Is the Clinical Database part of the Trial Master File (TMF)?

This was a question posed recently on the MHRA GCP Forum (see  I posted a response (though it wasn’t published by the Forum Moderator) pointing out that for several reasons the clinical database should not be considered an “essential document” and therefore not part of the TMF, though it would be advisable to retain the database for a number of other reasons (not least, for its ongoing research and development value). In my opinion, because the raw data (the CRFs/EDC data) and the derived data (the SAS datasets and the tables, figures and listings) are part of the TMF (and included in the TMF Reference Model), I would not consider the clinical database to be an essential inspectable component of the TMF, with the caveat that all of the supporting database records identified in the TMF Reference Model are retained (programming, specifications, validation records, QA/QC records etc). The MHRA take a different view.

On the Forum, the MHRA state:

“The MHRA will inspect all aspects of the database including the retained essential documents in the TMF and the database itself, which includes the clinical data and all the metadata (audit trails etc.). Direct (and guided) access to the database should be made available along with the TMF for inspection.”

This is a highly significant clarification, considering the retention period for clinical records per EU Regulation 536/2014 is a minimum of 25 years. I have great concerns regarding how a sponsor will be able to maintain all aspects of the clinical database – including “all the metadata (audit trails etc)” for a period of 25 years and provide direct access to the system for this period. Could you provide direct access to an MHRA inspector to the original clinical database from 1990? On the one hand, we’re directed to adopt a risk-based approach and to focus on issues that could potentially have a direct impact on patient safety. On the other hand, we get directives to retain derived/secondary data that has questionable GCP value when taken in the context of all other records that are maintained in the trial master file. I get the feeling we sometimes take one step forward and two steps back!


About rammellel

Records management consultant to the life sciences / pharmaceutical industry
This entry was posted in Compliance, e-records, Technology, TMF. Bookmark the permalink.

2 Responses to Is the Clinical Database part of the Trial Master File (TMF)?

  1. I do consider the clinical database part of the TMF (8.3.14; 8.3.15;
    – In case of eCRFs the database contains the original entries by the investigator
    – A monitor/data manager can implement self evident corrections (original entries) that are not acknowledged by the investigator
    – A data manager can make annotations
    – A data manager may need to change data table configuration different from the CRF for data analysis purpose.This may require dedicated instructions for data entry personnel
    – The validation rules (part of meta data) may have been programmed incorrectly resulting in data validation errors and unwanted data corrections.
    – The data manager will code AE’s and medication which are original entries and may contain incorrect interpretations
    – Do you need more reasons?

    If you do not retain the link between the CRF and the report there is no evidence of data integrity and your data will not be valid for submission to regulatory authorities. Therefore it is strange that they are not specified as essential documents.

  2. rammellel says:

    Thanks Jan. You are absolutely right for EDC trials, where the database IS the raw data i.e. it exactly matches the ICH requirement for a case report form. (8.3.14) and CRF corrections (8.3.15). I should have been more specific in my post but I was thinking of traditional paper-based studies where the paper CRF and paper data clarification forms meet the requirements of 8.3.14 and 8.3.15. The database is secondary data. Statistical programming converts the raw data into tables, figures and listings. It just appears off to me that elsewhere in the regulations, it is perfectly acceptable to transform and to transfer data into other format (e.g. for archival purposes…. see EU Notes for Guidance) and so long as the transformation process is validated there is no requirement to retain the original; that’s explicitly what the regulators have said. So how come they want the original clinical database retained for 25 years when we have the original paper CRFs, the programming, the validation records, and the final reported data. To my mind, there’s an inconsistency in the requirements.

    Incidentally, this scenario is not always just restricted to paper-based studies. Some sponsors take the EDC data which is collected and managed in one system and then import it into a separate clinical database for statistical analysis and reporting purposes. In this scenario, I would consider the EDC data sets to be the CRFs (8.3.14).

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.