Clinical Reasoning

Clinical reasoning is the thinking a clinician does when meeting a patient: gathering the history, examining the body, ordering tests, and deciding what is wrong and what to do.

15 frameworks, from 400 BCE to today · Clinical Medicine · Medicine & Health

still a working view ended · each leads to its entry below

Overview

Clinical reasoning is the thinking a clinician does when meeting a patient: gathering the history, examining the body, ordering tests, and deciding what is wrong and what to do. It is not a body of knowledge but a set of mental moves—pattern recognition, hypothesis testing, causal explanation, probability estimation—used under uncertainty and time pressure. Its history matters because each era's method decided not only what doctors concluded but what counted as a good reason: the humoral physician, the anatomist at the autopsy table, and the modern Bayesian all reasoned in ways that seemed self-evident to them and look strange to us. The story is one of methods invented to solve specific failures of the methods before them.

Humoral reasoning and the Hippocratic-Galenic method

The oldest surviving clinical method is the Hippocratic, recorded in the Greek corpus attributed to Hippocrates and his followers (5th–4th centuries BCE). Its reasoning was prognostic and holistic. The physician took a full historia—the patient's constitution, regimen, season, weather, and the "crises" of the illness—and explained symptoms as imbalance among the four humours (blood, phlegm, yellow bile, black bile). The Epidemics case histories show the method in action: careful day-by-day observation, but explanation by constitutional type and environment rather than by organ.

Galen (2nd century CE) systematized this into a deductive edifice12. Disease was a state of humoral dyscrasia; treatment opposed it (contraria contrariis). The method's strength was that it explained everything and integrated diet, temperament, and environment; its weakness was that its explanations could not be checked against the body after death. Galenic reasoning dominated learned medicine in Europe and the Islamic world for over a millennium, and its logic—reason from general principles to the individual case—remained the default long after its physiology was abandoned.

The anatomoclinical turn

The decisive break came in Paris in the early nineteenth century. The clinical method of the Paris school, associated with René Laennec, Gaspard Laurent Bayle, and later Pierre Charles Alexandre Louis, joined bedside observation to autopsy. The problem it solved was the humoral method's untestability: if symptoms came from organ lesions, then the lesion found at autopsy could be correlated with the symptoms observed in life. Laennec's invention of the stethoscope (1816) and his De l'auscultation médicale (1819) made internal lesions audible; the méthode anatomoclinique made the case the unit of reasoning, with the "fundamental lesion" as the explanation of the clinical picture.

This changed what counted as a reason. The clinician now reasoned from signs to pathological anatomy—a correlation later formalized as the "anatomoclinical" method. Louis's numerical method (his studies of bloodletting in the 1820s–30s) added a second strand: counting outcomes across comparable cases rather than trusting the individual impression. The anatomoclinical method spread with the rise of the hospital and the autopsy, and it remains the backbone of hospital diagnosis: the differential diagnosis is still, in essence, a list of candidate lesions.

Laboratory and pathophysiological reasoning

From the mid-nineteenth century, laboratory medicine—chemistry, microscopy, bacteriology—added a third layer. Rudolf Virchow's cellular pathology (1858) relocated disease to the cell, and the discovery of specific causative agents (anthrax bacillus, 1876; tubercle bacillus, 1882) made causal reasoning possible. The clinician could now reason from a laboratory finding to a mechanism to a diagnosis, and from mechanism to treatment.

The twentieth century built this into pathophysiological reasoning: explain the patient's findings by a disturbed physiological mechanism, then correct the mechanism. This is the reasoning taught in the classic internal-medicine tradition—reason from the sodium and water derangement to the volume state to the cause. It is powerful and satisfying, but it has a known failure mode: a plausible mechanism can make a wrong diagnosis feel right, and the mechanism can be elaborated without ever being tested against outcome.

Osler and the bedside apprenticeship

William Osler, at Johns Hopkins and later Oxford, made the bedside the center of clinical learning. His Principles and Practice of Medicine (1892) and his ward-round teaching established an apprenticeship model in which the student learns reasoning by doing it under supervision, at the bedside, with the patient as the text. Osler's method was observational and integrative: history and physical examination first, laboratory second, and the patient as a person throughout. The residency system he helped shape at Hopkins exported this model worldwide, and it remains the dominant pedagogy of clinical reasoning—reasoning learned as a craft, by imitation and correction, rather than from a textbook of logic.

The hypothetico-deductive model and its critics

The mid-twentieth century brought attempts to describe what clinicians actually do. The dominant model was hypothetico-deductive reasoning: the clinician generates a small set of early hypotheses from the first cues, then seeks data to confirm or refute them. This model, elaborated in the 1970s by researchers such as Arthur Elstein and colleagues studying physicians' problem-solving, described clinical thinking as scientific method applied to the case.

But empirical work complicated it. Studies of expert and novice diagnosticians found that experts did not generate more hypotheses or test them more rigorously; they recognized the answer quickly. This pointed to pattern recognition—non-analytic reasoning in which a familiar configuration of findings triggers a diagnosis directly, as a face is recognized34. The dual-process account, developed in psychology by Keith Stanovich and Richard West (2000) and applied to medicine by researchers including Pat Croskerry, described two modes: fast, automatic, intuitive (System 1) and slow, deliberate, analytic (System 2). The cognitive continuum theory of Kenneth Hammond (1980s) had already argued that clinical tasks lie along a continuum from intuitive to analytic, and that the right mode depends on the task's structure—a view that softened the dual-process dichotomy.

