Preventive Psychiatry

Abstract

Preventive psychiatry is a branch of community psychiatry concerned with stopping mental disorders before they begin, catching them early, and limiting the disability they cause. It imports into psychiatry the public-health logic of prevention, and its history is the story of two competing ways to classify a prevention effort: Gerald Caplan's disease-stage scheme of primary, secondary, and tertiary prevention, and the risk-based scheme of universal, selective, and indicated prevention that the Institute of Medicine adopted in 1994. The field's difficulty is quantitative: because most mental disorders are rare, a prevention program must enrol large samples to show it lowers incidence, and its yield depends sharply on the baseline risk of the group it targets. This article traces the classification schemes, the statistics that govern whether prevention can be demonstrated, and the modern turn toward clinical staging in young people.

Keywords: primary prevention, indicated prevention, clinical high-risk state

What Preventive Psychiatry Is

Preventive psychiatry is the application to mental disorders of the aim that defines all of preventive medicine: to reduce the occurrence, duration, and disabling consequences of illness rather than only to treat illness already established. Gerald Caplan, whose 1964 Principles of Preventive Psychiatry gave the field its name and its first systematic statement, defined it as the body of knowledge concerned with lowering the incidence of mental disorder in a community, shortening the course of the disorders that do occur, and reducing the impairment they leave behind (#ref-caplan-1964). Those three aims map onto the three classical stages of prevention, and the mapping is the field's organizing idea.

MeSH files Preventive Psychiatry (descriptor D011316) as a narrower kind of community psychiatry, and the placement is apt: prevention is inherently a population activity. A clinician treating a patient works one case at a time, but a prevention program acts on a defined group — a school cohort, a workplace, the children of parents with a disorder — and is judged by whether the rate of new cases in that group falls. This is what distinguishes preventive psychiatry from ordinary clinical psychiatry: its unit of analysis is the population, its outcome is an incidence rate, and its natural allies are epidemiology and public health rather than the consulting room.

The descriptor is also filed under preventive medicine in the MeSH tree, because the conceptual machinery is borrowed wholesale from general public health. The borrowing is not merely rhetorical: the levels of prevention, the logic of risk factors, and the arithmetic of incidence reduction all arrived in psychiatry from the epidemiology of physical disease, and much of the field's difficulty comes from the ways mental disorders fail to behave like the infectious diseases that logic was built for.

The Caplan Levels

The framework Caplan imported came from the preventive-medicine textbook of Hugh Leavell and E. Gurney Clark, who had arranged prevention along the natural history of a disease into three stages (#ref-leavell-1965). Primary prevention acts before the disorder appears, lowering incidence by removing causes or strengthening resistance in a still-healthy population. Secondary prevention acts once the disorder is present but early, shortening its course through prompt detection and treatment so that fewer cases become chronic. Tertiary prevention acts on established, often chronic disorder, limiting the disability and handicap it produces through rehabilitation. Caplan and Henry Grunebaum's 1967 review made primary prevention in particular the ambitious frontier of the field, cataloguing the ways a community might reduce the supply of new cases by acting on the social and developmental conditions that breed them (#ref-caplan-grunebaum-1967).

The appeal of the scheme is that it locates every prevention effort on a single timeline running from a healthy population, through early illness, to chronic disability, and assigns each effort a distinct target: incidence for primary, prevalence and chronicity for secondary, disability for tertiary. Its weakness, which the field spent the next thirty years confronting, is that the boundary between before and early illness is far blurrier for a mental disorder than for a fracture or an infection. Depression and psychosis emerge gradually, often through prodromal states that are neither clearly well nor clearly ill, and the stage-based scheme has no comfortable place for an intervention aimed at a person who is at high risk but not yet a case.

Gordon's scheme sorts prevention by the risk of the group targeted, not the stage of disease. Universal reaches everyone, selective a subgroup with a risk marker, and indicated individuals already showing minimal signs. Sort each case.

The three categories run from lowest to highest baseline risk. As the risk of the targeted group rises, the number needed to treat falls — which is why a more intensive, costlier intervention can be justified for an indicated group but not for a universal one.

