Analyzing Subjective Bias In A Clinical Trail By Comparing Open-label Trial And Double-blinded Trial

Background and Objective:In my research, I analyze the difference between the double-blinded group and the open-label group of a multicenter clinical trial to detect whether the open-label trail was influenced by the subjective bias compared with the double-blinded trial. So I can demonstrate whether it is essential to use double-blinded method in clinical trials and also can present how to control the subjective bias.Materials:The materials used in my research come from a mutilcenter clinical trial in which researchers want to detect the curative effect and the side effect of a new drug. I take the materials of five hospitals in which researchers carry out both the double-blinded trial and the open-label trail. There are 77 cases in the double-blinded group(39 cases in test group and 38 cases in controlled group) and 114 cases in open-label trail(all of them in test group). The therapy programs in the double-blinded trial and open-label trail are the same.Analysis methods:1 > Compare the difference between the double-blind trial and the open-label trail before the therapy begins.Variables: age, sex, infarct half ball, infarct section, the cardiovascular disease history, the concomitant cardiovascular disease, the score of nerve function, the infarct focus area under CT examination three days after the attack2^ Compare the therapy process and therapy result between the double-blinded test group and the open-label group which is also the test group. Variables: complete therapy, associated drug, side effect, score of nerve function after the therapy.Statistic methods:1. Single-variable analysisAt first, analyze every hospital's materials. Then, analyze the merged materials. For measurment materials, we use the Wilcoxon test; for enumeration materials, we use Chi-square test and Fisher's exact test. 2 Loglinear model analysisAccording to the result of the single-variable analysis, select the therapeutic variable and potential confounding factors to build loglinear model to analyze the interaction among the variables, especially the interaction between the group variable and the therapeutic variable.Results:1. From the comparison between the double-blinded group and the open-label group, we find out that there are four hospitals in which the historic cardiovasculardisease is severer in open-label group than in double-blinded group; there are four hospitals in which the nerve function score is lower in the open-label group than in double-blinded group; and there are three hospitals in which the infarct focus area under CT examination is larger in open-label group than in double-blinded group. These differences show us that the state of illness is severer in open-label group than in double-blinded group.2. From the comparison of the therapy process and therapy result between the double-blinded test group and the open-label group, we find out that the progress of nerve function score after the therapy is higher in open-label group than in double-blinded group and for other variables the two groups are equilibrint. We take the two hospitals in which the score has significant difference between the open-label group and the double-blinded group and using loglinear model and stratified analysis to analyze the materials. We find that the nerve function score before the therapy has .interaction with this after the therapy and adjust this interaction the score progress after therapy still interaction with the group variable.Conclusion:From the result above, we find out that the subjective bias can influence the clinical trial: 1. The researchers incline to choose the lower riskly cases to enter the double-blinded trial than to.the open-label trial. 2. When taking the semi-objective index to assess the curative effect the researchers tend to make better judgment to the open-label trial than to the double-blinded trial. So, it is essential to execute double-blinded trial in clinical trial.