{
  "id": "pattern-differentiation-tcm",
  "title": "Pattern Differentiation and Treatment",
  "lang": "en",
  "source_lang": "en",
  "summary": "Pattern differentiation and treatment is the traditional Chinese medical process of classifying clinical presentations and selecting treatment according to those classifications.",
  "category": "health",
  "subcategory": "medicine",
  "type": "concept",
  "aliases": [
    "Pattern identification and treatment",
    "syndrome differentiation and treatment",
    "treatment based on pattern differentiation",
    "bianzheng lunzhi"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/pattern-differentiation-tcm/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/pattern-differentiation-tcm/",
      "json": "https://aiwiki.page/zh/pattern-differentiation-tcm.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Pattern differentiation and treatment, commonly rendered in romanized Chinese as bianzheng lunzhi, is a central method of clinical reasoning in traditional Chinese medicine (TCM). It consists of interpreting a person’s symptoms, observable signs, and circumstances as a traditional diagnostic pattern, then using that pattern to determine a therapeutic approach. A pattern is understood within TCM as a configuration of disturbances at a particular stage of illness, rather than simply the name of a disease. The method draws on concepts such as qi, yin and yang, and the functional systems known as zang-fu.\n\nIts significance lies both in its historical role in Chinese medicine and in its continuing use in clinical practice, professional education, research, and health classification. The approach can assign different patterns to people with the same biomedical diagnosis, or identify a similar pattern across different diseases. These traditional classifications are not equivalent to established anatomical lesions, laboratory abnormalities, or biological mechanisms. Describing how the system operates therefore requires distinguishing its internal explanatory framework from independently demonstrated diagnostic accuracy, treatment effectiveness, and safety."
    },
    {
      "heading": "Meaning of a pattern and its relationship to disease",
      "text": "The Chinese expression combines bianzheng, identifying or differentiating a pattern, with lunzhi, determining treatment through clinical reasoning. English translations vary, particularly between “pattern” and “syndrome.” The latter can cause confusion because syndrome in contemporary medicine usually denotes an associated group of signs and symptoms without necessarily specifying its cause. A TCM pattern additionally incorporates traditional interpretations of location, pathogenic influence, functional disturbance, and the relationship between the person’s condition and presumed disease-producing factors.\n\nTraditional terminology distinguishes a symptom or sign, zheng written 症, from a pattern, zheng written 證 or 证, and a disease, bing. Although the first two terms share a pronunciation in modern Mandarin, their meanings differ. Thirst is a symptom; a configuration described as “stomach heat” is a pattern; and a named disease may be understood through either a traditional disease category or a biomedical diagnosis. In practice, terminology is not perfectly uniform across historical texts, translations, or contemporary institutions.\n\nA pattern is usually constructed from several kinds of information. These can include the main complaint, associated symptoms, tongue appearance, pulse qualities, perceived temperature preferences, appetite, sleep, and excretory changes. Patterns may also contain several components: a location such as a traditional organ system, a quality such as heat or cold, and a proposed disturbance such as deficiency, stagnation, or retention of fluids. Consequently, a label such as “spleen qi deficiency with dampness” is a condensed interpretation of a constellation of findings, not a direct measurement of a single physiological variable.\n\nTwo principles express the relationship between disease and pattern. “Different treatments for the same disease” means that people assigned the same disease diagnosis may receive different treatments because their patterns differ. “The same treatment for different diseases” means that different diseases may share a treatment principle when a common pattern is identified. These are propositions about how TCM organizes care; they do not establish that a shared traditional pattern predicts a shared biological mechanism or treatment response.\n\nPatterns are also considered changeable. A practitioner may revise a classification as symptoms evolve, a treatment is modified, or new findings emerge. This distinguishes pattern differentiation from a fixed constitutional label. It also differs from biomedical differential diagnosis, which compares possible causes or disease entities. The two processes can occur in parallel, but they answer different questions and rely on partly different standards of evidence."