The Risk-Based Reclassification

In 1983 Robert Gordon proposed replacing the disease-stage scheme with one built on who is targeted rather than what stage the disease has reached (#ref-gordon-1983). His three categories are defined by the risk profile of the recipient group. Universal prevention is offered to a whole population regardless of individual risk — a school-wide resilience curriculum, say — and is justified when the intervention is cheap, safe, and broadly beneficial. Selective prevention targets subgroups whose risk is elevated by a known marker, such as the bereaved or the children of a parent with a mental illness. Indicated prevention targets individuals who already show minimal but detectable early signs of disorder, or a biological marker of vulnerability, without yet meeting diagnostic criteria.

The Institute of Medicine's landmark 1994 report Reducing Risks for Mental Disorders, edited by Patricia Mrazek and Robert Haggerty, adopted Gordon's scheme with one decisive restriction: it reserved the word prevention for interventions delivered before a disorder's onset, placing treatment and maintenance in separate categories of a mental health intervention spectrum (#ref-mrazek-1994). Ricardo Muñoz and colleagues, summarizing the report, stressed why this mattered — a risk-based scheme lets prevention science borrow the full apparatus of risk-factor epidemiology, sorting causes into those that can be changed and estimating how much incidence would fall if they were (#ref-munoz-1996). The reclassification was not cosmetic. It shifted the field's question from at what stage do we intervene? to in whom, at what baseline risk, does an intervention repay its cost? — and that question, as the next section shows, is answered by arithmetic.

Table 1

Two Classifications of Prevention: Caplan's Disease-Stage Levels and Gordon's Risk-Based Categories

Note. The two schemes answer different questions. Caplan's levels sort an effort by the stage the disorder has reached (Caplan, 1964); Gordon's categories sort it by the risk profile of the group receiving it (Gordon, 1983). The IOM restricted prevention to the three risk-based categories delivered before onset (Mrazek & Haggerty, 1994), which is why all three carry incidence as their outcome.

The Statistical Problem

The hardest fact about preventive psychiatry is that most mental disorders are, over any given follow-up window, uncommon events, and preventing an uncommon event is statistically expensive to demonstrate. Pim Cuijpers made the point decisively in 2003: a trial that aims to show a reduction in the incidence of new cases needs far larger samples than a treatment trial, because incidence over a year or two is low and the number of new cases available to be prevented is correspondingly small (#ref-cuijpers-2003). A program that genuinely cuts risk by a quarter may still fail to reach statistical significance in any realistically sized trial, so that a real preventive effect is mistaken for no effect.

The arithmetic that governs this is the number needed to treat (NNT) — the count of people who must receive the intervention to prevent one new case. NNT is the reciprocal of the absolute risk reduction, and absolute risk reduction is the product of the baseline incidence and the relative reduction the program achieves. Because baseline incidence sits inside that product, the same program yields a very different NNT depending on whom it is offered to: cheap in a high-risk group where new cases are common, ruinously expensive in a low-risk general population where they are rare. This is the quantitative reason the field moved toward selective and indicated prevention. Thomas Insel, then directing the U.S. National Institute of Mental Health, made lowering the burden of mental illness through earlier, better-targeted intervention an explicit strategic priority, framing prevention as a research problem of finding the risk markers precise enough to make targeting pay (#ref-insel-2009). The worked example below makes the dependence of NNT on baseline risk concrete.

A prevention program with a fixed efficacy (its relative risk reduction) pays off very differently depending on the baseline risk of the group it reaches. Watch the number needed to treat as you move the controls.

Hold the relative reduction fixed and raise the baseline: the number needed to treat collapses. At a 25% reduction, a 4% general population needs 100 people per prevented case, but a 30% at-risk group needs only about 13 — the same program, roughly eight times more efficient where risk is concentrated.

