This chapter draws on Engel's original framing,¹ subsequent operationalizations in the major textbooks of psychiatry — Kaplan and Sadock,² the American Psychiatric Publishing Textbook of Psychiatry,⁴ and dedicated formulation texts including Sperry's Case Conceptualization⁵ and the Maudsley-derived clinical formulation literature.⁶ Guideline-level coverage of formulation is sparse — the framework is pedagogical rather than evidence-graded — so most citations point to textbooks and conceptual papers rather than randomized trials. Search terms across PubMed and Google Scholar (literature pass dated 2026-05-10) included "biopsychosocial formulation," "case formulation psychiatry," "4P model formulation," and "Engel biopsychosocial." Empirical work on inter-rater reliability and clinical utility is summarized in the limitations section.^7-8

Engel proposed the biopsychosocial model in 1977 as a corrective to a biomedical orthodoxy that, in his view, had become reductionist.¹ His argument was epistemological. The biomedical model treats disease as deviation from biological norms and treats psychosocial variables as either irrelevant or as residual noise to be controlled away. Engel countered that illness — the lived experience of being unwell — is shaped at biological, psychological, and social levels simultaneously, and that any of those levels can serve as the primary lever for change. The model rests on systems theory. A person is a hierarchy of nested systems — molecule, cell, organ, person, dyad, family, community — and a perturbation at one level propagates through the others. A serotonin transporter polymorphism does not cause depression in a vacuum; it changes the probability of depression given a particular childhood, a particular marriage, a particular job loss. The same logic runs in reverse: chronic social defeat changes glucocorticoid signaling and hippocampal volume.⁹ The biopsychosocial model is, at heart, a commitment to taking that bidirectional traffic seriously when reasoning about a single patient. Two clarifications are worth making early. First, the model is not a theory of any specific disorder. It is a framework for organizing what is known and unknown about a person. Second, the model is not anti-biological. A formulation that lists "genetic loading for bipolar disorder" under predisposing biological factors and "lithium response" under protective biological factors is doing exactly what Engel intended.

The standard implementation is a four-by-three grid: four temporal categories (the four Ps) crossed with three domains (biological, psychological, social).³ Each cell is a question the clinician must at least try to answer. Empty cells are findings, not failures — they tell the team where information is missing.

Predisposing factors

Predisposing factors are the longstanding vulnerabilities that loaded the dice before the index episode. They answer the question: why was this person, of all people, susceptible? They are typically distal in time — childhood, adolescence, family history — and stable. A useful test: would the factor still be true if the current crisis resolved tomorrow? If yes, it is probably predisposing rather than precipitating. Biological predisposing factors include family history of psychiatric illness (with attention to first-degree relatives and the specific phenotype, not just "depression runs in the family"), perinatal complications, neurodevelopmental history, head injury, chronic medical illness with known psychiatric sequelae (epilepsy, multiple sclerosis, hypothyroidism), and substance exposure during critical periods. Genetic risk should be specific where possible — a sibling with completed suicide on lithium-responsive bipolar I disorder is a different predisposing factor than a maternal aunt "with nerves." Psychological predisposing factors are the durable patterns of thinking, feeling, and relating that the person brings into adulthood. Early adverse experiences (the Adverse Childhood Experiences literature is the standard reference here¹⁰), Attachment style, temperament, cognitive schemas, and personality structure all belong here. Resist the temptation to file every personality trait under "perpetuating." A schema laid down at age six that the world is dangerous is predisposing; the way it now drives avoidance of treatment is perpetuating. Social predisposing factors include socioeconomic status of the family of origin, exposure to violence or discrimination, immigration history, educational trajectory, and cultural context. The point is not to compile a sociological profile. The point is to identify the social variables that meaningfully changed this person's risk for the presenting problem.

