A practitioner of Chinese medicine prescribes against a pattern, and the pattern is built from four kinds of finding. The WHO’s 2007 standard terminology for the Western Pacific calls them the four examinations: inspection, listening and smelling, inquiry, and palpation. It defines pattern identification as the analysis of those findings to determine the location, cause and nature of a patient’s disease. Patterns, symptoms and what you can look up explains why that conclusion, rather than the complaint, is what a classical prescription is written against. This page sets out what each examination looks for, as the WHO’s 2007 terminology and its 2022 successor for traditional Chinese medicine define it, and then what happened when researchers asked several practitioners to examine the same patients.
Looking
The 2007 terminology defines inspection as observing the patient’s mental state, facial expression, complexion, physical condition, tongue and secretions, and, in infants, the superficial venules on the palmar side of the index finger, read across three segments the terminology names the wind, qi and life bars. Inspection of the mental state is called inspecting the spirit. The 2022 terminology defines presence of spirit by bright, mobile eyes, clear consciousness, a lustrous face and even breathing, and false spirit as a sudden apparent improvement in a patient with a chronic critical condition, which it describes as an omen of death.
The tongue
Tongue inspection divides the tongue into two things, the body and the coating, and reads each separately. The 2007 terminology defines the tongue body as the musculature and vascular tissue of the tongue and the coating, which it also calls fur, as the layer on its surface; the 2022 terminology describes the coating as moss-like and generated by the ascending of stomach qi. The 2007 edition maps the surface into four regions, the tip, the margins, the centre and the root, and the 2022 edition adds the two sublingual veins under the tongue.
| Feature | What is read | Examples of defined terms |
|---|---|---|
| Body colour | The colour of the tongue body; the 2022 edition gives pale red and lustrous as normal | Pale, red, crimson (deep red), purple, bluish purple |
| Body form | Size, texture and surface of the body | Enlarged, thin, teeth-marked, fissured, spotted, peeled, mirror |
| Body movement | How the tongue moves when extended | Stiff, trembling, deviated, contracted, limp |
| Coating colour | The colour of the fur | White, yellow, grey, black |
| Coating texture | Thickness, moisture and consistency of the fur | Thin, thick, moist, dry, slippery, slimy, curdy, peeling |
The terminologies attach a meaning to many of these terms. The 2022 edition, for instance, reads a pale tongue as indicating blood or yang deficiency, and a thin coating through which the tongue surface is faintly visible as indicating normal stomach qi. Tongue findings also appear in AcuiQ’s index as search terms in their own right, such as tongue body red and tongue coating yellow, because published sources recorded them.
Listening and smelling
The 2007 terminology treats listening and smelling as one examination, wenzhen. It defines listening as attending to the voice, breathing, coughing and vomiting to judge whether a condition is cold or hot and deficient or excess. Its terms include a deep, muffled voice and a faint, low one; muttering, delirious speech and talking in sleep; sighing, panting, wheezing and audible phlegm; and rumbling in the bowels. The smelling half is represented in the same list by odours, such as smell from the mouth.
Asking
The 2007 terminology defines inquiry as asking the patient about the complaint and the history of the illness, and its inquiry section runs to more than 200 defined terms. They begin with sensations of heat and cold and with sweating, and go on through sleep, pain by its site and its quality, the eyes and nose, appetite and taste, the chest and abdomen, and stool and urine. The 2022 terminology gives the traditional order of questioning as the ten questions, a mnemonic verse that covers cold and heat, sweating, pain in the head and body, bladder and bowel movements, diet and appetite, the chest, hearing, thirst, previous illnesses and the cause of the illness. The Ming physician Zhang Jingyue (1563–1640) set out ten diagnostic inquiries in his complete works, whose opening volumes have been translated into English by Allen Tsaur.
Feeling the pulse and the body
Palpation covers pulse diagnosis and the touching and pressing of the body. The standard site for the pulse
is the radial artery at the wrist, called cunkou. The 2007 terminology places the middle finger at
the guan (bar) position, beside the radial styloid, the bony prominence on the thumb side of the wrist, the index finger at
cun (inch) on the side nearer the hand, and the ring finger at chi (cubit) on the side
nearer the elbow. Each position is felt with light, moderate and heavy pressure, which gives the
“three positions and nine indicators”. The same name also covers a scheme that takes the pulse at
three regions of the body, the head and the upper and lower limbs, and the 2022 terminology describes a related
method using three arteries: the carotid at ST09 (Renying), the radial
artery at the wrist, and the artery on the top of the foot at ST42
(Chongyang), with KD03 (Taixi) as an alternative for kidney qi.
The publisher’s description of Paul Unschuld’s translation of the
Nanjing, a text of the first century CE, names its discussion of pulse diagnosis and
needle treatment as the most important thing in it; The classical
texts covers where it sits among the others.
