Data-driven quantification and intelligent decision-making in traditional Chinese medicine: a review

Traditional Chinese medicine (TCM) originates from the practical experience of human beings’ constant struggle with nature. In five thousand years, TCM has gradually risen from empirical medicine to modern evidence-based medicine with complete scientific principles such as fundamental systematic theories, treatment principles and methods, classic prescriptions, famous medicines. The development of information science, data science, and computer technology has provided effective models, methods, and technologies for modern TCM’s quantitative and intelligent diagnosis and treatment decision-making. And it also has promoted the development of TCM from evidence-based medicine to intelligent TCM. Starting from the development of TCM, we introduce the rise and connotation of ancient, modern, and intelligent TCM. Moreover, we emphatically analyze the research status of quantification and intelligent decisions for the whole disease cycle, including data-driven modern TCM diagnosis, program optimization, and treatment program evaluation. In addition, we discuss the critical issues of data-driven TCM quantification and intelligent decision research and briefly elaborate on the new ideas of data-driven intelligent TCM research. In conclusion, compared with traditional research paradigms, the advantages of data-driven medical decision research paradigms are as follows: (1) From the perspective of decision-making subjects, the data-driven research paradigm describes the clinical decision-making mechanism in real scenarios with rigorous mathematical theories, which will break through the difference between the conclusions drawn by clinical design research methods and clinical practice. (2) By applying the results of basic theoretical research to clinical decision-making practice in real scenarios, the data-driven medical decision-making research paradigm will contribute to getting out of the dilemma that the conclusions drawn by traditional AI models are difficult to explain in clinical practical decision-making.