There is concern currently about safety in mental health in-patient services. During 2011-2021, over a quarter (28%) of patients died by suicide in acute mental health care settings, including in-patients (6%), post-discharge care (14%) and crisis resolution/home treatment (13%). Of the estimated 74 suicides by mental health in-patients in the UK, a quarter (28%) died whilst under enhanced nursing observation (i.e., frequent – every 15-30 minutes – checks on a patient or being with them constantly). The highest number of deaths after discharge from psychiatric in-patient care occurred on day three post-discharge.

We recommend clinical services need to focus on (1) creating a therapeutic ward environment, (2) the physical safety of the in-patient unit itself, (3) safe transition from ward to community, (4) pre- and post-discharge, (5) early follow-up after in-patient discharge, and (6) prompt access to crisis services.

In 2011-2021, there were 350 autistic people who died by suicide, 2% of all patient suicides and an average of 32 deaths per year. There were 159 people with ADHD who died by suicide, 1% of all patient suicides and an average of 15 deaths per year. 29 patients had received a diagnosis of both autism spectrum disorder and ADHD. The number of autistic people and those with ADHD increased over this ten-year period, likely a reflection of an increase in clinical recognition and diagnoses of these disorders.

Diagnoses of autism spectrum disorder and ADHD are becoming a larger part of suicide prevention in mental health services, especially among young people. We suggest clinicians may require specific training to recognise and support these patients. We also ask clinicians to be aware of the high rates of suicide-related internet use prior to suicide among autistic people and drug misuse among patients with ADHD.

There has been recent concern over in-patient safety for young people. There were 117 deaths by suicide in in-patients who were aged under 25 (10-24 years) in 2011-21, an average of 11 deaths per year; 20 were aged under 18. In-patients under 25 who died showed high rates of clinical risk factors associated with suicide, including self-harm, alcohol and/or drug misuse, and childhood abuse. Half had been detained under Mental Health Act powers. They were more often under enhanced nursing observation. In 43% the admission was at a non-local unit.

We ask services to be aware that in-patients under 25 who die by suicide may have different clinical characteristics to adult in-patients, with proportionately more deaths on the ward and Mental Health Act detention, and with enhanced nursing observations being more frequent. We suggest attention is needed to potential ligatures and ligature points used on the ward, and to the importance of admission to local units, where possible.

Our findings show the number of patients who were students (aged between 18 and 21) and died by suicide increased between 2011 and 2020, which may reflect improved contact with services, but fell in 2021. Of the 869 deaths in England and Wales by students between 2011 and 2021, 96 (11%) were mental health patients, a lower proportion than other young people in the general population who died by suicide (25%). These deaths were most common in October and April.

We suggest support should be enhanced at key times of risk, such as the start of the academic year and in the lead up to exams. We also recommend a clear pathway to mental health services.

In 2016-21, there were 1,001 deaths by patients whose contact with mental health services was a one-off assessment, an average of 167 deaths per year. Patients assessed only once were more likely to have a recent history of alcohol or drug misuse than other patients. The majority (71%) had experienced recent adverse life events, particularly financial problems and relationship break-up.

We recommend care is needed when discharge is considered following a single assessment. In considering follow-up plans, we ask clinicians to be aware of the potentially serious impact of recent life events such as financial difficulties and relationship break up, and that alcohol and drug misuse may indicate a greater need for follow-up.

Between 2011 and 2021, there were 3,894 patients who died by suicide in a public place, 29% of all patient suicides, and an average of 354 deaths per year. We found patients who died in a public place were younger than those who died in private locations. They appeared to be more acutely unwell (with higher rates of schizophrenia and other delusional disorders) and self-harm, drug misuse, relationship break-up and financial problems were more common.

We suggest the safety of the environment where mental health services are situated needs to be addressed. Local suicide prevention plans should reflect this risk, reflecting joint working between clinical leaders and local authorities. Sensitive discouragement of personal memorials and careful media reporting may also contribute to prevention.