A strong psychiatric risk assessment is not a number on a rating scale — it is a structured argument that links a patient's history, mental state, and circumstances to a defensible management plan. Exams reward structure: candidates who organise risk the same way every time score better than those who improvise, and that same discipline keeps you safe on the wards.

This guide gives you the frame examiners look for: the four risk domains, static versus dynamic factors, how to handle actuarial scales honestly, how to convert risk factors into a formulation, and how to plan, document, and escalate. Use it as your answer template for MRCPsych CASC, RANZCP, and residency OSCE cases.

The four domains of every psychiatric risk assessment

A structured assessment starts by asking four separate questions, not one vague "are they at risk?". Cover all four explicitly, even if three are low risk — an unmentioned domain looks like an unthought one.

  • Risk to self. Suicidal ideation, self-harm, and accidental harm: intent, plan, method specificity, access to means, history of attempts or self-harm, and current drivers such as hopelessness, panic, or command hallucinations.
  • Risk to others. Violence, aggression, and abuse: previous violence, current irritability and impulsivity, threats with a named target, intoxication, and whether psychosis identifies a victim or an excuse.
  • Risk from others. Victimisation: domestic abuse, sexual exploitation, bullying, hate crime, financial abuse, and neglect by carers. This domain is the one students most often forget, and examiners notice.
  • Self-neglect and vulnerability. Nutrition, hydration, medication adherence, hygiene, housing, money management, and the capacity-related consequences of those deficits.

For each domain, work through three layers: frequency (how likely in this person), severity (how bad if it happens), and imminence (how soon). A rare but catastrophic act and a frequent minor one demand different plans, and saying so shows mature reasoning.

Static, dynamic, and protective factors

Static factors are historical and unchangeable — they set base rates but can never be treated. Prior suicide attempts or self-harm, previous violence, male sex, older age in self-harm presentations, childhood adversity, and chronic psychotic illness all sit here. You can only ever act on what sits alongside them.

Dynamic factors are the mutable, present-state variables that treatment actually targets: hopelessness, impulsivity, aggression, acute intoxication, insomnia, persecutory delusions, recent loss, social isolation, and access to means. A good assessment names two or three dominant dynamic factors and links each to an intervention.

Protective factors deserve equal space: reasons for living, engagement with services, a supportive family, responsibility to dependents, treatment adherence, future orientation, and cultural or spiritual commitments to hope. Weak protective structure with strong dynamic risk is often the real finding.

Risk domain Static factors Dynamic factors Immediate mitigations
Risk to self Prior attempts or serious self-harm; chronic illness; male sex over 50 Hopelessness; impulsivity; insomnia; intoxication; recent loss; access to means Means restriction; safety plan; increased contact; treat the acute driver
Risk to others History of violence; antisocial traits; early conduct problems Agitation; substance use; paranoia; weapons access; humiliation or threat De-escalation; environment check; supervised observation; legal powers if needed
Risk from others Prior victimisation; dependence on an abusive partner; disability Current abuse; coercive control; isolation; financial dependence Safeguarding referral; private enquiry; document and hand over
Self-neglect / vulnerability Cognitive impairment; chronic disability; no next of kin Delirium; medication non-adherence; poverty; hoarding; carer burnout Capacity assessment; social care input; PRN or depot review; review schedule

Do rating scales decide risk?

Short answer: no. Scales such as SAD PERSONS, the Beck Scale for Suicide Ideation, and similar checklists are widely taught and widely mocked, and the critical literature on tools like SAD PERSONS has been blunt — studies have questioned their ability to predict individual outcomes (Harriss and colleagues reported this concern across editions of the scale; check current PubMed literature before quoting numbers in an essay).

The methodological problem is base rate: suicide and completed violence are rare, so even a decent-looking sensitivity and specificity produce poor positive predictive value at the bedside. Actuarial instruments perform better than unaided clinical intuition at group level — that is a genuine, replicated finding — but they still cannot tell you what this patient will do on Friday night.

Use scales for what they are good at: a structured prompt so you don't skip a domain, a shared vocabulary in handover, and a rough severity marker over time. NICE's self-harm guidance explicitly frames risk assessment as collaborative formulation and care planning rather than risk stratification alone (NICE CG16). Then do the real work: formulation and plan.

From risk factors to a risk formulation

A formulation is the sentence that turns your list into an argument: who is at risk, of what, driven by which dynamic factors, under what circumstances, and what would change the trajectory. "Previous overdose, now hopeless after a separation, intoxicated, with tablets in the house" is a formulation; "high risk on the scale" is not.

