JMIR Publications

ABSTRACT

Background:

There is a growing trend in the use of mobile health (mHealth) technologies in traditional Chinese medicine (TCM). Tongue diagnosis is an important procedure of TCM diagnosis, but its reliability in smartphone application has yet to be evaluated.

Objective:

This study aimed to determine the intra-rater and inter-rater reliability of tongue coating diagnosis using a quasi-Delphic process in a series of experiments.

Methods:

First, an operating classification scheme of tongue coating were formed through the analysis of 2023 tongue images by two groups of assessors. Then a representative tongue image set was evaluated by a group of clinicians based on their own interpretations as well as on an operating classification scheme using online questionnaires. Then clinicians in the study group each took tongue images of their patients and rated the tongue features according to their direct visual inspection and indirect tongue image inspection. Using the finalized operating classification scheme, clinicians of the study group (n=17) assessed representative tongue images (n=24) taken by themselves, and the intra-rater and inter-rater reliability of their assessments were evaluated.

Results:

Intra-rater reliability of tongue image assessments were good to very good (Cohen’s Kappa between 0.7-1.0), except for the color of tongue body and for slippery tongue fur. Inter-rater reliability was moderate for tongue coating (Gwet’s AC2 between 0.49-0.55), and fair for color and features of tongue body (Gwet’s AC2: 0.34).

Conclusions:

Taken together, our study has shown that tongue images collected via smartphone contain some reliable features that can be used in mHealth analysis, especially for those related to tongue coating. Our findings thus support the use of smartphone for detecting changes in tongue coating in telemedicine and mHealth.