Precipitating factors

Precipitating factors are the proximal triggers — what tipped a vulnerable person into the current episode now rather than six months ago or six months from now. They are the answer to "why now?" and they are usually the easiest cell to fill, because patients tell the clinician about them in the first ten minutes. The discipline is in not stopping there. Biologically: a recent medication change, a new substance, an intercurrent medical illness, sleep deprivation from a new baby or shift work, postpartum hormonal shift, perimenopause. Psychologically: a loss, a humiliation, an anniversary reaction, a developmental transition (leaving home, retirement, an empty nest). Socially: job loss, relationship breakdown, financial stressor, legal trouble, immigration, bereavement. A common error is to treat the precipitant as the cause. The precipitant explains the timing; the predisposition explains the susceptibility. A divorce precipitates a depressive episode in someone whose mother had recurrent unipolar depression and whose attachment history primed them to experience abandonment as catastrophic. The divorce is not the formulation; it is one cell in it.

Perpetuating factors

Perpetuating factors maintain the episode once it has begun. They are the cell most directly tied to the treatment plan, because they are usually the most modifiable. If the predisposition was set decades ago and the precipitant is already in the past, the perpetuating factors are what the clinician can actually change next week. Biological perpetuators include ongoing substance use, untreated medical comorbidity (sleep apnea in treatment-resistant depression is the canonical example), poor adherence, subtherapeutic dosing, drug interactions reducing exposure, and chronic pain. Psychological perpetuators include rumination, avoidance, catastrophic interpretations of symptoms, hopelessness about treatment, secondary gain, and unprocessed trauma. Social perpetuators include ongoing interpersonal conflict, high expressed emotion in the household,¹¹ housing instability, unemployment, isolation, and ongoing exposure to the original stressor.

Protective factors

Protective factors are the resources that have buffered the person so far and that will scaffold recovery. They are routinely underweighted in formulations written by trainees, who default to a deficit-focused account because that is how case presentations are typically structured. A formulation without protective factors is a formulation that cannot explain why the patient is still alive and in the room. Biological protective factors include physical health, absence of substance use, prior medication response, and lack of family history. Psychological protective factors include intelligence, psychological mindedness, capacity for therapeutic alliance, prior response to therapy, and specific coping strategies that have worked. Social protective factors include a stable partner or family, employment, housing, faith community, cultural identity, and access to care. Protective factors are also where the cultural formulation often does its real work — a strong religious framework can be a perpetuating factor for shame in one patient and a protective factor for meaning-making in another.

The information that fills the grid does not arrive pre-sorted. The clinical interview yields a narrative; the formulation is what the clinician builds from it afterward. The questions below are organized by domain rather than by interview phase, on the assumption that the reader already knows how to take a psychiatric history. They are prompts for what to listen for, not a script.

Biological domain — questions to ask

• Family psychiatric history: which first- and second-degree relatives, what specific diagnoses, what treatment responses, completed or attempted suicide, psychiatric hospitalizations.
• Developmental history: pregnancy and delivery complications, milestones, early temperament, learning difficulties, head injuries, seizures.
• Medical history with psychiatric relevance: thyroid disease, autoimmune disease, neurological disease, sleep disorders, chronic pain, traumatic brain injury.
• Substance history with timing: age of first use, escalation, periods of abstinence and what changed during them, current use, withdrawal history.
• Treatment history: every psychiatric medication ever tried, dose achieved, duration, response, reason for discontinuation. Trainees routinely accept "the antidepressant didn't work" without establishing dose or duration.

Psychological domain — questions to ask

• Childhood narrative: who raised the patient, what the household was like, how affection and discipline were handled, any abuse or neglect.
• Attachment and relational history: pattern of close relationships, friendships, romantic partners, capacity to trust, characteristic conflicts.
• Self-concept: how the patient describes themselves when well, what they are proud of, what they are ashamed of, what they believe others think of them.
• Coping repertoire: how the patient has handled previous setbacks, what worked, what did not, what defenses are characteristic (intellectualization, splitting, somatization).
• Beliefs about the illness: what the patient thinks is wrong, what they think caused it, what they think will help, what they fear about treatment.