    },
    {
      "heading": "Historical development and modern formulation",
      "text": "The intellectual foundations of pattern differentiation developed across many centuries rather than through a single invention. The Huangdi Neijing, a composite text whose principal layers took shape around the late centuries BCE and early centuries CE, discusses relationships among symptoms, seasons, bodily functions, pulse findings, and therapeutic choices. Its explanations use yin and yang, the five phases, and other correspondences. These materials supplied important categories for later diagnostic reasoning, but they should not be treated as a modern textbook system preserved unchanged from antiquity.\n\nThe Shanghan Zabing Lun, traditionally attributed to Zhang Zhongjing around the end of the Eastern Han period, became especially influential in connecting recognizable presentations with prescriptions. The original compilation does not survive intact. Its materials were transmitted through later recensions, notably the Shanghan Lun, or Treatise on Cold Damage, and the Jingui Yaolüe, or Essential Prescriptions of the Golden Cabinet. Their passages associate combinations of symptoms and pulse findings with particular formulas and modifications, creating an enduring model of presentation-based therapeutic reasoning.\n\nThe Song period, from 960 to 1279, was important for the editing, printing, and wider circulation of medical texts. Later physicians developed competing interpretations of disease causation and treatment. Liu Wansu emphasized fire and heat; Li Gao emphasized the spleen and stomach; Zhu Zhenheng became associated with theories concerning yin depletion and ministerial fire. These were not simply alternative labels for one invariant system. They changed which findings received diagnostic weight and which therapeutic strategies were favored.\n\nFrom the seventeenth through nineteenth centuries, warm-disease theory elaborated approaches to febrile illnesses that did not fit comfortably within some older interpretations of cold damage. Ye Tianshi, who lived from 1667 to 1746, is associated with differentiation through defensive, qi, nutritive, and blood levels. Wu Jutong, who lived from 1758 to 1836, developed an influential organization around the three burners. These frameworks remain part of TCM education, alongside older methods rather than as their complete replacements.\n\nThe present prominence of bianzheng lunzhi also reflects twentieth-century institutional development. The expansion of colleges, hospitals, professional curricula, and state-supported textbook production in the People’s Republic of China during the 1950s and subsequent decades helped establish it as a defining principle of standardized TCM. Modern accounts often present diverse historical practices through a unified explanatory vocabulary. Historical continuity is substantial, but contemporary pattern differentiation is also a product of selection, synthesis, and educational standardization."
    },
    {
      "heading": "Gathering findings: the four examinations",
      "text": "Clinical information is traditionally organized through the four examinations: inspection, listening and smelling, inquiry, and palpation. These categories overlap with parts of a medical history and physical examination, but they assign particular significance to observations interpreted through TCM theory. The four examinations are intended to be considered together. A single complaint or isolated tongue feature generally does not supply all the information required by the system’s own diagnostic logic.\n\nInspection includes general appearance, facial complexion, posture, movement, and the tongue. Tongue assessment commonly distinguishes the body of the tongue from its coating and records features such as color, shape, moisture, and coating thickness. A pale tongue may be interpreted as supporting certain deficiency or cold patterns; a red tongue may be interpreted as supporting heat-related patterns. These associations are traditional diagnostic conventions, not demonstrations that tongue appearance uniquely identifies a disease mechanism. Lighting, food pigments, hydration, smoking, oral conditions, and photographic settings can all affect observations.\n\nListening and smelling concern such features as the voice, breathing sounds, cough, and odors. Inquiry explores the chief complaint and its course, thermal sensations, sweating, thirst, appetite, bowel movements, urination, sleep, and other relevant experiences. Menstrual and reproductive history may also form part of the assessment. The diagnostic significance assigned to a finding depends on its combination with other findings: thirst with a preference for cold drinks, for example, is traditionally interpreted differently from dry mouth without a strong desire to drink.\n\nPalpation includes examination of painful or otherwise relevant body regions and assessment of the radial pulse. Traditional pulse diagnosis evaluates qualities such as depth, rate, strength, width, and contour, often at three positions on each wrist. Descriptors include floating, deep, wiry, slippery, weak, and rapid. These qualitative categories are not identical to pulse rate, blood pressure, or an electrocardiographic finding. Their interpretation varies with training tradition and examination technique.\n\nContext complicates all four examinations. Recent exercise can alter pulse rate; ambient conditions can affect temperature sensation; medications can change sweating or bowel habits; and symptoms may have several possible explanations. In contemporary clinical settings, traditional examination may accompany laboratory testing or imaging, but the addition of such information does not itself validate the traditional interpretation. Documentation can distinguish a reported experience, an observed feature, an instrumental measurement, and the pattern inferred from them; these are separate levels of clinical information."
