Clinical trials have come a long way from the randomized controlled trial of streptomycin in pulmonary tuberculosis, conducted by the UK Medical Research Council (MRC) in 1943.This trial was the first of its kind to deploy an organized trial design in addition to defining enrollment, data collection, and reporting criteria. Over the years, ethical standards in clinical research have grown more stringent to ensure patient safety, with tight regulations governing enrollment, data collection, and reporting. As a result, generating and handling the huge amount of data while ensuring compliance in clinical trials has become a herculean task. One of the most challenging situations arises when a sponsor pharmaceutical company is faced with the prospect of having to unlock a final database.
How database management works during clinical trials
The clinical trial process begins with finalizing the trial protocol and statistical analysis plan, followed by the actual trial and interim analysis. During clinical trials, a Clinical Data Management System (CDMS) is used to store the data generated from various sets of trials. The data is subjected to stringent scrutiny for validation by database managers, who query the site personnel or investigator in case of any discrepancies and prepare a clean database. This database is available to the clinical trial team for analysis and reporting, after database lock. A lot of effort goes into data cleaning, programming, and management before the final data base lock and freeze happens.
Sponsor companies ensure that all case report forms (CRFs) data is in the database and all queries are resolved before the decision to freeze is taken. Prior to the final database lock, a soft lock is declared. The clinical trial team reviews the dry run outputs which provide a glimpse of the final outputs. The team then prepares for further analysis, interpretation, and reporting of findings in a clinical study report (CSR). Any comments from the team on dry-run outputs are addressed and applied to the database for a final round of cleaning.
Some situations that call for a database unlock
Once the database is locked, it is considered final and cannot be unlocked as per regulations. However, there are some exceptions to this. For instance, if the trial team observes an erroneous or missing data point, it needs to be fixed before the CSR is developed. This requires the database to be unlocked.
Why is database unlock such a dreaded situation in a clinical trial? There are several reasons for this. The most important being questions about data integrity, besides other associated statistical concerns that can interfere with regulatory compliance. Due to procedural complexity and the regulatory concerns involved, small errors are often rectified by inserting a footnote or updating the output. However, in case of critical errors, unlocking and rectifying the database is the only option. Such situations include:
- Errors found in outputs, which were not part of the dry-run review. For instance, Public Key outputs that were not part of the dry-run, and were therefore not reviewed at all.
- Different units used by sites for the parameter, which goes unnoticed until after database lock.
- Errors found in individual subject-level data that became available only after database lock (such as individual subject listings and pharmacokinetic data).
- Any issue that qualifies as a protocol deviation or any data that is critical to patient safety.
- Protocol specified adverse events of special interest (AESIs) not flagged as adverse events.
There are no regulatory guidelines from health authorities on how to perform database unlock. However, it is expected that before that happens, exhaustive efforts must be made to ensure data quality, and that audits are completed and issues resolved. The sponsor company is expected to have a very clear and precise standard operating procedure (SOP) that details every step. Only authorized person(s) can make changes, and as soon as corrections are made, the database must be re-locked while maintaining audit trails for every step. All necessary documentation and note-to-files must be maintained by the sponsor to ensure that data integrity is not compromised.
Driving integrity and compliance in the database unlock process
A well-managed database unlock is critical to the success of the trial as well as the reputation of the sponsor. Following certain best practices can help. Ensure continuous monitoring of data and perform data cleaning at regular intervals for greater clarity and fewer errors. Conduct frequent interim analysis of study data with the help of statistical experts to identify outliers in the data, check cleanliness, and understand where the trial is headed to in terms of objectives and endpoints. Leveraging / internal and external expertise in perfect impact analysis, biostatistics, and database programming can result in a robust database unlock and relock process. Implementing standard operating procedures and clear workflows is also equally important.
Is your DB unlock process equipped to ensure data integrity, transparency, patient safety, and compliance?