Rules of uncover or
ceasing blinding:
|
This test uses the second unblinding method. After the blind state check, the data is locked, and the staff who save the blind bottom by the statistical center unveiled the first time. The group corresponding to each case number is informed by the code number of A and B, so as to carry out all the data. When the statistical analysis is over and the summary report is completed, a second unblinding session is made at the clinical summary meeting to announce the exact group of the A and B groups.
|
Statistical method:
|
Statistical description
1) Whether the measurement data conforms to the normal distribution: when the data is not met, the statistical method is modified or the data is converted. Because it is a random and open test, the sample size is large. Therefore, when the mean is close to the median, it is treated as a normal distribution.
2) With or without outliers: Conduct statistical and professional analysis to determine trade-offs.
3) Loss value processing of main efficacy index data: When individual subjects have a major efficacy data missing, the method of filling the gap is determined from a statistical and professional perspective. Those with missing cases were transferred with the previous measurement data.
4) Case analysis of unfinished trials: The cases of shedding should be analyzed one by one.
5) Descriptive statistics: indicate the mean, standard deviation, maximum, minimum, median, upper and lower quartiles (Q1 and Q3), confidence interval, frequency (composition ratio), etc.
2. Statistical inference method
1) Measurement data: using t test, paired t test, rank sum test, paired rank sum test, correlation analysis and other methods. For confounding factors that are difficult to control or uncontrolled before treatment, such as imbalance between groups before treatment, covariance analysis (ANCOVA) was used as a covariate to determine the difference in efficacy between groups and to eliminate the effect of these factors on efficacy.
2) Counting data: Chi-square test, Fisher's exact test, etc.; grade data were tested by CMH method. For confounding factors that are difficult to control or uncontrolled before treatment, such as imbalance between groups before treatment, logistic regression was used as a covariate to determine the difference in efficacy between groups and to eliminate the effects of these factors on efficacy.
3) Safety analysis: Firstly, according to the requirements of adverse reaction correlation, the list describes the adverse events and adverse reactions of the two groups (including the number of cases of various adverse events, laboratory test indicators "normal abnormality" or "abnormal" before and after the test Intensify the number of cases and the rate of diversion, and list the reasons and explanations. Statistical analysis of adverse reactions was performed using chi-square test.
3. Statistical expression
The report is represented by a table that is self-explanatory, with a title, notes, and main statistical indicators. The results of repeated measurement data are represented by a table with statistical charts to increase readability; supplemented by F and P values ??for repeated measures data analysis of variance. The general statistical test uses a two-sided test, and P is less than or equal to 0.05 is considered statistically significant.
|