Suicidality and self-harm

• Suicidality ranges from a wish to be dead through active suicidal thoughts, plans and attempts; self-harm is any intentional poisoning or injury, with or without an intent to die. ^[2][5]
• The most consistent global figures are that more than 740 000 people die by suicide each year and among people aged 15 to 29 only one cause of death is more frequent than suicide; about 73% of these deaths occur in low- and middle-income countries. ^[1]
• Ask directly about suicidal thoughts and plans with plain, respectful questions; asking does not create or worsen suicidal ideation and usually relieves it. ^[5]
• Do not rely on a global "low / medium / high" risk score or on a screening checklist total to decide care: neither predicts future suicide well. Use a clinical risk formulation tied to the person's needs, means, supports and imminent safety. ^[2][3][5]
• Management is the combination of engaging the person, removing or restricting access to lethal means, building and sharing a written safety plan, treating the underlying condition, and arranging close, early follow-up; lithium and clozapine have the strongest evidence for lowering suicide-related outcomes in bipolar disorder and schizophrenia respectively. ^[2][3][6]
• Admit (voluntarily when possible, or under the local legal framework) when the person cannot be kept safe in the community; never discharge after self-harm without a psychosocial assessment, a plan and named aftercare. ^[2]

By the end of this page you should be able to:

1. Define suicidality, self-harm and non-suicidal self-injury, and describe their overlap and differences.
2. Ask a person about suicidal thoughts and plans in a way that is direct, calm and effective, and explain why this helps rather than harms.
3. Carry out and record a clinical risk formulation instead of applying a global risk score.
4. Use the SAD PERSONS mnemonic and the Columbia Protocol (C-SSRS) appropriately, knowing their limits.
5. Build and share a safety plan and arrange means restriction.
6. Decide when voluntary or involuntary psychiatric admission is needed and arrange safe discharge and aftercare.
7. State the evidence for lithium and clozapine in reducing suicide-related outcomes and outline their place in care.
8. Recognise common pitfalls in the management of suicidal people.

A useful first distinction is between the outcome (suicide), the behaviour (suicide attempt and self-harm), and the thoughts (suicidal ideation). In this topic we group them under "suicidality and self-harm".

Suicidal ideation is a wish to die, ranging from passive thoughts ("better off dead", "wish I would not wake up") to active thoughts with intent, planning and preparation. A suicide attempt is a self-directed act carried out with at least some wish to die, regardless of whether injury occurs. Self-harm is intentional self-poisoning or injury, irrespective of the apparent purpose: the same act may be suicidal, non-suicidal, or ambivalent. ^[2][5] Non-suicidal self-injury is deliberate tissue damage without any intent to die, usually to regulate emotion or communicate distress. The underlying problems (depression, trauma, personality disorder, substance use, psychosis, social crisis) frequently overlap.

This matters because the management of suicidal thinking is not the same as the management of a self-harm episode, and neither is the same as a personality-disorder crisis. Keeping the distinction clear makes the assessment more honest and more useful.

Suicide is not a diagnosis. It is an outcome of the interaction between a person, their psychiatric and physical health, and their environment. That is why assessment and management here are organised as a pathway rather than as a treatment for one disease: the same clinical pathway applies whether the person is seen in primary care, in an emergency department, on a medical ward, or in a mental-health service. ^[3]

The World Health Organization estimated that more than 740 000 people die by suicide every year, with 746 000 deaths in recent estimates; among people aged 15 to 29 only one cause of death is more frequent than suicide, and roughly 73% of global suicides occur in low- and middle-income countries. ^[1] For every completed suicide there are many more attempts, and a prior attempt is an important risk factor. ^[1]

The central epidemiological point for a clinician is contact with services: many people who die by suicide have seen a primary care or mental health professional in the months before death, and they are more likely to see a general practitioner than a psychiatrist. ^[5] Every contact is therefore an opportunity to ask, to engage, and to put a plan in place. Stigma and under-reporting mean the true burden is almost certainly higher than these numbers suggest. ^[1]

The pathway here applies to adults, adolescents and children who present with suicidal thoughts, plans, attempts or self-harm. Risk is present across the life course, but vulnerability is concentrated in groups experiencing combinations of the following: ^[3][5]

• Mental disorder, especially depression, bipolar disorder, schizophrenia and related psychoses, post-traumatic stress, eating and substance-use disorders, personality disorder and anxiety disorders.
• Previous suicide attempt or self-harm.
• Hopelessness and pessimistic expectations.
• Chronic physical illness, chronic pain, neurological disease and functional decline.
• Acute stressors: bereavement, relationship breakdown, financial or legal trouble, housing loss, academic failure, exposure to violence or abuse, discrimination and isolation.
• Access to lethal means, particularly firearms and toxic substances.
• Personality factors such as impulsivity, aggression and problem-solving deficits.
• Social isolation, unemployment, broken family and living alone.