Illness scripts and the structure of expertise

A further refinement came from script theory. The illness script model, associated with Georges Bordage and later with Henk Schmidt and colleagues, proposed that expert clinicians store knowledge as scripts: enabling conditions (risk factors, context), the pathophysiology, and the clinical consequences (signs, symptoms, time course). Diagnosis is then a matter of matching the patient's presentation to a stored script. This explained why experts are fast and why their knowledge is organized around cases rather than lists, and it gave medical education a concrete target: build scripts, not just facts.

Situated cognition added that reasoning is not purely internal5. It is shaped by the setting—the ward, the clinic, the available tests, the team, the culture of the institution5. What a clinician notices and how they reason depends on where they stand, a point that connects clinical reasoning to the sociology of medical work and to the design of clinical environments.

Error, debiasing, and metacognition

As the cognitive account matured, attention turned to failure. The diagnostic error literature—given prominence by the Institute of Medicine's To Err Is Human (1999) and later by Mark Graber and colleagues—documented that errors are common and often cognitive rather than knowledge-based. The catalog of biases (anchoring, availability, premature closure, confirmation bias) became standard teaching. Metacognitive debiasing—strategies such as considering alternatives, deliberately seeking disconfirming evidence, and using checklists—was proposed as a corrective. The evidence that any single debiasing strategy reliably works remains mixed, which is itself an important finding: knowing about a bias does not reliably prevent it.

Bayesian and decision-analytic diagnosis

A parallel, quantitative tradition treated diagnosis as probability. The Bayesian framework—updating a prior probability with the likelihood ratio of a test result—gave a formal account of how findings should change belief. The decision-analytic approach, developed from the 1960s onward by researchers such as Lee Lusted and later in the clinical decision-making literature, extended this to treatment choice by weighing probabilities against utilities. Its practical legacy includes the likelihood ratio, the nomogram, and the modern diagnostic decision rule. Its limitation is that clinicians rarely have well-calibrated priors and do not naturally reason in odds; the framework describes an ideal more than a common practice.

The present

Today these strands coexist. Medical schools teach hypothetico-deductive reasoning and illness scripts; residency training still runs on Oslerian apprenticeship; hospital diagnosis still rests on the anatomoclinical differential; and evidence-based medicine supplies the probabilistic and outcome-oriented layer. The open questions are the ones the history has exposed: how to teach intuition without teaching its errors, whether debiasing can be made to work, how to reason well when the evidence is thin or the patient is atypical, and how to integrate the fast and slow modes under real time pressure. The field's history suggests that each new method answered a specific failure of the last—and that the next method will be invented at the bedside where the failures show up.

References

  1. Galenic corpus, Wikipedia
  2. Galen, Wikipedia
  3. Medical diagnosis (Pattern recognition), Wikipedia
  4. Pattern recognition (psychology), Wikipedia
  5. Situated cognition, Wikipedia

The frameworks of Clinical Reasoning

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 400 BCE-1800

    Hippocratic-Galenic humoral reasoning

    How do humours and environment explain illness?

    Fly to its tower in Nousfarer

  2. 1800-Present

    Anatomoclinical method

    Which organ lesion explains these signs?

    Fly to its tower in Nousfarer

  3. 1820-1900

    Numerical method

    Do outcomes differ across counted cases?

    Fly to its tower in Nousfarer

  4. 1850-Present

    Pathophysiological reasoning

    What mechanism produces these findings?

    Fly to its tower in Nousfarer

  5. 1890-Present

    Oslerian bedside apprenticeship

    How is reasoning learned at the bedside?

    Fly to its tower in Nousfarer

  6. 1960-Present
  7. 1960-Present

    Decision analysis

    Which action best weighs probabilities and utilities?

    Fly to its tower in Nousfarer

  8. 1970-Present

    Hypothetico-deductive model

    Do clinicians test hypotheses to diagnose?

    Fly to its tower in Nousfarer

  9. 1980-Present

    Cognitive continuum theory

    When should judgment be intuitive or analytic?

    Fly to its tower in Nousfarer

  10. 1980-2005

    Non-analytical pattern recognition

    How does a familiar picture trigger diagnosis?

    Fly to its tower in Nousfarer

  11. 1980-Present

    Pattern recognition

    How do experts recognize a diagnosis quickly?

    Fly to its tower in Nousfarer

  12. 1990-Present

    Illness scripts

    How is clinical knowledge stored as scripts?

    Fly to its tower in Nousfarer

  13. 1990-Present

    Situated cognition

    How does setting shape clinical reasoning?

    Fly to its tower in Nousfarer

  14. 1999-Present

    Metacognitive debiasing

    How can clinicians avoid cognitive bias?

    Fly to its tower in Nousfarer

  15. 2000-Present

    Dual-process theory

    When does fast or slow thinking govern judgment?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Hippocratic-Galenic humoral reasoningAnatomoclinical methodNumerical methodPathophysiological reasoningOslerian bedside apprenticeshipBayesian and decision-analytic diagno…Decision analysisHypothetico-deductive modelCognitive continuum theoryNon-analytical pattern recognitionPattern recognitionIllness scriptsSituated cognitionMetacognitive debiasingDual-process theory

Beyond the field

Where Clinical Reasoning's ideas came from in other fields, and what they led to there: the strongest 12 of 70 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Clinical Reasoning is a planet in Nousfarer

Its city has a tower for each of these 15 frameworks. Land, walk in, and climb one concept at a time.

Land on Clinical Reasoning

Related fields

Fields its ideas are linked to

More in Clinical Medicine

See where Clinical Reasoning sits on the Map of Human Knowledge