Application Of Item Response Theory In Randomized Clinical Trials

First, this paper introduced randomization methods in clinical trials,including the conception, classification and procedure of randomization ,as well as the balance between groups. It recommended the guideline reporting of randomization and baseline characteristics in a randomized controlled clinical trial. Simultaneously the equalization has also made the elaboration on the group, introduced the group the uniform request as well as the group the imbalanced processing method..Second, this article is to a project response theory exploration and the application. In the item response theory, we assume that identical is tried to each item response is mutually independent, this is the so-called partial independent supposition; Moreover, but also supposes the response which tries is differently mutually independent. According to previous research study, after in the system introduction project response theory model, to the rank responded in the model in the double parameter Logistic model parameter makes the further analysis, namely when project parameter known, estimates the estimated capacity parameter with the maximum likelihood; When ability parameter known, estimates the project parameter with the maximum likelihood estimate. Writes the iteration step.Finally, in the randomisation clinical test, the experimental examination table demonstrates more and more important position, then we applies the parameter estimate method to the randomisation clinical test, carries on the simulation test with the computer, has the random number to take the project parameter, uses these data estimated capacity parameter, then with confirms this method through the T-test the superiority. At last we draw the conclusion : When we use initial result as the latent characteristic, sample size in clinical trials may use the survey scales to respond in the theory.