Risk is also elevated in vulnerable and marginalised groups: migrants and refugees, indigenous peoples, LGBTI people, prisoners, and people in humanitarian crises. ^[1]

Protective factors are equally important: strong interpersonal relationships, religious or spiritual faith for some people, problem-solving and coping skills, reasons for living, and a strong therapeutic alliance. Assessment and care should strengthen these, not just reduce risk. ^[5]

Work through the steps in order. Do not skip ahead because of time pressure: a person who is about to leave the emergency department can usually be kept waiting for a few minutes while you finish step one.

1. Recognise and engage. Notice the verbal and non-verbal cues, create privacy and time, and build a working alliance.
2. Ask directly about death, self-harm, intent, plan, lethality and means.
3. Assess immediate safety and any need for restraint, observation or urgent medical care.
4. Formulate risk by gathering risk factors, protective factors and the story of this episode.
5. Manage in the least restrictive effective setting: means restriction and safety planning first, then treatment of the underlying condition, then consideration of admission.
6. Document and communicate the assessment, plan and contacts.
7. Follow up early and reassess after any change in mental state.

The flowchart at the end of this page puts the same steps into a single algorithm.

Many clinicians fear that asking about suicide will "put the idea into the person's head". The evidence does not support this: asking does not induce or increase suicidal ideation, and acknowledging and discussing suicide may reduce it. ^[5] Silence, by contrast, leaves the person alone with the thoughts.

Practical points:

• Raise the topic gently but directly, ideally while exploring mood or distress: "When people feel this low, they sometimes think life is not worth living. Have you had thoughts like that?"
• Start with open-ended questions and move to specific ones: ask about thoughts of death, then thoughts of self-harm, then whether there is a plan, then about means, timing and preparation.
• Ask without alarm or judgement. Reassure the person that these questions are routine and that you want to understand.
• If the person is intoxicated with alcohol or drugs at first interview, note the state and reassess later, when sober, because judgements about risk can change. ^[2][3]
• Speak with family or close friends (with the person's consent where appropriate, and without consent if there is imminent risk) because they may know about concealed plans or recent purchases. ^[2][3]
• In cultures where suicide is taboo, normalise: "These thoughts can come automatically when people are overwhelmed and are not always controllable."

A useful stepwise sequence is: general thoughts of death → passive suicidal ideation (no self-harm) → active suicidal ideation (thoughts of self-harm) → intent and a specific plan → preparation → prior attempts with their lethality. ^[5] A person with passive ideation and no plan usually needs mitigation of risk factors, strengthening of supports and monitoring; a person with active ideation, intent or a plan needs same-day psychiatric assessment and usually transport to the emergency department. ^[5]

A risk formulation is the process of drawing the gathered information together into a coherent narrative that informs a treatment and management plan. It is not a score. ^[2]

NICE NG225 recommends that mental health professionals undertake a risk formulation as part of every psychosocial assessment, and that assessment focus on the person's needs and how to support their immediate and long-term psychological and physical safety. ^[2]

• Historical factors that do not change easily: past attempts or self-harm, mental disorder, personality, family history of suicide, trauma, abuse, chronic physical illness, previous hospitalisation.
• Current (dynamic) factors: hopelessness, insomnia, agitation, panic, intoxication, acute losses, poor coping, non-adherence, escalating distress.
• Future factors: upcoming court dates, anniversaries, discharge from a hospital, exams, changes in medication.
• Protective factors: relationships, responsibilities, beliefs, problem-solving ability, engagement with help.

Risk is best thought of as the product of active ideation, the presence of a plan and means, capacity to act, lack of protective factors and limited engagement, weighed against stressors. ^[5]

Two tools are widely used and worth understanding. Neither should be used to predict future suicide or self-harm, and neither should decide who gets treatment or discharge. ^[2] This is a change in everyday practice for many services, and it is deliberate: tools were sold as predictive and largely were not. ^[2]

SAD PERSONS is a ten-item mnemonic for suicide-related risk factors, developed in 1983 as a teaching and prompt tool for medical students and non-psychiatrist doctors. ^[5] Each factor scores one point, up to a maximum of ten.

