Mining patterns of Chinese medicinal prescription for diabetes mellitus based on therapeutic effect

Traditional Chinese medicine (TCM) prescription comprises groups of Chinese herbs that embody thousands of years of history with respect to the treatment of diabetes mellitus (DM), a condition for which there are numerous prescriptions with different therapeutic effects. Existing studies on prescription patterns are based on the frequencies calculated using the traditional association rule algorithm. However, the most important concern for physicians is the efficacy of drug combinations in clinical practice, as no existing study has considered the efficacy of prescriptions. In this study, a weighted association rule algorithm called MWFPP (Mining Weighted Frequent Patterns of Prescription) was used to mine and analyze TCM prescriptions for the treatment of DM based on the therapeutic effect. As a result, the ranking of drug combinations with a low frequency but the good therapeutic effect in the randomized controlled trials (RCT) increased. These drug combinations were also effective in the treatment of DM according to TCM theory. Hence, effective drug combinations could be promising for prescription compatibility in clinical practice and drug discovery.