Clinical Staging and the At-Risk State

The most influential modern answer to Caplan's blurred boundary between before and early is the clinical staging model, which treats mental disorder as a progression through defined stages — from asymptomatic risk, through a subthreshold at-risk mental state, to a first episode and then chronic or recurrent illness — and matches the intervention to the stage. Patrick McGorry and colleagues have argued that the early, still-undifferentiated stages are transdiagnostic: a young person in an at-risk mental state may be on a path toward psychosis, mood disorder, or several other outcomes, and forcing an early diagnostic label onto that state is both inaccurate and needlessly alarming (#ref-mcgorry-2018). Staging reframes indicated prevention as intervention at an early, non-specific stage rather than as the prediction of a specific future disease.

The staging model gave indicated prevention a workable clinical object — the identified at-risk individual — but it inherited the statistical problem in a new form. Only a minority of those who meet at-risk criteria will transition to a full disorder, so even a group selected for elevated risk still has a base rate low enough that most of them would not have become cases regardless. The demo below shows how a screening step that concentrates risk raises the yield of a subsequent intervention, and why even good concentration leaves most of the at-risk group as people who would never have progressed.

Indicated prevention targets an identified at-risk group, but only a minority go on to a full disorder. Each square is one of 100 people meeting at-risk criteria. Set the transition rate and see how many would have become cases — and how many would not.

Even at the upper end of reported clinical-high-risk transition rates, most of the group — here 78 of 100 — never develops the disorder. They stand to gain nothing from a preventive intervention while bearing its cost or stigma, which is the ethical core of the caution against over-treating an at-risk label.

Worked Example

Consider one prevention program with a fixed efficacy — it reduces each recipient's relative risk of onset by 25% — offered to two different groups.

Offered universally, to a general population whose baseline incidence of the disorder over the follow-up is 4%, the intervention lowers incidence to:

4% × (1 − 0.25) = 4% × 0.75 = 3.0%.

The absolute risk reduction is 4% − 3% = 1 percentage point, so the number needed to treat is:

NNT = 1 / 0.01 = 100.

One hundred people must receive the program to prevent a single new case. Offered indicated, to a screened at-risk group whose baseline incidence is 30%, the same 25% relative reduction lowers incidence to:

30% × 0.75 = 22.5%,

an absolute risk reduction of 7.5 percentage points, and:

NNT = 1 / 0.075 ≈ 13.

The identical program is roughly eight times more efficient in the high-risk group — the ratio of the two NNTs is 100 / 13.3 ≈ 7.5, which is exactly the ratio of the two baseline rates (30 / 4 = 7.5), because relative risk reduction is held constant. This is the mathematical core of preventive psychiatry: efficacy alone does not determine whether prevention is worthwhile; efficacy multiplied by baseline risk does. It is also why universal prevention survives only when the intervention is so cheap and safe that a large NNT is acceptable, while indicated prevention can justify a more intensive, costlier intervention against a much smaller NNT. The demo for this section lets the baseline rate and the relative reduction be varied so the NNT can be read off any combination.

Discussion

Preventive psychiatry has always lived with a gap between the strength of its promise and the strength of its evidence. The promise is enormous: mental disorders are among the largest contributors to global disability, they typically begin in adolescence and early adulthood, and an intervention that shifted incidence even modestly would spare an immense burden. The World Health Organization's 2004 review argued on exactly these grounds that prevention of mental disorders should be a public-health priority, and that enough efficacious interventions already existed to justify policy action (#ref-who-2004). Shekhar Saxena and colleagues, writing the same year, pressed the practical corollary that the knowledge base, while incomplete, was sufficient to begin implementing and scaling programs rather than waiting for certainty (#ref-saxena-2006).

The counterweight is the evidentiary caution that Cuijpers's statistical argument demands. A field that intervenes on healthy or minimally symptomatic people carries a special obligation not to overstate benefit, because the people it enrols are, by construction, mostly those who would never have developed the disorder and who therefore stand to gain nothing while bearing whatever cost, stigma, or side effect the program imposes. The reclassification toward indicated prevention is partly an ethical response to this: concentrating the intervention on those at genuinely high risk improves not only the NNT but the balance of benefit and harm. The tension between the WHO's we know enough to act and the methodologist's we rarely have the power to be sure is not resolved; it is the productive disagreement that structures the field.