The original report suggested close monitoring for scores of 3 to 4, strong consideration of hospitalisation for 5 to 6, and hospitalisation for further assessment for 7 to 10. ^[5] Many later services condensed these to three bands (low 0 to 4, moderate 5 to 6, high 7 to 10).

These bands are now known to perform poorly: a systematic review found insufficient evidence to support the use of the scale to assess or predict suicidal behaviour, and the scale did not acutely predict future attempts. ^[5] A high score should not be necessary for action, and a low score is not reassuring. The mnemonic is still useful as a memory aid so that you do not forget to ask about a known risk factor, but the decision must rest on the formulation. ^[2][5]

The Columbia-Suicide Severity Rating Scale (C-SSRS), or Columbia Protocol, is a widely used screening scale that asks plain, direct questions and rates the severity and recency of ideation and behaviour. ^[4] It is short (the screening versions use between two and six questions, adapted to previous answers), free to use in community and health-care settings, and available in many translations. ^[4]

Its strengths are that it asks directly about suicidal thoughts and behaviours, separates ideation intensity from behaviour, and distinguishes interrupted, aborted and completed attempts from non-suicidal self-injury. ^[4] It is useful for triage and documentation. However, like every scale, it does not predict future suicide with confidence, and any positive endorsement must be followed by a full assessment and management plan rather than by a mechanistic next step. ^[2][4]

Assessment should include a full mental state examination, with attention to psychiatric symptoms that increase risk such as aggression, impulsivity, hopelessness, agitation, psychic anxiety, anhedonia, global insomnia and panic. ^[3] A practical summary list used across sources is: depressed mood, hopelessness, anxiety and agitation, psychosis or impaired reasoning, chronic physical illness, substance use, both older and younger age, unemployment, poor social support, and a family history of suicide. ^[3][5]

The American Psychiatric Association's legacy guideline notes that the presence of a psychiatric disorder is the most significant single risk factor, and that suicide assessment scales have very low predictive value and do not give reliable estimates of suicide risk. ^[3]

For a child or adolescent who self-harms, assessment should be by a clinician experienced in that age group and should cover the social, educational and home situation, caring responsibilities, peer relationships and the impact of social media and online contact. ^[2] Non-suicidal self-injury in adolescents is common; do not trivialise it, but do explore function (regulation, communication, punishment) and the absence or presence of suicidal intent, because the management differs.

The thresholds below are the practical decision points that recur across guidelines. They are given as guides to structure care, not as rules that override clinical judgement. ^[2][3][5]

Engagement is treatment. Establish and maintain a therapeutic alliance while attending to safety. ^[3] Take responsibility for care, without pretending that you can take responsibility for the person's life. ^[3]

Immediate steps:

• Move to a private area and remove the risk of interruption.
• Ensure one-to-one observation or continuous monitoring if the person is at significant risk, in an emergency or inpatient setting. ^[3]
• Screen for dangerous items and secure belongings; remove hazardous items from the room in inpatient settings. ^[3]
• Consider continuous observation, not mechanical restraint, in the emergency department; local policy on restraint should follow the NICE violence and aggression guideline. ^[2]
• Never leave an actively suicidal person alone in a room with available means.

The APA guideline advises treating the person in the least restrictive setting that is likely to be safe and effective, taking into account preferences as well as risk. ^[3] That principle frames the rest of management.

Means restriction is one of the few suicide-prevention interventions with evidence for population-level effect: limiting access to lethal means (pesticides, firearms, certain medications) reduces deaths, because suicidal crises are often brief. ^[1] At the clinical level:

• Ask specifically about means in the person's home, car and workplace, including firearms, ropes, ligatures, medications and toxic substances. ^[3]
• Discuss removing or securing these means collaboratively and therapeutically, not punitively. ^[2][3]
• Work with the person and, with appropriate consent, family or carers to remove or restrict access, and make the environment safe. ^[2]
• For self-poisoning presentations, follow local poisoning management pathways and secure medications on discharge. ^[2]
• Advocate and support public-health measures such as safe storage and disposal, safe dispensing, smaller pack sizes for dangerous medicines, and pesticide regulation; these are the responsibility of systems, not individuals. ^[1][2]

