Two experienced practitioners of Traditional Chinese Medicine examine the same patient. The odds they reach the same syndrome are barely better than chance: inter-rater agreement in the literature is κ ≈ 0.17. That silent fragility — diagnoses that shift from clinician to clinician — is the problem KDoctor was built to solve.
A vocabulary without a method
China did monumental work standardizing TCM terminology: national standards defining 2,060 syndrome terms. But standardizing what the syndromes are called is not the same as standardizing how you arrive at them. The regulator itself admits these category terms “do not apply to clinical diagnosis.”
The industry admits the gap — in numbers
In 2020, the China Association of Chinese Medicine surveyed 198 groups responsible for the national guidelines: 66% say syndrome criteria cannot be unified; 70% say a methodological tool suited to TCM is missing; 89% acknowledge the absence of high-quality clinical evidence. In formal clinical trials, only 12.7% apply syndrome differentiation in a documentable way.
The missing layer
This is the void KDoctor fills. Rather than competing with the terminology dictionaries, it consumes them and delivers what is missing: an auditable, reproducible syndrome-differentiation engine (辨证). The same patient produces the same reasoning — traceable from symptom to formula, following the classical herb hierarchy (君臣佐使). And every structured visit becomes clinical data, turning consultations into the evidence that does not exist today.
Regulation works in its favor
Best of all, demand does not need convincing — it is created by mandate. Since 2022, the WHO’s ICD-11 officially codes TCM syndromes; in China, recording the dual diagnosis (disease + syndrome) in structured form has been mandatory since 2021. The customer does not choose whether to record syndromes in coded form — only which tool to do it with.
The Chinese state standardized the vocabulary of Chinese Medicine but admits, in numbers, that it did not standardize the reasoning. KDoctor is the software layer that closes that gap.