A pulse is described by a quality. The 2022 terminology has an entry for the 28 pulse conditions, and the 2007 edition defines each quality separately, from floating and sunken to racing, plus a group of ten “strange pulses” it describes as signifying critical conditions. Rate is counted against the practitioner’s own breathing.
| Quality | Definition |
|---|---|
| Floating | Felt by light touch, grows faint on hard pressure |
| Sunken (deep) | Felt only when pressing hard |
| Slow | Fewer than four beats to one cycle of the practitioner’s breathing |
| Rapid | More than five or six beats to one breath |
| Racing | More than seven beats to one breath |
| Surging | Beats like dashing waves, forceful rise and gradual decline |
| Fine (thready) | Thin as a silk thread, soft, always perceptible on hard pressure |
The rest of palpation is the body itself. The 2007 terminology defines examination of the chest and abdomen by touching and pressing, examination of the skin of the forearm for texture and temperature, and palpation of acupuncture points, meaning pressing the body surface at points held to reflect the condition of the corresponding organs. The 2022 edition describes palpation as feeling for local temperature, moistness, softness, tenderness and lumps.
Putting the four together
The 2007 terminology has a term for combining them, the correlation of all four examinations, and treats the combined result as the input to pattern identification. Some pulse and tongue findings are indexed on AcuiQ, among them pulse thin, but a finding of that kind is a piece of the evidence for a pattern, not the pattern.
When several practitioners examine the same patient
If the four examinations lead reliably to one pattern, two practitioners examining the same patient should reach the same conclusion. A 2019 systematic review by Eric Jacobson and colleagues at Harvard Medical School collected 21 English-language studies that measured agreement of this kind, with practitioners examining real patients in person. Across the nine studies that reported pairwise agreement, practitioners agreed on average 57% of the time (median 65%, range 19% to 96%). Across the seven that reported Cohen’s kappa, a measure of agreement beyond what chance alone would produce in which 0 is chance and 1 is perfect agreement, the average was 0.34 (range 0.07 to 0.59). The authors judged agreement low to moderate with a few exceptions, and found that studies in which raters trained together beforehand scored higher.
| Study | Who examined whom | What they found |
|---|---|---|
| Hogeboom 2001, Seattle | 6 acupuncturists, the same 6 patients with chronic low back pain, on one day | 20 different diagnoses and 65 points used; most gave every patient qi and blood stagnation with kidney deficiency, and BL23 (Shenshu); agreement on the detail and the other points was poor |
| Zhang 2005, Baltimore | 3 practitioners, the same 40 patients with rheumatoid arthritis | Average agreement on the diagnosis 31.7%; the herbal prescription matched the textbook treatment for the diagnosis given 91.7% of the time |
| O’Brien 2009, Melbourne | 2 practitioners of Japanese toyohari, 62 people | Agreement 57%, 61% and 77% on pulse depth, speed and strength; 48% on the primary pattern |
| Birkeflet 2011, Oslo | 2 acupuncturists, 54 women, 30 of them infertile | 39 different patterns and 36 points used; poor to no agreement on the choice of points |
| Popplewell 2019, Sydney | 2 or 3 practitioners, 35 people from a teaching clinic, choosing from 56 patterns | 23% pattern agreement and 19% weighted agreement |
Jacobson’s review describes the Popplewell study as designed to approximate ordinary practice: its patients were not selected for any one condition and its raters had no joint training. The same group then had practitioners record their conclusions in a structured format, the Traditional Chinese Medical Diagnostic Descriptor, and simple agreement on the same 35 patients rose to 0.80, against 0.19 when the conclusions were written as TCM patterns.
Studies of single examinations show the same spread. In a 2008 tongue study at the University of Technology Sydney, 30 practitioners rated ten tongue slides twice; they reached 80% agreement with each other on only 17% to 19% of questions, most of them yes-or-no questions such as whether a coating was present, and on 5% of questions with more complex choices. The authors traced the result to the lack of working definitions for the tongue features and regions. A 2016 review of pulse diagnosis by Karen Bilton and Chris Zaslawski found 12 studies; methods with explicit operational definitions reached acceptable agreement, and poor agreement went with the unclear definitions inherited from the classical descriptions. An earlier 2009 review by Kylie O’Brien and Stephen Birch had reached a similar overall view: pulse studies ranged from low to very good agreement, tongue studies showed considerable variability, and studies of pattern diagnosis had not generally found high reliability.
What these studies cannot tell you
The studies measure agreement, not accuracy: two practitioners can agree and both be wrong, and none of the studies above tested whether a diagnosed pattern predicted how a patient fared. Most were small, and Jacobson’s review notes that a single meeting may understate the agreement reached over a course of consultations. They do show that the step between the examinations and the point prescription varies with the practitioner. Zhang found prescriptions followed the diagnosis closely, and Birkeflet found point selection tracked the pattern chosen, so a disagreement about the pattern carries through to the treatment. A published protocol on AcuiQ that names a pattern records the pattern its authors reached, not how they reached it, and What AcuiQ does not do explains why the site does not attempt the examination itself. Reading an acupuncture trial covers how individualised treatment is handled in trials.