    },
    {
      "heading": "Major systems of differentiation",
      "text": "The eight principles provide a broad organization through four paired categories: exterior and interior, cold and heat, deficiency and excess, and yin and yang. Exterior and interior concern the traditionally interpreted location or depth of illness. Cold and heat characterize a presentation rather than simply recording measured body temperature. Deficiency and excess distinguish, in traditional terms, insufficiency of normal functions or substances from the prominence of pathogenic influences or obstruction. Yin and yang serve both as specific categories and as higher-level ways of organizing the other distinctions.\n\nThese pairs are not always mutually exclusive. A person may be classified as having both deficiency and excess, or as showing heat in one aspect of the presentation and cold in another. “Deficiency heat,” for instance, is distinguished from heat attributed to an excess condition. Mixed patterns are therefore possible within the system, although increasing the number of qualifying categories also creates more opportunities for disagreement among practitioners. The eight principles provide an interpretive structure, not a validated numerical severity scale.\n\nZang-fu differentiation organizes findings through traditional functional networks labeled with organ names. “Liver qi stagnation,” “spleen qi deficiency,” and “kidney yin deficiency” are examples. These labels cannot be translated directly into disease of the anatomical liver, spleen, or kidney. Traditional “spleen” functions, for example, encompass aspects of food processing and fluid distribution that do not correspond to the anatomical spleen’s recognized functions. Confusing these categories can create false inferences about anatomy or physiology.\n\nOther frameworks organize changes in qi, blood, and body fluids, or identify disturbances associated with meridians. “Blood stasis” is a traditional category involving impaired movement of blood within TCM theory; it is not a synonym for thrombosis. “Dampness” and “phlegm” may refer to broad explanatory categories rather than visible moisture or respiratory secretions. The meaning of these terms depends on their technical context and does not necessarily match everyday English.\n\nThe six conformations associated with the Shanghan Lun are taiyang, yangming, shaoyang, taiyin, shaoyin, and jueyin. Warm-disease approaches include the four levels and the three-burner system. These frameworks organize illness differently and are not interchangeable staging scales. Practitioners may combine several systems, for example using an eight-principle description alongside a zang-fu pattern. Such combinations reflect the layered character of Chinese medical theory rather than a single universally accepted diagnostic hierarchy."
    },
    {
      "heading": "From pattern identification to treatment selection",
      "text": "Within TCM reasoning, identifying a pattern leads first to a treatment principle, zhifa, and then to a specific intervention. Traditional principles include supplementing deficiency, reducing excess, warming cold, clearing heat, promoting movement, and resolving retained fluids. These expressions describe the intended action within the traditional framework. They do not by themselves identify a demonstrated pharmacological effect, establish clinical benefit, or specify the safety of a treatment.\n\nIn herbal practice, a practitioner may select a classical formula whose associated presentation resembles the identified pattern, then modify its composition. Materials described in Chinese materia medica are traditionally classified by properties such as temperature, flavor, channel association, and therapeutic action. Formula organization may assign different ingredients principal, supporting, moderating, or coordinating roles. These categories are distinct from chemical composition, receptor activity, pharmacokinetics, or other measurements used in pharmacology.\n\nThe relationship between pattern and formula is not always one-to-one. Several formulas may be used for presentations placed within a broad pattern, while one formula may be applied across multiple disease categories. Some traditions give particular prominence to matching the patient’s presentation to a formula-associated configuration rather than first assigning a general organ-based pattern. Choice can therefore depend on theoretical school, the specificity of the classification, the patient’s changing presentation, and the practitioner’s interpretation of classical texts.\n\nPattern differentiation also informs acupuncture and moxibustion, although their treatment logic additionally involves channels, point functions, and local or regional findings. A treatment plan may combine points selected for the presenting complaint with points selected for a broader traditional pattern. The fact that an intervention is individualized does not establish that its pattern-specific component contributes to benefit; that is a separate empirical question.\n\nReassessment is integral to the traditional model. If symptoms change, the practitioner may alter the pattern label, treatment principle, formula, or point selection. This creates a dynamic treatment process rather than a single fixed prescription. It also complicates evaluation: an observed improvement could relate to the intervention, the natural course of illness, concurrent care, contextual effects, or other factors.\n\nPatient safety must be evaluated at the level of the actual intervention and clinical circumstances. Herbal preparations can have adverse effects, variable composition, contamination, or interactions with medicines; acupuncture and moxibustion have procedure-specific risks. A pattern label does not establish that a product is safe, exclude an urgent disease, or replace assessment of contraindications. These distinctions are particularly consequential when traditional care is used alongside treatment for serious illness."