Social domain — questions to ask

• Current living situation: who lives in the household, the quality of those relationships, financial stability, housing security.
• Occupational history and current functioning: longest job held, reasons for transitions, current work or school, performance.
• Social network: who the patient would call at 3 a.m., frequency of contact, isolation versus enmeshment.
• Cultural context: language, religion, immigration history, identity, community ties, experience of discrimination, idioms of distress.
• Legal and forensic: current charges, history of incarceration, child protective involvement, civil litigation.

The grid is the worksheet. The written formulation is a paragraph — sometimes two — that synthesizes the worksheet into a coherent story about the patient. The synthesis step is where most trainees stall. They produce twelve bullet points and call it a formulation. A list is not a formulation. A formulation is an argument. A serviceable structure for the written paragraph runs as follows. Open with a one-sentence summary statement. Identify the patient (age, gender, relevant identity) and the clinical problem in syndromic terms, not just diagnostic ones. Example opening: "Ms. A is a 34-year-old woman with a first lifetime episode of severe melancholic depression with active suicidal ideation, presenting two months after the unexpected death of her mother." The summary statement does the diagnostic work; the rest of the paragraph does the formulation work. Move to predisposition. Name the most important predisposing factors across all three domains in a single sentence or two, integrated rather than listed. "Her vulnerability is shaped by a strong maternal-line family history of recurrent unipolar depression, an early loss of her father, and an enmeshed mother-daughter relationship in which her identity was tightly bound to her caregiving role." Move to precipitation. Name the trigger and tie it explicitly to the predisposition. "Her mother's death — the very loss her early experience left her least equipped to absorb — precipitated the current episode." Move to perpetuation. Name what is keeping her stuck and signpost what the treatment plan will target. "Severe insomnia, withdrawal from her support network, and a guilty rumination about whether she could have prevented her mother's death are perpetuating the episode and will be the initial targets of treatment." Close with protective factors and prognosis. "She has prior good response to sertraline, an intact marriage, and engaged in therapy in her twenties — protective factors that support a guarded but reasonable prognosis with combined pharmacologic and psychotherapeutic treatment." The whole paragraph should run roughly 150 to 250 words. Longer formulations are usually under-edited rather than thorough. The reader — whether attending, board examiner, or future clinician picking up the chart — should finish the paragraph able to predict roughly what the treatment plan will say.

The case below is composite and anonymized. It is not a real patient. It is included to show the framework end-to-end on a case with enough texture to exercise each cell.

Presenting problem

Mr. B is a 42-year-old married man, employed as an accountant, presenting to the outpatient clinic with a six-week history of low mood, early morning awakening, fatigue, anhedonia, poor concentration, and passive suicidal ideation without plan or intent. He meets criteria for a major depressive episode, moderate severity, without psychotic features. He has had two prior episodes, the first at age 27 and a second at age 34, both treated with fluoxetine to remission.

Relevant history

His father had recurrent depression and died by suicide at age 58. His mother is alive with generalized anxiety. He is the elder of two siblings; his sister has anorexia nervosa in long-term remission. He grew up in a household he describes as cold and high-achieving. His father, when depressed, would withdraw to his study for weeks at a time. He met his wife in graduate school and has been married for 14 years; they have one child, age 9. He drinks two glasses of wine most evenings, more on weekends, and has had no other substance use. Six weeks ago he was passed over for promotion to partner at his firm, an outcome he had been preparing for over five years. On examination he is a well-dressed man with psychomotor slowing, full but constricted affect, linear thought process, no perceptual disturbance, intact cognition, and limited insight into the link between his current symptoms and the recent occupational stressor. PHQ-9 is 18.