NICE NG225 includes safer prescribing and dispensing recommendations (for example, limiting quantities of dangerous medicines) precisely because access to means matters. ^[2]

Safety planning is a brief, collaborative intervention that gives the person something to do when suicidal intensity rises. It is not the same as a "no-suicide contract", which is not supported by evidence and may encourage the person to hide intent. ^[5]

A written safety plan should include: ^[2]

• recognition of the personal triggers and warning signs of worsening distress, further self-harm or a suicidal crisis;
• individually chosen coping strategies and problem-solving of obstacles to using them;
• social contacts and social settings that can distract from suicidal thoughts;
• family members or friends who can support, and how to ask them;
• contact details for services, including out-of-hours and emergency contacts;
• environmental safety, by collaboratively removing or restricting lethal means.

The plan should be held by the person, shared with family, carers and professionals as the person decides, and accessible at times of crisis. ^[2] It is developed in partnership, using a problem-solving approach, and should also be given as information at first contact. ^[2]

If a safety plan already exists, discuss and update it whenever the risk level changes. ^[5]

NICE recommends that treatment following self-harm take into account associated coexisting conditions and the psychosocial assessment, and it (rightly) recommends that drug treatment not be offered as a specific intervention to reduce self-harm itself. ^[2] Treating the illness, however, is central to reducing risk.

• Depression: optimise treatment with evidence-based psychological and pharmacological therapies. ^[5] Refer specifically to the relevant NICE guidelines on depression. ^[2]
• Psychological therapies: offer a structured, person-centred, CBT-informed intervention tailored for adults who self-harm, starting as soon as possible and typically four to ten sessions; use DBT adapted for adolescents with significant emotional dysregulation and frequent self-harm. ^[2]
• Harm minimisation strategies may be considered, as part of an overall recovery plan and after collaborative discussion, when a person is not yet able to stop self-harming but further harm can be reduced. ^[2]
• Substance use should be addressed directly, because intoxication both increases impulsivity and makes assessment unreliable. ^[2][3]
• Anxiety: because anxiety is a significant and modifiable risk factor, treating it may reduce risk, but benzodiazepines can occasionally disinhibit aggressive or impulsive behaviour, particularly in personality disorder, so they require caution. ^[3]
• Electroconvulsive therapy may reduce suicidal ideation, at least in the short term, and may be appropriate for severe depression. ^[3]
• Do not use diagnosis, age, substance misuse or comorbidity as a reason to withhold psychological intervention. ^[2]

Two pharmacological measures stand out for reducing suicide-related outcomes, and they are (unusually) supported by more than just theory. ^[3][6]

Lithium. Long-term maintenance lithium in people with recurring bipolar disorder and major depressive disorder is associated with substantial reductions in the risks of suicide and suicide attempts. ^[3] In a 2026 pooled analysis of observational studies, lithium was associated with lower suicide risk in bipolar disorder (odds ratio 0.38; 95% confidence interval 0.28 to 0.50) and with lower suicide attempt risk (odds ratio 0.60; 95% confidence interval 0.44 to 0.82). ^[6] There was evidence of publication bias toward studies showing lower risk, so treat the size of the effect as probably optimistic, but the direction of effect is consistent with older evidence. ^[6]

Place in practice: lithium is considered for mood stabilization in bipolar disorder and for augmentation in treatment-resistant unipolar depression, after weighing the person's risk profile. It requires monitoring (renal function, thyroid function, serum level, hydration, interacting medicines). Because lithium toxicity is dangerous and the benefit is long term, it is not an emergency treatment for acute suicidality and it should be prescribed only by clinicians able to monitor it safely. ^[6]

Clozapine. In schizophrenia spectrum disorders, clozapine has repeatedly been associated with reductions in rates of suicide attempts and suicide. ^[3] In the same pooled analysis, clozapine was associated with lower suicide mortality in schizophrenia spectrum disorders (odds ratio 0.40; 95% confidence interval 0.36 to 0.60). ^[6] It also has a specific regulatory indication in several jurisdictions for reducing the risk of suicidal behaviour in schizophrenia (check your local product information). Other first- and second-generation antipsychotics may also reduce suicide risk, particularly in highly agitated patients. ^[3]

Place in practice: clozapine is reserved for treatment-resistant schizophrenia (failure of at least two adequate antipsychotic trials) and requires regular haematological monitoring because of agranulocytosis, usually through a supervised monitoring programme. It is therefore a specialist treatment, and its anti-suicidal effect is a reason to consider it earlier than habit in the right patient, not a reason to start it unsupervised. ^[3][6]

Two honest caveats. Both sets of evidence come mainly from observational and maintenance studies; firm causal claims are not justified, and confounding by indication (people who are better engaged are more likely to get and adhere to these drugs) is a real threat. ^[6] Neither drug should delay, replace or be mistaken for the immediate, non-pharmacological steps above. Antidepressants, antipsychotics and other medicines remain indicated for their primary psychiatric indications; the clinical decision is about which agent to choose in a person for whom suicide is a concern.

