Statistical method:
|
The analysis content includes case distribution, demographic data and baseline analysis, and curative effect analysis.
Statistic data was described by frequency (composition ratio), and the changes before and after treatment in the two groups were analyzed by x2 test or non-parametric test. When the total number of cases is less than 30, the Yates continuity correction x2 test is used instead. When the theoretical frequency of any grid in the four grid table is less than 1 case, use Fisher's exact probability test instead.
Hierarchical data uses the frequency of each level for statistical description. Ridit test or wncoxon rank sum test was used for comparison between groups.
Missing data: The effectiveness analysis is filled in with the LOCF (last observation carried forward) method, that is, the case data that cannot be observed in all the treatment processes are transferred to the final result of the trial with the last observation data; the safety evaluation does not carry out the missing data estimate.
Drop-off analysis: The comparison of the total drop-off rate between the two groups and the drop-off rate due to adverse events will use the Pearson ?2 test.
Descriptive statistics: the measurement data are given the mean, standard deviation and confidence interval, and if necessary, the minimum, maximum, P25, median and P75 are given; the paired measurement data should also be given the mean sum of the difference Standard deviation; when using the non-parametric method, the median and average rank are given. Counting data gives relative indicators such as ratio, rate, etc. The rank information gives the average rank.
Balance analysis of the two sets of baseline data: including demographic indicators, general conditions before intervention, and primary and secondary indicators. The measurement data were analyzed by one-way ANOVA, and the Welch method was used for correction when the variance was uneven; the count data was used by the Pearson ?2 test; the grade data was used by the Kruskal-Wallis H test.
Compliance analysis: Compliance analysis is carried out according to the treatment amount and treatment time.
Safety analysis: Use the Pearson ?2 test to compare the incidence of adverse events between the two groups, and list the adverse events that occurred in this test; describe the normal/abnormal changes in laboratory test results before and after the test, and the treatment when abnormal changes occur The relationship between the methods, and a list of the "normal/abnormal" changes that occurred in this test.
|