Efficacy and Safety of Chinese Medicine JCM-16021 for Diarrhea-predominant Irritable Bowel Syndrome

Sample size calculation:

Sample size is calculated on the basis of primary efficacy variables. From our previous study, the global assessment of improvement (GAI) were 52% in herbal medicine group and 32% in western medicine group, respectively. According to the references (Corazziari E, Bytzer P, Delvaux M, et al. Clinical trial guidelines for pharmacological treatment of irritable bowel syndrome. Alimentary pharmacology \& therapeutics 2003; 18 (6): 569-580), the investigational drug is more effective than placebo (the overall improvement rate of symptoms is 15%), using StudySize2.0 software to calculate the sample size, assuming the improvement in the treatment group is 52%. In order to detect a difference with a two-side p value <0.05 and 80% statistical power, we will need to recruit 166 patients per arm. Further assuming a 15% dropout rate, we conclude that a total of 392 patients (196 per arm) will be recruited to ensure statistically significant results.For the number of cases distribution between centers, according to the references (Lai D,Chang KC, Rahbar MH, Moye LA. Optimal Allocation of Sample Sizes to Multicenter Clinical Trials. Journal of biopharmaceutical statistics 2013; 23 (4): 818-828) , from the following equation, we will consider the center about the patient flow, traffic, treatment,costs and other possible factors, we will use the formula below to estimate the number of cases that will be recruited in each center.

Research medical record and Electronic Database:

All patients should be observed and assessed based on clinical trial protocol and the investigators need to document in the medical record accurately and clearly. Research medical record is the source document which cannot be altered. Any correction should not change the original record and can only be added in a way of narration with reasons. The doctor participated in the clinical trial needs to sign and date the record. An electronic database will be created. Each study site will input their own data and be responsible for its accuracy. A chief statistician will be responsible for data cleaning and data analysis.

Analysis parameters:

All parameters and study elements will be analyzed. The statistical analysis will be performed using SAS 9.1 and SPSS software.

Analysis sets:

Full analysis set (FAS): The analysis will be conducted according to the intention-to-treat (ITT) principle which means to eliminate the participants with a minimum and reasonable method. ITT population refers to all participants who go through randomization, enter double-blind treatment period, and receive IMP at least one time. Missing values of efficacy will be imputed by the last-observation-carried forward (LOCF) method. Per-protocol set (PP): PP population refers to all participants who complete relative observation according to protocol requirement and are confirmed to meet following conditions: ① compliance between 80% and 120%; ② not taking probihited medications during the process of trial; ③ meeting inclusion criteria and not fitting any exclusion items; ④ completing all planned visits and necessary items of CRF. Missing values of this set will still be processed as missing data and not be imputed. Safety analysis set: Population for safety analysis refers to all participants who enter the trial, receive medication at least one time and have suitable follow-up data for safety analysis. All safety data including AEs and laboratory results from participants will be assessed.16.3 Statistical analysis technique Baseline data (gender, age, race, weight, height, vital signs, course of IBS, history of smoking and alcohol) will be descriptively summarized. Differences of measurement data between the groups will be assessed with the use of t-test for normally distributed continuous variables and Wilcoxon signed rank test for non-normally distributed. Differences of enumeration data between the groups will be assessed with the use of chi-square test or CMH test when considering multicenter character. Measurement data of different groups in each visit will be reported as mean ± standard deviation (SD). Intra-group comparisons between baseline and each visit will be conducted by using paired t-test (or Wilcoxon signed rank test). Comparisons between groups will be conducted by using an analysis of variance (ANOVA), with other confounding factors like multicenter character conducting the covariate analysis. Statistical analysis for the data which do not meet above conditions (e.g. non-normal) will be conducted with the use of non-parametric test. Enumeration data of different groups in each visit will be reported as frequency (proportion). Comparisons between groups will be assessed with the use of X2 test (CMH test) or non-parametric test. Dropout analysis: Dropout analysis will be conducted with the use of chi-square test. If the data do not conform to chi-square test (data include 0, or theoretical frequency is below 1), Fisher's exact test will be used. Compliance analysis: Compliance analysis will be conducted with the use of chi-square test. If the data do not conform to chi-square test (data include 0, or theoretical frequency is below 1), Fisher's exact test will be used.

Hypothesis testing:

This trial will conduct superiority analysis firstly. Other difference test will be conduct by two-sided test. The statistical significance will be defined as two-sided P-value of ≤0.05 without any special explanation.

Efficacy analysis:

The efficacy analysis will be conducted with the use of PP analysis and ITT analysis in the meantime. Comparisons of measurement data will be conducted by using analysis of covariance (ANCOVA), with treatment group and trial center as a factor in the model and baseline as the covariate. Comparisons of measurement data will be conducted by using chi-square test or CMH chi-square test when considering multicenter character. Meanwhile, superiority analysis between experimental group and control group will be conducted based on primary efficacy variables. Superiority test depends on interval method.

Safety analysis:

Extent of exposure: Descriptive statistics will be conducted according to the exposure dose and time of medication in different groups. AEs analysis: Comparisons of incidence rate of AEs between groups will be conducted with the use of X2 test. And investigators need to list and describe the AEs happened in this trial. If the data do not conform to X2 test (data include 0, or theoretical frequency is below 5), Fisher's exact test will be used.

Data management:

CRFs are filled in by investigators and study coordinators, other assessment forms by every participant (including dropout cases). Data processing will be conducted in accordance with the following protocol:

1. Verification of CRFs: Study coordinators need to verify CRFs before inputting.
2. Data verification needs to be conducted successively in the following two steps:
3. Verify the consistency and logicality of data: Review contents of data range and logicality are determined by the range of each indexes and the interrelation. Corresponding software will also be written to correct the incorrect data.
4. Compare database and CRFs by manual testing. Selectively counter check 10% CRFs with participants' medical notes to know the quality of inputting and analyze and handle the existing problems.
5. Data inspection and closure of database: After verifying the validity of established database and statistical protocol, principal investigators, will lock the data. The locked data are not allowed to change. Confirmed problems found after locking will be handled in the process of statistical analysis. All mistakes and modification should be recorded and kept properly.