Source data and source files
Source data and source files are two concepts that the clinical trial industry must face. Both concepts have clear concepts in ICH-GCP. . The author does not intend to confuse the public. I just want to share with you my thoughts on source data and source files under the concept of ICH-GCP. If there is anything wrong, I hope you will criticize and correct me.
1. Source data:
According to the definition of ICH-GCP, source data are original records or copies of clinical findings, observations and other activities during the clinical trial process (verification all information necessary to reconstruct and evaluate the clinical trial. It can be seen from this definition:
1. From the source, the source data is generated in clinical trials, and generated in medical routine or practice does not belong to the source data discussed in ICH-GCP.
2. From a functional point of view, the source data is used to reconstruct and evaluate clinical trials. Some other medical information may also be generated in clinical trials, such as blood routine test values ??that are not within the scope of research data collection. . But there are also special circumstances. Once the test values ??are abnormal, researchers need to judge the clinical significance of these values. Regardless of whether the judgment results are meaningful or not, it is all for the purpose of evaluating the safety information of patients participating in clinical trials. The evaluation also belongs to the source data. This also points out that the source data is relative and is relative to the research data collected.
3. From the perspective of the carrier, source data is the information on the original records of clinical findings, observations and other activities during the clinical trial or their copies (certified copies).
4. In essence, source data is... information, so it can be any form of information such as numerical values, images, text, dates, voices, videos, sound waves, etc. However, in clinical trials, the most commonly seen or routinely used items are numerical values, images, text, and dates. There may be more and more voices or videos in the future, and sound waves or ultrasound or other visits may also be explored by humans. The continuous development of technology in the world or universe is gradually added.
5. From the perspective of predictability, source data is all the information necessary to reconstruct and evaluate clinical trials, covering the data collected in clinical trials. The source data of the collected data can and should be determined initially in the program design. Come.
6. From the classification point of view, a. According to the carrier classification, it can be divided into paper data, electronic data or other media data (such as optical disks, U disks, tapes, etc.). b. According to the type of generation, it can be divided into objective data and subjective data (refer to Guangdong Pharmaceutical Association's "Drug Clinical Trial Source Data Management Guangdong *** Knowledge"). Objective data is generated by objective circumstances and cannot be modified. However, objective data may be biased due to problems with instruments and equipment. In this case, it is necessary to use calibrated and verified instruments and equipment to obtain correct objective data. Once the data is generated , changes cannot be made unless there are sufficient reasons and written explanations. Subjective data refers to the diagnosis, assessment or scale response or information exchange made by researchers, patients or other clinical trial participants based on objective data. It will be biased due to the influence of human knowledge, experience and objective environment. and inconsistent, so subjective data can and may be modified, but the modification must have a certain basis and reason, and there must be a modification track that meets the requirements of clinical trials.
7. From the perspective of originality. Source data is the original record or a photocopy thereof (certified copy). However, because the questionnaires of some studies are directly interviewed by the researchers and recorded in the research data collection system (CRF or eCRF), ICH-GCP also recognizes these data as source data. If this situation does arise, the plan should be explained in advance when designing the plan, or at least a filing should be made to the ethics committee before taking this measure.
2. Source files:
According to the definition of ICH-GCP, source files are original documents, data and records. Then a series of examples are given to illustrate which source files are. The author was confused when I first saw it: it is understandable that the source file is an original file, and the source file can also be original data and records. Isn't it no different from the source data? But when I checked the words, I found that the data and records here are plural concepts.
Therefore, the author believes that the source data corresponds to one research collection data item, and the source file can cover multiple research collection data items at the same time, and is the storage carrier of the source data.
The source file is the carrier of the source data. It must be the source file that records or carries the source data, but not all the source data is carried on the source file. This can also be done using the blood routine example given in the source data above. The blood routine also contains some test values ??that are not within the scope of the research collection data. These test values ??are routine medical test values, but are not seen in reconstruction studies and evaluation studies. It has to have some meaning.
Source files are the carrier of source data, which can be paper, electronic, or other media (such as optical disks, U disks, tapes, etc.). This reminds me of a question I discussed with my colleagues before: Can the printed version of medical records in Word be used as a source file? According to the definition of ICH-GCP, it does not specify which types of files cannot be used as source files, but the requirements for source files and source data have been clearly stated - the ALCOACCEA principle. If the word printed medical record confirmed by the researcher's signature and date is positioned as the source document, then it must: 1. The unsigned electronic version of the medical record in the computer has a time series track to prove that the document is at the same time as the recorded content It is generated first-hand; 2. The modified part must be modified on the paper version, and the modification track is clear and traceable; 3. The provenance document record must be recognized within the institution. If the institution does not Admitted, is illegal data. As a reminder, in the newly revised version of China's GCP, it is clearly stated that researchers should record research-related information in the hospital's outpatient or inpatient system.