Voluntary (informal) admission is preferred whenever the person agrees and it is safe. It preserves the therapeutic alliance, which is itself protective. ^[3]

Involuntary (compulsory) admission is justified when the person is at substantial risk and refuses or lacks capacity. The specific legal criteria, thresholds and processes differ between countries and even between regions; this page gives principles, not legal advice. In general, compulsory admission requires that the person has a mental disorder, that there is serious risk to the person's health or safety, and that the least restrictive option cannot safely be used. ^[2][3]

Practical application:

• Admit to a general hospital after an episode of self-harm when there are safety concerns but psychiatric admission is not indicated, when safeguarding planning is needed, or when the person cannot engage in a psychosocial assessment, for example because of distress or intoxication. ^[2]
• Refer to a psychiatric (or adolescent, older-adult or liaison) service for assessment of need for specialist inpatient care when there is high or acute risk, severe mental illness, or lack of a safe environment.
• Before discharge from any hospital after self-harm, a psychosocial assessment must have taken place, a management plan must be agreed, and aftercare must be specified in writing with the primary care team. ^[2]
• Discharge when the person can be safely supported in the community, with a named contact and early follow-up.
• No mechanical restraint should be used in the emergency department to prevent self-harm or to stop a person leaving; use policy, de-escalation and appropriate observation. ^[2]

Agree the purpose, format and frequency of aftercare with the person and (as appropriate) their family and carers, and record it. ^[2] Where there are ongoing safety concerns, initial aftercare should be provided within 48 hours of the psychosocial assessment. ^[2] Reassess safety after any change in mental state, with a change in level of observation or before discharge, and whenever the person was intoxicated at first interview. ^[3] Consider closer follow-up in the early stages of treatment, because some medications take weeks to work and because early improvement in energy can briefly increase risk while depressed mood persists. ^[3]

Escalate without delay if any of the following are present:

• expressed suicidal intent, a specific and practicable plan, or preparation;
• a recent attempt of high lethality;
• command hallucinations or severe psychosis;
• severe depression, especially with agitation or insomnia;
• manic or mixed states;
• increasing frequency or severity of self-harm;
• the assessing clinician is concerned, even if no single indicator explains why;
• rising or sustained distress in the person or in family members and carers, despite support. ^[2][3]

Escalation pathways:

• Primary care should refer to mental health or social care services for a psychosocial assessment after an episode of self-harm, and make it a priority when the conditions above are present. ^[2]
• Emergency departments should refer to age-appropriate liaison psychiatry, with a psychosocial assessment alongside physical care, and speak to the person at every attendance. ^[2]
• Ambulance services should use care and safety plans already in place and seek advice from mental health services where necessary. ^[2]
• Social care should seek advice from, and refer to, urgent and emergency mental health services when self-harm is identified. ^[2]
• Older adults should be assessed by a clinician experienced with older people, with attention to depression, cognitive impairment, physical health and frailty. ^[2]

For patients seen in primary care, also ask, document and share the risk, the reasons for it, mental state, history, medication, and social contacts when transferring someone for urgent psychiatric care. ^[5]

Documentation protects the person and the next clinician. It should include: ^[5][2]

• the means of self-harm, if any, and what was done about them;
• the presence or absence of suicidal ideation, its type and intensity, and any plan, intent or preparation;
• mental state examination findings and any psychiatric symptoms relevant to risk (aggression, impulsivity, hopelessness, agitation, anxiety, anhedonia, insomnia, panic);
• risk and protective factors and the narrative formulation;
• capacity, consent, and any use of the legal framework for admission;
• the safety plan, its contents, to whom it was given, and when it is to be reviewed;
• any referrals and arrangements for aftercare, including named contacts and dates;
• communication with family, carers and other professionals, and the reasons for any disclosure without consent.

