A mechanic asked to fix a rattle does not look up rattles. They work out which part is loose, and the repair follows from the part. The rattle is what got the car into the workshop; it is not what gets worked on. Chinese medicine is organised the same way, and it is the single largest source of confusion for anyone arriving at a point prescription from outside it.
Two different kinds of term
A symptom is what the patient reports: a headache, loose stools, ringing in the ears. A pattern is what the practitioner concludes: damp-heat in the lower burner, Liver qi stagnation, Spleen qi deficiency. The pattern is a compressed description of how a set of findings hang together, and it is the thing a classical prescription is actually written against.
The gap between them is not a translation problem. Two patients with the same headache can receive different points because they were differentiated into different patterns, and two patients with unrelated complaints can receive the same points because they landed in one. A prescription indexed by complaint alone throws away the step that produced it.
What is actually in the index
AcuiQ indexes 14,240 terms, of which 11,074 are canonical concepts and the rest are variant spellings resolved onto them. Three categories say in their own name that they are not symptoms:
| Category | Terms | What it holds |
|---|---|---|
tcm-action |
1,392 | What a prescription is meant to do – move qi, dispel damp, astringe intestines |
tcm-pattern |
1,055 | Differentiated patterns – damp-heat diarrhea, cold counterflow, wind-cold cough |
tcm-diagnostic |
163 | Tongue and pulse findings – tongue coating yellow, pulse slippery |
The counts above are read from the database as the page is served, so they are what the index holds now rather than what it held when this was written. The same is true of the corpus figures on how AcuiQ is built.
The largest category is not a category
4,914 terms, about 35 in every hundred, are filed under
general. It is the biggest group in the index by a wide margin, and it is a
residual rather than a clinical grouping: it is where a term went when the extraction pass
could not place it. It holds ordinary biomedical diagnoses (copd), and it also
holds the exact three kinds of term the table above is supposed to isolate.
descend qi and disperse accumulation & disinhibit damp are
therapeutic actions.
exuberant heat stirring wind is a pattern. A handful of rows are bare digits and
are not terms at all.
This matters more than a tidiness complaint would. The three labelled categories are the
site’s only machine-readable warning that a term is not a complaint, and the label is
what the code acts on: the popular-symptom lists drop tcm-action by name, and a
point page hides action terms from its indications for the same reason. A therapeutic action
filed under general gets none of that. It is treated as a symptom because nothing
in the record says otherwise.
Why they are in there at all
Because that is how the sources are written. AcuiQ builds its index by reading published prescriptions and recording what each one says it treats, and a great many of them say “damp-heat diarrhea” rather than “diarrhea”. Dropping the pattern terms would discard the prescriptions attached to them. Rewriting them into plain complaints would assert a differentiation nobody made.
An action term is stranger still. “Benefit bladder” is not a complaint at all; it
is a therapeutic intent, and it appears in the index because the source recorded it in the
indications field. These terms are useful for reading a prescription and close to useless for
searching from a complaint, which is why the ones that carry the tcm-action label
are excluded from the popular-symptom lists the site surfaces.
What this means for a search
Search from the complaint. It is the term that has the most prescriptions behind it and the one the fuzzy matcher is tuned for. If you already know the pattern, search for it directly – the index will find it, and the results will be tighter and fewer. If a result comes back that reads like an instruction rather than a condition, you have landed on an action term. Nothing on the page will tell you so, which is the reason this one exists.
What AcuiQ will not do is get you from one to the other. It does not differentiate patterns, and What AcuiQ does not do says so among the other things the tool deliberately leaves alone. Differentiation is the part of the consultation that requires the patient in the room: the pulse, the tongue, the history, the questions whose answers change the conclusion. A search box has none of that, and a tool that inferred a pattern from a typed complaint would be inventing the clinical step it skipped.
The index is a record of what other people prescribed and what they said they were prescribing for. Read that way it is a genuinely large piece of evidence. Read as a diagnosis engine it is the wrong instrument, and confidently so.