Data mining on prescription, herbal pairs, and pattern identification of pulmonary tuberculosis cases

Purpose: Adapting the features of TCM, this study establishes a new methodology of clinical decision making, like online diagnosis and treatment. Methods: Firstly, using major symptoms as the keywords got 26 cases of He, Hongfang, an ancient pulmonary tuberculosis specialist, in database. The following methods in IBM SPSS modeler premium were applied: Data understanding, original data preparation, data base establishment, proportion of the pathological mechanism, coordinate plots of prescription(Figure 1), associate rules of herbal pairs(Table 1), and a decision tree of pattern identification. Results: Discovered were one major mechanism, one special prescription, 4 herbal pairs, and pattern identification. It got positive feedbacks after peer reviewing by over 10 clinical specialists. Modern literatures support it as well. Conclusion: A new methodology emerged on clinical decision of pulmonary tuberculosis. Other contagious diseases can apply it as well. It increases the possibility and accuracy of online diagnosis and treatment especially on contagious diseases.