    },
    {
      "heading": "Reliability, validity, and diagnostic measurement",
      "text": "Evaluating pattern differentiation requires separating reliability from validity. Reliability concerns whether different practitioners, or the same practitioner on different occasions, assign the same classification under comparable conditions. Validity concerns whether the classification measures what it claims to measure and whether it supports accurate prediction or useful clinical decisions. Agreement among practitioners would not by itself establish a biological mechanism; conversely, disagreement makes it harder to interpret studies that treat patterns as stable patient characteristics.\n\nInter-rater research can compare individual observations, such as tongue color, with higher-level judgments, such as a complete pattern diagnosis. These are different tasks. Practitioners might agree that a tongue has a particular appearance but disagree about its diagnostic significance. Agreement can also be affected by whether participants choose one dominant pattern, several concurrent patterns, or a ranked list. Studies using different rules are not directly comparable.\n\nStatistical measures include percentage agreement and kappa statistics, which adjust observed agreement for an estimate of agreement expected by chance. For two raters, Cohen’s kappa is commonly expressed as\n\n\\[\\kappa=\\frac{p_o-p_e}{1-p_e},\\]\n\nwhere \\(p_o\\) is observed agreement and \\(p_e\\) is expected agreement under the model used. Kappa depends on the distribution of categories, so rare or very common labels can complicate interpretation. Reporting only one agreement coefficient can conceal whether disagreement concerns a minor modifier or an entirely different pattern.\n\nResearch has found variable agreement across practitioners, diagnostic tasks, and clinical settings. Important methodological differences include practitioner training, familiarity with the patient population, use of structured checklists, access to medical records, and whether assessments are independent. Consensus meetings and standardized criteria can improve operational consistency, but agreement produced after discussion is not equivalent to agreement obtained independently. Written case vignettes also remove some of the information used in an in-person examination.\n\nA central problem is the absence of an independent reference standard for many traditional patterns. Expert consensus can define a classification for research, but it cannot alone prove that the classification represents a distinct biological condition. Associations with a biomarker, prognosis, or treatment response may contribute evidence, provided the pattern is defined independently and the finding is replicated. Diagnostic sensitivity and specificity are meaningful only when the target condition and reference standard are clearly specified.\n\nQuestionnaires and structured scoring systems can make pattern assignments more reproducible. Their interpretation nevertheless depends on how items were selected, how thresholds were set, whether the instrument works across languages and populations, and whether validation used data independent of instrument development. A reproducible score is a useful measurement property, not a complete demonstration of clinical utility."
    },
    {
      "heading": "Clinical evidence and the problem of testing individualization",
      "text": "Clinical research can ask several distinct questions: whether a particular intervention helps a defined condition, whether an individualized TCM care package helps, and whether matching treatment to a pattern improves outcomes beyond an alternative selection method. Evidence for one question does not automatically answer the others. A trial of a fixed herbal product, for example, may provide information about that product without testing the diagnostic value of pattern differentiation.\n\nA randomized controlled trial can evaluate individualized care by specifying the allowable interventions and decision rules while permitting case-by-case variation. This preserves part of clinical practice but makes the intervention more complex to describe and reproduce. Alternatively, a study may enroll only patients meeting a predefined pattern and use a fixed treatment. That design estimates effects in a selected population; without an appropriate comparison, it does not establish that patients outside the pattern would respond differently.\n\nTesting the value of pattern matching requires a design that can distinguish treatment effects from treatment-selection effects. One approach examines whether treatment benefit differs between prespecified pattern groups. Another compares a pattern-guided strategy with a clearly defined alternative strategy. A statistically significant improvement in one subgroup and a nonsignificant improvement in another is not, by itself, proof that the subgroups respond differently; the difference requires a direct interaction analysis.\n\nBlinding can be difficult when practitioners must examine patients and modify treatment. Herbal preparations may differ in taste and smell, and procedural controls for acupuncture may have effects of their own. Blinded outcome assessment, allocation concealment, prespecified analyses, and complete reporting can still address important sources of bias. The CONSORT statement and intervention-specific reporting extensions provide frameworks for describing trial methods, although adherence to reporting standards does not itself guarantee a well-designed study.\n\nOutcomes also require careful selection. A reduction in a pattern score may be relevant to the traditional framework, but it is not automatically evidence of improved function, fewer complications, or longer survival. Patient-reported outcomes can be important when they are valid for the population and interpreted alongside adverse events and other clinically relevant measures. Multiple outcomes and repeated subgroup analyses create opportunities for chance findings unless accounted for in the study design.\n\nThere is no single efficacy estimate for pattern differentiation across all interventions and diseases. It is a clinical reasoning strategy applied to many different treatments. Its contribution must therefore be assessed in defined contexts. Evidence that a specific preparation or procedure has an effect cannot establish all traditional explanatory claims, while limitations of those claims do not substitute for evaluating the specific intervention. Evidence-based medicine separates these questions rather than treating traditional practice as either wholly validated or wholly disproved by one result."
