|
Data collection and Management (A standard data collection and management system include a CRF and an electronic data capture：
|
In order to improve the quality of research and reduce research bias, real-time input of subject data, third-party data management, and third-party blind statistical analysis will be carried out during the research process. The third-party will conduct data analysis and issue statistical analysis reports.
1. Data Management
The requirements are as follows:
(1) The researcher observes the original data according to the actual conditions of the subjects and records the CRF table. Researchers need to ensure that the data is true, complete and accurate. All items must be filled in, no blanks or omissions are allowed. When making any corrections, only lines can be drawn. The revised content shall be marked and dated by the investigator. The original records shall not be erased or overwritten.
(2) Data entry and management are handled by dedicated personnel. Write the corresponding computer program, and use the independent double entry method for data entry. The data administrator checks the data, and promptly asks the investigator for any questions found, and the data administrator confirms and revises the data based on the investigator's answers. The content of the error and the result of the modification shall be truthfully recorded and properly kept.
(3) The inspector regularly checks the research data and confirms that all the questionnaires are completed and correct. If there are errors or omissions, the investigator shall be asked to correct it in time, save the modified content, and the investigator shall sign and indicate the date.
(4) Electronic data files include databases, inspection procedures, analysis procedures, analysis results, codebooks and explanatory documents, etc., which should be stored in categories, and multiple backups should be stored on different disks or recording media, properly stored to prevent damage.
(5) For adverse events that occurred during the study period, their symptoms, severity, time of occurrence, duration, treatment measures, and progress should be recorded on the pathology report form, and their relevance to the study drug should be evaluated, and the investigator should record it in detail. Sign and date. At the same time, it must report to the subject in charge and the base office of the main research unit as soon as possible (within 2 hours), and the base office of the main research unit should report to the Drug Administration and the ethics committee within 24 hours. The researcher should sign and date the report. Save the data for future reference.
(6) After the trial is completed, all materials in the clinical trial, including the research protocol, the patient’s signed informed consent form, the patient’s original hospital record, and the completed CRF form, etc., will be reviewed and signed by the researcher and supervisor, according to China The requirements of GCP shall be preserved and managed until 2 years after the termination of clinical trials. Except for the national or local regulatory authorities, or the ethics committee to consult as required, it shall not be provided to others in any form.
2. Data statistical analysis
(1) Statistical software: SPSS21.0 statistical software is used for data statistical analysis.
(2) Basic principles: All statistical inferences use two-sided tests, the statistically significant test level is set at 0.05, and the confidence interval of the parameters is estimated to be 95% confidence interval. Use the parametric method as much as possible. When the data does not meet the parametric method conditions, the data conversion method can be used to make it meet the conditions. If it still does not meet the conditions, the non-parametric method can be considered.
(3) Descriptive statistics: the measurement data are given the mean and standard deviation for description; the median and interquartile range are used for statistical description for non-normality; the count data is given the frequency distribution and the corresponding percentage.
(4) Balance analysis of baseline characteristic indicators: statistically describe the patient's demographic information, course of disease, degree of atrophy, and baseline characteristics of pathological characteristics. Measurement data conforming to normal variances were analyzed by analysis of variance or two independent samples T test, otherwise, rank sum test was used; count data were used chi-square test or Fisher's exact test.
(5) Missing data processing: Mean imputation or mean imputation of the same kind is used for missing values.
(6) Efficacy analysis: Comparison of efficacy indicators before and after treatment in the same group of patients: quantitative variables were tested by paired t-test; qualitative ratios were tested by McNemar X2. Comparison between groups: measurement data conforming to normal variances were analyzed by analysis of variance or two independent samples T test, otherwise, rank sum test was used; count data was analyzed by chi-square test. Rank-sum test is used for rank variables. In the experimental design, the confounding bias that may be foreseen was controlled. In the data processing, the confounding factors that may affect the outcome of the efficacy were controlled by stratified analysis and Logistic regression analysis.
|