The source file is the carrier of the source data. Any modification of the source data means a change in the content of the source file. However, the source file is the original record carrier of the source data and should not be changed at will. Changes for any reason should be supported by corresponding reasons or explanations. Some hospital medical records are recorded in the hospital His system, electronically signed by the researcher and then archived, and then printed and archived with a handwritten signature. When a medical dispute occurs, or in the statement of some hospitals' medical systems, the paper version of the handwritten signed medical record is regarded as a legally binding document. In such cases, relevant instructions or statements must be archived. Due to paper document storage space issues, some hospitals even scan the paper version of the medical record and provide the scanned version of the electronic medical record as the source document to the sponsor. At this time, the supervisor must also be informed of these situations in a timely manner during the visit to the selection center, and make corresponding reports and archives.
When the source data is directly recorded in the CRF or eCRF, this part of the CRF or eCRF becomes the source document, but this must be explained in the plan, or submitted to ethics for filing in advance.
When the source document of the original record is unavailable, a verified copy may also be accepted as the source document, but it must be accompanied by corresponding verification (signature of copywriter and date of signature) or certification, and the content of the document must be It must be clear and legible, and there must be a traceable track for file modifications.
A brief reflection on the design principles and use of source data positioning tables or positioning protocols
Based on the above thoughts on source data and source files, the author will also discuss the methods currently used by many companies. I would like to share my thoughts on the usage significance and design principles of source data location table files with you, and please comment on them.
1. Brief thoughts on the meaning of use:
Although some companies have begun to use this form, due to the panic about the national bureau’s on-site verification and the different participants’ understanding of the source data and source documents Inconsistency in understanding, you can always hear doubts about whether the use of this form is necessary: ??the use of the form is likely to dig a hole for yourself, because regardless of whether the form is used or not, the on-site verification of the National Bureau will be based on its own understanding of the source conduct. The author originally used this form in a confused way, thinking that this form was a measure for major pharmaceutical companies or CROs to deal with the 7.22 storm, and it was a last resort. As everyone knows, many experienced and relatively standardized large pharmaceutical R&D companies or CROs have begun to use such a form as early as 2015. Most of their basis is the sponsor's responsibility for clinical trial data quality management, which is also the sponsor's responsibility. It is a practical measure and evidence for parties to evaluate the center’s source data and source file management when selecting a center.
As I communicate and discuss with leaders and colleagues from different backgrounds and fields, and as I think about various problems encountered in the process of revising this table driven by work responsibilities, the author is constantly sorting out my own ideas: < /p>
1. The use of this type of table is not to define the source data or source files, but to confirm the source data and source files on the basis of adhering to the definition of source data and source files in ICH-GCP. form. Therefore, the concepts of source data and source documents adhered to in this form are consistent with those of the national bureau's verifiers, and there is no problem of self-digging.
2. The use of this type of form can clearly outline the source data and source files required for research in the research center, provide a basis for risk analysis for clinical trial data quality management, and can also provide a basis for monitoring Members of the SDV clarify the scope and provide SDR with the convenience of searching based on the map.
3. Setting and signing the form in advance can also remind researchers to record source data consistently during the research process, avoiding unnecessary casual recording, repeated recording, and delayed recording.
4. The signing of the form can also provide a basis for efficient and rapid query of source data and source documents during verification and audit, and provide a reference for the credibility and integrity of the research data. Therefore, the author supports the use of this form and will continue to work hard to improve the content of this form.
