5 Organisation and delivery of psychiatric services5.2 Preventative strategies
5.2 Prevention of psychiatric disorder
5 min read1 figure
Written by Resident Doctor Committee · Published
Caplan's three levels
Leavell and Clark's levels of prevention, applied to psychiatry by Caplan (1964), are indexed to the stage of disease.
Level | When it acts | What it does | Example |
|---|---|---|---|
Primary | Before onset, on people who are well and not selected by risk | Lowers incidence | Rubella vaccination, which lowers congenital intellectual disability; public education; alcohol minimum unit pricing |
Secondary | Early or existing disorder | Detects and treats it promptly to shorten or halt it, which lowers prevalence | Screening mothers for depression at the six-week baby check; a brief intervention after a positive alcohol screen |
Tertiary | Established, diagnosed disorder | Limits disability, relapse and complications | Rehabilitation, supported employment, lithium to prevent bipolar relapse, treating established alcohol dependence |
Screening everyone is still secondary prevention, not universal prevention, because nothing is prevented before onset.
Secondary prevention catches a disorder early; tertiary prevention limits the harm of a disorder already established.
Tertiary prevention makes up most preventive activity in present-day psychiatry, because causal knowledge is too limited to support much primary or secondary prevention.
Prevention acts on perinatal factors, childhood adversity, substances, physical, sensory and cognitive impairment, and social conditions such as isolation and disadvantage.
Extended levels
Primordial prevention changes a population's social and environmental conditions so that risk factors never arise, whereas primary prevention protects people already exposed to a risk.
Quaternary prevention, a term from Marc Jamoulle adopted by the World Organization of Family Doctors, protects patients from harm caused by medical care, such as repeated tests, false positives, incidental findings and needless procedures.
Institute of Medicine categories
The Institute of Medicine framework (Mrazek and Haggerty, 1994) builds on Gordon's three tiers of primary prevention. It divides prevention before onset by who receives it, set against Caplan's secondary and tertiary prevention once illness has begun.
Category | Who receives it | What decides eligibility | Example |
|---|---|---|---|
Universal | A whole population or year group | Nothing about individual risk | A law limiting access to a common means of self-poisoning; a school course; routine antenatal care |
Selective | A subgroup sharing a risk factor, without symptoms | Membership of the group, not anything observed in the individual | Children of a parent with bipolar disorder or severe mental illness; teenagers with a parent with psychosis |
Indicated | An individual with early, subthreshold symptoms short of diagnostic criteria | The person's own symptoms | An at-risk mental state with attenuated psychotic experiences, addressed by at-risk mental state services |
Acceptable cost and risk rise as targeting narrows: universal measures must be cheap and very safe, while indicated measures may justify some risk.
In suicide prevention, universal measures reach everyone, selective measures target a vulnerable group without suicidal behaviour, and indicated measures target people with suicidal ideation, plans or recent self-harm.
What is not prevention
Mental health promotion builds general wellbeing and resilience, through social connection, secure housing, meaningful work, sleep and exercise, without aiming at a specific disorder; universal prevention aims at one named disorder.
Case identification, finding people who already meet criteria and bringing them to care early, is the first step of the treatment band on the Institute of Medicine spectrum.
Measuring prevalence, for example with anonymised scores that identify and treat no one, describes the burden of disease without changing onset or course.
Population strategy and Rose's paradox
Rose's prevention paradox: a population strategy gives a large collective benefit but only a small benefit to each individual, because many people at modest risk generate more cases than a few at high risk.
A population strategy therefore prevents more cases overall than a selective approach, even though each high-risk recipient of the selective approach gains more.
Restricting suicide methods reaches people whom services never identified as high risk.
Measure | Observed effect |
|---|---|
Paracetamol pack size limit (1998) | A 43% fall in paracetamol-related suicide in a long-term study |
Removing carbon monoxide from domestic gas (1960s) | UK suicide fell despite rising unemployment (Kreitman's coal gas story) |
Scottish minimum unit pricing (2018; 65p a unit from 2024) | An estimated 13.4% fall in deaths wholly attributable to alcohol; a 4.1% fall in related admissions that was not statistically significant |
The Lancet Commission life-course model (2024 update) lists 14 modifiable dementia risk factors, including hearing loss and hypertension, each with a population attributable fraction and a life stage.
Social isolation is the leading later-life factor, while some larger factors belong to midlife.
Addressing all 14 might theoretically prevent or delay around 45% of dementia, an estimate rather than a demonstrated effect.
Early intervention in psychosis
Screening helps only when earlier treatment improves outcomes and the benefits outweigh the harm of false positives.
A longer duration of untreated psychosis predicts poorer outcomes, but the association does not prove causation.
Insidious onset is the most consistent predictor of a long duration of untreated psychosis in first-episode psychosis.
Quiet features (insidious onset, negative symptoms, poor insight) lengthen untreated psychosis, whereas conspicuous, risky or affective presentations shorten it.
Early at-risk mental state cohorts reported about 40% transition to psychosis within 1 to 2.5 years; recent cohorts report 14 to 15% at 12 months, partly because broader referral brings in lower-risk people.
Most people at risk never develop psychosis, and the label can cause anxiety, stigma and needless antipsychotic exposure.
Relapse prevention
Relapse prevention combines psychoeducation about early warning signs, maintenance medication or psychological treatment, and a crisis plan agreed while the person is well.
It also addresses high expressed emotion through family work, barriers to adherence, and housing and employment.
Evaluation and ethics
Outcomes must fit the aim: incidence for primary prevention, and duration, relapse, disability or mortality for later stages.
Evaluations must check other methods for substitution, because the fall in car-exhaust suicide after catalytic converters did not rule out changes in other methods.
When predictive value is poor, anxiety, stigma and unnecessary treatment from false positives may outweigh the benefit.
Lifestyle advice cannot fix poverty, insecure housing or discrimination, so prevention should address structural causes without blaming individuals.