Consent matters. Seek consent to involve family and carers early. Even without consent, family members can provide information; and when there is imminent risk, safety overrides confidentiality, provided the person is informed where possible that disclosure may occur. ^[2][5]

• Using a screening scale total or a "low / medium / high" band to admit, discharge or deny treatment. ^[2][5]
• Avoiding the question about suicide for fear of causing harm. ^[5]
• Accepting "I'm fine now" without exploring whether lethality or intent has actually changed, especially as depression lifts. ^[3]
• Reassuring yourself with a no-suicide contract. ^[5]
• Delaying psychosocial assessment until after medical treatment, or until the person is sober. ^[2]
• Discharging someone after self-harm without a psychosocial assessment, plan and written aftercare. ^[2]
• Prescribing sedatives to an acutely suicidal person without assessing suicidality and planning follow-up. ^[3]
• Missing means in the home, especially firearms and toxic medicines. ^[3]
• Failing to reassess a person who was intoxicated at first contact. ^[2][3]
• Managing high risk alone, without consultation or escalation. ^[3]
• Ignoring cultural and age-specific factors, or invalidating the person's experience.
• Treating self-harm as purely manipulative or attention-seeking; it is also a serious marker of distress and risk. ^[2]

Suicidal ideation in pregnancy and the perinatal period must be taken seriously and linked with perinatal mental health services. The broader approach above applies; medicines (including lithium) should be chosen with regard to pregnancy and breastfeeding, but that discussion is beyond this page. Involve specialist and obstetric input early.

Children and adolescents need age-appropriate services and sensitive assessment of abuse, bullying, social media and home difficulties; CAMHS or equivalent crisis access should be available, including out of hours. ^[2] Older adults should be assessed with attention to physical illness, frailty, cognitive impairment and social isolation. ^[2]

• Early recognition is the highest-value step in under-resourced settings; train all front-line staff (primary care, emergency, maternity, HIV, refugee, prison) to ask directly. ^[1][2]
• Restricting access to the most common local means (pesticides, certain medicines, firearms) is often the most effective national action and is recommended by WHO LIVE LIFE, the 2021 suicide prevention implementation guide. ^[1]
• Many treatments in these guidelines assume access to mental health professionals and crisis services. Where these do not exist, rely on general health workers trained in mhGAP-style care: brief psychological support, engagement of family, means restriction, safety planning, and referral to the nearest higher-level service. ^[1]
• Decriminalisation of suicide attempts (see the WHO policy brief) supports help-seeking; treat suicide attempts as a health priority rather than a crime. ^[1]
• When inpatient psychiatric care is unavailable, risk mitigation must be community-based, with clear family roles, daily follow-up, and rapid referral routes agreed in advance.

• Suicidality is not a diagnosis. It is an interaction of psychiatric illness, physical health, psychology and social context. ^[3]
• Self-harm is intentional poisoning or injury, whatever its apparent purpose; suicide attempt is self-harm with intent to die. ^[2][5]
• Among people aged 15 to 29 only one cause of death is more frequent than suicide; more than 740 000 people die by suicide each year, and roughly 73% of those deaths occur in low- and middle-income countries. ^[1]
• Asking about suicide does not plant the idea; it usually reduces distress and improves detection. ^[5]
• A psychiatric disorder is the strongest single risk factor, and previous attempt is one of the strongest behavioural predictors. ^[3][5]
• Do not use global risk scores, and do not use screening tool totals, to predict suicide or to decide care. Use a clinical risk formulation plus the person's needs, means and supports. ^[2][3][5]
• Means restriction, safety planning and early follow-up are the practical cores of management. ^[1][2]
• No-suicide contracts have no evidence base and may hide intent; avoid them. ^[5]
• Lithium is associated with reduced suicide-related outcomes in bipolar disorder and major depression; clozapine is associated with reduced suicide-related outcomes in schizophrenia. ^[3][6]
• Discharge after self-harm requires a completed psychosocial assessment, an agreed plan and named, written aftercare. ^[2]
• The least restrictive effective setting is the guiding principle for admission decisions. ^[3]
• Every contact with a suicidal person is an opportunity to ask, engage and plan.

Depression and bipolar disorder; schizophrenia and other psychotic disorders; substance-use disorders; personality disorders and complex trauma; poisoning and self-poisoning management; psychopharmacology (antidepressants, antipsychotics, mood stabilisers); first-episode psychosis; older adult psychiatry. (These are listed as plain text because the final link targets are being verified; integrate links once the topics are verified.)