The grid

Synthesis paragraph

Mr. B is a 42-year-old married man with a third lifetime episode of moderate major depressive disorder, presenting six weeks after being passed over for partnership at his firm. His vulnerability is shaped by a strong paternal-line history of depression and completed suicide, an upbringing that linked self-worth tightly to professional achievement, and a characterological tendency to withdraw under distress that mirrors his father's depressive presentations. The denied promotion functioned as a precipitant precisely because it struck the contingency on which his self-esteem was constructed; the loss was not only of a role but of the organizing narrative of his adult life. The episode is now perpetuated by daily alcohol use disrupting sleep, ruminative self-blame, and withdrawal from his wife — the person most able to challenge his catastrophic appraisal. Two prior robust responses to fluoxetine, an intact marriage, demonstrated capacity for psychotherapeutic engagement, and the absence of medical comorbidity or active suicidal planning support a favorable short-term prognosis. The initial treatment plan will combine reinstatement of fluoxetine, alcohol reduction, structured behavioral activation, and a course of cognitive therapy targeting the achievement-contingent self-worth schema; longer-term work on identification with the paternal depressive pattern is a second-stage goal once the acute episode has remitted. Three observations about this synthesis are worth naming. First, every sentence ties a fact to an inference; the paragraph never simply lists. Second, the treatment plan flows out of the formulation rather than from the diagnosis alone — fluoxetine is chosen partly because the formulation identifies prior response, and behavioral activation is chosen because the perpetuating cell pinpoints withdrawal. Third, the formulation makes a falsifiable prediction: if the achievement-contingent schema is central, recurrence risk will be highest at future career inflection points and prophylaxis decisions should weight that.

The biopsychosocial model remains the dominant teaching framework in North American psychiatry residency training and is endorsed in the major textbooks.^2,4 Its strengths are real. It is comprehensive, in the sense that it forces consideration of three domains rather than defaulting to one. It is theoretically agnostic, accommodating biological, psychodynamic, cognitive-behavioral, systems-based, and cultural contributions without committing to any one. It is teachable — the four-by-three grid is a stable scaffold for trainees to populate as their interview matures. And it generates a treatment plan with multiple entry points, rather than confining the clinician to one modality.

Strengths

• Forces a multi-domain assessment in a field where uni-domain reasoning ("this is a serotonin problem," "this is a trauma case") is a persistent failure mode.
• Provides an organizing structure for case presentations, oral examinations, and chart documentation.
• Generates a treatment plan with multiple, modality-diverse leverage points.
• Accommodates cultural, spiritual, and developmental variables that purely descriptive diagnosis omits.
• Travels across patient populations and settings — emergency, inpatient, outpatient, consultation-liaison.

Limitations

The model has been substantively criticized, and a competent user should know the criticisms. First, the model has been called a slogan rather than a theory. Ghaemi and others have argued that the biopsychosocial model, as commonly used, lacks a clear method for weighting the three domains, leaves prioritization to individual judgment, and risks becoming an eclectic license to include whatever the clinician finds interesting.⁷ The criticism has force. Two clinicians presented with the same case can produce biopsychosocial formulations that look similar in form but generate different treatment plans, with no principled way to adjudicate between them. Second, empirical work on inter-rater reliability of psychiatric formulation is limited and the reliability that has been measured is modest.⁸ Formulations are clinical artifacts, not measurements; they are not validated in the way a diagnostic instrument is validated. The user should hold the formulation as a working hypothesis rather than a finding. Third, the four-by-three grid invites a tickbox approach. A trainee under time pressure can populate twelve cells with phrases that are technically true and clinically inert ("family history of depression," "recent stressor," "social isolation," "engaged with treatment") and produce a document that satisfies a supervisor without illuminating the patient. The grid is a worksheet, not a deliverable. Fourth, the model can underweight the biological in patients whose presentations are predominantly biological (delirium, neurodegenerative disease, autoimmune encephalitis, treatment-resistant melancholia) and underweight the social in patients whose presentations are predominantly social (refugee mental health, intimate partner violence, structural poverty). Even-handedness across three domains is not the same as accuracy.

