings reiterate the practical question. What is the responsibility of society towards persons who are in a severe personal crisis and who need support? Is it acceptable that a person leaves involuntary admission with a suicide risk many times that of other psychiatric service user groups and of the general population? And if not, what should be done? And another factor, social isolation marks higher risk, and urge policy-makers to look at strengthening family and social support and community integration for people who have been through involuntary admission.

For now, the Swedish data establish that people discharged after involuntary admission face a suicide risk above that of any comparison group, and that the reasons for this excess risk are not yet understood. Grossmann says their group’s aim is to “gradually” build a more precise picture of “how we can reduce both unnecessary use and potential harms of coercion while still providing effective care to people in severe psychiatric crises.”

If you or someone you know is affected by suicidal thoughts, please contact a local crisis service or emergency number.