Current Directions

The most active contemporary program tries to make indicated prevention in young people both rigorous and scalable. Paolo Fusar-Poli and colleagues set out a detailed blueprint for preventive psychiatry built on detecting and treating the clinical high-risk state for psychosis and other disorders, arguing that improved risk prediction, earlier detection, and stage-matched intervention could move the field from aspiration to routine service (#ref-fusarpoli-2021). Celso Arango and colleagues reviewed the evidence across disorders and concluded that preventive strategies with a reasonable evidence base already exist — particularly those acting on early developmental and social risk factors — while cautioning that the effect sizes are modest and the highest-risk windows are early (#ref-arango-2018). Both lines of work accept the statistical lesson of the previous two decades: prevention succeeds by targeting, and targeting succeeds only as well as its risk markers.

A complementary strand works upstream of any clinical service, on the population's own capacity to recognize and respond to mental disorder. Anthony Jorm's concept of mental health literacy — the public's knowledge and beliefs about mental disorders that aid their recognition, management, and prevention — reframes part of primary and secondary prevention as an educational task, improving the odds that early problems are noticed and help is sought before a disorder consolidates (#ref-jorm-2012). Between better risk prediction at the clinical high-risk end and better literacy at the population end, the field is converging on a division of labour that Caplan's original three levels anticipated but could not yet operationalize.

Common Misconceptions

Prevention means the same thing as early treatment.

It does not. Since the 1994 Institute of Medicine report, prevention has been reserved for interventions delivered before a disorder's onset; detecting and treating an established early disorder is treatment, filed in a separate band of the intervention spectrum (#ref-mrazek-1994).

Primary, secondary, and tertiary are the current standard categories.

They are the older Caplan/Leavell scheme, organized by disease stage. Modern prevention science more often uses Gordon's universal, selective, and indicated categories, organized by the risk profile of the group targeted (#ref-gordon-1983).

A program that fails to lower incidence in a trial has been shown not to work.

Often it has only been shown that the trial lacked the statistical power to detect an effect, because incidence is low and new cases are few; a real preventive effect can hide inside a non-significant result (#ref-cuijpers-2003).

A high-risk label predicts that the person will develop the disorder.

Most people who meet at-risk or clinical high-risk criteria never transition to a full disorder; the criteria concentrate risk without determining outcome, which is why the early stages are treated as transdiagnostic rather than as a specific diagnosis-in-waiting (#ref-mcgorry-2018).

Glossary

Absolute risk reduction.

The arithmetic difference between the incidence in an untreated group and the incidence in the treated group; the quantity whose reciprocal is the number needed to treat.

At-risk mental state.

A subthreshold condition, often in a young person, marked by minimal or attenuated symptoms that signal elevated but not certain risk of a full disorder.

Clinical high-risk state.

An operationalized at-risk mental state, typically for psychosis, defined by specified attenuated or brief symptoms and used to select individuals for indicated prevention.

Clinical staging.

A model that arranges mental disorder into progressive stages from asymptomatic risk to chronic illness and matches each stage to a stage-appropriate intervention.

Community psychiatry.

The branch of psychiatry concerned with the mental health of defined populations; the parent kind under which MeSH files preventive psychiatry.

Incidence.

The rate of new cases of a disorder arising in a population over a defined period; the primary outcome a prevention program aims to lower.

Indicated prevention.

Gordon's category of prevention aimed at individuals already showing minimal detectable signs of a disorder without yet meeting diagnostic criteria.

Mental health literacy.

Jorm's term for the knowledge and beliefs about mental disorders that aid a community in their recognition, management, and prevention.

Number needed to treat.

The count of people who must receive an intervention to prevent one adverse outcome; the reciprocal of the absolute risk reduction.

Primary prevention.

Caplan's level acting before a disorder appears, aiming to lower incidence by removing causes or strengthening resistance in a healthy population.

Relative risk reduction.

The proportional decrease in risk an intervention produces, independent of the baseline rate; efficacy expressed as a fraction of the untreated risk.

Secondary prevention.