2. Brief Thoughts on Design Principles
Clinical trials are conducted to collect effectiveness data and safety data of drugs. Therefore, a reasonably designed clinical trial should clearly understand the needs Collected data, and any data collected must be corresponding to a source data. The authenticity of research data that is not supported by source data is questioned, and source data needs a carrier-source file to be saved. The specific file to be saved will vary depending on the hospital's subjective and objective reasons. Therefore, many companies have adopted a form or protocol similar to source data positioning (Guangdong Pharmaceutical Association also has special discussions and suggestions in "Drug Clinical Trials Source Data Management Guangdong *** Knowledge"). The form of this type of table generally contains three parts: source data, source data carrier and remarks. Some will also set some other columns according to the company's data management purpose or research purpose, such as whether it can be queried in the hospital system, and the department where the source file is stored. or specific location, etc., but in any case, the form should reflect the following principles:
1. Collaboration: The design and completion of this type of form is jointly completed by the sponsor and the researcher. The sponsor should clearly list the source data of the research data collected by the institute in advance, and explain the data items in special situations, such as instructions for the source data to be filled directly into the CRF or eCRF, for example, what kind of source documents must be collected as medical history information evidence. The researcher is responsible for confirming the source files of the source data collected in the center in the center, and explaining special situations, such as the identification and clarification of the source data when multiple source files contain the same data collected for the study, or An explanation of the situation where the same source data item appears in different source files due to differences in patient study visit methods, etc. In the author's opinion, the determination and completion of the source data location table cannot be separated from the institutional office of the clinical trial manager of the research center. Because researchers may not be familiar with the positioning of some source data, the addition of institutions can provide more precise positioning from the hospital. Moreover, the source files of some hospitals are not necessarily the first to be generated. For example, hospitals that use scanned copies as source files or printed signed versions of His documents as source files. If there is no clear negotiation with the institution, the researcher will not be able to do a good job. Be aware that institutions sometimes will not recognize unilateral signature forms between the sponsor and the investigator. Therefore, the author understands that this form is best called an agreement. Even if the institution does not participate in signing, the auditor should clearly record the communication with the institution and the reasons for not signing in his report.
2. Openness: The source data of this type of table is provided by the sponsor's project team based on the research data collection, but the extent to which the source data should be listed and how many should be based on the actual situation of each study. It depends. In principle, the more detailed the source data is, the more accurate the positioning will be.
But the author does not recommend listing everything in detail. For example, for example, laboratory test values, is it necessary to list blood routine test values ??or biochemical routine test values? Or what about the platelet or creatinine values ??in more detail? The author recommends setting source data items based on the same type, collected at the same time, or reported on the same source file at the same time. There is no need to subdivide, otherwise there will be too much data to list. Of course, these must be consistent with the actual data collected for the study. On the other hand, the researcher's positioning of the source data - the exploration of the source files is also open. When the researcher understands the source data items provided by the sponsor, the researcher confirms the first recorded file of the source data of the center - the source file. .
Openness is also reflected in other aspects. Generally speaking, this form is relatively stable to use, and it is best to be consistent. However, if there is any change in the source document in a center, the researcher is obliged to remind the inspector of the situation and update the form in a timely manner. The reason for the update of the form should be recorded in writing. During the monitoring process, if the supervisor discovers that the source file where the actual source data is located is inconsistent with the one confirmed by the researcher, he should confirm the reason with the researcher in a timely manner. If necessary, he should initiate the update of this form and record the reason in writing. . If it is not necessary to update this form, the inconsistency should be recorded in writing and filed with this form.
3. Complementarity: Regarding the settings of source data items, the sponsor’s project team should set them according to the data collection situation of this study, but should adhere to the complementarity of the settings of each source data item and try to There should be no duplicate source data items. If there are duplicate source data items, the project team should add a description of the source data collected by the source data item. For example: the admission criteria can be set to the admission criteria assessment, admission criteria records that cannot be supported by corroboration or objective data, and admission criteria that are supported by objective data or corroborating data (test orders or imaging, pathology data, etc.) . If the entry criteria, laboratory results, imaging reports and pathology reports are set, it will often cause the source data to appear repeatedly. At this time, the entry standard source data items should be explained and the specific source data covered should be explained. Do not duplicate laboratory results, imaging reports, and pathology reports.
4. Uniqueness: According to the definition of source data and source files, source data and source files should be unique and correspond one-to-one. However, in reality, some inconsistencies often occur due to the complexity of the research situation or the research center situation. For example, the patient notification process records source data items. Because some patients are informed during outpatient visits and some are informed through inpatient medical records, the source data items in a center may have different carriers due to different patient visit methods - source document. The research visit of the same patient may be an inpatient visit or an outpatient visit, so the records of the research process or related content may be inpatient medical records or outpatient medical records. At this time, it is necessary to make notes on the correspondence between the source data items and the source files to clearly understand the inconsistencies and specific reasons.