Objective: To investigate the efficacy of applying automatic tongue diagnosis system (ATDS) to the training of novice Chinese medicine (CM) doctors in the tongue diagnosis.
Method: A two-round experiment composed of pre- and after-training tests was performed. Two groups of CM doctors were formed by experienced doctors (ED) and novice doctors (ND). The intra- and inter-observer agreements for the ED and ND groups, and the inter-observer agreements between ATDS and ED, and ATDS and ND were derived first. After a 1-month training of the ND, the intra- and inter-observer agreements of the ND group, and the inter-observer agreement between ATDS and ND were obtained.
Result: The kappa value of agreement data for ND after training was significantly higher than ND before training with a level of P<0.05, indicating an obvious improvement of intra-observer agreement for ND after training with ATDS. The inter-observer agreements for the ED group, and ND group after training were significantly higher (P<0.05) than the ND group before training. In contrast, no significance is observed between the ED group and the ND group after training, indicating that physicians in the ND group achieve an inter-observer agreement to the same level as that of ED after 1-month training.
Conclusion: This finding validates the utilization of ATDS as a mean of education in forging consensus during the learning process of tongue diagnosis.