Regulation and Reproducibility: Can You Reproduce Your Clinical Trial Results?
Imagine that it’s been three years since the completion of a trial, and that suddenly a regulatory body calls into question the findings:
Was a particular trial site operating properly?
Can you clarify an aspect of the results?
Why did you make a particular decision at an interim look?
Suddenly, your somewhat old data needs to be able to reproduce your initial findings. In such a case, how long would it take you to satisfy the regulatory body?
Reproducibility is a rarely discussed, but absolutely critical aspect of trial implementation and data management. It’s a mistake to believe that once your statisticians have analysed and presented your data, the job of data management is complete. Regulatory bodies and others can technically call into question the validity of your data at any given time.
While it may take some time to track down your data, there are a few things you can do to simplify the process:
Format your data consistently: Data stored in different formats will create headaches for accessibility. While this may sound obvious, there is a tendency amongst sponsors to store data in the format most convenient for a project team. Moreover, consistent data formatting across studies (e.g. across two or more phase 2 trials) sometimes appears to slow the process of trial completion. However, in the long run, ensuring that all data is available in the same format will make it easier to reassess and reproduce.
Leave a clear audit trail: When implementing a trial, try if you can to leave a clear audit trail of who had access to what evidence at which point in the trial. Certain software is built with the foresight that you may have to communicate this information to a regulator at some point. A comprehensive audit trail can only protect the findings of your trial.
Streamline data management and biometrics: Ensure that your biometrics team is involved with data management (or vice versa): Ultimately, the more closely aligned biometrics is with data management, the more likely each team will understand the needs of the others. Streamlining their processes will make it easier to track down and reproduce data later.