Caplan's level acting on early, established disorder to shorten its course through prompt detection and treatment.

Selective prevention.

Gordon's category of prevention aimed at subgroups whose risk is raised by a known marker, such as bereavement or a family history.

Tertiary prevention.

Caplan's level acting on chronic, established disorder to limit the disability and handicap it produces, largely through rehabilitation.

Universal prevention.

Gordon's category of prevention offered to a whole population regardless of individual risk, justified when the intervention is cheap, safe, and broadly beneficial.

Key Researchers

Celso Arango (living). Psychiatrist at Hospital General Universitario Gregorio Marañón, Madrid; lead author of a 2018 review synthesizing preventive strategies across mental disorders and their evidence base. ORCID - Google Scholar - Wikipedia

Gerald Caplan (1917-2008). Founder of the field; his 1964 Principles of Preventive Psychiatry imported the primary/secondary/tertiary scheme into psychiatry and, with Henry Grunebaum, made primary prevention its central ambition. Worked at Harvard and later Hadassah-Hebrew University. Library of Congress authority - Obituary (American Psychologist)

Pim Cuijpers (living). Clinical psychologist at Vrije Universiteit Amsterdam whose 2003 analysis showed that most prevention trials lack the statistical power to detect reductions in disorder incidence, reshaping how the field designs and reads its evidence. ORCID - Google Scholar - Faculty Page

Paolo Fusar-Poli (living). Professor of preventive psychiatry at King's College London; lead author of a 2021 blueprint for detecting and treating the clinical high-risk state to prevent psychosis and other disorders in young people. ORCID - Google Scholar - Wikipedia

Anthony F. Jorm (living). Mental health researcher at the University of Melbourne who coined mental health literacy, recasting part of prevention as improving the public's capacity to recognize and respond to mental disorder. ORCID - Wikipedia - Faculty Page

Patrick McGorry (living). Psychiatrist at the University of Melbourne and Orygen; pioneer of the clinical staging model and early intervention in youth mental health, and advocate of a transdiagnostic view of the early at-risk state. ORCID - Google Scholar - Wikipedia

Frequently Asked Questions

What is preventive psychiatry? It is the branch of psychiatry that aims to stop mental disorders before they begin, catch them early, and limit the disability they cause, applying to mental health the public-health logic of prevention. MeSH classifies it as a kind of community psychiatry because it acts on populations rather than one patient at a time.

What are the primary, secondary, and tertiary levels of prevention? They are Gerald Caplan's three levels, borrowed from preventive medicine: primary prevention acts before a disorder appears to lower incidence, secondary prevention acts on early disorder to shorten its course, and tertiary prevention acts on chronic disorder to limit disability.

How is that different from universal, selective, and indicated prevention? The universal/selective/indicated scheme, proposed by Robert Gordon in 1983 and adopted by the Institute of Medicine in 1994, classifies prevention by the risk profile of the group targeted rather than by the stage of disease. It reserves the word prevention for interventions delivered before onset.

Why is it so hard to prove a prevention program works? Because most mental disorders are uncommon over any given follow-up period, so few new cases are available to be prevented. A trial then needs very large samples to detect a reduction in incidence, and a genuinely effective program can fail to reach statistical significance.

What is the number needed to treat in prevention? It is the number of people who must receive the intervention to prevent one new case, calculated as one divided by the absolute risk reduction. Because absolute risk reduction depends on baseline risk, the same program has a much lower number needed to treat in a high-risk group.

What is the clinical high-risk state? It is an operationalized early stage, often for psychosis, defined by minimal or attenuated symptoms that mark elevated risk. It gives indicated prevention a concrete target, though only a minority of those identified go on to develop a full disorder.

Does preventive psychiatry have a parent category in MeSH? Yes. MeSH files preventive psychiatry as a narrower kind of community psychiatry, and also under preventive medicine, reflecting that its methods are borrowed from public health and applied to defined populations.

What is mental health literacy? It is Anthony Jorm's term for the public's knowledge and beliefs about mental disorders that help a community recognize, manage, and prevent them, treating part of prevention as an educational task rather than only a clinical one.

References