Anchor the violence component on the factors most consistently associated with aggression in the literature — prior violence, substance misuse, younger male sex, acute psychosis, and antisocial traits are consistently reported across guidelines and reviews, though their relative weights differ by population and instrument. For self-harm, treat hopelessness as a primary cognitive target, impulsivity as the state variable most responsive to short-term intervention, and aggression — internalised as self-attack or externalised toward others — as the behavioural bridge between the two.

Ask also what the risk means. Behaviour that communicates distress, protest, or a request for help carries a different formulation — and a different plan — than behaviour that is instrumental or predatory. That distinction between function and form is exactly what separates an adequate answer from an excellent one, and it maps directly onto how psychiatric case formulation structures any presentation.

Worked practice makes this automatic. Preppsy's AI tutor lets you run a case through the four domains and get feedback on missing factors and weak links between evidence and plan — far more useful than re-reading a chapter about risk.

Safety planning, documentation, and escalation

The assessment is only finished when the plan exists. A collaborative safety plan should contain: (1) personalised warning signs that relapse or crisis is starting; (2) internal coping strategies the patient has actually used before; (3) people and settings that provide distraction; (4) contacts for help — friends, family, crisis lines, the team; (5) concrete means-restriction steps, agreed, not imposed; and (6) reasons for living, in the patient's own words. Keep it short enough to be read during a crisis.

Documentation should mirror the domains: risk to self, to others, from others, and self-neglect, each with evidence, an explicit level (low / moderate / high — state your criteria), the plan, who was informed, and a review date. If you decided against leave, against discharge, or against involving the family, write why. Record the patient's own account of risk alongside yours.

Escalation and handover follow the same logic: state what would trigger re-assessment or a change in level of observation (escalation criteria), and hand over using a structured format so the next clinician receives the argument, not just the conclusion. Where risk touches a third party, follow your local safeguarding and threat-assessment pathways without delay.

Finally, hold your cultural humility. Idioms of distress, ideas about family honour or spirituality, and differing willingness to disclose abuse all shape what a patient tells you. Use interpreters properly (never the family for abuse questions), ask openly about victimisation in private, and treat cultural context as formulation data rather than an obstacle to disclosure.

Exam trap

Presenting a numeric scale total as "the risk assessment." SAD PERSONS scores and similar totals are not the answer. Formulation plus plan is the answer: scales inform, they don't decide — and an unwarranted "low risk" label based on a score loses marks (and worse, patients).

Frequently asked questions

How do you structure a psychiatric risk assessment?

Structure it across four domains — risk to self, risk to others, risk from others, and self-neglect — then within each domain separate static historical factors, current dynamic factors, and protective factors, before converting the dominant dynamic factors into a formulation with frequency, severity, and imminence, and finishing with a safety plan, documentation, and escalation criteria.

What is the difference between static and dynamic risk factors?

Static factors are fixed history that sets base rate but cannot be treated — prior attempts, previous violence, childhood adversity. Dynamic factors are present-state and mutable — hopelessness, impulsivity, substance use, insomnia, acute psychosis, access to means — and every effective intervention targets them. Protective factors are the third category: engagement, reasons for living, and social support that lower risk alongside dynamic change.

Are risk assessment scales useful in psychiatry?

Only as adjuncts. Scales like SAD PERSONS prompt you not to skip domains and standardise language in handover, and actuarial tools outperform unaided intuition at group level. But rare base rates mean poor positive predictive value for the individual patient, so NICE frames risk work as collaborative formulation rather than stratification. Scales inform; formulation decides.

What should be included in a psychiatric risk documentation note?

Each of the four domains with supporting evidence, an explicit risk level with stated criteria, the dynamic factors driving risk, the agreed plan including means restriction and safety planning, who was informed, escalation triggers, the patient's own view, and a review date. Record reasons whenever you deviate from a request or a standard pathway.

Key takeaways

  • Cover four domains every time: self, others, from others, and self-neglect — unmentioned domains look unthought.
  • Separate static (historical, untreatable) from dynamic (mutable, treatable) factors, and give protective factors equal space.
  • Rating scales prompt and structure but never decide; formulation plus plan is the real assessment.
  • Name two or three dominant dynamic targets — hopelessness, impulsivity, aggression — and link each to an intervention.
  • Finish with a collaborative safety plan, explicit documentation with a review date, and stated escalation criteria.

Build the habit on every case you see and the structure will run itself on exam day — then drill it against MRCPsych CASC scenarios with Preppsy's MRCPsych preparation tracks.

Cover image: “Psychiatric Risk Assessment (cover)” by Preppsy Editorial Team (Original artwork) via Preppsy.