数据采集和管理(说明:数据采集和管理由两部分组成,一为病例记录表(Case Record Form, CRF),二为电子采集和管理系统(Electronic Data Capture, EDC),如ResMan即为一种基于互联网的EDC:
|
数据管理: (1)研究者根据受试者的原始观察记录,将数据及时、完整、正确、清晰地载入病例报告表。 (2)监查员监查研究的进行是否遵循研究方案。确认所有病例报告表填写正确完整,并与原始资料一致。如有错误和遗漏,及时要求研究者改正。修改时需保持原有记录清晰可见,改正处需经研究者签名并注明日期。 (3)经过监查员检查后的病例报告表,由监查员核查签字后,及时送交临床研究数据管理员。对于完成的病例报告表在研究者、监查员、数据管理员之间的传送应有专门的记录,收到时应有相应的签名,记录需妥善保存。 (4)数据管理员在数据录入前再次核查,发现问题及时通知监查员,要求研究者作出回答。他们之间的各种疑问及解答的交换应当采用疑问表形式,疑问表应保存备查。 (5)数据管理员在进行数据录入前,要了解观察表格各项目的内容及编码情况,将编码工作过程记录于编码本保存。数据库命名应规范、易读、易查找。并保证其正确、安全和保密。 (6)数据录入员录入数据采用二次录入。录入过程发现问题或意外情况,应做好登记并及时报告,以便迅速处理问题,数据录入结束后应抽查部分观察表格,了解录入质量,分析并处理存在的问题。 (7)数据管理员应与主要研究者一起,按病例报告表中各指标数值的范围和相互关系拟定数据范围检查和逻辑检查内容。并编写相应的计算机程序,在输入前控制错误数据输入,找出错误原因加以改正,所有错误内容及修改结果应有记录并妥善保存。 (8)原始病例报告表在按要求完成数据录入和核查后,按编号的顺序归档保存,并填写检索目录等,以备查考。电子数据文件包括数据库、检查程序、分析程序、分析结果、编码本和说明文件等,应分类保存,并有多个备份保存于不同磁盘或记录介质上,妥善保存,防止损坏。所有原始档案应按我国《药物临床试验质量管理规范》的规定期限保存。 统计分析: (1)详细统计分析方法和内容将另有一份统计分析计划,并在数据锁定前由主要研究者、统计学家、申办者的签字确定。 (2)所有的统计检验均采用双侧检验,P值小于或等于0.05将被认为所检验的差别有统计意义。 (3)计量资料将采用均数±标准差进行统计描述。两两比较采用t检验。三组的变化采用F检验或非参数检验进行比较。 (4)计数资料采用频数进行统计描述。三组的变化采用X2检验和Wilcoxon秩和检验。 (5)脱落分析。 (6)基础值的均衡性分析:采用F检验或X2检验,以衡量均衡性如何。
|
Data collection and Management (A standard data collection and management system include a CRF and an electronic data capture:
|
Data management: (1) The investigators recorded the data in a timely, complete, correct and clear manner, based on the original observations of the subjects. (2) whether the inspector conducts the research program to follow the research program. Confirm that all case reports are filled in correctly and consistent with the original data. If there are errors and omissions, promptly ask the researchers to correct. Modify the original record to keep clear, corrections need to be signed by the researchers and specify the date. (3) after the inspector to check the case report form, by the inspector to verify the signature, the timely delivery of clinical research data administrator. There should be a special record for the delivery of the completed case report form between the investigator, the inspector and the data manager, and the corresponding signature should be received and the record should be kept properly. (4) the data administrator in the data before the re-verification, found the problem promptly notify the inspector, asked the researchers to answer. The interrogation between them and the exchange of questions should be in the form of questionable forms. The question form should be kept for future reference. (5) data manager in the data before the entry, to understand the contents of the observation form of the project and the code, the coding process is recorded in the code of the preservation. Database naming should be standardized, easy to read, easy to find. And to ensure that it is correct, safe and confidential. (6) data entry data entry using the second entry. The entry process should identify the problem or accident, should be registered and timely report in order to quickly deal with the problem, after the end of the data entry should be part of the observation table to understand the quality of input, analysis and processing of existing problems. (7) The data manager shall, together with the principal investigator, formulate data range checks and logically check the contents of the range and the interrelationships of the indicators in the case report form. And the preparation of the corresponding computer program, before the input control error data input, find out the cause of the error to be corrected, all the error content and modify the results should be recorded and properly saved. (8) the original case report form in accordance with the requirements to complete the data entry and verification, according to the order of the order of preservation, and fill in the search directory, for inspection. The electronic data files include the database, the inspection program, the analysis program, the analysis result, the codebook and the documentation, etc., should be classified and saved, and there are multiple backups to be stored on different disks or recording media, so as to save and prevent damage. All original archives should be in accordance with China's "drug clinical trial quality management practices," the provisions of the period of preservation. Statistical Analysis: (1) detailed statistical analysis methods and content will be a statistical analysis plan, and before the data locked by the main researchers, statisticians, the sponsor's signature to determine. (2) All statistical tests are conducted on a bilateral basis, and a P value less than or equal to 0.05 will be considered statistically significant. (3) The measurement data will be described by mean ± standard deviation. Comparison of the two t test. Changes in the three groups were compared using F-test or nonparametric tests. (4) count data using the frequency of statistical description. Three groups of changes using X2 test and Wilcoxon rank sum test. (5) shedding analysis. (6) the balance of the basic value of the analysis: the use of F test or X2 test to measure the balance.
|