    },
    {
      "heading": "Standardization, classification, and computational research",
      "text": "Contemporary standardization aims to make traditional diagnostic terms more consistent across teaching, clinical records, research, and translation. International terminology publications from the World Health Organization have addressed traditional medicine vocabulary. Such work helps distinguish terms that might otherwise be translated inconsistently—for example, “pattern,” “syndrome,” “deficiency,” and “vacuity.” Terminological agreement concerns how concepts are represented; it is not the same as confirming their biological validity.\n\nThe eleventh revision of the International Classification of Diseases includes a supplementary chapter for traditional medicine conditions, initially addressing diagnostic concepts derived from East Asian traditions. ICD-11 was adopted by the World Health Assembly in 2019 and came into effect on 1 January 2022. Traditional medicine coding supports the recording and comparison of encounters within participating systems. Inclusion does not constitute WHO endorsement of a treatment’s efficacy or scientific validation of the concepts recorded.\n\nClinical databases increasingly combine pattern labels with symptoms, prescriptions, laboratory findings, and outcomes. In an electronic health record, structured fields can make it possible to distinguish a pattern assigned at the initial visit from one assigned later. This matters because patterns may change and because a classification entered after treatment could be influenced by the observed response. Differences in documentation practice can otherwise appear to be differences in disease distribution or treatment effectiveness.\n\nMachine learning has been applied to tongue images, pulse signals, questionnaire responses, and textual records. These methods can classify observations or reproduce labels assigned by practitioners. Their meaning depends on the target used for training: a model trained on expert pattern labels primarily learns to predict those labels. High agreement with them does not independently establish a disease mechanism, and a model may exploit camera settings, clinical-site differences, or documentation habits rather than clinically meaningful features. Independent testing across institutions is therefore distinct from performance within the original dataset.\n\nBiological research has also examined associations between traditional patterns and gene-expression profiles, metabolites, inflammatory markers, or other measurements. Such associations may generate hypotheses, but they can be influenced by age, medication, underlying disease, illness severity, and the criteria used to define the pattern. A pattern correlated with inflammation in one selected sample cannot automatically be equated with inflammation across conditions. Studies that use the same small dataset both to discover and to evaluate a signature are particularly vulnerable to overfitting.\n\nStandardization also involves a practical tension. Fixed criteria make research and communication easier, while some clinical traditions emphasize flexible interpretation of combinations that resist rigid checklists. Digital tools can formalize either approach, but they do not remove the need to specify who assigned a pattern, which criteria were used, when the assessment occurred, and what outcomes the classification is intended to predict. These details determine whether a recorded pattern functions as a traditional description, a reproducible research variable, or a clinically useful treatment-selection tool."
    }
  ],
  "links": [
    "traditional-chinese-medicine",
    "yin-yang",
    "zang-fu",
    "syndrome",
    "diagnosis",
    "differential-diagnosis",
    "huangdi-neijing",
    "wuxing",
    "shanghan-zabing-lun",
    "medical-history",
    "physical-examination",
    "cough",
    "laboratory-test",
    "spleen",
    "kidney",
    "anatomy",
    "physiology",
    "meridians-tcm",
    "thrombosis",
    "chinese-materia-medica",
    "pharmacokinetics",
    "pharmacology",
    "acupuncture",
    "moxibustion",
    "patient-safety",
    "biomarker",
    "randomized-controlled-trial",
    "blinding-clinical-trials",
    "allocation-concealment",
    "consort-statement",
    "patient-reported-outcome",
    "evidence-based-medicine",
    "world-health-organization",
    "icd-eleven",
    "world-health-assembly",
    "machine-learning",
    "inflammation"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}