Incorporating traditional Chinese medicine syndrome differentiation in randomized trials: Methodological issues

Introduction: In traditional Chinese medicine (TCM), decision on prescription is based on a process called Bian Zheng Lun Zhi (syndrome differentiation guided treatment decision). Syndrome differentiation process may not be recognized in conventional standards of randomized controlled trial (RCT), limiting model validity and generalizability of results. Method: Three major international databases were searched up to July 2016 for articles which discussed how syndrome differentiation, a classical TCM approach in diagnosis, can be incorporated into RCT design. Results: Four methodological solutions were proposed: (i) Lessons learnt from the design of patient reported outcome questionnaire can inform how a TCM diagnosis instrument can be developed. A proper TCM diagnostic tool with sound psychometric properties can reduce variation in the syndrome differentiation process. (ii) Treatment strategies for a specific TCM diagnosis could be highly diversified. Delphi technique can inform the design of an optimal treatment program by facilitating consensus among experts. (iii) Subgroup analysis is often needed in RCTs recruiting patients with several TCM diagnoses. It is highlighted that investigators should consider whether the design, analysis and context of the trial are robust enough to support a reliable claim of the subgroup effect associated with a particular TCM diagnosis. (iv) Finally, we discuss alternative research and analysis approaches for handling misalignment of Western and TCM diagnoses, including the possibility of unifying TCM syndrome with Western phenotypes using latent class analysis. Conclusion: Further methodological advances are needed for better alignment of classical TCM theories and diagnostic instrument development, as well as in reducing bias during expert consensus processes. (C) 2016 Elsevier GmbH. All rights reserved.