Comparison with other frameworks

Several alternative or complementary frameworks deserve brief mention. The Perry-derived neurodevelopmental formulation places adverse childhood experience and developmental sequencing of brain regions at the center, and is the dominant lens in trauma-focused practice and child psychiatry.¹⁰ It is more theoretically committed than the biopsychosocial model and more powerful in trauma-loaded cases; it is less useful when developmental history is unremarkable. Cognitive-behavioral case conceptualization, in the Persons or Beck tradition, organizes the formulation around situations, automatic thoughts, schemas, and maintenance cycles.⁵ It generates a more directly testable treatment plan and is the standard within CBT, but it brackets biological and macro-social variables that the biopsychosocial model insists on. Psychodynamic formulation, in the Cabaniss-derived structure, organizes information around developmental conflict, defenses, object relations, and the self.¹³ It complements the biopsychosocial model in psychotherapy-oriented practice and is often the framework that fills out the psychological cell of the BPS grid in depth. The DSM-5-TR Cultural Formulation Interview is not a competing framework but a structured supplement, and should be integrated into the BPS grid rather than relegated to a separate section.¹² The Power Threat Meaning Framework, developed by the British Psychological Society as an explicit alternative to diagnosis-led practice, reframes distress in terms of operations of power, threats experienced, meanings made, and threat responses enacted.¹⁴ It is controversial, more useful in some settings than others, and worth knowing about as a contemporary critique of medicalized formulation.

Three live debates are worth knowing. Whether the biopsychosocial model is a model or a slogan: addressed above. Ghaemi's critique has been influential and has prompted the development of more theoretically committed alternatives.⁷ Defenders argue that the model's flexibility is a feature, not a bug, and that the alternative — uni-domain reductionism — is empirically worse. Whether formulation should drive treatment: there is no high-quality evidence that patients formulated by a biopsychosocial framework have better outcomes than patients managed on the basis of diagnosis and guideline alone. The argument for formulation rests on clinical utility, communicability, and educational value rather than on outcome data. The user should be honest about this. How to weight the three domains: there is no algorithm. In emergency and consultation-liaison settings, biological dominates by triage logic. In long-term outpatient psychotherapy, psychological often dominates. In community mental health, social often dominates. The clinician's task is to weight the domains in proportion to where the leverage and the threats actually are for this patient, not to weight them equally on principle.

• The biopsychosocial model was articulated by George Engel in a 1977 paper in Science, framed as a corrective to biomedical reductionism.
• The standard implementation is a four-by-three grid: predisposing, precipitating, perpetuating, protective × biological, psychological, social.
• Predisposing factors are distal and stable; precipitating factors answer "why now?"; perpetuating factors maintain the episode and are usually the most modifiable; protective factors are routinely underweighted by trainees.
• A formulation is a hypothesis, not a verdict. It is a paragraph, not a list of bullets, and it should make a falsifiable prediction.
• The synthesized written formulation runs roughly 150-250 words, opens with a one-sentence diagnostic summary statement, and closes with prognosis and protective factors.
• The model is theoretically agnostic and accommodates psychodynamic, cognitive-behavioral, cultural, and systems-based content within its grid.
• The major published critique (Ghaemi) argues the model lacks a method for weighting domains and risks becoming eclectic license. Inter-rater reliability of psychiatric formulation is modest.
• The Cultural Formulation Interview in DSM-5-TR is a structured supplement to be integrated into the grid, not a separate document.
• The model does not replace differential diagnosis; medical and substance-induced contributors must be ruled out independently.
• High-expressed-emotion family environments are a classic perpetuating social factor, particularly in schizophrenia and recurrent mood disorders.

No external funding. No conflicts of interest declared. Drafted with Claude (Anthropic) and human-edited by the Psychpedia editorial team. Peer-review status: pending.