Instrumental Variable Methods and Its Applications in Outcome Research of Traditional Chinese Medicine / 世界科学技术-中医药现代化

ABSTRACT

This paper was aimed to discuss the feasibility and attentions of application of instrumental variable (IV) methods in traditional Chinese medicine (TCM) outcome research. First, the application of IV was introduced, which included the basic principles and hypothesis, statistical model, estimator of IV and weak IV. Then, an example was given to illustrate the evaluation criteria and attentions of IV. The resultsshowed that IV method was proposed as a potential approach to the problems of confounding in statistics. But using IV methods should be based on a series of statistical hypotheses. It was concluded that the IV analysis was a method controlling confounding bias, but generally it was not chosen as the preferred analytical method. The issue of searching for valid and plausible IV seemed to be the biggest obstacle in the outcome of TCM researches.